102126 Live Long and Prosper WO RL D BA N K E A S T A SI A A N D PAC I FIC REG I O N A L REP O R T S Known for their economic success and dynamism, countries in the East Asia and Pacific region must tackle an increasingly complex set of challenges to continue on a path of sustainable devel- opment. Learning from others within the region and beyond can help identify what works, what doesn’t, and why, in the search for practical solutions to these challenges. This regional flagship series presents analyses of issues relevant to the region, drawing on the global knowledge and experience of the World Bank and its partners. The series aims to inform public discussion, policy formulation, and development practitioners’ actions to turn challenges into opportunities. T I T L ES IN T H E SE R IES Live Long and Prosper: Aging in East Asia and Pacific East Asia Pacific At Work: Employment, Enterprise, and Well-being Toward Gender Equality in East Asia and the Pacific: A Companion to the World Development Report Putting Higher Education to Work: Skills and Research for Growth in East Asia All books in this series are available for free at https://openknowledge.worldbank.org / handle/10986/2147 World Bank East Asia and Pacific Regional Report Live Long and Prosper Aging in East Asia and Pacific © 2016 International Bank for Reconstruction and Development / The World Bank 1818 H Street NW, Washington, DC 20433 Telephone: 202-473-1000; Internet: www.worldbank.org Some rights reserved 1 2 3 4 18 17 16 15 This work is a product of the staff of The World Bank with external contributions. The fi ndings, interpre- tations, and conclusions expressed in this work do not necessarily reflect the views of The World Bank, its Board of Executive Directors, or the governments they represent. The World Bank does not guarantee the accuracy of the data included in this work. The boundaries, colors, denominations, and other information shown on any map in this work do not imply any judgment on the part of The World Bank concerning the legal status of any territory or the endorsement or acceptance of such boundaries. Nothing herein shall constitute or be considered to be a limitation upon or waiver of the privileges and immunities of The World Bank, all of which are specifically reserved. Rights and Permissions This work is available under the Creative Commons Attribution 3.0 IGO license (CC BY 3.0 IGO) http:// creativecommons.org/licenses/by/3.0/igo. Under the Creative Commons Attribution license, you are free to copy, distribute, transmit, and adapt this work, including for commercial purposes, under the following conditions: Attribution—Please cite the work as follows: World Bank. 2016. Live Long and Prosper: Aging in East Asia and Pacifi c. Washington, DC: World Bank. doi:10.1596/978-1-4648-0469-4. License: Creative Commons Attribution CC BY 3.0 IGO Translations—If you create a translation of this work, please add the following disclaimer along with the attribution: This translation was not created by The World Bank and should not be considered an offi cial World Bank translation. The World Bank shall not be liable for any content or error in this translation. Adaptations—If you create an adaptation of this work, please add the following disclaimer along with the attribution: This is an adaptation of an original work by The World Bank. Views and opinions expressed in the adaptation are the sole responsibility of the author or authors of the adaptation and are not endorsed by The World Bank. Third-party content—The World Bank does not necessarily own each component of the content con- tained within the work. The World Bank therefore does not warrant that the use of any third-party-owned individual component or part contained in the work will not infringe on the rights of those third parties. The risk of claims resulting from such infringement rests solely with you. If you wish to re-use a component of the work, it is your responsibility to determine whether permission is needed for that re-use and to obtain permission from the copyright owner. Examples of components can include, but are not limited to, tables, figures, or images. All queries on rights and licenses should be addressed to the Publishing and Knowledge Division, The World Bank, 1818 H Street NW, Washington, DC 20433, USA; fax: 202-522-2625; e-mail: pubrights@ worldbank.org. ISBN (paper): 978-1-4648-0469-4 ISBN (electronic): 978-1-4648-0470-0 DOI: 10.1596/978-1-4648-0469-4 Cover photo: © Qilai Shen / Panos Pictures. Used with the permission of Qilai Shen / Panos Pictures. Further permission required for reuse. Library of Congress Cataloging-in-Publication Data has been requested. Contents Foreword . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . xv Acknowledgments. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . xvii Abbreviations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . xix Overview . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1 Understanding the rapid demographic and epidemiological transition . . . . . . . . . . . . . . . . . 3 Situation and sources of support of older people . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6 Managing the effect of aging on growth and its drivers . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9 Developing pension, health, and long-term care systems for aging societies . . . . . . . . . . . . 13 Addressing cross-cutting challenges in aging . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 17 Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 18 References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 18 I The Demographic Transition and Well-Being of Older People in East Asia and Pacific. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 21 1. The Demographic and Epidemiological Transitions in East Asia and Pacific . . . . . . . . . . . 23 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 23 The demographic transition in East Asia and Pacific . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 24 The epidemiological transition in East Asia and Pacific . . . . . . . . . . . . . . . . . . . . . . . . . . . 33 Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 40 Notes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 43 References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 43 2. The Welfare and Labor Market Behavior of Older People in East Asia and Pacific . . . . . . 47 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 47 Poverty among older people in East Asia and Pacific . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 48 Other measures of well-being . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 51 Living arrangements and access to support among older people . . . . . . . . . . . . . . . . . . . . . 54 v vi CONTENTS Sources of fi nancial support and labor market behavior of older workers . . . . . . . . . . . . . 57 Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 73 Notes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 74 References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 75 II Aging and the Economy . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 79 3. Aging and the Macroeconomy in East Asia and Pacific . . . . . . . . . . . . . . . . . . . . . . . . . . . 81 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 81 Aging and growth in East Asia and Pacific . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 81 Channels to augment the labor force . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 83 Labor force quality and productivity . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 87 Saving in East Asia and Pacific: Will aging play an influential role? . . . . . . . . . . . . . . . . . . 91 Fiscal effects of aging . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 96 Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 102 Notes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 103 References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 104 4. Labor Market Policies to Address Aging in East Asia and Pacific. . . . . . . . . . . . . . . . . . . 109 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 109 Fertility incentives . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 110 Enhancing female labor force participation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 110 Extending productive working lives . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 114 Promoting immigration into aged East Asian and Pacific economies . . . . . . . . . . . . . . . . 122 Policy effects on the labor force . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 125 Upgrading skills of older workers and integrating strategies. . . . . . . . . . . . . . . . . . . . . . . 129 Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 133 Notes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 133 References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 134 III Preparing for an Aging East Asia and Pacific . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 139 5. Pension Systems and Aging: Reforms for Winning the Race. . . . . . . . . . . . . . . . . . . . . . . 141 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 141 Overview of East Asian pension systems . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 141 Assessment of the performance of East Asia’s pension systems . . . . . . . . . . . . . . . . . . . . . 147 Reforms to date and options looking forward . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 166 Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 177 Notes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 177 References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 180 6. Population Aging and Health Services in East Asia and Pacific . . . . . . . . . . . . . . . . . . . . 185 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 185 Population aging and health service delivery challenges in East Asia and Pacific . . . . . . . 185 Universal health coverage in the context of population aging and NCDs . . . . . . . . . . . . . 195 Population aging and health expenditures in East Asia and Pacific . . . . . . . . . . . . . . . . . 195 Health policies and strategies for an aging population . . . . . . . . . . . . . . . . . . . . . . . . . . . 202 Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 220 Notes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 221 References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 221 7. Long-Term Care in Aging East Asia and Pacific . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 227 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 227 LTC demand and expenditures: Current trends and projections . . . . . . . . . . . . . . . . . . . . 228 CONTENTS vii Providing LTC. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 231 Financing LTC . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 236 Coordinating LTC services . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 241 Ensuring capacity and quality for LTC provision . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 243 Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 245 Notes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 246 References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 246 IV Looking Ahead . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 251 8. To Live Long and Prosper: Looking Ahead . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 253 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 253 Demographics . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 254 Macro and fiscal issues . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 254 Labor markets. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 256 The situation of older people and households in East Asia and Pacific . . . . . . . . . . . . . . . 258 Health and long-term care . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 259 Pensions and social security . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 261 Cross-cutting issues . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 262 Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 262 Notes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 263 References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 263 Boxes 1.1 Uncertainty and assumptions in United Nations population data . . . . . . . . . . . . . . . . 30 1.2 Alternative measures of dependency . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 34 1.3 Is longevity accompanied by more good health or less? . . . . . . . . . . . . . . . . . . . . . . . . 35 1.4 Very high prevalence of smoking and related death rates in the region . . . . . . . . . . . . 41 1.5 Indonesia: Increasing and often unequal exposure to risks . . . . . . . . . . . . . . . . . . . . . 41 1.6 Malaysia: Undiagnosed noncommunicable disease risks and modest social economic status differentials in management of risk . . . . . . . . . . . . . . . . . . . . . . . . . . 42 1.7 Adverse childhood socioeconomic conditions and increased mortality risks later in life . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 42 3.1 The scope for improving skills of an aging workforce . . . . . . . . . . . . . . . . . . . . . . . . . 88 3.2 A labor productivity paradox? . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 90 4.1 Fertility and public policy in East Asia and Pacific and the OECD . . . . . . . . . . . . . . 113 4.2 The lump of labor fallacy . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 116 4.3 Employment services targeted to older workers in Japan and the Republic of Korea. . . 117 4.4 Reforming seniority wage systems in the Republic of Korea, Singapore, and Japan . . . 120 4.5 East Asia and Pacific’s distinctive approach to retirement age and continued work . . . . 121 4.6 Assumptions underlying labor force scenarios . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 127 4.7 Second Basic Employment Promotion Plan (2012–16) for the aged in the Republic of Korea. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 132 5.1 Automatic adjustments for changes in life expectancy in defined benefit schemes . . . 170 5.2 Integrating noncontributory and contributory pensions: Chile’s solidarity pillar. . . . 174 6.1 Poor qualifications of staff to provide effective diagnosis and treatment . . . . . . . . . . 190 6.2 Limited supply-side readiness to manage and treat noncommunicable diseases at the primary health care level in Indonesia . . . . . . . . . . . . . . . . . . . . . . . . 192 6.3 The role of technological change and interactions with aging in health care cost escalation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 200 viii CONTENTS 6.4 Care coordination strategies in Organisation for Economic Co-operation and Development countries . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 211 6.5 The effect of care coordination on key health system outcomes. . . . . . . . . . . . . . . . . 212 6.6 Cost containment through provider payment and purchasing reforms in Thailand and Turkey . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 215 6.7 The power of smart purchasing: Pharmaceutical procurement in the Republic of Korea, Sri Lanka, and Thailand versus Vietnam. . . . . . . . . . . . . . . . . . . 217 6.8 Good examples of decision-making processes for new technologies in East Asia and Pacific . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 219 7.1 LTC hospitals and LTC facilities in the Republic of Korea: In direct competition . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 233 8.1 The Active Aging Index: A tool for monitoring the situation and potential of older people . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 260 Figures O.1 East Asia and Pacific has more older people than any other region . . . . . . . . . . . . . . . . 3 O.2 East Asian and Pacific economies are aging more rapidly than economies elsewhere . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .4 O.3 Three distinct aging patterns are apparent in East Asia and Pacific . . . . . . . . . . . . . . . . 5 O.4 In the richer economies of East Asia and Pacific, the working-age population is projected to shrink between 2010 and 2040. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6 O.5 Labor force participation rates in East Asia and Pacific vary, but across the region, rural people work significantly longer than urban people . . . . . . . . . . . . . . 7 O.6 Work remains the primary source of old-age support in most countries in East Asia and Pacific . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8 O.7 Attitudes on the preferred source of fi nancial support in old age and on the expected source of personal care are shifting in East and Southeast Asian economies . . . . . . . . . 9 O.8 East Asia and Pacific’s developing economies are getting old before getting rich . . . . . 10 O.9 Greater participation by women, older people, and migrants can mitigate the projected decline in the labor force in richer East and Southeast Asian economies . . . 12 O.10 Across APEC, pensions will eat up a larger and larger share of GDP in coming decades without reforms . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13 1.1 East Asia and Pacific is a diverse region in terms of the share of older people . . . . . . . 24 1.2 East Asia, led by China, has more old people than any other region . . . . . . . . . . . . . . 25 1.3 East Asian and Pacific economies are aging more quickly than other economies in the past . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 26 1.4 The share of elderly population in East Asia and Pacific will converge with that of OECD countries by 2060 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 26 1.5 The population of “oldest old” will increase more in East Asia and Pacific between 2010 and 2060 than in any other region . . . . . . . . . . . . . . . . . . . . . . . 27 1.6 Three distinct aging patterns are apparent among East Asian and Pacific economies . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 28 1.7 People in aging East Asian and Pacific countries see aging as a major challenge for their countries . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 29 1.8 Fertility rates have declined significantly across all East Asian and Pacific economies but are projected to flatten in orange and red economies. . . . . . . . . 29 B1.1.1 UN population projections for East Asia and Pacific have been subject to significant uncertainty . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 30 CONTENTS ix 1.9 Declines in age-specific mortality rates in East Asia and Pacific in recent decades have been substantial, particularly in the early years of life and for women . . . . . . . . 31 1.10 The typology of East Asian and Pacific economies is robust in scenarios with low-, medium-, and high-fertility assumptions. . . . . . . . . . . . . . . . . . . . . . . . . . . 31 1.11 The working-age population will shrink over coming decades in richer East Asian and Pacific economies . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 32 1.12 Total dependency ratios conceal contrasting trends in youth and elderly population shares . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 33 1.13 Life expectancy and health-adjusted life expectancy in East Asia and Pacific both increased between 1990 and 2010 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 36 1.14 Large differentials in healthy years of life are evident across the region, but men generally live shorter, less healthy lives than women . . . . . . . . . . . . . . . . . . . 37 1.15 NCDs have increased sharply as leading causes of disease burden in East Asia and Pacific in recent decades . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 38 1.16 NCDs and injuries dominate the leading causes of mortality in China . . . . . . . . . . . . 39 1.17 In the Republic of Korea both men and women 65 years of age and older are more likely to have multiple chronic conditions . . . . . . . . . . . . . . . . . . . . . . . . . . . 39 1.18 In China, in 2006, socioeconomic status had an inverse relationship with noncommunicable diseases . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 40 1.19 NCD-related risk factors grew in East Asia and Pacific between 1990 and 2010 . . . . 40 2.1 Poverty rates increase with age in some East and Southeast Asian countries, but elderly poverty has fallen over time . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 48 2.2 Poverty among the rural elderly remains significantly higher than for urban elderly . . . . 49 2.3 Difficulties with physical function and instrumental activities of daily living increase with age . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 51 2.4 Poorer and less educated elderly are more likely to face difficulty performing instrumental activities of daily living . . . . . . . . . . . . . . . . . . . . . . . . . . . . 52 2.5 Rates of depression tend to increase with age and are higher among poor households . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 53 2.6 Elderly co-residence rates are generally high in East Asia and Pacific but vary across countries and by gender and location within countries . . . . . . . . . . . . . . . 55 2.7 Co-residence rates of older adults with adult children have declined over time in China and Thailand . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 56 2.8 As countries get richer, elderly co-residence rates tend to decline. . . . . . . . . . . . . . . . . 56 2.9 Labor is the leading income source for people ages 60 to 85 in various East and Southeast Asian countries . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 57 2.10 Source of fi nancial support can affect poverty rates in various East and Southeast Asian countries . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 59 2.11 The relative roles of public and private transfers and of savings and assets across the life cycle varies notably across East and Southeast Asian economies and between East and Southeast Asia and other regions . . . . . . . . . . . . . . 61 2.12 Attitudes on the preferred source of fi nancial support in old age and on the expected source of personal care are shifting in East and Southeast Asian economies . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 62 2.13 Labor force participation rates vary by location and gender in various East and Southeast Asian countries . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 63 2.14 Better-educated people tend to withdraw from the labor force earlier in East and Southeast Asia . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 64 2.15 Self-employment is an increasingly important source of work as people age in East Asia and Pacific . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 65 x CONTENTS 2.16 Those who continue to be employed in older age work long hours until advanced ages in East and Southeast Asia . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 66 2.17 Pension availability and mandatory retirement rules have strong effects on labor force behavior of older Chinese workers . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 67 2.18 Labor force participation is influenced by pension receipt in select East and Southeast Asian countries . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 68 2.19 Older people in East and Southeast Asia are a significant source of volunteer work . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 73 3.1 East Asia and Pacific’s developing economies are getting old before getting rich . . . . . 82 3.2 Aging’s influence on GDP per capita is the product of the employment rate, the working-age share of the population, and GDP per employed worker . . . . . . . . . . 84 3.3 Older women in East Asia and Pacific will be able to work later in life by virtue of rising health-adjusted longevity . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 86 3.4 Prevalence of stunting among children under age five could undermine investments in education and affect cognitive skills in East Asian and Pacific economies of the green group . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 89 3.5 Productivity depends on the age distribution of the population, but mitigation channels exist . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 91 3.6 Although in many countries, entrepreneurship seems to decline with worker age, business ownership in East Asian countries is significant among older workers . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 92 3.7 Compared with household savings rates in other regions, the rate in East Asia and Pacific is relatively high and stable . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 93 3.8 Projected increases in pension spending to 2030 point to a significant increase in China and Malaysia, with more modest increases for Indonesia, the Philippines, and Thailand . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 96 3.9 Estimated increases in East and Southeast Asian pension spending for 2014 to 2050 fi nd significant fiscal risks from pensions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 97 3.10 Pension spending projections by stylized groupings for APEC economies, 2010–70, illustrate rapid increase in pension spending in a no-reform scenario . . . . . 97 3.11 Pension spending increases can be substantially reduced with key reforms . . . . . . . . . 98 3.12 Projected increases in public spending on health caused by aging, though significant, are not as substantial as for pensions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 98 3.13 Projected increases in public health spending reflect the importance of policy choices in sustainability of health care systems . . . . . . . . . . . . . . . . . . . . . . . . . 99 3.14 Projected increases in public long-term care spending indicate additional fiscal pressure in countries where the aging effect is strong . . . . . . . . . . . . . . . . . . . . . 99 3.15 Schematic overview of the links between demographics and fiscal accounts . . . . . . . 101 3.16 Initial debt and spending levels and future debt dynamics have important interactions with demographics that can sharply increase the fiscal impacts of aging . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 102 4.1 The Kurumin mark is given to fi rms in Japan that pursue child-friendly policies . . . 112 4.2 The relationship between women’s work and fertility has reversed sharply in recent decades in most member countries of the OECD . . . . . . . . . . . . . . . . . . . . 113 4.3 Japanese workers show a desire to work in old age well above OECD counterparts . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 115 B4.2.1 In OECD countries, higher employment rates of older workers are associated with modestly higher rates of youth employment . . . . . . . . . . . . . . . . . . . . . . . . . . . . 116 4.4 Tax force to retire is strongly correlated with labor force participation of older men in OECD countries . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 119 CONTENTS xi 4.5 The potential channels for mitigating labor force decline vary across aging countries in East and Southeast Asia . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 128 4.6 Greater participation by women, older people, and migrants can mitigate the projected decline in the total labor force in richer East and Southeast Asian economies. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 129 4.7 Migration leads to a 5 percent decline in the total labor force size of the migrants’ origin countries . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 129 4.8 Japan and the Republic of Korea show low rates of on-the-job training for middle-aged and older workers relative to countries such as the United States . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 130 5.1 East Asian and Pacific economies can be divided into globally average adopters, late adopters, and global outliers . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 143 5.2 East Asian and Pacific economies currently spend less on pensions than other economies after taking into account their demographics . . . . . . . . . . . . . . . . . 148 5.3 Coverage of contributory pension schemes is around par globally for income levels and demography . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 150 5.4 Elderly pension coverage is low in the bottom income deciles in China, Indonesia, and the Republic of Korea . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 151 5.5 Increasing contributory pension coverage among the poorer segments of populations is difficult . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 152 5.6 Elderly pension coverage is higher for men than for women . . . . . . . . . . . . . . . . . . . 153 5.7 A wide range of design choices in the pension benefit is available . . . . . . . . . . . . . . . 153 5.8 In most economies, target replacement rates are significantly higher for lower-income workers . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 154 5.9 In most economies, actual replacement rates in DB schemes tend to be lower than would be expected . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 155 5.10 Real returns on DC accounts and growth of real wages . . . . . . . . . . . . . . . . . . . . . . 155 5.11 Returns on funds in DC accounts in China have barely exceeded inflation and have lagged far behind wage growth . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 156 5.12 Social pension benefits are modest in East Asia and Pacific . . . . . . . . . . . . . . . . . . . . 156 5.13 Older economies have higher pension contribution rates than do younger economies . . . 157 5.14 Retirement ages in national pension schemes are relatively low in East Asia and Pacific compared with life expectancy at age of retirement . . . . . . . . . . . . . . . . . 159 5.15 Ability to work at older ages is not closely linked to labor force participation rates . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 159 5.16 Pension spending is closely linked to labor force participation rates . . . . . . . . . . . . . 160 5.17 Only higher-income economies have pension assets that are large relative to GDP. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 163 5.18 Pension agencies in East Asia and Pacific generally have conservative investment practices . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 164 5.19 Except in Japan, contribution rates required to pay for defi ned benefits are significantly higher than actual rates . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 169 5.20 Retirement ages for men and women are converging in OECD countries . . . . . . . . . 170 5.21 Expanding coverage has become difficult in recent years . . . . . . . . . . . . . . . . . . . . . . 171 5.22 The increase in coverage of contributory pension systems in East Asia and Pacific is unlikely to keep pace with population aging . . . . . . . . . . . . . . . . . . . . . . . . 172 5.23 In the Republic of Korea, a dramatic change in coverage took place . . . . . . . . . . . . . 173 B5.2.1 Design of the Chilean pension system . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 174 6.1 Health care in East Asia and Pacific faces major health system and service delivery challenges . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 186 xii CONTENTS 6.2 Hospital share of health expenditures is often greater than nonhospital share . . . . . 187 6.3 Tertiary hospitals experienced the fastest growth of inpatient and outpatient visits in China . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 187 6.4 Average length of stay in East Asian and Pacific acute care hospitals is higher than or comparable to OECD averages . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 188 6.5 Inpatient admission rates increase with mean age of household . . . . . . . . . . . . . . . . . 189 6.6 Health share of GDP has been rising in East Asia and Pacific with income growth . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 196 6.7 Age-specific expenditure profi le exhibits J-shape for Japan . . . . . . . . . . . . . . . . . . . 197 6.8 Spikes in death rates vary with the speed of aging across East Asia and Pacific . . . . . 197 6.9 Years lived with disability are increasing in East Asia and Pacific as life expectancy increases . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 198 6.10 Total and public health spending has increased more rapidly than GDP in much of East Asia and Pacific in the 1990s and 2000s . . . . . . . . . . . . . . . . . . . . . 199 6.11 Age-specific expenditure profi les vary for OECD countries . . . . . . . . . . . . . . . . . . . . 200 6.12 The rising tide of cancer in East Asia and Pacific . . . . . . . . . . . . . . . . . . . . . . . . . . . . 208 6.13 A health system is transformed from a pyramidal model into a primary care–driven coordinated care model . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 209 6.14 PHARMAC has positively affected drug expenditure over time in New Zealand . . . 218 B6.8.1 National Health Security Office’s decision process for new technologies involves extensive criteria in Thailand . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 219 7.1 Long-term care users in OECD countries are expected to increase with the aging population . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 229 7.2 In China, ADL and IADL disabilities may increase rather than decline as in OECD countries . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 230 7.3 Care gaps are greater for the rural elderly, the female elderly, and those living in poverty in East Asian and Pacific countries . . . . . . . . . . . . . . . . . . . . . . . . . 230 7.4 Long-term care expenditures as a share of GDP are relatively low but growing faster than health expenditures . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 231 7.5 Long-term care services include both social and health care provided in a variety of settings . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 232 B7.1.1 Average length of stay in long-term care hospitals in the Republic of Korea has declined only marginally . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 233 7.6 The health and long-term care systems in England are complex and riddled with perverse incentives . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 241 Tables 2.1 Lower elderly poverty rates when older people receive pensions . . . . . . . . . . . . . . . . . 50 2.2 Private transfers to elderly people are not significantly affected by level of transfers from public sources in East and Southeast Asia. . . . . . . . . . . . . . . . . . . . . . . 60 2.3 Poor health results in lower probability of work at older ages in East and Southeast Asia . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 71 2.4 Hours spent providing care for children and elderly among Chinese cohorts ages 45 and above . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 72 2.5 Influence of employment of a spouse on probability of own employment . . . . . . . . . . 73 3.1 The educational attainment of future generations in East Asia and Pacific will be considerably higher than for the current working-age population (secondary and above) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 87 CONTENTS xiii 3.2 The educational attainment of future generations in East Asia and Pacific will be considerably higher than for the current working-age population (tertiary) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 87 4.1 There is enormous variation in the importance of emigration in East Asian and Pacific countries . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 123 4.2 The share of immigrants ranges from very low to very high by global standards in East Asia and Pacific . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 123 4.3 Policies to maximize the benefits of migration in East Asia and Pacific for sending and receiving countries . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 126 5.1 Combined elements of pension systems in East Asia . . . . . . . . . . . . . . . . . . . . . . . . . 142 5.2 Implications of pension scheme designs . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 144 5.3 Defi ned benefit schemes in East and Southeast Asia face significant sustainability challenges . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 149 5.4 Retirement rules for East Asian contributory pension schemes . . . . . . . . . . . . . . . . . 161 5.5 Retirement rules and incentives for male workers in selected countries . . . . . . . . . . . 162 5.6 Assessment of pension system performance and level of reform needed . . . . . . . . . . . 165 5.7 Pension reforms in East Asia since 2000 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 167 6.1 Hospital patients in Vietnam are being treated at higher levels than is necessary . . . 188 6.2 Significant shares of elderly are experiencing catastrophic out-of-pocket payments for health care relative to total household spending . . . . . . . . . . . . . . . . . . 193 6.3 Significant shares of elderly are experiencing catastrophic out-of-pocket payments for health care relative to total nonfood household spending . . . . . . . . . . . 194 6.4 Effective strategies to reduce exposure to key NCD-related risks in East Asia and Pacific . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 203 6.5 Tobacco prices and taxes in East Asia and Pacific are not high enough . . . . . . . . . . . 203 6.6 Effective strategies for the management of key risks such as cardiovascular disease . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 205 6.7 Status of availability of essential package of cardiovascular drugs in selected East Asian and Pacific countries . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 206 6.8 Costs of treating diabetes in Vanuatu, 2012 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 207 6.9 Key design features of an effective coordinated care strategy. . . . . . . . . . . . . . . . . . . 209 6.10 Effective strategies for slowing functional decline among the elderly. . . . . . . . . . . . . 212 6.11 Effective strategies for getting better value for money from health systems . . . . . . . . 214 6.12 Price comparisons of cholesterol-lowering drugs in selected countries . . . . . . . . . . . . 216 6.13 Effective strategies outside of the health system to prepare for functional and cognitive decline in old age . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 220 B7.1.1 Comparison of LTCHs and LTCFs in the Republic of Korea . . . . . . . . . . . . . . . . . . . 233 B7.1.2 Distribution of resource utilization groups in the Republic of Korea and Ontario by type of institution . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 234 7.1 Features of long-term care insurance in Japan, the Republic of Korea, and Taiwan, China. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 237 Foreword A s a region, East Asia and Pacific is and its citizens, which will influence policy aging rapidly. The region is home to responses. over a third of the global population This report finds that East Asia and Pacific ages 65 and older—mostly in China—and is well positioned to manage the risks from to more old people than any other region. aging. First, people in the region already have More significantly, the region is aging more long working lives. Second, entitlements to rapidly than any region in history—a trend pensions, health care, and long-term care driven both by sharp declines in fertility and for most of the population are still modest, by steady increases in life expectancy—and and there is scope to act now to put in place in many countries, aging is occurring at systems that can be sustainable in the future. relatively low-income levels. Nevertheless, Third, household savings in the region are the pattern and pace of aging across East already high, and people tend to save until Asia and Pacific is diverse, ranging from later in life, suggesting that they may be aged richer countries such as Japan and the better prepared for old age. Finally, people in Republic of Korea; to rapidly aging middle- East Asia and Pacific have in recent decades income countries such as China, Thailand, seen a steady increase in the number of years and Vietnam; to younger and poorer coun- lived in good health. tries such as the Lao People’s Democratic The key message of the report is that Republic, which will only start to age rapidly it will be possible to manage rapid aging two or three decades from now. in East Asia and Pacific while sustaining Across these countries, aging raises economic dynamism. This effort will require policy challenges and economic and social politically difficult policy choices, including risks. Policy makers express concerns about dealing with associated fi scal risks. Several the impact that shrinking and aging labor complementary policy reforms are required forces will have on economic growth. More to manage these risks. In the labor market, immediately, there are fiscal challenges to a range of measures are needed (a) to encour- providing sustainable health, pension, and age increased female labor force participa- long-term care systems in the face of rapid tion while stemming the sharp decline in aging. Aging is occurring in a wider context fertility, (b) to ensure that older workers in of shifting relationships between the state the formal sector do not retire too early as xv xvi FOREWORD a result of social security and wage-setting that are already experiencing or are about systems, (c) to accommodate older workers in to experience rapid aging. In the meantime, the workplace through greater use of flexible countries with younger populations will need work arrangements, and (d) to open up aging to prepare now for rapid aging in the future labor markets to greater inflows of young by setting up sustainable health and pension immigrants. systems while maximizing the demographic Pension systems also need reform. Formal dividend from their still-expanding labor sector pension schemes will need to be forces. more fiscally sustainable, which will require As we have seen in the richer economies reforms such as gradual increases in the in East Asia and Pacific and elsewhere in retirement age. Such changes are needed to the world, the political economy of reform create fiscal space for expanding pension in aging societies can be difficult. Societal coverage to the large and currently uncovered expectations change rapidly, especially as informal sector—an effort that will require incomes grow, and people expect more sup- greater use of noncontributory approaches port and better services from the state in such as social pensions. old age. Managing aging is not only about In the health sector, the impacts of aging “old people” or about “the old” versus will likely be significant, particularly given “the young.” Rather, it requires a life-cycle the increased incidence of noncommunicable perspective on policy design that recognizes, diseases. This trend underscores the urgency for example, that the health of children to shift from a hospital-centric model to one affects their health as adults, that saving in which primary care plays a bigger role for old age needs to start early, and that and in which the treatment of older patients the burden of taxation to provide services with chronic conditions is managed afford- and benefits needs to be spread fairly across ably at the right levels of the system. As in generations. the case of pensions, the financing model for This report argues that impending demo- health care will need to rely more on budget graphic changes are important and will financing to achieve universal health cov- become even more so in the countries of erage as the needs of an aging population East Asia and Pacific in coming decades. But grow. Aging is also creating a rapidly grow- demography is not destiny. Through the right ing demand and market for long-term care. mix of policies, governments can help soci- Innovation is urgently needed in this area as eties adapt to rapid aging and improve the traditional family networks become increas- well-being of citizens. ingly stretched. I hope this report contributes to a consid- Given the diverse stages of demographic ered debate of the policy choices needed to transition across the region, lessons can be achieve healthy and productive aging across drawn from the experience of richer East the region. Asian and Pacific economies. Their expe- riences, as well as those of high-income Axel van Trotsenburg countries in other regions, highlight that Regional Vice President this policy agenda will be challenging for East Asia and Pacific Region developing East Asian and Pacific countries The World Bank Acknowledgments T his report has been prepared as part Production and editing of the report were of the World Bank East Asia and coordinated by the World Bank’s Publish- Pacific Regional Report series by a ing and Knowledge Department under the team consisting of Philip O’Keefe (team task direction of Susan Graham, Mark Ingebret- leader), Nithin Umapathi, Aparnaa Somana- sen, and Patricia Katayama. than, Robert Palacios, John Giles, Thomas The team would like to acknowledge a Flochel, Ralph Van Doorn, Juan Pradelli, number of people who have provided inputs Yang Huang, Harry Moroz, Shonali Sen, and guidance during the course of report Malathi Velamuri, Naoko Miake, Çaglar preparation. Professor Albert Park and par- Özden, Mauro Testaverde, Minna Han ticipants in a joint workshop held in 2014 Tong, Yuqing Hu, Risa Nakayama, Andrew by the World Bank and Hong Kong Uni- Beath, and Yuki Ikeda. It has been prepared versity of Science and Technology in Hong under the initial oversight and manage- Kong SAR, China, provided useful feedback. ment of Bert Hofman (then regional chief Professor John Campbell provided invalu- economist, East Asia and Pacific region) able insights on long-term care and Japan’s and Xiaoqing Yu (then sector director for health system. Professors Hidehiko Ichimura Human Development, East Asia and Pacific and Yasuyuki Sawada also gave important region) and more recently under the guid- input. Professor David Canning shared his ance of Sudhir Shetty (regional chief econo- insights on aging in East Asia and Pacific. mist, East Asia and Pacific region) and Jehan Richard Jackson shared recent work on atti- Arulpragasam (practice manager for Social tudes toward old-age support in Asia. The Protection and Labor, East Asia and Pacific team would also like to acknowledge assis- region). The peer reviewers for the report tance provided by Yuki Murakami of the were Professor Andrew Mason, University Organisation for Economic Co-operation of Hawaii at Manoa and East-West Center; and Development and by Eduardo Klien and Professor John Piggott, University of New his colleagues at the HelpAge International South Wales, director of the ARC Centre Regional Office for Asia. The team also ben- of Excellence in Population Aging Research efited from insights from regional staff at the (CEPAR); and Emily Sinnott, senior econo- United Nations Population Fund, the United mist, Poverty Global Practice, World Bank. Nations Economic and Social Commission xvii xviii ACKNOWLEDGMENTS for Asia and the Pacific, and the International in Chiangmai, Thailand, gave insights on Labour Organization. Professor John Knodel work from the report that were most help- participated in upstream discussions of the ful. The team also acknowledges the assis- study in Bangkok, Thailand. Yvonne Sin of tance provided by John Beard of W HO Towers Watson provided useful input on the Geneva and by Rafal Chomik of CEPAR. pensions chapter, and Professor Atsushi Seike Many World Bank colleagues provided gave generously of his time in Japan and valuable assistance, including Axel van Trot- beyond, as did Professor Soonman Kwon, senburg, Arup Banerji, Yasusuke Tsuka- Professor Rachel Lu, Professor Winnie Yip, goshi, Rocio Castro, Omar Arias, Makoto Kate Mandeville, and Katie Leach-Kemon. Nakagawa (now returned to the government The team’s official counterparts from of Japan), Dewen Wang, Elena Glinskaya, C h i na , T ha i la nd , a nd Viet na m gave Mitchell Wiener, Mark Dorfman, Toomas useful guidance to the team as it framed Palu, Ajay Tandon, Gerard La Forgia, Eko the research agenda for the study, and Pambudi, Sutayut Osornprasop, Johannes offi cials from the Republic of Korea pro- Koettl, Brooks Evans, Michele Gragnolati, vided insights on Korea’s labor market Maurizio Bussolo, Hans Timmer, Victoria and social security reforms. Offi cials from Kwakwa, Joyce Msuya, Carl Hanlon, Alek- the Ministry of Finance and the Ministry sandra Posarac, Nga Nguyet Nguyen, Reena of Health, Labor and Social Welfare of Badiani-Magnusson, Lars Sondergaard, Japan facilitated a study visit of the team Nikola Spatafora, and participants in a joint in late 2013. Participants in World Health workshop on aging and the macroeconomy Organization (WHO) conferences in Kobe held by the East Asia and Pacific and Europe and Hanoi and in a HelpAge conference and Central Asia regions of the World Bank. Abbreviations AAI Active Aging Index ADL activity of daily living APEC Asia-Pacific Economic Cooperation ASEAN Association of Southeast Asian Nations BMI body mass index CES-D 10 Center for Epidemiologic Studies Depression Scale CHARLS China Health and Retirement Longitudinal Study CPF Central Provident Fund (Singapore) CVD cardiovascular disease DALY disability-adjusted life year DB defi ned benefit DC defi ned contribution ELSA English Longitudinal Study of Ageing EPF Employees Provident Fund (Malaysia) EU13 Bulgaria, Croatia, Cyprus, the Czech Republic, Estonia, Hungary, Latvia, Lithuania, Malta, Poland, Romania, the Slovak Republic, and Slovenia FFS fee-for-service (payment method) GDP gross domestic product GNI gross national income GPF Government Pension Fund (Thailand) GSIS Government Service Insurance System (the Philippines) HALE health-adjusted life expectancy HR health and retirement HTA health technology assessment IADL instrumental activity of daily living IFLS Indonesia Family Life Survey IMF International Monetary Fund LASI Longitudinal Aging Study in India LCH life-cycle savings hypothesis LDL low-density lipoprotein LFPR labor force participation rate xix xx ABBRE VIATIONS LTC long-term care LTCF long-term care facility LTCH long-term care hospital MPFA Mandatory Provident Fund Authority (Hong Kong SAR, China) NCD noncommunicable disease NDC notional defi ned contribution NHS National Health Service (United Kingdom) NHSO National Health Security Office (Thailand) NPS National Pension Service (Republic of Korea) NSSF National Social Security Fund (China) NTA National Transfer Accounts OECD Organisation for Economic Co-operation and Development OOP out-of-pocket (payment) PHARMAC Pharmaceutical Management Agency (New Zealand) PHC primary health care PISA Programme for International Student Assessment POEA Philippines Overseas Employment Administration PROST Pension Reform Options Simulation Toolkit SAGE Study on Global Ageing and Adult Health (World Health Organization) SAR Special Administrative Region SHARE Survey of Health, Ageing, and Retirement in Europe SJSN Sistem Jaminan Sosial Nasional, or National Social Security System (Indonesia) SPICE Singapore Programme for Integrated Care for the Elderly SSO Social Security Office (Thailand) SSS Social Security System (the Philippines) STEP Skills Toward Employability and Productivity TDR total dependency ratio TFR total fertility rate TVET technical and vocational education and training UN United Nations WHS World Health Survey Overview Introduction working lives, high savings rates at all ages, and social security systems that thus far have The population in East Asia and Pacific largely avoided unsustainable legacy com- is aging rapidly, raising concerns about mitments. But, the region is at a crossroads. the implications for the region’s economic Managing rapid societal aging while sustain- future. Although the region is aging rapidly, ing economic dynamism calls for politically proactive public policy can play an impor- challenging policy choices that require strong tant role in helping it adapt to changing leadership and social consensus. demographics and mitigating the downside The effect of and policy responses to aging risks of aging. in East Asia and Pacific have seven overarch- This report aims to crystallize what is ing themes: known about the rapidly changing and diverse demographics of East Asian and 1. The very rapid demographic and epide- Pacific countries. It documents policy frame- miologic al transition heightens the works on aging and explores the implications urgency of policy responses in many of alternative reform options. Its purpose is developing countries in the region. The not only to provide a comprehensive review speed of aging is accompanied by other of aging in the region and associated policy rapid socioeconomic transitions that responses, but also to encourage policy debate have significant implications for the by facilitating comparison of policy regimes effects of aging on household and soci- across the region. etal welfare and the capacity of countries Aging raises many challenges and risks, to manage the demographic transition. which East Asian and Pacific countries are They include rapid urbanization and well positioned to manage—provided they migration, with accompanying mobility make policy choices that promote appropriate and dispersion of extended families; rap- behavioral change by households and employ- idly changing social values, which are ers. Initial conditions are poised to promote shifting expectations of support between healthy and productive aging. The region’s generations; and rapidly rising incomes, countries already have populations with long which create not only new possibilities 1 2 LIVE LONG AND PROSPER but also new expectations from citi- 5. Although commonalities exist, initial zens. The speed of all of these changes conditions vary enormously across the matters. region, and emerging policy responses 2. Several middle-income countries in the already show considerable diversity. region are growing old before growing Some countries are already very old; oth- rich. Unlike most Organisation for ers have decades of potential demo- Economic Co-operation and Development graphic dividend to realize. However, (OECD) member countries, which aged whether aging is a current, emerging, or at a gradual pace as they grew wealthy, more distant reality, nearly all societies in several East Asian and Pacifi c countries the region need to frame public policy are experiencing—or will experience in with the realization that rapid aging will coming decades—accelerated aging at come at some point. relatively low levels of per capita income. 6. Responding to aging is not just about old This pattern has major implications for people. Effective responses require policy the welfare of older people and the possi- and behavioral change across the life bilities for public policy and institutions cycle of individuals. Actions to effect to help manage the aging process. these changes must consider fertility and 3. Aging is occurring in a wider context of child care policies; education and skills shifting relationships between state and systems; tax, social security, and transfer citizens in many East Asian and Pacifi c systems; labor market policies across the countries. In much of the region, the life cycle; and policies directed at older state has traditionally provided more people themselves. At the most basic limited direct support than in other level, evidence increasingly points toward parts of the world. Traditional values the importance of early childhood health have placed greater emphasis on the sup- and nutrition as predictors of adult and port role of families and other networks. old age health. Although such values have strong resid- 7. Aging is multidimensional and requires ual influence, citizens are increasingly new modes of policy coordination. The expecting more from the state. In rapidly fact that aging cuts across all dimensions aging countries in the region, citizens of people’s lives and public policy, whereas view aging as a major challenge: about public institutions tend to operate in silos, 80 percent of the people in the Republic creates new challenges to craf ting of Korea and two-thirds of the people in responses that are more than the sum of China regard aging as a “major prob- their parts. The expansion of national lem” for their country (Pew Global aging policies and cross-cutting institu- Attitudes Survey 2013). tions in East Asia and Pacific is an 4. Demographics are largely a given (at acknowledgment of the challenge. least for the foreseeable future), but pol- icy and behavioral responses are within The overview is organized as follows. the control of governments and citizens. The first section reviews the demographic However, the behavioral changes required and epidemiological transitions in East Asia of people, employers, and societies do not and Pacific. It is followed by a discussion of happen overnight. In a number of cases, the living standards and other indicators of they go to the heart of sociocultural welfare of older people in the region, such beliefs about gender and the obligations as sources of support and living arrange- of generations toward one another. Policy ments. The third section discusses the effects changes to address aging and realize the of aging on growth, the labor force, savings, potential of older people are challenging and government budgets, and the ways in and often unpopular. which countries can manage them. The fourth OVERVIEW 3 section looks at challenges facing pension, FIGURE O.1 East Asia and Pacific has more older people than any health, and long-term care systems. This sec- other region tion is followed by a short discussion of cross- Number of people ages 65 and above by region and East Asian and Pacific cutting challenges. The final section provides subregion some concluding thoughts. OECD countries Understanding the China rapid demographic and epidemiological transition South Asia Although much of East Asia and Pacific has ECA emerging still to experience its most rapid period of countries aging, it already has the largest regional pop- LAC developing ulation of people ages 65 and older, primar- countries ily in China. In 2010, East Asian and Pacific EAP high-income economies countries had about 185 million people ages EAP developing 65 and older, about 36 percent of the global countries population of this age group (figure O.1). China alone had almost 114 million old peo- Sub-Saharan Africa ple, more than any developing region. MENA developing More significantly, East Asia and Pacific is countries aging more rapidly than any region in history. 40 0 0 0 0 20 60 80 10 12 14 The phenomenon is driven primarily by a rapid decline in fertility but also by increased Number of people age 65 and older, 2010 (millions) longevity. Nearly all middle-income countries Source: World Bank estimates based on data from UN 2013. in the region are in the midst of or will soon Note: EAP = East Asia and Pacific; ECA = Eastern Europe and Central Asia; LAC = Latin America experience a pace of aging that is unprec- and the Caribbean; MENA = Middle East and North Africa; OECD = Organisation for Economic Co-operation and Development. EAP developing countries exclude China. edented, transitioning from young to old societies in 20 to 25 years—a transition that took 50 to more than 100 years in OECD countries (figure O.2). Although the starting Declining fertility has combined with steady points vary, nearly all East Asian and Pacific increases in both life expectancy and health countries move from aging to aged societies life expectancy, in contrast to regions such as relatively quickly when the aging threshold is Eastern Europe and Central Asia. reached (exceptions are the Philippines and The demographic diversity of East Papua New Guinea). Asian and Pacific economies suggests three The main driver of aging in East Asia and groupings. The first group (red)—comprising Pacific has been the sharp decline in fertility, Hong Kong SAR, China; Japan; the Republic which fell from 5.91 children per woman in of Korea; and Singapore—includes the 1960 to 2.46 in 2005. This decline was signifi- wealthiest economies. People ages 65 and cantly faster than the global decline (from 5.51 older represented 14 percent of the total children per woman to 3.03). Fertility rates in population of this group in 2010. The second several East Asian economies are now among group (orange) includes China, Indonesia, the lowest in the world, with richer economies Malaysia, Mongolia, Thailand, and Vietnam. in the region averaging only 1.28 children These countries are aging very quickly. The per woman in 2010. In major middle-income third group (green) comprises Cambodia, the countries, such as China, Thailand, and Lao People’s Democratic Republic, Myanmar, Vietnam, the fertility rate is well below replace- Papua New Guinea, the Philippines, ment fertility and continues to fall (UN 2013). Timor-Leste, and the Pacific Island countries 4 LIVE LONG AND PROSPER FIGURE O.2 East Asian and Pacific economies are aging more rapidly than economies elsewhere Years to move from 7 to 14 percent population share 65 years and older and the start and end years of transition Vietnam (15) Lao PDR (20) Malaysia (20) Indonesia (20) Brazil (20) Thailand (20) Korea, Rep. (20) Timor-Leste (25) Mongolia (25) Myanmar (25) Cambodia (25) Turkey (25) China (25) Singapore (25) Japan (25) Hong Kong SAR, China (30) Philippines (35) Papua New Guinea (40) South Africa (40) United Kingdom (45) Russian Federation (50) United States (69) France (115) 1860 1885 1910 1935 1960 1985 2010 2035 2060 2085 Sources: World Bank estimates based on data from UN 2013 and Kinsella and He 2009. Note: Figure shows starting and ending year for transition from 7 percent (aging) to 14 percent (aged) of population ages 65 and older. Aging and aged thresholds are based on United Nations definitions. East Asia and Pacific economies rounded to five-year increments. (not shown in figure O.3). These countries are ratios. The record of demographic projec- still young—just 4 percent of the population tions in the region is mixed, suggesting that was age 65 or older in 2010—but they will projections should be treated with caution. begin to age quickly in coming decades. Fertility decline, for example, has tended Demographics drive divergent trends in to consistently outpace projections in the working-age populations. Several East Asian region; assumptions about future reversals and Pacific countries are projected to expe- in fertility for richer East Asian and Pacific rience sharp declines in their labor forces economies risk extrapolating from unstable (figure O.4). Between 2010 and 2040, the short-run dynamics. share of the working-age population will The basic challenges are further compli- fall by more than 15 percent in Korea and cated by the speed of urbanization, internal by at least 10 percent in China, Japan, and and foreign migration, and other factors Thailand. In contrast, in green and some that will inevitably affect demographic orange countries, the working-age share of trajectories. Moreover, even in cases where the population is not expected to shrink until aggregate projections are reliable, rethinking after 2040. In absolute terms, the Philippines traditional measures such as elderly depen- and Indonesia will account for the largest dency ratios will be needed. The static defini- share of the regional increase. tion of the working-age population as people Although the directions of the demo- ages 15 to 64, for example, is increasingly graphic transition in East Asia and Pacific are being questioned. Alternative dependency clear, considerable uncertainty remains about measures that have been proposed (such as the speed of the transition and the way to years of remaining life or ratios based on age think about key measures such as dependency profiles of actual labor force participation) OVERVIEW 5 FIGURE O.3 Three distinct aging patterns are apparent in East Asia and Pacific Percentage of population age 65 or older in East Asian and Pacific economies 40 36% 35 Share of population age 65 and older (%) 30 24% 25 20 14% 13% 15 10 6% 5 4% 0 50 60 70 80 90 00 10 20 30 40 50 60 70 80 90 00 19 19 19 19 19 20 20 20 20 20 20 20 20 20 20 21 High-income economies, aged population Middle- and lower-middle-income countries, aging population Low-income countries, young population Source: World Bank estimates based on data from UN 2013. Note: The high-income “red” economies, which have an aged population, include Hong Kong SAR, China; Japan; the Republic of Korea; and Singapore. The middle- and lower-middle-income “orange” countries with an aging population include China, Indonesia, Malaysia, Mongolia, Thailand, and Vietnam. The low-income “green” countries with a young population include Cambodia, Lao PDR, Myanmar, Papua New Guinea, the Philippines, and Timor-Leste. may be more relevant for policy purposes as bulk of disability-adjusted life years among people live longer and young people spend people ages 60 and over in 2010, with more time acquiring education. comorbidities complicating approaches to Together with the demographic transition, treatments and care. Underlying these trends East Asia and Pacific is experiencing a rapid is growing exposure to risk factors, to which epidemiological transition toward noncom- the poor are most vulnerable. Between 1990 municable diseases (NCDs), in part driven and 2010, premature death and disability by population aging. Healthy life expectancy associated with NCD-related risk factors in the region increased between 1990 and (such as high salt intake, blood pressure, 2010, albeit by slightly less than total life total cholesterol, and blood sugar) increased expectancy. At the same time, the burden everywhere in the world. The increase was of disease related to NCDs has grown rap- particularly pronounced in East Asia and idly, with initial onset increasingly occurring Pacific. As a result of earlier onset of lifestyle- in middle age. NCDs caused 76 percent of related conditions such as hypertension and deaths in the region in 2008, and the share obesity, in many countries in the region, the is projected to rise to 85 percent by 2030 current cohort of middle-aged people is more (IHME 2013). Cardiovascular disease, can- vulnerable to poor health in old age than cer, and diabetes already accounted for the were earlier cohorts (IHME 2013). 6 LIVE LONG AND PROSPER FIGURE O.4 In the richer economies of East Asia and Pacific, the Co-residence of older people with adult working-age population is projected to shrink between 2010 and children is high in East Asia and Pacific, 2040 although it varies across countries and has Percentage change in population ages 15–64 between 2010 and 2040, declined significantly over time in some coun- selected economies tries. It generally increases between middle age and the mid-60s. For people ages 60 Hong Kong SAR, China and older, co-residence rates range from 25 Korea, Rep. to 30 percent to more than 80 percent, with Singapore significant variation by gender and by urban Thailand and rural location. Rates are particularly high Japan in low-income countries, and, consistent with China global patterns, lower at higher-income levels within countries. Co-residence with children Vietnam has been declining rapidly in China, Korea, Mongolia and Thailand. In China, the co-residence rate Malaysia for people ages 65 to 70 fell from almost Myanmar 66 percent in the early 1980s to roughly Indonesia 43 percent by 2011. In Korea, co-residence of Cambodia people ages 65 and older fell from more than Philippines 80 percent in 1980 to well below 30 percent Papua New Guinea by 2010 (Giles and Huang 2015). Older people in East Asia and Pacific often Lao PDR work until very old age. However, significant Timor-Leste differences exist between urban and rural areas –20 –15 –10 –5 0 5 10 15 and between men and women (figure O.5). Percentage change On average, people work until late in life, but many people in urban areas stop working rel- Source: World Bank estimates based on data from UN 2013. atively early, and some groups (such as urban Chinese female workers) retire very early (only about one-third are still working at age 60). Situation and sources of support Withdrawal of urban people from work is of older people clearly correlated with access to a formal sec- The situation of older people in East Asia and tor pension, though caregiving responsibilities Pacific is diverse across and within countries. for grandchildren and other elderly also play a Across developing countries, older people role in the early withdrawal of urban women shared in the significant decline in poverty, from formal work. Nearly everywhere, urban but poverty in a number of the region’s coun- female workers are the least likely to work tries tends to rise with age after age 45, before into old age, and gender gaps in participation flattening or even declining at the oldest ages. are substantial. Among people who continue As with the general population, older people working into old age, self-employment is the with higher levels of education (and people dominant form of work, and self-employment with children with higher levels of educa- rates of 90 percent and higher are common tion), access to a formal sector pension, and in rural areas of the region (Giles, Hu, and better health status tend to be better off than Huang 2015). their peers. In contrast to the situation of the Older people in East and Southeast Asia younger adult population, however, employ- tend to continue working long hours, particu- ment status has a more complex relation- larly in rural areas. In the economies analyzed ship with poverty among the elderly, in part for this report, men who continue to work past because many poorer, often rural, older peo- age 65 work, on average, 40 hours or more a ple have to “work till they drop.” week, and women work 30–45 hours a week. OVERVIEW 7 FIGURE O.5 Labor force participation rates in East Asia and Pacific vary, but across the region, rural people work significantly longer than urban people Percentage of population participating in the labor force, by age, gender, and urban or rural location, selected countries a. Cambodia, 2011 b. China, 2011 100 100 90 90 80 80 % of cohort working % of cohort working 70 70 60 60 50 50 40 40 30 30 20 20 10 10 0 0 40 50 60 70 80 40 50 60 70 80 Age Age c. Indonesia, 2007 d. Republic of Korea, 2010 100 100 90 90 80 80 % of cohort working % of cohort working 70 70 60 60 50 50 40 40 30 30 20 20 10 10 0 0 40 50 60 70 80 40 50 60 70 80 Age Age e. Philippines, 2009 f. Thailand, 2011 100 100 90 90 80 80 % of cohort working % of cohort working 70 70 60 60 50 50 40 40 30 30 20 20 10 10 0 0 40 50 60 70 80 40 50 60 70 80 Age Age Urban, men Urban, women Rural, men Rural, women Sources: Giles, Hu, and Huang 2015, based on data from CHARLS 2011; IFLS 2007; KLoSA 2010; World Bank East Asia and Pacific Standardized Household Surveys, various years; and ThaiSES 2011. 8 LIVE LONG AND PROSPER FIGURE O.6 Work remains the primary source of old-age support in most countries in East Asia and Pacific Percentages of total income of people ages 60 and above, in urban and rural areas, from labor, public and private transfers, and other sources, selected countries a. Urban 100 90 80 % of total income 70 60 50 40 30 20 10 0 p. a sia a te lia m s d in di ne Re es na an go ne Ch bo pi -L a, et ail on do ilip m or re Vi Th M Ca In Ko Tim Ph b. Rural 100 90 80 % of total income 70 60 50 40 30 20 10 0 p. a sia a lia te m s d in di ne Re es na an go ne Ch bo pi -L a, et ail on do ilip m or re Vi Th M Ca In Ko Tim Ph Labor Public transfer Private transfer Other income Sources: Giles and Huang 2015, based on data from CHARLS 2011; IFLS 2007; KLoSA 2010; ThaiSES 2011; and World Bank East Asia and Pacific Standardized Household Surveys, various years. In rural areas, men who continue to work do support of their families, especially in rural so for 30–40 hours a week even at age 75, and areas. Labor income plays a critical role in old- women that age work 20–35 hours a week. In age support in the region, with work income addition to what is formally counted as work, dominating other sources of support in rural older people in East Asia and Pacific often areas in nearly all countries and in urban areas bear significant caregiving responsibilities, a in most countries (figure O.6). In most coun- social and economic role that is not formally tries, private transfers to elderly people are also recognized in national accounts. more significant than public transfers. Familial The roles of the state, families, and indi- support is even more pronounced when one viduals in old-age support in developing East factors in nonfinancial support. Asian and Pacific countries are distinctive. In Diversity is also growing across and within developing regions such as Latin America and countries, rooted in initial conditions such the Caribbean and Eastern Europe and Central as labor market formality, urban and rural Asia, the state plays an important role through shares, and formal social security coverage. generous and widespread pension systems. In Public transfers play an important role in contrast, older people in East Asia and Pacific urban China and in Mongolia, for example, rely more heavily on their own labor and the but only a negligible role in Cambodia and OVERVIEW 9 FIGURE O.7 Attitudes on the preferred source of financial support in old age and on the expected source of personal care are shifting in East and Southeast Asian economies Percentage of adults reporting government, selves, or family as likely sources of support, selected economies 70 60 % of respondents 50 40 30 20 10 0 p. m na ia a sia re s d ne in Re ys an na po hi ne Ch pi ala ail et ,C a, ga do ilip re Vi Th M an Sin In Ko Ph iw Ta Retirees themselves Government Grown children or other family members Source: Jackson and Peter 2015. Indonesia, where informality is high and pen- Thailand provides a case in point. sion and social assistance systems are under- Although it has enjoyed solid growth in developed. Within China, differences between recent decades, the pace of aging is outstrip- rural and urban areas are sharp. Although ping per capita income growth. It is following public transfers play a relatively modest role a path closer to that of Poland, which is now in old-age support in most of the region’s struggling with fiscal pressures from aging. countries, social expectations of the relative Mongolia is experiencing stagnant GDP per roles of the state, families, and retirees them- capita and rapid aging. selves in postretirement support are changing The demographic dividend has played a rapidly, with opinion surveys showing expec- substantial role in the region’s remarkable tations of future state (rather than familial) growth story in recent decades. However, support rising sharply (figure O.7). its younger countries will need to deepen reforms if they are to enjoy the same growth effects as older countries did at similar Managing the effect of aging on stages of the demographic transition. They growth and its drivers can adopt productivity-enhancing reforms Developing countries in East Asia and Pacific that will result in more manageable elderly are getting old before getting rich. In par- dependency and GDP trajectories. Around ticular, orange (largely aging middle-income) one-third of the high per capita income countries have been aging—in some cases growth in East Asia from 1960 to the 1990s rapidly—at low levels of gross domestic prod- can be attributed to the demographic divi- uct (GDP) per capita (figure O.8). This trend dend, with some estimates of the effect as boosts the pressure to increase productivity high as 44 percent of growth in per capita growth in the region’s developing countries. income (Bloom, Canning, and Malaney Even with sustained productivity growth, 2000; Bloom, Canning, and Sevilla 2003; however, orange countries will not reach the Bloom and Williamson 1998; Kelley and income levels of red (richer and older) East Schmidt 2005; this literature generally Asian and Pacific economies or OECD and excludes the Pacific). In most East Asian middle-income countries such as Mexico at countries, improvements in educational similar points in the demographic transition. attainment, capital investment, and other 10 LIVE LONG AND PROSPER FIGURE O.8 East Asia and Pacific’s developing economies are getting old before getting rich Elderly dependency ratio by GDP per capita (2005 PPP), various economies, 1980–2010 60,000 50,000 GDP per capita (US$2005 PPP) 40,000 30,000 20,000 10,000 0 0.05 0.10 0.15 0.20 0.25 0.30 0.35 0.40 Elderly dependency ratio Singapore Hong Kong SAR, China United States United Kingdom France Russian Federation Korea, Rep. Malaysia Cambodia Timor-Leste Thailand Vietnam Lao PDR Philippines China Poland Japan South Africa Mongolia Indonesia Mexico Brazil Sources: GDP data in 2005 US$PPP for 1980 to 2010 from the World Bank’s World Development Indicators database; elderly dependency ratios calculated as ratio of population over age 65 to population ages 15–64, using data from UN 2013. Note: PPP = purchasing power parity. factors compounded the pure demographic to growth are real, even rapidly aging coun- effect. For younger East Asian and Pacific tries have significant scope to act. Managing countries currently in the midst of their the growth and other macroeconomic effects youth bulges, for which rapid aging will of aging will require public policy action and come only in future decades, the key is to wider behavioral change in a range of areas make productivity-enhancing reforms and that span the life cycle. Priorities differ across investments now, so that they both maximize country groupings: the demographic dividend and are prepared for rapid aging when it comes. • For red economies, slowing the structural The tailwind to growth provided by demo- decline in the size of the labor force will be graphics in older East Asian and Pacific coun- the most critical challenge, because most tries has raised fears of a headwind as the size economies are closer to the productivity of the labor force declines and aging accel- frontier and therefore lack easy options erates, but policy and behavioral responses for increasing labor productivity sig- can mitigate these forces. Although the risks nificantly. A sharper focus on increasing OVERVIEW 11 female labor force participation, extend- in a variety of ways. Various labor mar- ing productive working lives, and, in ket institutions and policies, as well as some countries, increasing immigration of social attitudes, constrain productive labor younger workers from within the region force participation of groups that need will be vital. to be mobilized to mitigate the structural • Orange countries will need to sustain high decline in the workforce. The experience of productivity growth and undertake struc- red economies points to a distinctive East tural reforms of social security, health and Asian approach to increasing labor force long-term care, and labor market policies. participation that combines incentives and Even if they make such changes, however, mandates for employers. Channels for these countries will become aged societies mitigating the effects of aging on the labor at much lower levels of income than the force include the following: red economies. They will need to mitigate the labor supply and fiscal effects of rapid • Extending productive working lives aging through ongoing reforms of pension through labor market and social secu- and health systems and labor policies to rity policies. Countries could gradually extend the working lives of their urban raise retirement ages and reverse incen- and formal sector workers as they undergo tives for early retirement; incentivize rapid urbanization. There are reasons for workplace adjustments to accommodate cautious optimism, because more edu- older workers (through, for example, flex- cated cohorts will be better prepared for ible and part-time work arrangements the prospect of longer working lives than and midcareer skill upgrading); and pro- previous generations. mote attitudinal change among employ- • Green countries will enjoy favorable ers. They could also reform labor market demographics as the youth bulge con- institutions such as seniority wage sys- tinues to feed their labor forces for the tems, which create barriers to productive next few decades. The priorities are to employment of older workers. establish conditions to realize maximum • Raising female labor force participa- GDP growth from the demographic tion, with an emphasis on making child dividend (through investments that care available and providing subsidies raise productivity and maximize youth to make it affordable. The experience of employment) and to avoid adopting poli- richer East Asian and Pacific economies cies in areas such as pensions and health that adopted measures to promote female that are affordable now but would rap- labor force participation demonstrates the idly become unsustainable when aging challenges of shifting strong gender norms accelerates. These countries’ success in in the region. doing so will determine their readiness to • Increasing immigration, which will manage aging when it accelerates over the require proactive efforts in both sending coming two to three decades. and receiving countries. The different tim- ing of aging across East Asia and Pacific Aging has potentially deleterious effects on offers opportunities for “chronological the labor force, capital formation and savings, arbitrage.” Singapore and Hong Kong and the budget, each discussed in turn below. SAR, China, offer examples of the eco- Policies are needed to mitigate these effects. nomic benefits of more open immigration policies. Other countries are exploring ways to overcome traditionally limited Mitigating the effects on the labor force reliance on immigration. Although this East Asian and Pacifi c countries can help area is politically challenging, the benefits reduce constraints to labor force par- for sending and receiving countries are ticipation and thus mitigate the demo- clear in the context of rapid but asynchro- graphic effect of aging on labor force size nous aging across the region. 12 LIVE LONG AND PROSPER FIGURE O.9 Greater participation by women, older people, and More flexible employment arrangements, migrants can mitigate the projected decline in the labor force in support for parents through subsidized child richer East and Southeast Asian economies care services, and broader shifts in gender and Projected changes in labor force size of destination economies with family relations are more likely to affect work changes in female, elderly, and migrant labor force participation and fertility decisions. Efforts to mitigate the shrinking size of the 130 labor force will be supplemented by improve- ments in labor force quality as a result of expanded access to higher levels of educa- 120 tion in recent decades and ongoing efforts to improve educational quality. Capital deepen- Millions of people ing, which would increase per worker labor 110 productivity, is also likely in response to labor force decline. 100 The effects on savings and capital formation 90 Debates on aging often focus on poten- 2010 2015 2020 2025 2030 2035 2040 2045 2050 tial declines in savings rates. The effects on Baseline household savings depend on the relative Female increased participation strength of two offsetting effects: (a) the compositional effect from a higher share of Elderly increased participation older people in the population and (b) the Permanent migration behavioral effect as people save more to Temporary migration finance a longer expected period of retire- ment. Predictions on which of these effects Source: Özden and Testaverde 2014. will dominate in East Asia and Pacific are Note: Destination economies are Hong Kong SAR, China; Japan; the Republic of Korea; Malaysia; and Singapore. Sending economies are Cambodia, Indonesia, Lao PDR, Mongolia, Myanmar, Papua New mixed, but on balance they suggest that Guinea, the Philippines, Thailand, Timor-Leste, and Vietnam. Green line shows increase in labor concerns about the effects of aging on sav- force relative to baseline when women participate at same level as men. Red line shows increase given increased labor force participation by the elderly. Orange line shows increase given migration ings and capital formation in the region of 10 percent of labor force ages 25–35 (migrants stay permanently). Light blue line shows increase may be overstated, for three reasons. First, given temporary increase in migration of 20 percent of labor force ages 25–35 (migrants stay 10 years). Dark blue line shows the baseline. household and corporate savings rates are high in the region and provide a more solid foundation than they do in other regions. Second, survey evidence points to flatter sav- The relative importance of these chan- ings and age profiles in the region than in nels varies across countries. In the richest other parts of the world and to an increase East Asian and Pacific economies, the effect in savings rates at all ages in the region in is likely to be greatest from higher female recent decades. Third, significant inefficien- labor force participation and, in some cases cies in financial markets in the region’s devel- (for example, Japan and Korea), migration. In oping countries suggest that scope exists for some middle-income countries, the potential more efficient mediation of savings into capi- role of increased participation by the elderly tal formation and increased productivity. labor force is also significant (figure O.9). In contrast to these measures, policies to The effects on fiscal balances increase the birth rate through direct finan- cial incentives such as birth grants or favor- Even if East Asia and Pacific avoids severe able tax treatment do not appear to have adverse growth effects from rapid aging, been effective in stemming fertility decline. the fiscal risks of business as usual in core OVERVIEW 13 programs such as pensions, health, and long- FIGURE O.10 Across APEC, pensions will eat up a larger and larger term care are substantial. Major fiscal risks share of GDP in coming decades without reforms from aging are already manifesting them- Projected pension spending as a percentage of GDP in APEC economies selves in the region and will require policy (no-reform scenario), 2009–75 leadership to mitigate. Projections of pension and health spend- 18 16 ing reveal significant fiscal pressure in 14 coming decades in the absence of reform. 12 Stylized projections by broad country type in % of GDP 10 Asia-Pacific Economic Cooperation (APEC) 8 member countries through 2070 show 6 economies converging to ratios of pension 4 spending to GDP that are 8 to 12 percent- 2 age points higher than current levels (figure 0 O.10). Cash flow deficits are expected to 2009 2019 2029 2039 2049 2059 2069 2079 emerge in most defined benefit schemes in High-income economies, mature pension system developing East Asian and Pacific countries, Middle-income economies, maturing pension system with deficits of 1.4 to 4.5 percent of GDP High-income economies, maturing pension system by 2040. Health care spending projections also point to higher public spending (though Low-income economies, developing pension system lower than increases in pension spending). Source: Hinz and Zviniene 2011. Aging accounts for about one-third of the Note: APEC = Asia-Pacific Economic Cooperation. projected increase in public health spending in several developing countries in the region, broadening pension coverage, and explor- and the effect is even higher in China (Hinz ing the role of the state in long-term care. and Zviniene 2011). A growing commitment to shifting some A more complex (and uncertain) picture of the burden of old-age financial sup- of fiscal pressures from aging emerges when port and health care costs from families debt dynamics are incorporated into fiscal and individuals to the state through pool- projections. Where increased age-related ing mechanisms can be seen. The gradual spending drives up deficits and they are shift of financing from private to public funded through debt, the potential for an sources exerts significant pressure on pub- upward spiral in public spending is substan- lic spending. tial. The risk is higher where significant shares • The financing models for pensions and of debt are denominated in foreign currency. health are increasingly mixed, illustrating The importance of factoring in demographics the limitations of contributory financing in scenarios of public debt should therefore models in settings with high labor mar- not be underestimated. ket informality. More nuanced blends of contributory and general revenue financ- ing are emerging that also aim to priori- Developing pension, health, and tize inclusion of the poor and near-poor. long-term care systems for aging For health care coverage in particular, societies the emphasis has been almost entirely on expanding coverage through general rev- The key areas of pensions, health, and long- enue financing. term care face several common reform • Pension and health systems face the chal- challenges: lenges of balancing coverage, adequacy, • Most orange and some green countries and sustainability. To date, the emerging are expanding policy-driven coverage— approach in developing East Asian and achieving universal health coverage, Pacific countries has been to prioritize 14 LIVE LONG AND PROSPER coverage expansion, albeit often with The good news is that considerable inno- shallow financial protection. vation is already evident within East Asia and • Pension systems in many East Asian and Pacific. In response to the persistent failure of Pacific countries are fragmented, with contributory pension schemes in closing the separate schemes for the civil service, the coverage gap, Chile (2008), China (2011), military, and state-owned enterprises; Korea (2004), Mexico (2012), and Thailand often fragmented pooling of contribu- (the 2000s) all rapidly expanded their social tions, which leads to variable benefits and pensions and budget-financed matching con- undermines the insurance function; and tribution schemes for informal sector work- distinct rural and urban systems in China. ers (see Holzmann, Robalino, and Takayama Health financing is characterized by simi- 2009 for regional and global experience with lar challenges. social pensions; Hinz et al. 2013 for experi- ence with matching defined contributions). Countries in the region have also expanded Pensions health insurance coverage, through heavily subsidized health insurance premiums. Use of Pension systems in East Asia and Pacific are consumption taxes in Japan and budget sup- quite heterogeneous. The pension challenges port for the contributory pension scheme in facing countries in the region vary not only China are driven partly by reluctance to raise because of their disparate demographic payroll tax rates further. The trend is not uni- situations, but also because of past policy form, however, as evidenced by the introduc- choices and the nature of their labor and tion of a social insurance scheme in Lao PDR capital markets. Richer (red) economies have and legislation in Indonesia and Myanmar achieved wide coverage of their systems with that make these countries among the last in better financial sustainability than other rich the world to rely on contributory approaches economies, but they face challenges of pen- to expanding pension provision. The expe- sion adequacy and hence old-age financial rience of dozens of countries that had once protection. Middle-income (largely orange) faced the same choice suggests that it may not countries are grappling with the dual chal- be the appropriate path. lenges of improving the sustainability of Reform should be based on the premise their existing or legacy systems while try- that the redistributive element of the pen- ing to expand coverage to large uncovered sion system should be financed from general informal sectors. For poorer and younger revenues and clearly distinguished from the (green) countries, the key issues are choosing insurance or savings component, which ide- a pension system model that can achieve sig- ally should be financed by individuals on nificant coverage over time. These countries an actuarially fair basis (that is, identified need to avoid overreliance on contributory sources of financing are sufficient to pay pension models, which have proven ineffec- the expected expenditures). The simplest tive in extending coverage to informal sector approach would be to eschew the mandated workers. scheme altogether and instead broaden the Addressing the coverage gap, particularly coverage of a social pension (for example, in countries with high levels of informality, by making it universal or means tested) and will require policies that move away from sole complement it with a voluntary retirement reliance on payroll taxes to finance pensions. savings scheme. In New Zealand, which This paradigm shift in the way pension sys- adopted this approach, the opt-in default tems have historically been financed is moti- for a voluntary defined contribution scheme vated by the limitations of a traditional social led to relatively high participation rates. insurance approach, which risks losing the This option is not likely to be realistic in race between pension coverage expansion and countries that already have contributory rapid societal aging. mandates, however, and it would be difficult OVERVIEW 15 even for countries such as Myanmar and in healthy life expectancy in assessing aging’s Timor-Leste, which have nascent contribu- effects over time is important. tory systems. These countries, along with The structural reform challenges fac- Indonesia, should aim for actuarially fair ing health delivery systems in the region and modest schemes, leaving redistribution will become more acute as populations age to be financed through general revenues or because the increased incidence of chronic through social assistance or social pensions. NCDs with age means both that the share The wider group of East Asian and Pacific of people living with these conditions will countries with mature, but low-coverage man- increase and that larger shares of people will dated contributory systems should also con- be living with multiple chronic NCDs. Health sider shifting the balance of pension system delivery systems in East Asia and Pacific are financing toward general revenue–financed ill prepared for the NCD epidemic, especially redistribution and modest target benefit lev- in countries with rising shares of older peo- els, but in a way that does not compromise ple, for several reasons. First, these systems incentives to participate in the contributory tend to provide low-value, high-cost services scheme. Doing so requires strong coordina- that are likely hospital-centric; preventive tion between social and contributory pension and primary care services tend to be inef- design and implementation. The most feasible fective and of poor quality. Too often, bet- path for such countries is likely to mix resid- ter health is incorrectly attributed to more ual mandatory contributory systems with consultations, admissions, drugs, and pro- subsidized voluntary contributory approaches cedures. Second, too little attention is paid and social pensions targeted to wider or nar- to the prevention, early diagnosis, treatment, rower populations, depending on the level and control of health conditions. When a of coverage achieved through contributory diagnosis is made, care is rarely coordinated mechanisms. Countries such as China, Korea, across provider levels, resulting in service and Thailand provide distinctive models of duplication and lack of continuity. In addi- this approach. Although the adequacy of tion, health care is often sought too late, pensions for the majority is limited currently, leading to high-cost treatment in expensive sustained growth should allow for deepening acute care hospitals. Third, the lack of effec- financial protection over time. tive referrals, gatekeeping, and postdischarge care contributes to costly (and avoidable) readmissions. Fourth, incentives embedded Health in provider payment systems encourage phy- In the health sector, the effects of aging are sicians to overprovide services or provide likely to be significant. In East Asia and Pacific unnecessary care. and globally, age-specific profiles of expen- Health system shortcomings and wel- ditures are typically a J-shape, with health fare consequences are more severe for older spending rising steeply toward the end of people, who are much more likely to receive life. Consequently, a common assumption is inpatient care (often unnecessarily), use that aging will generate substantial increases more pharmaceuticals than average (and in health care spending. In fact, the effects hence are more affected by poor pharmaceu- of aging on health expenditure are com- tical purchasing and prescription practices), plex. Globally, technology and policy choices and experience substantially higher out-of- regarding the financing and delivery of health pocket spending and higher rates of cata- care, rather than aging itself, are the biggest strophic health spending. In China in 2011, drivers of health spending. Aging does add to about 32 percent of households with a mean health costs (accounting for about one-third age of 60 or more incurred out-of-pocket of spending increases in several East Asian health payments of more than 25 percent of and Pacific countries) (Somanathan 2015). total nonfood spending; in Vietnam in 2012, However, taking into account improvements the figure was 30 percent of households 16 LIVE LONG AND PROSPER (Somanathan 2015). Older people have Health systems will need to spend more higher comorbidity rates, which require a efficiently if the benefits of system transforma- mixed set of services and more complex tion are to be realized and afforded. Fee-for- management of movement between levels service provider payments will need reform of the health system. They also experience to reduce overprovision of services and over- functional and cognitive decline, which puts charging for care. Global experience with new demands on health systems. For all of payment reforms points toward case-based or these reasons, countries with aging popula- diagnosis-related group approaches for sec- tions need to create health systems in which ondary and tertiary care, mixed-method pay- prevention and treatment are offered in a ments for primary care, and global budgets to coordinated manner over a sustained period contain costs. Having a large single purchaser of time while enabling individuals to assume of health services with strategic purchasing greater responsibility for managing their capacity has also proven critical for cost con- own care. tainment (Somanathan 2015). To reorient health systems to meet the rein- Many countries in the region will need to forcing challenges of universal health care, continue reforms of pharmaceutical procure- NCDs, and aging, countries will need to pur- ment to rein in the high prices paid by govern- sue several reforms in parallel that together ments and patients. These countries can learn will take at least a generation to achieve. from experiences both within the region and An overarching need is to transform health globally in purchasing and other practices. delivery systems by strengthening primary Practices to control health technology care services, shifting care away from acute diffusion, such as health technology assess- care hospitals, reducing overprovision, and ments, are also crucial to control health improving coordination among providers. spending growth. East Asian and Pacific The quality of the health workforce also countries could begin by instituting trans- needs to be substantially improved. These parent and evidence-based processes for reforms will help contain costs in the long prioritizing new technologies and drugs, term, but they will require additional spend- using the findings from more established ing in the medium term. health technology assessment agencies to Because mobilizing new resources to meet inform the prioritization processes. This all of the additional requirements of service challenging and interrelated set of reforms delivery reforms would simply not be sus- is necessary even without aging. Accelerated tainable, generating efficiency savings to aging makes reform even more essential and gain better value for money from existing urgent. health systems is imperative in East Asia and Pacific. Although the reformed service deliv- Long-term care ery model will generate significant efficiency savings over time, the initial investment Formal long-term care (LTC) systems in costs are likely to be substantial. Meanwhile, developing East Asian and Pacific coun- other policy priorities—not least of which is tries remain nascent. However, a growing the policy commitment to achieve universal number of aging countries are grappling health coverage—will continue to place signif- with the appropriate role of the state in an icant demands on East Asian and Pacific gov- area that has traditionally been the domain ernments’ budgetary allocations to the health of families, communities, and the health sector. Health system inefficiencies are already system. exerting pressure on tax and social health Rapid aging and social change have insurance revenues. Service delivery reforms exposed the limitations of traditional infor- are thus critical to ensure the sustainability of mal modes of LTC for frail elderly people health systems for aging populations and jus- in the region. Part of the response has been tify additional investment. default reliance on health systems, but this OVERVIEW 17 approach is costly and complicates health better placed than some others institutionally. reform efforts. Proactive policy choices in the Its consolidated health, labor, and welfare LTC domain are therefore important. They ministry brings a critical mass of age-related require careful planning with respect to their policies and programs under a single body. interaction with informal care systems and formal health and welfare systems. Whatever The political economy of reform and the design choice, countries need to craft behavioral change explicit (rather than costly default) LTC strat- egies, with a strong emphasis on home- and Many of the policy reforms needed to pro- community-based care. “Aging in place” is mote healthy and productive aging are politi- important because it not only provides posi- cally difficult and may challenge societal tive experiences for older people, but also is norms. Extending working lives by raising more sustainable than other options as the the retirement age has proven politically elderly population grows. challenging in many parts of the world. Increasing female labor force participation involves deeply rooted cultural norms and, Addressing cross-cutting in some cases, trade-offs (grandparents who challenges in aging work late into life will not be available to In addition to sector-specific challenges, a provide the informal child care and aged care range of cross-cutting considerations under- on which working-age households currently pins policy choices and societal responses. rely, for example). Such considerations will require more inte- The performance on immigration from grated ways of approaching not only public younger to older countries in the region is policy and service delivery systems but also mixed. Health system reform is notoriously a broader societal need to forge consensus difficult, for both providers and the wider within and across generations. population, and changing ingrained lifestyle habits that drive the NCD epidemic is a gen- erational challenge. All these measures will The multisectoral nature of aging require patient building of social consensus. One of the key challenges in shaping public Intergenerational equity and political econ- policy responses to aging is its inherently mul- omy is also a wider challenge. In other parts tisectoral nature. Like areas such as early child- of the world, older people have successfully hood, aging presents inherent challenges to the lobbied for their interests, at times at the cost institutional and policy setups of governments of younger generations. Red economies in because it cuts across many sectors and stages East Asia and Pacific seem to have balanced of life. Governments are typically not orga- the interests of different generations and nized to respond multisectorally. Although a shared the costs and benefits of an aging pop- number of East Asian and Pacific countries ulation more equitably than have countries have recognized this challenge and established in other parts of the world. The evolution of national aging commissions to promote more the political economy dynamics around aging holistic thinking on aging, these agencies tend will be an important consideration in shaping to be primarily advocacy bodies that may lack policy responses. the political and bureaucratic clout to drive whole-of-government approaches. The roles of public policy and market Some richer countries have sought to bring forces a wider view to sectoral policies. One exam- ple is the periodic intergenerational reports Another broad issue is the appropriate roles of produced for the Australian Treasury, which public policy and market forces in shaping the help set long-term expenditure and policy response to rapid aging. Proactive public pol- priorities for age-related programs. Japan is icy will be required, but it will not necessarily 18 LIVE LONG AND PROSPER be the correct response to every challenge. A hands of policy makers, communities, and careful balance of mandates, incentives, and citizens. Demography is a powerful force reliance on market forces will be required. in development, but it is not destiny. Policy The experience of red economies in East makers in the region have the potential to Asia and Pacific in extending working lives shape responses that increase the chances for illustrates how supportive public policies are healthy and productive aging and promote gaining more traction as the market forces of societies in which the compact between gen- supply and demand have increased the appeal erations is one that is fair and realizes people’s of older workers. The most effective public potential at all ages. policies will sometimes involve a retreat from In this process, East Asian and Pacific state direction where policies (such as rigid economies can learn from the experiences— seniority wages or unnecessary restrictions good and bad—of richer and older coun- on women’s work) hinder the play of market tries, as well as from one another. Crafting forces that would be expected to increase the appropriate policies will inevitably involve employment of older people and women. experimentation and course correction. It will also require strong leadership in the face of inevitable reluctance to embrace change. Knowledge gaps on aging Fortunately, the region’s strong economic and Much remains unknown about aging, in East cultural traditions place it in a favorable posi- Asia and Pacific and globally. Understanding tion to create environments in which citizens long-term trends and policy effects is often will live long and prosper. speculative. Significant knowledge gaps remain—on the interaction of labor and savings market behavior and rapid social References security reforms; the societal and individ- Bloom, David E., David Canning, and Pia N. ual behavioral change in the face of rapid Malaney. 2000. “Population Dynamics and growth, urbanization, and changing family Economic Growth in Asia.” Population and structures; the way to reverse ultralow fertil- Development Review 26 (suppl.): 257–90. ity rates in richer East Asian and Pacific econ- Bloom, David E., David Canning, and Jaypee omies; the productivity of older workers and Sevilla. 2003. The Demographic Dividend: the cost-effectiveness of policies to extend A New Perspective on the Economic working lives; the cross-country dimensions Consequences of Population Change. Santa Monica, CA: RAND. of aging in the region with respect to migra- Bloom, David E., and Jeffrey G. Williamson. 1998. tion, capital flows, and other factors; and “Demographic Transitions and Economic more broadly, the roles of market forces and Miracles in Emerging Asia.” World Bank public policy in promoting healthy and pro- Economic Review 12 (3): 419–55. ductive aging. CHARLS (China Health and Retirement Even an understanding of the most funda- Longitudinal Study). 2011. National School of mental question of the trajectory of fertility Development, Peking University. http://charls and aging has been imperfect. The speed of .ccer.edu.cn/en. aging in the region and its diverse cultural Giles, John, Yuqing Hu, and Yang Huang. 2015. and political economy contexts suggest that “Understanding the Retirement Decision in caution is therefore needed in predicting Aging East Asia.” Background paper for the East Asia and Pacific regional report on aging, future trends. World Bank, Washington, DC. Giles, John, and Yang Huang. 2015. “Are the Conclusion Elderly Left Behind in a Time of Rapid Demographic and Economic Change? Aging will fundamentally reshape societies A Comparative Study of the Poverty and economies in East Asia and Pacific, but and Well-Being of East Asia’s Elderly.” the shape of the future is very much in the Background paper for the East Asia and OVERVIEW 19 Pacific regional report on aging, World Bank, Kelley, Allen C., and Richard M. Schmidt. 2005. Washington, DC. “Evolution of Recent Economic-Demographic Heston, Alan, Robert Summers, and Bettina Aten. Modeling: A Synthesis.” Journal of Population 2012. Penn World Table Version 7.1. Center Economics 18 (2): 275–300. for International Comparisons of Production, Kinsella, Kevin, and Wan He. 2009. An Aging Income, and Prices, University of Pennsylvania, World: 2008. U.S. Census Bureau, International Philadelphia. Population Report P95/09-1. Washington, DC: Hinz, Richard, Robert Holzmann, David Tuesta, U.S. Government Printing Office. and Noriyuki Takayama, eds. 2013. Matching KLoSA (Korean Longitudinal Study of Contributions for Pensions: A Review of Ageing). 2010. Korea Labor Institute, Seoul. International Experience. Washington, DC: http://www.kli.re.kr/klosa/en/about/introduce World Bank. .jsp. Hinz, Richard, and Asta Zviniene. 2011. Özden, Çaglar, and Mauro Testaverde. 2014. “The Potential Fiscal Effects of Population “International Migration in Aging Societies: Aging: Simulation of Aggregate Pension Impacts and Benefits in East Asia and Pacific and Health Expenditure for Four Stylized Countries.” Background paper for the East Countries at Various Levels of Economic Asia and Pacific regional report on aging, Development.” Social Protection and Labor World Bank, Washington, DC. Practice, World Bank, Washington, DC. Pew Global Attitudes Survey. 2013. Pew Holzmann, Robert, David A. Robalino, and R e s e a r c h C e n t e r, Wa s h i n g t o n , D C . Noriyuki Takayama, eds. 2009. Closing the http://www.pewglobal.org/search/global Coverage Gap: The Role of Social Pensions +attitudes+survey+2013/. and Other Retirement Income Transfers . Somanathan, Aparnaa. 2015. “Health and Health Washington, DC: World Bank. Care in an Ageing East Asia and Pacific.” IFLS (Indonesia Family Life Survey). 2007. RAND, World Bank, Washington, DC. Santa Monica, CA. http://www.rand.org/labor ThaiSES (Thai Household Social and Economic /FLS/IFLS.html. Surveys). 2011. National Statistical Office, IHME (Institute for Health Metrics and Bangkok. http://web.nso.go.th/survey/house Evaluation). 2013. The Global Burden of _seco/socio.htm. Disease: Generating Evidence, Guiding Policy. UN (United Nations). 2013. “World Population Seattle, WA: IHME. Prospects: The 2012 Revision, Highlights Jackson, Richard, and Tobias Peter. 2015. “From and Advance Tables.” Working Paper ESA/P/ Challenge to Opportunity: Wave 2 of the WP228, Department of Economic and Social East Asia Retirement Survey.” Global Aging Affairs, Population Division, United Nations, Institute, Alexandria, VA. New York. The Demographic Transition and Well-Being of Older People in East Asia and Pacific I The Demographic and Epidemiological Transitions 1 in East Asia and Pacific Introduction In parallel, East Asia and Pacific is experi- encing an accelerated epidemiological transi- In recent decades, East Asia and Pacific tion, driven in part by population aging but has undergone perhaps the most dramatic also by factors such as rapid urbanization demographic transition ever seen, and it is and rising incomes. On the positive side, both on course to continue that trend. Very fast life expectancy and healthy life expectancy1 drops in fertility rates, coupled with rapid have continued to increase steadily, and extensions of life expectancy beyond age communicable diseases have been brought 60, have resulted in an unprecedented pace largely under control in most economies in and scale of population aging in East Asia. the region, except for some low-income coun- At the same time, the region is character- tries and Pacific Island countries that have a ized by remarkable variation, ranging from double burden of disease. However, the rise of the “super-aging” societies of Japan and noncommunicable diseases (NCDs) has been the Republic of Korea, to middle-income dramatic, and they now dominate the disease countries with rapidly growing aging pop- burden throughout the region. Older people ulations, to the youthful societies of the are more prone to NCDs and experience Philippines and Pacific Island countries. This much higher comorbidities, which will place diversity provides rich lessons—both posi- new demands on health systems. Another tive and cautionary—from within the region problem is that in many East Asian and on managing rapid aging. Whether such Pacific countries today, middle-aged popula- unprecedented trends are alarming in large tions exhibit worse lifestyles and NCD preva- part reflects our understanding of depen- lence than the current elderly did at the same dency among growing elderly populations, ages. This situation raises concerns about because current definitions rather mechani- future deterioration of health in the elderly cally conflate chronological age with func- in the absence of major lifestyle and health tional limitation. system reforms. This chapter is based on background papers by Thomas Flochel, Yuki Ikeda, Harry Moroz, and Nithin Umapathi (2014) and by Aparnaa Somanathan (2015), with inputs from Ying Ho. 23 24 LIVE LONG AND PROSPER To provide context for the remainder of modest shares of elderly similar to those found the report, this chapter discusses both the in the Middle East and North Africa and in demographic and the epidemiological tran- South Asia. Overall, in East Asia and Pacific, sitions taking place in East Asia and Pacific. the share of the population ages 65 and above The next section describes the demographic accounted for an average of around 7 percent transition. It highlights aging trends across the of the total population in 2010 (figure 1.1). region and compares the transition to those in This share is similar to that in Latin America other regions of the world. It also examines and the Caribbean (6 percent) and somewhat the factors underlying such aging trends and higher than the shares in the Middle East and discusses trends in working-age population North Africa and in South Asia (5 percent). and dependency ratios. The remainder of the However, the regional average for East Asia chapter focuses on the epidemiological tran- and Pacific conceals enormous diversity: sition, providing more in-depth discussion high-income countries have shares of older of trends in life expectancy and the growing people that are close to the OECD average burden of NCDs. (around 14 percent, which means that they are aged societies according to the United Nations definition), whereas many young and The demographic transition in poorer countries have very modest shares of East Asia and Pacific older people (for example, only 3 percent of East Asia and Pacific is a diverse region in the populations of Papua New Guinea and terms of the share of older people across coun- Timor-Leste are 65 years old and above—the tries, ranging from levels found in member same as the average for Sub-Saharan Africa). countries of the Organisation for Economic Despite the mixed picture on shares of Co-operation and Development (OECD) to older people, in terms of absolute numbers East Asia and Pacific has easily the largest regional population of individuals who are 65 years old and above. China alone is home FIGURE 1.1 East Asia and Pacific is a diverse region in terms of the share of older people to more old people than any other developing Share of population ages 65 and above, 2010, selected regions and East region. In 2010, East Asian and Pacific econo- Asian and Pacific subregions mies were home to around 187 million people ages 65 and above (around 150 million of 18 them in developing countries). This number Share of population ages 65 and older (%) 16 represents around 36 percent of the global population in this age group. China is respon- 14 sible for the bulk of this elderly population, 12 with nearly 114 million people ages 65 and 10 older (figure 1.2). 8 East Asia and Pacific is also aging more 6 quickly than any other region in history. 4 The population ages 65 and over in many 2 countries in the region will increase from 0 7 percent to 14 percent of the total popula- tion in just two or three decades—a change CD rie e em a ia r g tri ng P h tr ing es a tri ng OE unt om in LA unt pin ric As co s de ies es ig ies tri un pi So es that took 45 years in the United Kingdom, Ch un pi EA un g Af co -inc co er un co lo co velo EA uth co lo e an e ev h v 69 years in the United States, and 115 years ar de Pd A h EC C A Sa EN in France (Kinsella and He 2009; see also b- M Su figure 1.3).2 Five-year growth rates of the Source: World Bank estimates based on data from UN 2013. population ages 65 and above have been Note: EAP = East Asia and Pacific; ECA = Eastern Europe and Central Asia; LAC = Latin America and the Caribbean; MENA = Middle East and North Africa; OECD = Organisation for Economic higher in East Asia and Pacific than in Co- operation and Development. EAP developing excludes China. OECD and emerging Eastern European and T H E D E M O G R A P H I C A N D E P I D E M I O LO G I C A L T R A N S I T I O N S I N E A S T A S I A A N D PAC I F I C 25 Central Asian countries in each 20-year FIGURE 1.2 East Asia, led by China, has more old people than any period from 1955 to 2055. Between 2015 other region and 2034, five-year growth rates of the Number of people ages 65 and above by region and East Asian and Pacific population ages 65 and above will aver- subregion age 22 percent in East Asia, second only to rates in the Middle East and North Africa. OECD countries As a result, by 2060, the average share of population ages 65 and over in East Asia China (22 percent) will approach that of the OECD countries (24 percent) and Eastern South Asia Europe and Central Asia (26 percent), with ECA emerging the share in high-income East Asian and countries Pacific economies and in China being even LAC developing higher (figure 1.4). Whereas in 2010 only 1 countries of the 25 “oldest” economies (by share of EAP high-income economies population ages 65 and older) was in East EAP developing Asia (Japan), by 2060 five East Asian econ- countries omies are projected to be among the 25.3 Although most of the region will age Sub-Saharan Africa quickly, the onset of aging varies, and some MENA developing countries will experience especially rapid countries aging. Figure 1.3 shows the different points 40 0 0 0 0 20 60 80 10 12 14 in time and speeds at which East Asian and Pacific economies have moved or will move Number of people age 65 and older, 2010 (millions) from being classified as aging to aged (that Source: World Bank estimates based on data from UN 2013. is, moving from 7 percent of the population Note: EAP = East Asia and Pacific; ECA = Eastern Europe and Central Asia; LAC = Latin America ages 65 and older to 14 percent). At one and the Caribbean; MENA = Middle East and North Africa; OECD = Organisation for Economic Co-operation and Development. EAP developing excludes China. extreme, Japan completed its transition to an aged society decades ago. Others, includ- ing China, Korea, and Thailand, are already aging, and the poorer societies in the region the population across all world regions, have decades to go before they cross that including in East Asia and Pacific. However, threshold (although Cambodia and Myanmar the region’s average population of those will do so much sooner). In terms of speed of ages 80 years and above will increase by aging, the typical East Asian and Pacific econ- 6.2 percentage points on average between omy shown in figure 1.3 moves from aging to 2010 and 2060, more than in any other aged in around 25 years, but countries such region. For high-income East Asian and as Indonesia, Korea, and Vietnam are notable Pacific economies, both the rate and ulti- for their even more accelerated demographic mate level are even more pronounced, rising transitions. Equally notable is that nearly all by 14 percentage points to 17 percent of East Asian and Pacific countries move from the population. This group is important for aging to aged societies relatively quickly future trends in demand for long-term care, once the 7 percent threshold is reached (with with 50 and 64 percent of long-term care exceptions such as Papua New Guinea and clients in Korea and Japan, respectively, the Philippines). being 80 years old and above. The population of those ages 80 and The demographic diversity of econo- above—the “oldest old”—will also grow mies in East Asia and Pacific suggests three faster in East Asia and Pacific than in any groups, which are used for much of the analy- other region (figure 1.5). In 2010, the old- sis in this report.4 Hong Kong SAR, China; est old represented a very small portion of Japan; Korea; and Singapore—the wealthiest 26 LIVE LONG AND PROSPER FIGURE 1.3 East Asian and Pacific economies are aging more quickly than other economies in the past Years to move from 7 to 14 percent population share 65 and older and the start and end years of transition Vietnam (15) Lao PDR (20) Malaysia (20) Indonesia (20) Brazil (20) Thailand (20) Korea, Rep. (20) Timor-Leste (25) Mongolia (25) Myanmar (25) Cambodia (25) Turkey (25) China (25) Singapore (25) Japan (25) Hong Kong SAR, China (30) Philippines (35) Papua New Guinea (40) South Africa (40) United Kingdom (45) Russian Federation (50) United States (69) France (115) 1860 1885 1910 1935 1960 1985 2010 2035 2060 2085 Sources: World Bank estimates based on data from UN 2013 and Kinsella and He 2009. Note: Figure shows starting and ending year for transition from 7 percent (aging) to 14 percent (aged) of population ages 65 and older. Aging and aged thresholds are based on United Nations definitions. East Asia and Pacific economies rounded to five-year increments. FIGURE 1.4 The share of elderly population in East Asia and Pacific will converge with that of OECD countries by 2060 Share of population 65 and older by region and East Asian and Pacific subregion, 1950–2060 40 Share of population ages 65 and older (%) 35 30 25 20 15 10 5 0 80 00 10 20 50 60 70 90 30 40 50 60 19 19 19 19 20 20 20 19 20 20 20 20 China EAP high-income economies EAP developing countries ECA emerging countries LAC developing countries MENA developing countries OECD countries South Asia Sub-Saharan Africa Source: World Bank estimates based on data from UN 2013. Note: EAP = East Asia and Pacific; ECA = Eastern Europe and Central Asia; LAC = Latin America and the Caribbean; MENA = Middle East and North Africa; OECD = Organisation for Economic Co-operation and Development. EAP developing excludes China. T H E D E M O G R A P H I C A N D E P I D E M I O LO G I C A L T R A N S I T I O N S I N E A S T A S I A A N D PAC I F I C 27 economies in the region—are called the “red” FIGURE 1.5 The population of “oldest old” will increase more economies. Red economies have the highest in East Asia and Pacific between 2010 and 2060 than in any other number of people of advanced age, with an region average of 14 percent of their total popula- Percentage of population 80 years old and above by region and East Asian and Pacific subregion, 2010 and 2060 tions 65 years old and above (figure 1.6). China, Indonesia, Malaysia, Mongolia, Share of population ages 80 and older (%) Thailand, and Vietnam represent a middle 18 group of economies that are currently aging 16 very quickly. An average of 6 percent of the 14 population in these “orange” economies was 12 65 years old and above in 2010, and across 10 the group, an acceleration of aging can be 8 seen from the middle of this decade. The 6 third group consists of Cambodia, the Lao 4 People’s Democratic Republic, Myanmar, 2 Papua New Guinea, the Philippines, and 0 Timor-Leste. These “green” economies are tri e ric n de Asia tri ng a ig tries CD ries g tri ng tri ng in un om still quite young—an average of 4 percent of Af ara t n es a de es un pi es es un gi un pi un pi Ch n h co nc co velo EN uth co mer co velo co velo EA cou Sa their populations was 65 years old and above i h- o b- e e S Su Pd A Ph A EC C in 2010—but they will begin to age in com- OE EA LA M ing decades. This typology also holds for the 2010 2060 population 80 years of age and above: by 2060, the oldest old population will average Source: World Bank estimates based on data from UN 2013. Note: EAP = East Asia and Pacific; ECA = Eastern Europe and Central Asia; LAC = Latin America 17 percent in red economies, 7 percent in and the Caribbean; MENA = Middle East and North Africa; OECD = Organisation for Economic orange economies, and just 3 percent in green Co-operation and Development. EAP developing excludes China. economies. The pace and progression of aging are decline from 5.51 to 3.03 children per reflected in the relative strength of concerns woman. Fertility rates in several East Asian about aging among East Asian and Pacific countries are now among the lowest in the populations. This comparison can be seen world. The decline in fertility can be seen in figure 1.7, which shows findings of the across the region but has occurred unevenly 2013 Pew Global Attitudes Project survey across countries. The older East Asian coun- on the share of people reporting aging as tries have extremely low total fertility rates a “major problem” in their countries (Pew (TFRs), averaging 1.28 children per woman Research Center 2014). Three countries (figure 1.8) in 2010. Although younger coun- in the region easily rank highest, and even tries in the region have also experienced Indonesia (with relatively young demo- significant declines in TFR, their rates are graphics) shows results similar to those of more than twice as high (3.45 children per the United States. In one sense, the results woman). The orange economies fall between are positive in terms of the societal aware- the two extremes with an average TFR of ness of the challenges of aging populations, 1.93. In this intermediate group, 2010 TFRs but in another sense they reflect genuine had fallen to 1.66 in China, 1.41 in Thailand, concerns about the readiness of countries to and 1.75 in Vietnam. Fertility rates are pro- manage aging effectively. jected to continue to decline in the youngest In most countries in East Asia and countries and to flatten in the middle ones. Pacific, the decline in fertility rates has Notably, the United Nations (UN) projects a been the major driver of population aging. modest increase in fertility rates in the oldest Fertility rates declined dramatically from countries. However, as the UN itself recog- 5.91 children per woman in 1960 to 2.46 nizes, the recovery of TFRs from extremely in 2005—significantly faster than the global low levels in the oldest countries in the region 28 LIVE LONG AND PROSPER FIGURE 1.6 Three distinct aging patterns are apparent among East Asian and Pacific economies Percentage of population 65 years old and above 40 36% 35 Share of population ages 65 and older (%) 30 24% 25 20 14% 13% 15 10 6% 5 4% 0 50 60 70 80 90 00 10 20 30 40 50 60 70 80 90 00 19 19 19 19 19 20 20 20 20 20 20 20 20 20 20 21 High-income economies, aged population Middle- and lower-middle-income countries, aging population Low-income countries, young population Source: World Bank estimates based on data from UN 2013. Note: The high-income “red” economies, which have an aged population, include Hong Kong SAR, China; Japan; the Republic of Korea; and Singapore. The middle- and lower-middle-income “orange” countries with an aging population include China, Indonesia, Malaysia, Mongolia, Thailand, and Vietnam. The low-income “green” countries with a young population include Cambodia, Lao PDR, Myanmar, Papua New Guinea, the Philippines, and Timor-Leste. is far from certain and is driven to a signifi- years in the red economies and is currently cant extent by the convergence assumption much higher than that of their peers in the used in the UN models (see box 1.1 regarding orange or green economies. Average gains population projections).5 in life expectancy in the green, or youngest, Although not as dramatic, changes in economies have been much less dramatic, so age-specific mortality are also a notable the gap in life expectancy at age 60 between part of the aging story, with sharp improve- the region’s youngest and oldest economies ments in early life conditions increasing life is projected to grow from around 3 years in expectancy in East Asia and Pacific faster 1950 to 7 years in 2010 and to 10 years by than in any other region of the world. 2060. The reverse is true of the gap in life Dramatic gains in life expectancy at birth in expectancy at birth between the green and the region, from less than 45 years in 1950 red economies: the gap is projected to nar- to 74 years today, mirror the decline in mor- row from 19 years in 1950 to 15 years in tality rates (figure 1.9). The three groups of 2010 and to 14 years by 2060. This inver- East Asian and Pacific economies exhibit sion is consistent with the youngest econo- distinct patterns of life expectancy. Life mies becoming younger because mortality expectancy of individuals ages 60 and 80 rates are lower at younger ages, whereas has increased much more quickly in recent the oldest economies become older through T H E D E M O G R A P H I C A N D E P I D E M I O LO G I C A L T R A N S I T I O N S I N E A S T A S I A A N D PAC I F I C 29 extensions in life expectancy at the oldest FIGURE 1.7 People in aging East Asian and Pacific countries see ages. Bloom, Canning, and Finlay (2010) aging as a major challenge for their countries calculate that the proportion of individuals Percentage of people by country who rate aging as a “major problem” for 0 to 5 years of age in East Asia and Pacific their country, selected countries would have been 7.0 percentage points higher in 2005 had fertility rates remained 100 at 1960 levels rather than declined. This 90 contrasts with a 0.1 percentage point 80 decline in the same population had age- 70 specific mortality rates remained at 1960 60 Percent levels.6 In the future, the low fertility and 50 mortality rates that evolved between 1960 40 and 2005 will continue to drive population 30 aging as those birth cohorts move through the age distribution, even as the future rates 20 themselves are projected to stabilize. 10 These uncertainties and assumptions 0 mean that all population figures and projec- p. a n a ce m sia y ey il es in in pa an Re az do an at rk ne nt Ch Ja Br rm ng St a, Tu Fr tions should be treated with caution. The ge do re Ge d Ki Ar In Ko ite d selection of fertility scenarios, for instance, ite Un Un affects conclusions about the evolution of population aging in the region. This Source: Pew Research Center 2014, based on the 2013 Pew Research Center Global Attitudes Project survey. report—in line with common practice—uses the medium fertility scenario throughout. FIGURE 1.8 Fertility rates have declined significantly across all Figure 1.10 shows the effect of different East Asian and Pacific economies but are projected to flatten in TFR assumptions on the population share of orange and red economies people 65 years of age and older by country Total fertility rates, 1950–2060 group. Although the division of East Asian and Pacific economies into three typologies is largely robust in the fertility scenarios, the 8 variation between scenarios within the red 7 and orange economies is nearly as large as the variation across the two groups, showing 6 the significant uncertainty about projections Total fertility rate 5 further into the future. The UN extrapolations of fertility rates 4 are particularly controversial in the East Asian context. The lack of parallelism in the 3 projected recovery of East Asian TFR, as 2 occurred in Europe, is debated for the fol- lowing reasons. First, the impact of relax- 1 ing China’s one-child policy is not known. Evidence shows that fertility norms have 0 50 60 70 80 90 00 10 20 30 40 50 60 shifted to the extent that desired fertility is 19 19 19 19 19 20 20 20 20 20 20 20 below replacement level, the one-child policy Orange economies Red economies ECA emerging countries notwithstanding (Gu and Cai 2009). The sec- OECD countries Green economies ond reason for doubt is the nature of gender relations in the East Asian context, in which Source: World Bank estimates based on data from UN 2013. traditional family values exert pressure on Note: ECA = Eastern Europe and Central Asia; OECD = Organisation for Economic Co-operation and women for childbearing and child rearing and Development. 30 LIVE LONG AND PROSPER BOX 1.1 Uncertainty and assumptions in United Nations population data The population data used in this chapter, which are tion of fertility rates, which range from low to high drawn from the 2012 revision of the United Nations fertility. It makes additional assumptions in a sixth (UN) World Population Prospects, involve important scenario that projects population with constant mor- uncertainties and assumptions that affect long-term tality rates, and in a seventh scenario that projects projections (UN 2013). First, historical fertility and population with no migration. An eighth scenario life expectancy rates are estimated with significant projects population holding fertility and mortality error rates, introducing imprecision into current constant. Although the UN invests significant effort population figures. This uncertainty means that in providing the best projections possible, the com- even past fertility, mortality, and population data plexity of the endeavor results in projections that should be used with caution. Especially for devel- can be considerably inaccurate. This can be seen for oping countries, the UN draws from a variety of selected East Asian and Pacific countries in figure sources to develop its estimates. Second, UN projec- B1.1.1, which compares 1978 projections with 2000 tions require nontrivial assumptions about fertility, outcomes, highlighting both the significant average mortality, and international migration. In response, margins of error and the variability of the errors the UN provides five different scenarios for evolu- across the age distribution. FIGURE B1.1.1 UN population projections for East Asia and Pacific have been subject to significant uncertainty Comparison of 1978 projections with 2000 outcomes 100 80 60 40 20 Percent error 0 –20 –40 –60 –80 –100 + 4 9 4 9 4 9 4 9 4 9 4 9 4 9 0– 5– –1 –1 –2 –2 –3 –3 –4 –4 –5 –5 –6 –6 70 10 15 20 25 30 35 40 45 50 55 60 65 Age group Indonesia Malaysia Philippines Singapore Thailand Vietnam Source: UN 2013. T H E D E M O G R A P H I C A N D E P I D E M I O LO G I C A L T R A N S I T I O N S I N E A S T A S I A A N D PAC I F I C 31 often care of the elderly (Jones 2011). The FIGURE 1.9 Declines in age-specific mortality rates in East Asia third reason is the UN’s practice of project- and Pacific in recent decades have been substantial, particularly in ing fertility, which is strongly influenced by the early years of life and for women the small increases in total fertility observed Percentage of decline in age-specific mortality rates in East Asia and Pacific by gender, 1970–2010 in recent years. Some research has suggested that the recent increases were driven by 90 important but short-lived effects of auspi- cious animal years (for example, the Golden 80 Decline in mortality rate (%) Pig year in 2007) and by catch-up from later 70 initiation of childbearing. These are likely to 60 be short-lived phenomena compared with the 50 European upturn, which arises from a desire 40 for increased lifetime fertility (Choe and 30 Retherford 2009; Jones 2011). The demographic trends in East Asian 20 and Pacific economies result in divergent 10 future trends in working-age populations, 0 with a number of economies projected to <1 20 9 40 9 70 9 + 4 10 9 15 4 30 9 50 9 55 4 60 59 9 25 4 35 4 75 4 65 4 45 4 –1 –3 –6 1– 5– –1 –2 –4 –5 –7 –2 –3 –7 80 –6 –4 – experience sharp declines in their labor forces in coming years. Figure 1.11 shows Age group the change in shares and the total size of Male Female the working-age population between 2010 and 2040. As expected, several economies Source: IHME and World Bank 2013. FIGURE 1.10 The typology of East Asian and Pacific economies is robust in scenarios with low-, medium-, and high-fertility assumptions Effect of different total fertility rate assumptions on share of people 65 years of age and older by country group 0.45 0.4 Share of population ages 65 and older (%) 0.35 0.3 0.25 0.2 0.15 0.1 0.05 0 90 10 20 50 60 50 60 70 80 00 30 40 19 20 20 20 20 19 19 19 19 20 20 20 Low fertility Medium fertility High fertility Source: World Bank estimates based on data from UN 2013. 32 LIVE LONG AND PROSPER FIGURE 1.11 The working-age population will shrink over coming will experience substantial declines in their decades in richer East Asian and Pacific economies shares of working-age population: over 15 Percentage change and change in absolute terms in population, ages percent in Korea and 10 percent or above 15–64, in selected economies, 2010–40 in China, Japan, and Thailand. Whereas in relative terms China fares somewhat better a. Percentage change than richer East Asian and Pacific econo- Hong Kong SAR, China mies, in absolute terms China dwarfs all Korea, Rep. other economies, with an expected reduction Singapore of 90 million people of working age between Thailand 2010 and 2040. In contrast, in lower-income Japan green economies and some orange econo- China mies with younger populations, the share of Vietnam the working-age population is not expected to shrink until after 2040. In absolute terms, Mongolia Indonesia and the Philippines will account Malaysia for a large share of the regional increase, Myanmar but in relative terms, Timor-Leste, Lao PDR, Indonesia Papua New Guinea, and the Philippines will Cambodia lead the way. Philippines The three groups of East Asian and Papua New Guinea Pacific economies also exhibit distinct pat- Lao PDR terns of dependency ratios. As shown in Timor-Leste figure 1.12, changes in total dependency –20 –15 –10 –5 0 5 10 15 ratios (TDRs) will be driven largely by ris- ing elderly dependency ratios in rich and Percentage change middle-income economies of the region, and both rising elderly and falling youth depen- dency ratios will drive changes in younger b. Absolute numbers and poorer East Asian and Pacific econo- China mies. In the youngest economies, TDRs have Japan been declining and are projected to continue Thailand to decline until 2045. TDR in the oldest red Korea, Rep. economies, in contrast, started increasing Hong Kong SAR, China sharply from 2010 and will reach 94 depen- Singapore dents for every 100 working-age individu- Timor-Leste als by 2060. TDR has also begun a steep Mongolia increase in orange economies and will reach Lao PDR 67 dependents for every 100 working-age individuals by 2060. The rates of change Papua New Guinea in TDRs in East Asia and Pacific are much Cambodia steeper than those experienced by OECD Myanmar and emerging Eastern European and Central Vietnam Asian economies. Malaysia The classic measures of dependency in Philippines terms of the rigid definition of working Indonesia age have important limitations. In short, –100 –80 –60 –40 –20 0 20 40 60 standard dependency ratios conflate age Working-age population (millions) with dependency, both for those 65 years of age and above and for those under Source: World Bank estimates based on data from UN 2013. age 15. Economically, use of this measure T H E D E M O G R A P H I C A N D E P I D E M I O LO G I C A L T R A N S I T I O N S I N E A S T A S I A A N D PAC I F I C 33 FIGURE 1.12 Total dependency ratios conceal contrasting trends in youth and elderly population shares Elderly, youth, and total dependency ratios in three groups of East Asian and Pacific economies, 1950–2060 a. Red economies b. Orange economies 100 100 their sum to the population ages 15–64 (%) Ratio of population ages 0–14 and 65+ and their sum to the population ages 15–64 (%) Ratio of population ages 0–14 and 65+ and 90 90 80 80 70 70 60 60 50 50 40 40 30 30 20 20 10 10 0 0 50 60 70 80 90 00 10 20 30 40 50 50 50 60 70 80 90 00 10 20 30 40 60 60 19 19 19 19 19 20 20 20 20 20 20 20 19 19 19 19 19 20 20 20 20 20 20 20 c. Green economies their sum to the population ages 15–64 (%) 100 Ratio of population ages 0–14 and 65+ and 90 80 70 60 50 40 30 20 10 0 50 60 70 80 90 00 10 20 30 40 50 60 19 19 19 19 19 20 20 20 20 20 20 20 Total dependency ratio Youth dependency ratio Elderly dependency ratio Source: World Bank estimates based on data from UN 2013. Note: The elderly dependency ratio (EDR) is calculated as the population 65 years and older to the population ages 15–64. The youth dependency ratio (YDR) is calculated as the population ages 0–14 to the population ages 15–64. The total dependency ratio (TDR) equals EDR plus YDR. is justified for the young who are truly work, savings and assets, and other factors. dependent, in most cases—though whether Some alternative definitions of dependency 15 years of age remains an appropriate are discussed in box 1.2. threshold in societies where the majority of young people complete secondary school and many go on to postsecondary studies The epidemiological transition in is a question. However, for individuals 65 East Asia and Pacific years of age and older, it presents a prob- Together with the demographic transition, lem, given the variation in actual depen- East Asia and Pacific is experiencing a rapid dency as a result of differential patterns of epidemiological transition, driven in part by 34 LIVE LONG AND PROSPER BOX 1.2 Alternative measures of dependency Many attempts have been made to rethink depen- remaining life expectancy as the threshold. Empirical dency ratios to make them more relevant to policy evidence linking this particular 15-year threshold to making (Sanderson and Scherbov 2005, 2007, 2013; dependency is limited. Another alternative depen- Spijker and MacInnes 2013). In a recent example, dency ratio makes additional progress in overcoming Spijker and MacInnes (2013) propose using 15 years the arbitrary dependency cutoff. This measure uses or less of remaining life expectancy as a proxy for the life-cycle pattern of consumption described in dependency, arguing that remaining life expectancy is National Transfer Accounts to directly compute the closely linked to health and active behaviors in addi- ratio of people weighted by age-specific earnings and tion to being an important “second” component of age-specific labor force participation rates to people age. The authors also propose the employed popu- weighted by age-specific consumption (Prskawetz lation as a more accurate indicator of who works and Sambt 2014). The results are the opposite of than the more traditional definition of working-age those found in Spijker and MacInnes (2013): in the population as 15 to 64 years of age. The results of future, the Prskawetz and Sambt measure declines applying this alternative old-age dependency ratio to more (that is, implies more dependency) than the tra- the United Kingdom and several member countries ditional economic support ratio on which it is based, of the Organisation for Economic Co-operation and which divides the working-age population by overall Development are striking. The traditional old-age population and uses arbitrary age cutoffs. dependency ratio results in far fewer working-age The contrasting results of Spijker and MacInnes individuals for every person 65 years of age and older (2013) and Prskawetz and Sambt (2014) illustrate and implies that this trend will continue. The Spijker that measures of aging should be approached with and MacInnes measure, in contrast, finds that depen- an open mind. Nonetheless, both these alternatives dency fell in the past, will continue to fall in the near are improvements on the traditional measure because future, and will rise only gradually in the long term. they acknowledge that dependency ratios have policy As the authors put it, “over the past four decades the implications: dependency is not only about demo- population, far from aging, has in fact been getting graphic structure but also about behavior that policy younger” (Spijker and MacInnes 2013, 21). can influence. These refined dependency ratios sug- However, the bright picture of aging presented in gest potential policy remedies—such as incentives to Spijker and MacInnes has its own issues. Although increase labor force participation of older people, their alternative dependency ratio incorporates investments in health that are associated with length- changes in education, female labor force participa- ier working lives, and adjustments to pension eligi- tion, and early retirement, it still relies on an arbi- bility age—when policy makers consider dependency trary cutoff for dependency, setting 15 years of ratios to be too high. population aging but also by factors such as health, takes into account age-specific mortal- higher incomes and urbanization. Apart from ity, morbidity, and functional health status.7 aging itself, a vital element of understanding The main hypotheses regarding how healthy changing demographics is knowing whether life expectancy evolves with the increase in people in the region are experiencing healthy life spans, and the empirical evidence to date, aging and how the disease profile is chang- are summarized in box 1.3. ing. “Healthy aging” implies not only living Health-adjusted life expectancy (HALE), longer but also spending those additional or healthy life expectancy, increased in East years of life in good, active health. Thus, Asia and Pacific between 1990 and 2010, examining healthy life expectancy is useful. although slightly less than total life expec- This measure, which estimates the number of tancy (figure 1.13). Each year of increase in years that a person can expect to live in good life expectancy at birth has been accompanied T H E D E M O G R A P H I C A N D E P I D E M I O LO G I C A L T R A N S I T I O N S I N E A S T A S I A A N D PAC I F I C 35 BOX 1.3 Is longevity accompanied by more good health or less? In the past 100 to 150 years, medical technologies and prevalence of chronic illnesses and morbidity have been effective at saving lives from a range of are compressed toward the later period of life at a fatal diseases. Often, however, this success has meant faster pace than reductions in mortality. People do that the life years saved are spent in poor health and not just live longer; they live longer in better health. characterized by chronic illness and disability. As • The dynamic equilibrium hypothesis (Manton preventive health care throughout the course of life 1982). Under this hypothesis, the effects of com- becomes more widespread, and as medical technolo- pression and expansion of morbidity offset one gies improve in their ability to reduce mortality and another, with, on one hand, decreasing prevalence morbidity, the additional years of life saved may be or incidence of chronic illness, and on the other spent in good health. Better health goes along with hand, decreasing fatality rates of diseases, leading fewer health care needs and may drive down use of to longer prevalence of disability. health services and health expenditure (Rechel et al. The empirical research is not clear on whether 2009). Therefore, an important way to measure compression or expansion of morbidity occurs progress is to determine if longevity is accompanied in countries that have already experienced rapid by more good health or less. aging. For example, higher levels of some disabling Three hypotheses have been put forward to pre- conditions (dementia and musculoskeletal diseases) dict possible future interaction between the evolu- go along with falling rates of prevalence of other tion in life expectancy and changes in prevalence of conditions (cardiovascular and chronic respiratory disability and bad health. This demonstrates that diseases). The growing consensus from Organisa- predicting changes in health status and morbidity is tion for Economic Co-operation and Develop- problematic, not least because epidemiological data ment countries is as follows: the prevalence of are subject to greater uncertainty. The three hypoth- severe disability has been reduced substantially, eses are as follows: the prevalence of mild to moderate disability has • The expansion of morbidity hypothesis (Gruen- increased, and overall, health status has improved berg 1977; Olshansky et al. 1991; Verbrugge (Christensen et al. 2009). 1984). Under this hypothesis, the decline in mor- Trends in disability may reflect underlying trends tality is largely due to a reduction in disease fatality in other domains, such as the rising use of assistive rather than a reduction in prevalence or incidence technology, improved accessibility at home and in of those diseases. Consequently, falling mortality public spaces, and changes in social perceptions of is accompanied by an increase in morbidity and disability. The decline in disability might also reflect disability. improvements in living conditions, not only of those • The compression of morbidity hypothesis (Fries currently elderly, but also of their childhood living 1980, 1989). Under this hypothesis, the incidence conditions (Christensen et al. 2009). by an 11-month increase in healthy life expec- region at age 50 by country and gender tancy at birth. The biggest improvements in and shows the gender gap. The difference healthy life expectancy in the region were between life expectancy and healthy life in low-income countries, and the slowest expectancy can be interpreted as the aver- improvements were in middle-income coun- age number of years of healthy life lost to tries (IHME and World Bank 2013). poor health. HALE at age 50 is also a use- Large differentials exist across East Asian ful indicator in thinking about appropriate and Pacific countries in the number of years retirement ages in pension systems, as dis- that people age 50 can expect to live in cussed in chapter 5. A significant gap exists good health. Figure 1.14 compares healthy between total and healthy life expectancy life expectancy with life expectancy in the in the region, with important implications 36 LIVE LONG AND PROSPER FIGURE 1.13 Life expectancy and health-adjusted such as Indonesia and China. At the age of life expectancy in East Asia and Pacific both 50, men and women in Japan can expect increased between 1990 and 2010 to live free of disability approximately 10 Years of life expectancy and health-adjusted life years more than their counterparts in Papua expectancy, by gender New Guinea. The gender gap also is signifi- cant in all East Asian and Pacific countries, a. Male with women spending more years free of 80 disability than men. The policy and imple- mentation implications of those systems are 70 discussed in chapter 6. 60 The leading causes of mortality and dis- ability in East Asia and Pacific have shifted Number of years 50 significantly toward NCDs over the past two decades, in part because people are 40 living to much older ages. NCDs gener- 30 ally increase in prevalence with age. The epidemiological transition is characterized 20 by a decline in premature mortality and morbidity from communicable, maternal, 10 and child health causes and an increase 0 in deaths attributable to NCDs and inju- 1990 2010 ries. NCDs include cardiovascular disease, diabetes, and cancer, which are chronic b. Female conditions that require primary and sec- 80 ondary prevention and pharmacological 70 management over several years (see chap- ter 6 for a discussion on health systems 60 implications of aging and NCDs). The region as a whole has transitioned epide- Number of years 50 miologically over the past 20 years, with 40 China and Vietnam leading the way. The NCD burden has grown rapidly, with ini- 30 tial onset increasingly happening in middle age. Cardiovascular disease, cancers, and 20 diabetes accounted for the bulk of all dis- 10 ability-adjusted life years (DALYs8) among those ages 60 years and above in East Asia 0 and Pacific in 2010 (figure 1.15). In high- 1990 2010 income countries such as Australia, Japan, Life expectancy Health-adjusted life expectancy and Korea, cancers have taken over as the leading cause of DALYs, particularly for Source: IHME and World Bank 2013. men. Cancers are the second major cause of disease burden among those ages 60 years for health delivery systems. Years spent and above in China and Mongolia and are in good health are considerably lower in among the top five causes of disease bur- Mongolia, Papua New Guinea, and some den in most other countries (WHO 2014). Pacific Islands compared with not only Nevertheless, in many low-income coun- high-income countries such as Australia tries in the region, such as Cambodia, Lao and Japan but also middle-income countries PDR, Myanmar, Papua New Guinea, and FIGURE 1.14 Large differentials in healthy years of life are evident across the region, but men generally live shorter, less healthy lives than women Additional years of life and healthy life expectancy at age 50, by gender, and gender gap a. Life expectancy and health-adjusted b. Life expectancy and health-adjusted c. Gender gap at age 50 life expectancy for women at age 50 life expectancy for men at age 50 Papua New Guinea Kiribati Timor-Leste Solomon Islands Papua New Guinea Fiji Kiribati Solomon Islands Papua New Guinea Marshall Islands Marshall Islands New Zealand Micronesia, Fed. Sts. Mongolia Solomon Islands Vanuatu Micronesia, Fed. Sts. Indonesia Fiji Vanuatu Marshall Islands Myanmar Myanmar Lao PDR Lao PDR Fiji Australia Timor-Leste Lao PDR Vanuatu Mongolia Philippines Singapore Cambodia Cambodia Samoa Samoa Tonga Micronesia, Fed. Sts. Tonga Samoa Thailand Indonesia Timor-Leste Malaysia Philippines Korea, Dem. People’s Rep. Cambodia Korea, Dem. People’s Rep. Indonesia Myanmar Malaysia Malaysia Korea, Dem. People’s Rep. Thailand Vietnam Tonga Vietnam Thailand Japan China China Taiwan, China New Zealand Taiwan, China China Australia Korea, Rep. Korea, Rep. Taiwan, China New Zealand Vietnam Singapore Australia Kiribati Korea, Rep. Singapore Philippines Japan Japan Mongolia 0 10 20 30 40 0 10 20 30 40 0 2 4 6 8 Years Years Years Life expectancy Health-adjusted life expectancy T H E D E M O G R A P H I C A N D E P I D E M I O LO G I C A L T R A N S I T I O N S I N E A S T A S I A A N D PAC I F I C Source: IHME and World Bank 2013. 37 38 LIVE LONG AND PROSPER FIGURE 1.15 NCDs have increased sharply as leading causes of hearing loss (WHO 2014). Musculoskeletal disease burden in East Asia and Pacific in recent decades conditions, which include lower back pain Percentage change in total DALYs from 1990 to 2010 for East Asia and and neck pain, are major sources of dis- Pacific by condition ability both globally and in East Asia and Pacific for those ages 45 years and above. Stroke The regional pattern of increased preva- Ischemic heart disease lence of NCDs is mirrored in individual Road injury countries. Figure 1.16 shows the changes in COPD Low back pain the leading causes of death in China from Major depressive disorder 1990 to 2010 as an example of how NCDs Lung cancer have increased as causes of mortality. Only Lower respiratory infections one communicable disease remains among Diabetes Liver cancer the leading causes of death in China, with Tuberculosis 83 percent of total mortality attributable to Falls NCDs and injuries (Bloom et al. 2014; IHME Neck pain Other musculoskeletal disorders and World Bank 2013). Preterm birth complications As people age and develop chronic Self-harm illnesses, they are increasingly likely to have Stomach cancer several comorbid conditions, as is observed Congenital anomalies Neonatal encephalopathy in East Asia and Pacific. Comorbidities can Diarrheal diseases interact with each other and complicate Cirrhosis treatment. Among the elderly in Korea, for Iron-deficiency anemia example, over half of all women and one- HIV/AIDS Chronic kidney disease third of men had three or more chronic Drowning conditions (figure 1.17). Similarly, findings –80 –60 –40 –20 0 20 40 60 80 100 120 140 160 180 200 from a 2008 Australian study reveal the % change in total disability-adjusted life years, 1990–2010 extent of the problem among the elderly: Communicable, newborn, nutritional, Noncommunicable diseases over half of elderly patients with arthritis and maternal diseases also had hypertension, 20 percent had car- Injuries diovascular disease, 14 percent had dia- betes, and 12 percent had mental health Source: IHME and World Bank 2013. problems. Of those with cardiovascu- Note: COPD = chronic obstructive pulmonary disease; DALYs = disability-adjusted life years; NCDs = noncommunicable diseases. lar disease, 60 percent also had arthritis, 20 percent had diabetes, and 10 percent had asthma or mental health problems Timor-Leste, communicable diseases still (Caughey et al. 2008). account for a significant share of prema- The burden of disease caused by chronic ture mortality and morbidity, resulting in a conditions is also characterized by large double burden of disease. inequalities based on social determinants Sensory organ diseases (for example, such as poverty and lack of education. cataracts) and musculoskeletal disorders In China, for instance, men and women (for example, osteoarthritis) also feature from the poorest quintiles of the popula- prominently among major causes of dis- tion were three to four times more likely to ease burden in the region, especially in have a chronic illness in 2006 than those in higher-income countries with a greater pro- the top three quintiles (figure 1.18) (World portion of older people. The demographic Bank 2012). These inequalities persist even shift toward older age groups and projected in high-income countries that have made increases in longevity will inevitably lead significant progress in reducing the burden to an increase in disability caused by sen- of cardiovascular disease. For instance, in sory organ diseases, such as cataracts and Australia (Brown 2013) and New Zealand T H E D E M O G R A P H I C A N D E P I D E M I O LO G I C A L T R A N S I T I O N S I N E A S T A S I A A N D PAC I F I C 39 (Wang et al. 2014), people from ethnic FIGURE 1.16 NCDs and injuries dominate the leading causes of minorities are at significantly higher risk of mortality in China experiencing cardiovascular disease than the Percentage change in total disability-adjusted life years (DALYs) by cause, rest of the population. 1990–2010 The burden of NCDs across East Asia Stroke and Pacific reflects growing exposure to Ischemic heart disease risk factors, which are quite pronounced COPD compared with the rest of the world Lung cancer (figure 1.19). These risk factors include Liver cancer the very high prevalence of smoking (see Stomach cancer box 1.4) and the rapid rise in predisposing Road injury Lower respiratory infections conditions such as hypertension, obesity, Esophageal cancer and diabetes. A review of 33 cohort stud- Hypertensive heart disease ies in the region found continuous positive Self-harm associations between baseline body mass Diabetes index (BMI) 9 and the risks of ischemic Colorectal cancer Falls stroke, hemorrhagic stroke, and ischemic Cirrhosis heart disease as well as diabetes in several Other cardio and circulatory disease countries in the region (Asia-Pacific Cohort Chronic kidney disease Studies Collaboration 2004, 2006). 10 Drowning Between 1990 and 2010, premature death Leukemia and disability attributable to NCD-related Pancreatic cancer Rheumatic heart disease risk factors—such as dietary risks (for Aortic aneurysm example, high sodium intake), high blood Breast cancer pressure, high levels of total cholesterol and Alzheimers disease fasting plasma glucose (blood sugar), and Brain cancer high BMI—increased worldwide but was –150 –100 –50 0 50 100 150 200 quite pronounced in East Asia and Pacific % change in total disability-adjusted life year, 1990–2010 (figure 1.19). For instance, DALYs attrib- Communicable, newborn, nutritional, Noncommunicable diseases Injuries utable to high BMI increased 82 percent and maternal diseases globally but rose 198 percent in East Asia Source: IHME and World Bank 2013. and Pacific (IHME and World Bank 2013). Note: COPD = chronic obstructive pulmonary disease; NCDs = noncommunicable diseases. The Pacific Islands alone have some of the highest rates of obesity and overweight FIGURE 1.17 In the Republic of Korea both men and women incidence in the world. In four countries 65 years of age and older are more likely to have multiple of the Pacific, 11 at least half of adults are chronic conditions Percentage of people ages 65 and older by number of chronic conditions obese, and over 40 percent of students ages and gender 13–15 are overweight (Anderson 2013). Stark contrasts exist between and within 60 countries. In upper-middle-income countries ages 65 and older (%) Share of population 50 such as Malaysia and Thailand, the NCD- 40 related risk factors were the dominant causes 30 of DALYs, whereas in low-income and lower-middle-income countries, iron defi- 20 ciency and childhood undernutrition were 10 dominant causes. Country-level evidence, 0 None One Two Three and more although somewhat limited, points to signifi- Chronic conditions cant socioeconomic gradients in exposure to Total Men Women and treatment of NCD-related risks (boxes 1.5 and 1.6). Source: Sunwoo 2013. 40 LIVE LONG AND PROSPER FIGURE 1.18 In China, in 2006, socioeconomic status had an Adverse initial conditions in life, including inverse relationship with noncommunicable diseases low birthweight, have a significant impact on Proportion of population with a chronic illness, by gender and income mortality and morbidity later in life, a find- quintile ing that has significant implications for East Asia and Pacific. First, current and future 40 elderly cohorts in low- and middle-income Share of population (%) 30 countries in the region are likely to have experienced more economic distress and dis- 20 advantage as children than their counterparts 10 in the developed world. During the 1950s 0 to 1970s, East Asian and Pacific countries 1st (poorest) 2nd 3rd 4th 5th (richest) experienced low economic growth and, in Quintiles some cases, war (Cambodia and Vietnam) or Men Women extreme social and economic distress of other forms. The evidence on the early origins of Source: World Bank 2012. adult disease summarized in box 1.7 suggest that current and future elderly cohorts in the region are likely to be especially vulnerable to cardiovascular diseases and cancers as a FIGURE 1.19 NCD-related risk factors grew in East Asia and Pacific result of the adverse circumstances they faced between 1990 and 2010 in childhood. Second, low birthweight and Percentage change in total DALYs, 1990–2010 childhood undernutrition, a critical factor in the early origins of adult disease, contin- Dietary risks ues to be a major cause of premature death High blood pressure and disability in low-income countries such Smoking as Cambodia, Lao PDR, and Timor-Leste (IHME and World Bank 2013). Therefore, Household air pollution adverse initial conditions will continue to put Ambient PM pollution future elderly populations at risk for debili- High fasting plasma glucose tating NCDs even in the relatively young Alcohol use countries of the region. High body mass index Unless these risks are addressed urgently, Occupational risks the population cohorts that will survive the Physical inactivity next two decades and become the majority of the elderly in the 2030s will be more vulnera- High total cholesterol ble to poor health and morbidity than today’s Iron deficiency elderly. Addressing these risks, in particular Drug use identifying and managing risks through pre- Suboptimal breastfeeding vention and treatment, is therefore vital for Intimate partner violence healthy aging in these populations. Some –80 –60 –40 –20 0 20 40 60 80 100 120 140 160 180 200 risks, such as hypertension and hypercholes- % change in total disability-adjusted life year, 1990–2010 terolemia, are silent killers that can go undi- agnosed for a long time. Air pollution Alcohol and drug use Physical inactivity Smoking Physiological risks Occupational risks Undernutrition Dietary risks Sexual abuse and violence Conclusion Source: IHME and World Bank 2013. The demographic and epidemiological Note: DALY = disability-adjusted life year; NCDs = noncommunicable diseases; PM = particulate matter. trends in East Asia and Pacific point to the T H E D E M O G R A P H I C A N D E P I D E M I O LO G I C A L T R A N S I T I O N S I N E A S T A S I A A N D PAC I F I C 41 BOX 1.4 Very high prevalence of smoking and related death rates in the region East Asia and Pacific has the highest prevalence of smoking among men ages 15–69, at 54 percent, smoking in the world, and smoking is among the is among the highest in the world. However, the top three risk factors in both lower- and middle- number of daily cigarettes consumed per smoker income countries. Two-thirds of the world’s tobacco in Asian countries is low compared with that in users live in 15 countries (WHO 2011b), and 5 of developed countries (WHO 2011a). Despite the these countries with high-risk burdens are in Asia: lower rates of consumption, smoking death rates China, India, the Philippines, Thailand, and Viet- per smoker appear to be as extreme as those seen in nam. China alone is the world’s largest producer the high-income countries, where smoking has been and consumer of tobacco: prevalence of tobacco prevalent for decades. BOX 1.5 Indonesia: Increasing and often unequal exposure to risks • High body mass index (BMI) has increased. In correlation with the probability of having either 2007, around 31 percent of elderly women had high total cholesterol or low HDL. a BMI of 25 or over, more than double the per- • Levels of hypertension are substantial but have centage in 1993. Among elderly men in 2007, not changed over time. Among men ages 45 and 17 percent were overweight, compared with over older, around 44 percent had hypertension in 8 percent in 1993. 2007, the same percentage as in 1997. Similarly • BMI increases with both education and income. among women, 53 percent were hypertensive, Completing primary school (for women) and and the number did not change much over the junior high school (for men) was positively asso- years. These are considered high prevalence rates ciated with increased BMI, as was per capita at the population level. Overall, a strong posi- income. tive relationship exists between age and being • Waist circumference has also increased, lead- hypertensive. ing to higher cardiovascular risks. For women • Education and income are associated with a more than men, mean waist circumference has higher probability of the hypertension being diag- increased while BMI has remained constant. nosed and of adherence to treatment. Probability • Cholesterol levels are not correlated with educa- of underdiagnosis is larger for lower-income and tion or income. Socioeconomic status has little uneducated individuals, particularly women. importance of better understanding the house- already have proved technically and politi- hold and individual behavior and preferences cally challenging, both in global contexts and of the region’s older populations to inform in richer economies in the region. The fol- policy choices. The “big picture” trends pre- lowing chapter moves from the big picture sented in this chapter point to the need for and aims to provide insights into these policy reforming policies in a number of areas that reforms. 42 LIVE LONG AND PROSPER BOX 1.6 Malaysia: Undiagnosed noncommunicable disease risks and modest social economic status differentials in management of risk • Diabetes is largely undiagnosed or is not con- among men. Prevalence is higher among better- trolled. Only 47 percent of diabetics in Malaysia off men, but no socioeconomic status gradient are currently diagnosed. Of the 7 percent who exists for women. reported being diagnosed as diabetic, only 41 • Significant socioeconomic status disparities can percent (that is, 3 percent of the population) had be seen in the rates of diagnosis and treatment controlled blood glucose levels. of hypercholesterolemia. A person in the richest • Little variation by social economic status is quintile is 2.2 times more likely than a person in evident in diabetes prevalence and treatment. the poorest quintile to have been diagnosed and Overall patterns of actual diabetes prevalence and 2.1 times more likely to be successfully treated. the likelihood of being diagnosed do not vary by • Hypertension remains largely undiagnosed. Over income quintile. 32 percent of the population was found to have • Hypercholesterolemia (high cholesterol in the blood hypertension, compared with the self-reported stream) remains largely undiagnosed. Prevalence rate of 13 percent. among adults in 2011 was 35 percent, with the • Modest socioeconomic status disparities can be rate higher among females (40 percent) than males seen in the rates of diagnosis and treatment of (30 percent). Only 8 percent had been advised by hypertension. A person in the richest quintile is medical personnel about having the condition. 1.2 times more likely than a person in the poor- • Large variation by socioeconomic status is evi- est quintile to have been diagnosed, and 1.5 times dent in prevalence of hypercholesterolemia more likely to be successfully treated. BOX 1.7 Adverse childhood socioeconomic conditions and increased mortality risks later in life The links in the chain The evidence The fetal origins hypothesis of adult diseases. Fetal • Analysis of a sample of 14,520 Dutch individu- undernutrition, as reflected by low birthweight, is als found that exposure to recessions in early associated with susceptibility to development of isch- life significantly increases cancer mortality risks emic heart disease and other chronic noncommuni- of older individuals (both male and female). The cable diseases in later life (Barker 1990, 1995; Barker residual life expectancies are about 8 percent and Osmond 1986). Exposure to adverse nutritional and 6 percent lower for male and female cancer conditions during pregnancy results in alterations mortality, respectively, and about 5 percent lower in the development of vital organs, tissues, or other for female cardiovascular mortality (Yeung et al. human systems. These alterations, though advanta- 2014). geous for short-term survival, may be detrimental in • Early malnutrition in Latin America is highly cor- the long term and may increase the predisposition to related with self-reported diabetes (WHO 2012). chronic diseases during adulthood. The mechanism is • A Swedish study (Finch et al. 2004) with excellent mainly biological. historical data concluded that reduced early expo- The chain of risk model. Exposure to an adverse sure to infectious diseases was related to increases environment in early life may set individuals on unfa- in life expectancy. vorable life trajectories. For instance, individuals • Adverse childhood socioeconomic status condi- born in poor families may be more likely to be ill tions are positively correlated with increased and may also go to school less often than others. This mortality risks in later life, such as stomach pattern may affect future educational attainments, cancer, hemorrhagic stroke, coronary heart dis- labor market skills, later earnings, and adult health ease, and chronic obstructive pulmonary disease (Almond and Currie 2011; Yeung et al. 2014). (Galobardes et al. 2008). T H E D E M O G R A P H I C A N D E P I D E M I O LO G I C A L T R A N S I T I O N S I N E A S T A S I A A N D PAC I F I C 43 Notes 10. The paper reviewed 33 cohort studies, including 310,283 participants, who 1. Healthy life expectancy (also known as HALE, contributed 2,148,354 person-years of or health-adjusted life expectancy) is defined follow-up, during which 3,332 stroke events as the average number of years that a person and 2,073 ischemic heart disease events were can expect to live in full health, excluding the observed. years lived in less than full health because of 11. For example, obesity rates in the Cook Islands disease or injury. are 63 percent, in Nauru 71 percent, in 2. The 7 and 14 percent thresholds are based Samoa 54 percent, and in Tonga 57 percent on the United Nations definition of what (Anderson 2013). constitutes an aging society and an aged society, respectively. 3. The five economies are Hong Kong SAR, China (37 percent elderly); Japan (37 percent); References Korea (37 percent); Singapore (32 percent); Almond, Douglas, and Janet Currie. 2011. “Killing and Thailand (33 percent). Me Softly: The Fetal Origins Hypothesis.” 4. Country groupings do not include a number Journal of Economic Perspectives 25 (3): of Pacific Island states. 153–72. 5. In contrast, a simulation by the Korean Anderson, Ian. 2013. “The Economic Costs of National Assembly Research Service in Non-communicable Diseases in the Pacific 2014 projected that the population would Islands: A Rapid Stocktake of the Situation decline from around 50 million currently to in Samoa, Tonga, and Vanuatu.” Health, 5 million in 2172, to 100,000 in 2379, and Nutrition, and Population Discussion Paper to extinction by 2750 (Mundy 2014). Such 86522, World Bank, Washington, DC. projections highlight the risks of projecting Asia-Pacific Cohort Studies Collaboration. 2004. current trends into the long-term future. “Body Mass Index and Cardiovascular Disease 6. As explained in Bloom, Canning, and Finlay in the Asia-Pacific Region: An Overview of (2010), the decline in mortality rates across all 33 Cohorts Involving 310,000 Participants.” ages is one explanation for their limited effect International Journal of Epidemiology 33 (3): on population aging compared to the decline 751–58. in fertility rates. ———. 2006. “Body-Mass Index and the Risk of 7. Internationally comparable measures of Diabetes Mellitus in the Asia-Pacific Region.” health status are less widely available than Asia-Pacific Journal of Clinical Nutrition 15 mortality statistics. Data on life expectancy (2): 127–33. and health-adjusted life expectancy were Barker, David J. 1990. “The Fetal and Infant obtained from the Global Burden of Disease Origins of Adult Disease.” British Medical 2010 project. Years lived with disability were Journal 301 (6761): 1111. calculated from numerous sources. Data gaps ———. 1995. “Fetal Origins of Coronary Heart in many countries, however, prompted the use Disease.” British Medical Journal 311 (6998): of a modeling tool to estimate prevalence of 171–74. disability, which flattens variability between Barker, David J., and Clive Osmond. 1986. “Infant countries. Mortality, Childhood Nutrition, and Ischaemic 8. The DALY is a summary measure of mortality Heart Disease in England and Wales.” Lancet and morbidity, calculated by adding years 327 (8489): 1077–81. lived with disability and years of life lost Bloom, David, Elizabeth Cafiero, Mark to premature death. In most countries, McGovern, Klaus Prettner, Anderson cardiovascular disease is the leading cause of Stanciole, Jonathan Weiss, Samuel Bakkila, and death and disability. Larry Rosenberg. 2014. “The Macroeconomic 9. BMI is calculated from a person’s weight Impact of Non-communicable Diseases in and height. The index provides a reliable China and India: Estimates, Projections, and indicator of body fat for most people and is Comparisons.” Journal of the Economics of used to screen for weight categories that may Ageing 4: 100–11. lead to health problems. For adults over 20, Bloom, David, David Canning, and Jocelyn Finlay. a BMI of 25.0–29.9 implies that the adult is 2010. “Population Aging and Economic Growth overweight, and a BMI of 30.0 implies that in Asia.” In The Economic Consequences the adult is obese. of Demographic Change in East Asia, 44 LIVE LONG AND PROSPER edited by Takatoshi Ito and Andrew Rose, IHME (Institute for Health Metrics and Evaluation) 61–89. Chicago: University of Chicago Press. and World Bank. 2013. The Global Burden of Brown, Alex. 2013. “Essential Service Standards Disease: Generating Evidence, Guiding Policy. for Equitable National Cardiovascular Care Seattle, WA: IHME. for Aboriginal and Torres Strait Islander Jones, Gavin W. 2011. “Recent Fertility Trends, People: An Exemplar Approach to Closing the Policy Responses, and Fertility Prospects in Gap.” Presentation made at the World Heart Low Fertility Countries of East and Southeast Federation’s World Congress of Cardiology, Asia.” Expert Paper 2011/ 5, Population Melbourne, Australia, May. Division, Department of Economic and Social Caughey, Gillian, Agena Vitry, Andrew Gilbert, Affairs, United Nations, New York. http:// and Elizabeth Roughead. 2008. “Prevalence of www.un.org/esa/population/publications Comorbidity of Chronic Diseases in Australia.” / expertpapers/2011-5_Jones_Expert-Paper BioMed Central Public Health 8: 221. http:// _FINAL_ALL-Pages.pdf. www.biomedcentral.com/1471-2458/8/221. Kinsella, Kevin, and Wan He. 2009. An Aging Choe, Minja Kim, and Robert D. Retherford. World: 2008. International Population Reports 2009. “The Contribution of Education to PS95/09-1. Washington, DC: U.S. Government South Korea’s Fertility Decline to ‘Lowest- Printing Office. Low’ Level.” Asian Population Studies 5 (3): Manton, Kenneth G. 1982. “Changing Concepts 267–87. of Morbidity and Mortality in the Elderly Christensen, Kaare, Gabriele Doblhammer, Population.” Milbank Memorial Fund Roland Rau, and James Vaupel. 2009. “Aging Quarterly 60: 183–244. Populations: The Challenges Ahead.” Lancet Mundy, Simon. 2014. “South Koreans Alarmed 374 (9696): 1196–208. by Prospect of Extinction by 2750.” Financial Finch, Caleb E., and Eileen M. Crimmins. 2004. Times, August 25, 2014. “Inflammatory Exposure and Historical Olshansky, S. Jay, Mark A. Rudberg, Bruce Carnes, Changes in Human Life Spans.” Science 305 Christine K. Cassel, and Jacob A. Brody. 1991. (5691): 1736–39. “Trading Off Longer Life for Worsening Flochel, Thomas, Yuki Ikeda, Harry Moroz, and Health” Journal of Aging and Health 3 (2): Nithin Umapathi. 2014. “Macroeconomic 194–216. Implications of Aging: Demographics, Labor Pew Research Center. 2014. “Attitudes about Markets, and Productivity in East Asia Pacific.” Aging: A Global Perspective.” Pew Research Background paper for the East Asia and Center, Washington, DC. Pacific regional report on aging, World Bank, Prskawetz, Alexia, and Jože Sambt. 2014. “Economic Washington, DC. Supports Ratios and the Demographic Dividend Fries, James F. 1980. “Aging, Natural Death, and in Europe.” Demographic Research 30 (34): the Compression of Morbidity.” New England 963–1010. Journal of Medicine 303 (3): 130–35. Rechel, Bernd, Jonathan Erskine, Barrie ———. 1989. “The Compression of Morbidity: Dowdeswell, Stephen Wright, and Martin Near or Far?” Milbank Memorial Fund McKee. 2009. “Capital Investment for Health: Quarterly 67 (2): 208–32. Case Studies from Europe.” Observatory Study Galobardes, Bruno, John Lynch, and George Davey 18, World Health Organization on behalf of Smith. 2008. “Is the Association between European Observatory on Health Systems and Childhood Socioeconomic Circumstances and Policies, Copenhagen. Cause-Specific Mortality Established? Update Sanderson, Warren, and Sergei Scherbov. 2005. of a Systematic Review.” British Medical “Average Remaining Lifetimes Can Increase as Journal 62 (5): 387–90. Human Populations Age.” Nature 435 (7043): Gruenberg, Ernest M. 1977. “The Failure of 811–13. Success.” Milbank Memorial Fund Quarterly ———. 2007. “A New Perspective on Population 55 (1): 3–24. Aging.” Demographic Research 16 (2): Gu, Baochang, and Yong Cai. 2009. “Fertility 27–58. prospects in China.” UN Document UN/POP/ ———. 2013. “The Characteristics Approach EGM-FERT/2009/6, Population Division, to the Measurement of Population Aging.” Department of Economic and Social Affairs, Population and Development Review 39 (4): United Nations, New York. 673–85. T H E D E M O G R A P H I C A N D E P I D E M I O LO G I C A L T R A N S I T I O N S I N E A S T A S I A A N D PAC I F I C 45 Somanathan, Aparnaa. 2015. “Health and Health “Comparing Contemporary Coronary Artery Care in an Ageing East Asia and Pacific.” Bypass Grafting across Six Main Ethnic Background paper for the East Asia and Groups of New Zealand: Another Example Pacific regional report on aging. World Bank, of Inequalities in Cardiovascular Disease.” Washington, DC. Presented at the World Heart Federation’s Spijker, Jeroen, and John MacInnes. 2013. World Congress of Cardiology, Melbourne, “Population Ageing: The Time Bomb That Australia, May 4–7. Isn’t?” British Medical Journal 347 (7933): World Bank. 2012. “Toward a Healthy and 20–22. Harmonious Life in China: Stemming the Rising Sunwoo, Duk. 2013. “Economic Costs and Tide of Non-communicable Diseases.” Report Opportunities for Aging and Health Care.” 62318-CN, World Bank, Washington, DC. Presented at the Forum on Innovation for WHO (World Health Organization). 2011a. Ageing Populations, Kobe, Japan, December MPOWER: A Policy Package to Reverse the 10–12. Tobacco Epidemic. WHO, Geneva. UN (United Nations). 2013. “World Population ———. 2011b. Noncommunicable Diseases Prospects: The 2012 Revision, Highlights, Country Profiles 2011. Geneva: WHO. and Advance Tables.” Working Paper ESA/P/ ———. 2012. “Global Health and Aging.” WHO, WP228, Department of Economic and Social Geneva. Affairs, Population Division, UN, New York. ———. 2014. The Health of Older People in Verbrugge, Lois M. 1984. “Longer Life but Selected Countries of the Western Pacific Worsening Health? Trends in Health and Region. Geneva: WHO. Mortality of Middle-Aged and Older Persons.” Yeung, Gary Y. C., Gerard J. van den Berg, Milbank Memorial Fund Quarterly 62 (3): Maarten Lindeboom, and France R. M. 475–519. Portrait. 2014. “The Impact of Early-Life Wang, Tom Kai Ming, Tharumenthiran Economic Conditions on Cause-Specific Ramanathan, Sue Crengle, Greg Gamble, Mortality during Adulthood.” Journal of Ralph Stewart, and Harvey White. 2014. Population Economics 27 (3): 895–919. The Welfare and Labor Market Behavior of Older People in 2 East Asia and Pacific Introduction and to older people in East Asia and Pacific and expectations regarding future roles of the Across East Asia and Pacific and even within state and families. All of these factors have the region’s countries, diverse patterns of important implications for policies on health well-being, labor market behavior, and care, labor, pensions, and other related areas. sources of support among the elderly can To help inform policy making with a bet- be seen. Declining poverty rates among the ter understanding of the situation and behav- elderly suggest that in most of the region’s ior of older people in East Asia and Pacific, countries, the elderly have benefited from this chapter discusses key aspects of well- economic growth. However, nonincome indi- being, sources of support, and labor supply cators of their well-being such as physical of the elderly. The first section in this chap- limitations and depression indicate cause for ter documents trends in poverty among the concern even in richer East Asian and Pacific region’s elderly, and the second section looks economies. Overall, labor income remains the at other measures of well-being for the elderly dominant source of support for the elderly such as functional limitations and mental in the region (even where co-residence with health conditions. For an understanding of children remains high), often well into old the importance of public and private mecha- age and especially in rural areas, where retire- nisms in support of older people, the third sec- ment is often not an option for financial rea- tion describes living arrangements and access sons. In contrast, the retirement behavior of to support for them. The fourth section then urban formal sector workers is significantly discusses the sources of financial support for influenced by pension receipt and raises con- older people in the region and their relative cerns over premature withdrawal from work importance in preventing incomes from falling in rapidly aging countries. At the same time, below the poverty line, as well as patterns of rapid social and economic change is affecting labor force participation among older workers traditional patterns of informal support from and their main drivers. This chapter is largely based on two background papers, John Giles, Yuqing Hu, and Yang Huang (2015) and John Giles and Yang Huang (2015), both of which contain fuller bibliographical references. 47 48 LIVE LONG AND PROSPER Poverty among older people in At the same time, in general, poverty East Asia and Pacific increases with time after age 45 and then declines somewhat at older ages in the region, Similar to the population as a whole, poverty a pattern likely driven by both mortality selec- rates of older people have declined across tion (poorer elderly die at younger ages) and all East Asian and Pacific countries over the movement of elderly over age 80 into homes period from 2006 to 2012.1 Poverty rates with income-earning adult children. The pat- among the rural elderly have shown a steeper terns of national poverty rates across the decline. Elderly poverty rates in Mongolia adult life cycle can be seen in figure 2.1 and and Thailand are very low (2–3 percent). vary notably. The significant increase in pov- In Thailand, they are likely influenced by erty rates with age in China, Indonesia, and near universal access to a low social pen- Vietnam contrasts with flat rates with age in sion, whereas moderate poverty rates are Mongolia, the Philippines, and Thailand. The evident for the elderly in China, Indonesia, flattening of the income poverty line with age and Vietnam. in Thailand between 2006 and 2011 is also FIGURE 2.1 Poverty rates increase with age in some East and Southeast Asian countries, but elderly poverty has fallen over time Percentage below poverty line by consumption and by income, by age, selected countries a. Consumption poverty, 2006–08 b. Consumption poverty, 2011–12 60 60 55 55 50 50 45 45 Poverty rate (%) 40 Poverty rate (%) 40 35 35 30 30 25 25 20 20 15 15 10 10 5 5 0 0 25–45 46–60 61–70 71–80 81+ 25–45 46–60 61–70 71–80 81+ Age group Age group Indonesia 2008 Mongolia 2007 Indonesia 2011 Mongolia 2011 Thailand 2006 Vietnam 2006 Thailand 2011 Vietnam 2012 c. Income poverty, 2006–08 d. Income poverty, 2009–12 60 60 55 55 50 50 45 45 Poverty rate (%) Poverty rate (%) 40 40 35 35 30 30 25 25 20 20 15 15 10 10 5 5 0 0 25–45 46–60 61–70 71–80 81+ 25–45 46–60 61–70 71–80 81+ Age group Age group China 2006 Philippines 2006 China 2011 Philippines 2009 Thailand 2006 Vietnam 2008 Thailand 2011 Vietnam 2012 Sources: Giles and Huang 2015 based on data from CHNS 2006, 2011; ThaiSES 2011; World Bank East Asia and Pacific Standardized Household Surveys, various years; and VHLSS 2012. Note: Poverty is based on per capita daily income less than US$1.25 (2005 dollar purchasing power parity). T H E W E L FA R E A N D L A B O R MA R K E T B E H AV I O R O F O L D E R P E O P L E I N E A S T A S I A A N D PAC I F I C 49 notable because it occurred during the period For Mongolia and Thailand, this difference is when the social pension was universalized. relatively small, whereas for China, Indonesia, In addition, a comparison of the profiles across the Philippines, and Vietnam with regard to consumption (panels a and b) and income consumption poverty, the rural-urban divide (panels c and d) shows that consumption- is much more pronounced. This variation based poverty rates are generally smoother deserves more exploration but may reflect fac- across the life cycle, as one would expect given tors such as the generally low level of poverty the reliance on savings and assets. and much wider coverage of social pensions in As one would expect, poverty among the Mongolia and Thailand, and perhaps factors rural elderly is higher than for urban elderly such as relative differences in patterns of land in nearly all cases (figure 2.2). Although and asset ownership across countries. higher rural poverty among the elderly mir- The correlates of old age poverty in East rors the spatial pattern of the general popula- Asia and Pacific are not surprising but provide tion, the size of the difference in rural-urban grounds for optimism regarding the future of elderly poverty rates across countries varies. poverty among older people. The correlates FIGURE 2.2 Poverty among the rural elderly remains significantly higher than for urban elderly Percentage below poverty line by age and rural or urban location, selected countries Consumption poverty a. Consumption-based poverty rates by age, urban, 2011–12 b. Consumption-based poverty rates by age, rural, 2011–12 40 40 35 35 30 30 Poverty rate (%) Poverty rate (%) 25 25 20 20 15 15 10 10 5 5 0 0 25–45 46–60 61–70 71–80 81+ 25–45 46–60 61–70 71–80 81+ Age group Age group Indonesia 2011 Mongolia 2011 Indonesia 2011 Mongolia 2011 Thailand 2011 Vietnam 2012 Thailand 2011 Vietnam 2012 Income poverty c. Income-based poverty rates by age, urban, 2009–2012 d. Income-based poverty rates by age, rural, 2009–2012 40 40 35 35 Poverty rate (%) 30 30 Poverty rate (%) 25 25 20 20 15 15 10 10 5 5 0 0 25–45 46–60 61–70 71–80 81+ 25–45 46–60 61–70 71–80 81+ Age group Age group China 2011 Philippines 2009 China 2011 Philippines 2009 Thailand 2011 Vietnam 2012 Thailand 2011 Vietnam 2012 Sources: Giles and Huang 2015 based on data from CHNS 2006, 2011; ThaiSES 2011; World Bank East Asia and Pacific Standardized Household Surveys, vari- ous years; and VHLSS 2012. Note: Poverty is based on per capita daily income less than US$1.25 (2005 dollar purchasing power parity). 50 LIVE LONG AND PROSPER of poverty for China, Indonesia, Thailand, also declined sharply in both Thailand and Vietnam are as follows: and Vietnam, where increases in average educational attainment of adult children • Not surprisingly, older people with high are associated with declines in the prob- school education or above are less likely ability that both urban and rural residents to be poor. Educated older people will will be income poor or consumption poor. have earned higher incomes over their • Pension receipt is associated with signifi- lives than their less educated peers and cant poverty reduction in both rural and will thus have higher savings and other urban areas. Table 2.1 shows the reduc- sources of wealth to support consump- tion in income poverty rates when older tion. Reflecting this, attainment of sec- people in a household are receiving pen- ondary education is associated with a sions. Reductions in income poverty are 4.0 percentage point reduction in the greater in rural areas of all countries, probability that consumption falls below where the probability of falling into the poverty line in rural China, and 8.4 poverty is greater and pension receipt is and 11.5 percentage point reductions in lower. As one might expect, the effect on urban and rural Indonesia, respectively. poverty is lower in Thailand given that Significant reductions in consumption the large majority are receiving modest poverty among the elderly are also associ- social pensions. Research conducted to ated with educational attainment in urban identify a causal relationship finds that and rural Thailand and rural Vietnam. the New Rural Pension Program in China If one considers the higher educational contributes to poverty reduction among attainment of current working-age adults, rural residents over age 60 (Zhang, Giles, the relationship between own education and Zhao 2014). and likelihood of falling into poverty sug- • Unlike the general adult population, one gests that the elderly are likely to be better should not expect a priori to observe off in the future. Although returns among positive or negative relationships between older workers may decline with increases employment status and poverty for older in the elder share with high school edu- people. Older workers with few resources cation or more, the more highly educated may be unable to retire, and continued future elderly will likely be employed in work may be important for staying out less physically demanding jobs and are of poverty. Alternatively, some elderly likely to adjust more readily to new work may continue to be poor even if they environments and tasks than will less edu- are actively farming. In urban China cated workers. and Indonesia, increases in the number • Apart from own education, older people who have had more children and more TABLE 2.1 Lower elderly poverty rates when older educated children are less likely to be people receive pensions poor. Declining family size with demo- Percent reduction in poverty rate among elderly graphic transition may lead to fewer receiving pensions by rural or urban location, selected sources of financial support for the countries elderly, but increases in average educa- tion promise to more than offset the Country Urban Rural consequences of demographic change. China −10.5 −12.7 Earlier work demonstrates this in urban Indonesia −8.8 −13.9 China and suggests that declines in sup- Thailand −0.6 −1.2 port as a result of the One Child Policy in Vietnam −5.6 −22.1 China will not necessarily lead to finan- Sources: Giles, Hu, and Huang 2015, based on data from CHARLS 2011, IFLS 2007, ThaiSES 2011, and VHLSS 2012. cial impoverishment of the elderly (Cai, Note: Results are significant at 1 percent level, except Thailand at 5 percent Giles, and Meng 2006). Family size has level. T H E W E L FA R E A N D L A B O R MA R K E T B E H AV I O R O F O L D E R P E O P L E I N E A S T A S I A A N D PAC I F I C 51 of other household members working physical functioning abilities in China and after age 60 are associated with a some- Indonesia, where these variables are avail- what higher probability of consumption able. Although individuals with serious poverty. Conditional on employment, disability may in some cases not be able however, the probability of being poor to earn income through work, this evi- decreases as the number of hours worked dence suggests that public transfers and increases in both countries. Increases in informal support, through financial and employment of and hours worked by both in-kind transfers, contribute to keeping younger (ages 16–59) and older (ages 60 some infirm elderly out of poverty. and above) members of the household are associated with decreases in income poverty in urban and rural areas of Other measures of well-being China, Indonesia, and Thailand (Giles, Functional limitations not only increase with Lei et al. 2015; Zhang, Giles, and Zhao age, but also are greater among the poor and 2014). This finding underscores a com- less educated elderly, as well as older women. mon observation that rural residents must Household surveys analyzed for this report, often continue working until relatively as well as other studies conducted in China, late ages because they lack the resources Thailand, and Vietnam, show that difficulties sufficient to retire (see Pang, de Brauw, with ADLs and instrumental ADLs (IADLs) and Rozelle 2004). increase with age, including sharper increases • As one might expect, difficulties in physical among the older-old (ages 75 and above) who functioning are associated with a higher experience difficulties with multiple ADLs probability of income poverty. Interest- ( figure 2.3) (Knodel, Prachuabmoh, and ingly, however, this finding does not carry Chayovan 2013; Qin 2007; Van Minh, Byass, over to consumption poverty. To examine and Wall 2010).2 In a range of East Asian the relationship between health status and and Pacific economies (such as China; Japan; likelihood of poverty, we include z-scores Singapore; Taiwan, China; Thailand; and calculated from questions about activities Vietnam), lower levels of income and education of daily living (ADLs) that characterized together with mountainous dwelling have been FIGURE 2.3 Difficulties with physical function and instrumental activities of daily living increase with age Instrumental activity of daily living z-scores among elderly, selected countries a. Men b. Women 1.0 1.0 Instrumental activities of 0.8 0.8 Instrumental activities of daily living (z-score) daily living (z-score) 0.6 0.6 0.4 0.4 0.2 0.2 0.0 0.0 −0.2 –0.2 −0.4 –0.4 50 60 70 80 50 60 70 80 Age (years) Age (years) China Indonesia Japan Korea, Rep. Sources: Giles, Hu, and Huang 2015 based on data from CHARLS 2011, IFLS 2007, JSTAR 2012, and KLoSA 2010. 52 LIVE LONG AND PROSPER found to be associated with higher incidence of As populations age, mental health functional limitations, and women consistently conditions also increase, in East Asia and report greater functional limitations than men Pacific and globally. Poor mental health, (Knodel, Prachuabmoh, and Chayovan 2013; particularly depression, is another aspect Liang, Liu, and Gu 2001; Lim et al. 2007; of older age and one that is more likely to Liu et al. 1995; Van Minh, Byass, and Wall affect not only the poorer, less educated 2010; Zimmer and Amornsirisomboon 2001; elderly, but also women. Depression is Zimmer et al. 1998). As shown in figure 2.4, due partly to psychosocial stressors such the household data analyzed for this report as loss, loneliness, and lack of social sup- confirm socioeconomic gradients with respect port and partly to increasing frailty and to functional limitations, including the decline illness. A correlation also exists between in IADL limitations for those with higher depression and comorbid physical con- incomes and higher educational attainment ditions. Worldwide, the prevalence of (with similar results for ADL not shown).3 depression in the elderly is estimated at FIGURE 2.4 Poorer and less educated elderly are more likely to face difficulty performing instrumental activities of daily living Instrumental activity of daily living z-scores among elderly, by education and income, selected countries a. Educational attainment, men b. Educational attainment, women 1.0 1.0 Instrumental activities of Instrumental activities of 0.8 0.8 daily living (z-score) daily living (z-score) 0.6 0.6 0.4 0.4 0.2 0.2 0.0 0.0 −0.2 −0.2 −0.4 −0.4 Elementary Elementary and High school Elementary Elementary and High school middle school and above middle school and above Education Education China Indonesia Japan Korea, Rep. c. Income per capita, men d. Income per capita, women 1.0 1.0 Instrumental activities of Instrumental activities of 0.8 0.8 daily living (z-score) daily living (z-score) 0.6 0.6 0.4 0.4 0.2 0.2 0.0 0.0 −0.2 −0.2 −0.4 −0.4 1 2 3 4 5 6 7 8 1 2 3 4 5 6 7 8 (poorest) (richest) (poorest) (richest) Quintile Quintile China Indonesia Korea, Rep. Sources: Giles and Huang 2015 based on data from CHARLS 2011, IFLS 2007, JSTAR 2011, and KLoSA 2010. T H E W E L FA R E A N D L A B O R MA R K E T B E H AV I O R O F O L D E R P E O P L E I N E A S T A S I A A N D PAC I F I C 53 about 10 percent for those living indepen- and age is much stronger in countries such dently, 25 percent for those with concur- as China and the Republic of Korea than in rent chronic diseases, and even higher for Indonesia and Japan. In addition to being a those in institutional care (McDougall et al. type of disability, depression is also a cause 2007). 4 Surveys in the region show that of higher suicide rates among the elderly rates of depression tend to increase with (Reynolds and Kupfer 1999). The sharp age and are greater among poor and less increase in depression rates with old age in educated elderly people (figure 2.5).5 Rates Korea is especially notable because suicide also tend to be higher for older women than rates among elderly Koreans are some of the older men, in part perhaps because of the highest in the world at 80 per 100,000 in higher likelihood of women living alone. 2011: the rate has increased fivefold in the The correlation between depression rates past two decades (Eyres 2014). Suicide rates FIGURE 2.5 Rates of depression tend to increase with age and are higher among poor households Percentage depressed, according to the CES-D 10 measure, by age, gender, and income, selected countries a. Age, men b. Age, women 60 60 Depression, CES-D 10 (%) Depression, CES-D 10 (%) 50 50 40 40 30 30 20 20 10 10 0 0 50 60 70 80 50 60 70 80 Years Years China Indonesia Japan Korea, Rep. c. Income per capita, men d. Income per capita, women 60 60 Depression, CES-D 10 (%) Depression, CES-D 10 (%) 50 50 40 40 30 30 20 20 10 10 0 0 1 2 3 4 5 1 2 3 4 5 (poorest) (richest) (poorest) (richest) Quintiles Quintiles China Indonesia Korea, Rep. Sources: Giles and Huang 2015 based on data from CHARLS 2011, IFLS 2007, JSTAR 2011, and KLoSA 2010. Note: CES-D 10 = Center for Epidemiologic Studies Depression Scale. 54 LIVE LONG AND PROSPER are also high among the elderly in Japan Consistent with descriptive results, individu- and are linked to loneliness and income- als with secondary education and above in related stressors. China, Indonesia, and Korea are less likely We a k e n i n g c o g n i t i v e f u n c t i o n , to be depressed, as are the elderly with ter- particularly dementia, is another aspect tiary education or greater in Japan. Whether of old age and a major source of disability because of a source of pride and a sense of and dependence. Dementia is not easy success in raising their children or simply to diagnose, and estimates of prevalence as a reflection of the prospect of higher pri- vary considerably because of measurement vate transfers, the elderly in both China and issues. In 2000, just under 7 percent of peo- Indonesia are happier when they have more ple ages 65 and above in Organisation for educated children.6 Economic Co-operation and Development In addition to their income support role (OECD) member countries were estimated for the elderly, public pensions also have to be affected by dementia, with incidence the potential to support other dimensions of the disease rising sharply with age: fewer of well-being. Receipt of a pension is asso- than 3 percent of those ages 65–69 years, ciated with higher body mass index in rural but almost 30 percent of those ages 85–89 Indonesia and with significant reductions years (WHO 2012). Alzheimer’s disease, the in the likelihood of depression among the most common form of dementia, accounted elderly in both rural and urban areas of China for between two-fifths and four-fifths of all and Indonesia. dementia cases in the OECD study. Evidence on dementia in East Asia and Pacific is more limited. In Korea, the prevalence of demen- Living arrangements and access tia among elderly people was estimated to to support among older people be 9.4 percent in 2013, of which over two- Co-residence of older people with adult chil- thirds were described as having mild to mod- dren is significant, although it varies consid- erate dementia. In Indonesia, analysis shows erably within East Asia and Pacific. Across a strong negative relationship between age the developing world, co-residence with adult and memory for men and women (for Korea, children is frequently viewed as an impor- Sunwoo 2013; for Indonesia, Witoelar, tant indicator of whether the elderly have Strauss, and Sikoki 2012). With life expec- access to both instrumental and financial tancy increases, a substantial rise in preva- support. In East Asia and Pacific, co-residence lence of dementia is anticipated, although of older people is quite high and gener- there is some evidence of compression of ally increases with age between age 50 and cognitive decline at older ages in OECD mid-60s (figure 2.6). For the elderly ages 60 countries (Christensen et al. 2009). and above, co-residence rates vary signifi- If one looks at multivariate findings, cantly across countries in the region and by living in poverty, having low education gender and urban-rural location within coun- levels, and being female are associated with tries. Rates are particularly high in countries deterioration in the subjective well-being of such as the Lao People’s Democratic Republic older people (higher CES-D 10 [Center for and Timor-Leste and, consistent with the Epidemiologic Studies Depression Scale] global pattern, tend to be lower in countries scores) in urban and rural areas of China, at higher per capita income levels (Evans and Indonesia, Japan, and Korea. In rural areas of Palacios 2015). China, Indonesia, and Korea, one observes a At the same time, in a number of East Asian significant association between increasing age and Pacific countries, elderly co-residence has and decreasing subjective well-being. In addi- declined significantly over time in the face of tion, women tend to have a higher incidence smaller families, higher incomes, and factors of depression than men in rural areas of such as migration (figure 2.7). Compounding China and Indonesia, in Japan, and in Korea. the demographic facts, privacy may be viewed T H E W E L FA R E A N D L A B O R MA R K E T B E H AV I O R O F O L D E R P E O P L E I N E A S T A S I A A N D PAC I F I C 55 FIGURE 2.6 Elderly co-residence rates are generally high in East Asia and Pacific but vary across countries and by gender and location within countries Share of elderly by age and gender living with an adult child, selected countries a. Cambodia, 2011 b. Indonesia, 2007 100 100 90 90 80 80 Share of population (%) Share of population (%) 70 70 60 60 50 50 40 40 30 30 20 20 10 10 0 0 40 50 60 70 80 40 50 60 70 80 Age (years) Age (years) c. Mongolia, 2010 d. Philippines, 2009 100 100 90 90 80 80 Share of population (%) Share of population (%) 70 70 60 60 50 50 40 40 30 30 20 20 10 10 0 0 40 50 60 70 80 40 50 60 70 80 Age (years) Age (years) e. Timor-Leste, 2007 f. Vietnam, 2012 100 100 90 90 80 80 Share of population (%) Share of population (%) 70 70 60 60 50 50 40 40 30 30 20 20 10 10 0 0 40 50 60 70 80 40 50 60 70 80 Age (years) Age (years) Urban, men Urban, women Rural, men Rural, women Sources: Giles and Huang 2015 based on data from IFLS 2007; World Bank East Asia and Pacific Standardized Household Surveys, various years; and VHLSS 2012. Note: Adult child is defined as child age 24 and above. as a normal good by both older and younger which finds both declining co-residence as generations. When the elderly and the young countries get richer (figure 2.8) and lower have the ability to live separately, they may co-residence among the richer elderly across simply choose to do so, without any con- countries (median co-residence in a sam- sequence for well-being, as suggested by ple of 61 developing countries in the mid- research from China (Giles and Mu 2007; Lei 2000s was over 83 percent for people ages et al. 2012). This view is also consistent with 60 and above in the poorest quintile but fell global evidence from developing countries, to only 64 percent for the richest quintile) 56 LIVE LONG AND PROSPER FIGURE 2.7 Co-residence rates of older adults with adult children have declined over time in China and Thailand Percentage by age of older adults living with adult children, various years a. China b. Thailand 100 100 90 90 80 80 Share of population (%) Share of population (%) 70 70 60 60 50 50 40 40 30 30 20 20 10 10 0 0 40 50 60 70 80 40 50 60 70 80 90 Age (years) Age (years) 1982 1990 2000 2011 1990 1996 2006 2011 Sources: Giles and Huang 2015 based on data from CHARLS 2011; China Census 1982, 1990, and 2000; and ThaiSES, various years. Note: Adult children are defined as ages 24 and above. FIGURE 2.8 As countries get richer, elderly co-residence rates tend by 2011, while in Korea, co-residence of to decline those ages 65 and above decreased from over Share of older people co-residing with non-elderly and GNI per capita, 80 percent in 1980 to well under 30 percent latest available year by 2010 (see Lee 2009). The speed of declines in co-residence contrasts with trends in a 18,000 number of developed countries, such as the GNI per capita (US$ PPP-adjusted) 16,000 United States, where elderly co-residence was 14,000 72 percent in 1880 and fell below 30 percent 12,000 only during the 1960s (Costa 1998). Most 10,000 East Asian and Pacific countries also have 8,000 notable variations between different groups 6,000 R2 = 0.509 of older people by rural-urban location and 4,000 gender. Cai, Giles, and Meng (2006) also find 2,000 that adult children in China are more likely 0 to reside with older parents who have higher 40 50 60 70 80 90 100 incomes and more housing wealth. Share of co-residing elderly (%) Although co-residence trends are impor- tant, they need to be interpreted with caution Source: Evans and Palacios 2015. in terms of potential effects on the welfare Note: GNI = gross national income; PPP = purchasing power parity. Elderly are defined as people 60 years and over, living with working-age adults whether or not related. of older people. On the one hand, evidence shows that co-residence has welfare benefits in East Asia and Pacific. In both China and (Evans and Palacios 2015). Since the mid- Indonesia, for example, poverty and depres- 1980s, co-residence with adult children in sion rates and ADL scores are higher for older Thailand has declined steadily to only 56 people who live separately from children liv- percent in 2011 (Knodel 2014b). Rates of co- ing in the same area than for those who live residence have also been declining in China with adult children or have migrant children and Korea. In China, the co-residence rate for living farther away. On the other hand, the those ages 65 to 70 fell from nearly 66 per- differences are not dramatic and, nota- cent in the early 1980s to roughly 43 percent bly, separate residence does not necessarily T H E W E L FA R E A N D L A B O R MA R K E T B E H AV I O R O F O L D E R P E O P L E I N E A S T A S I A A N D PAC I F I C 57 mean a decline in intergenerational support, income for urban and rural people ages 60 to as demonstrated in China by research using 85, respectively.7 With the notable exceptions CHARLS (China Health and Retirement of urban China and Mongolia, labor contin- Longitudinal Study) and noted by researchers ues to be the dominant source of income for in Southeast Asia (Giles, Lei et al. 2015; older people across the region, even in OECD Knodel 2014a). countries and in Korea (owing in part to its still immature pension system). Labor income also maintains its overall importance with Sources of financial support and regard to the elderly ages 70 and above in labor market behavior of older the region (see discussion below). In contrast, workers public transfers are a relatively minor share Labor income remains the dominant income of elderly income in most cases, including in source for people ages 60 and above in nearly rural China where the relative importance all East Asian and Pacific countries, and even of labor and transfers is very different from more so in rural areas in most of those coun- that in urban areas. The relative importance tries. Figure 2.9 shows the main sources of of public and private transfers is also notable FIGURE 2.9 Labor is the leading income source for people ages 60 to 85 in various East and Southeast Asian countries Percentage of total income of people ages 60 and above, in rural and urban areas, from labor, public and private transfers, and other sources, selected countries a. Urban 100 90 80 % of total income 70 60 50 40 30 20 10 0 p. a sia a te lia m s nd in di ne Re es na go ne Ch bo ila pi -L a, et on do lip a m or re Vi Th M Ca In Ko i Tim Ph b. Rural 100 90 80 % of total income 70 60 50 40 30 20 10 0 . a ia a lia ste m ep s d in di ne es na an go ,R e Ch bo pi n -L et ail on ea do ilip m or Vi Th r M Ca In Ko Tim Ph Labor Public transfer Private transfer Other income Sources: Giles and Huang 2015 based on data from CHARLS 2011; IFLS 2007; KLoSA 2010; ThaiSES 2011; World Bank East Asia and Pacific Standardized Household Surveys, various years; and VHLSS 2012. 58 LIVE LONG AND PROSPER in China, including very low private trans- additional public transfers for older people fers in several cases, but significant in other interact with private transfers and whether countries like Korea and the Philippines, such transfers simply displace private trans- and generally plays a slightly more signifi- fers at different income levels of older peo- cant role in rural areas in most countries. ple (so-called crowding out). Analysis of the Overall, the commonly held assumption responsiveness of private transfers to low that familial support is the dominant source incomes in households with elderly residents of income for elderly people in East Asia for this report shows the following: and Pacific is not borne out by the evidence, • For low-income elderly in China and although such support clearly remains impor- Indonesia, private transfers are on aver- tant, especially in settings like rural China age insufficient to lift older people out of and the Philippines where remittances are poverty. significant.8 • Transfers from adult children tend to be The role of private and public transfers significantly higher when some children in reducing elderly poverty varies across the are migrants (around twice the level on region, although private transfers remain average in China and Thailand for elderly important in nearly all East Asian and Pacific households at half the poverty line). countries. To gauge the relative importance • A larger number of adult children is of public and private transfers for older peo- associated with higher private transfers ple in the region, we follow an approach used to older people, but higher educational by Park, Lee, and Mason (2012) to examine attainment of children is also associated the likely rates of poverty in the absence of with higher transfers, suggesting that public and private transfers.9 Figure 2.10 smaller family size can be offset in part by shows poverty rates by age, including with better-educated (and paid) children. all earnings and transfers, without pub- • Most important, in many cases the effect lic transfers, without private transfers, and of increased transfers from nonfamily without all transfers. For China, Korea, and sources (including the state) on transfers Mongolia, public transfers play a more sig- from adult children is modest. nificant role in reducing poverty. In China, for example, eliminating public transfers These findings can be seen in table 2.2, would lead to an increase in the poverty rate which shows the elasticities of private transfers of a 70-year-old person by nearly 20 percent- to older family members to increased transfers age points (a 70 percent increase in poverty). from other sources at different income levels In contrast, elimination of private transfers of the elderly household. For both China and would lead to only a five percentage point Indonesia, the evidence of crowding out is increase in the elderly poverty rate. In the weak, with a reduction of only 3 to 9 percent Philippines, Thailand, and Vietnam, private in private transfers to the poorest elderly transfers play a significant role in reducing households in response to increased transfers poverty, though public transfers in Thailand from other sources, including social pensions. and Vietnam also play an important role The pattern also holds for elderly households in supporting the elderly. However, not all at and above the poverty line in both coun- elderly in the region benefit from substantial tries and is robust for households with and private and public transfers. In Cambodia without adult migrant children. Thailand is and Indonesia, both public and private trans- distinguishable in terms of the scale of crowd- fers play a relatively minor role in poverty ing out at low incomes, though in most cases, alleviation. the reduction in private transfers is no more Available evidence from East Asia and than one third. These findings have important Pacific suggests that the risk of public transfers implications for pension and transfer policies significantly crowding out private transfers as countries throughout the region seek to is modest. Policy makers must consider how expand coverage of public pension programs. T H E W E L FA R E A N D L A B O R MA R K E T B E H AV I O R O F O L D E R P E O P L E I N E A S T A S I A A N D PAC I F I C 59 FIGURE 2.10 Source of financial support can affect poverty rates in various East and Southeast Asian countries Percentage below the income poverty line with and without public and private transfers, selected countries a. Cambodia, 2011 b. China, 2011 c. Indonesia, 2007 100 100 100 90 90 90 80 80 80 Share of population living on or Share of population living on or Share of population living on or below the poverty line (%) below the poverty line (%) below the poverty line (%) 70 70 70 60 60 60 50 50 50 40 40 40 30 30 30 20 20 20 10 10 10 0 0 0 40 50 60 70 80 40 50 60 70 80 40 50 60 70 80 Age (years) Age (years) Age (years) d. Republic of Korea, 2010 e. Mongolia, 2010 f. Philippines, 2009 100 100 100 90 90 90 80 80 80 Share of population living on or Share of population living on or Share of population living on or below the poverty line (%) below the poverty line (%) below the poverty line (%) 70 70 70 60 60 60 50 50 50 40 40 40 30 30 30 20 20 20 10 10 10 0 0 0 40 50 60 70 80 40 50 60 70 80 40 50 60 70 80 Age (years) Age (years) Age (years) g. Thailand, 2011 h. Vietnam, 2012 100 100 90 90 80 80 Share of population living on or Share of population living on or below the poverty line (%) below the poverty line (%) 70 70 60 60 50 50 40 40 30 30 20 20 10 10 0 0 40 50 60 70 80 40 50 60 70 80 Age (years) Age (years) Income Income without public transfers Income without private transfers Income without all transfers Sources: Giles and Huang 2015 based on data from CHARLS 2011; IFLS 2007; KLoSA 2010; ThaiSES 2011; World Bank East Asia and Pacific Standardized Household Surveys, various years; and VHLSS 2012. Note: Poverty for China is defined as daily per capita income less than Y 6.3 (official poverty line). Poverty line for the Republic of Korea is defined as half of daily mean national per capita income in 2009: 20.8 won. Poverty lines for other countries are defined as per capita daily income (2005 dollars purchasing power parity) less than US$1.25. 60 LIVE LONG AND PROSPER TABLE 2.2 Private transfers to elderly people are not significantly affected by level of transfers from public sources in East and Southeast Asia Elasticities of private transfers to the elderly to increased transfers from other sources (percent change in private transfers), selected countries Country 0.5 × Poverty line Poverty line 2.0 × Poverty line 4.0 × Poverty line China With migrants −0.027 −0.103 −0.046 −0.058 Without migrants −0.062 −0.049 −0.009 −0.006 Indonesia With migrants 0.018 −0.086 0.114 0.035 Without migrants −0.092 0.010 0.010 −0.022 Thailand With migrants −0.177 −0.144 0.048 0.156 Without migrants −0.194 −0.278 −0.126 −0.034 Vietnam With migrants −0.379 −0.067 −0.105 −0.060 Without migrants −0.196 −0.077 −0.042 −0.041 Sources: Giles and Huang 2015 based on data from CHARLS 2011, IFLS 2007, and ThaiSES 2011. Another way to assess the relative impor- 2004 and precede the major expansion of tance of public and private transfers and self- social pensions, and data for the Philippines support is to look across the life cycle, using are from 1999 and precede the introduction the life-cycle deficit approach pioneered of a major conditional cash transfer pro- within the National Transfer Accounts gram that now covers around 20 percent of (NTA). NTA is a system of national accounts households.) Notably, NTA does not capture decomposed by age profile. It reflects a cohort differences. whole-of-life approach that captures the net Although public transfers currently play contribution of public and private trans- a relatively modest role in old age support in fers and asset-based reallocations (through East Asia and Pacific relative to other regions, channels such as housing, land, household social expectations of the relative roles of durables, asset income, and savings) to the the state, families, and retirees themselves in life-cycle deficit. The system is being more postretirement support are changing rapidly. widely adopted in Asia and globally and Policy makers and citizens in the region point provides a rich tool for cross-country com- to the important role of family support net- parisons.10 As figure 2.11 shows, asset-based works, though the earlier analysis shows limi- reallocations are the dominant source of tations of such networks in providing elderly support in Indonesia, the Philippines, and financial support. However, opinion surveys Thailand, whereas public transfers assume a also show that expectations are changing more significant role in richer Japan; Korea; rapidly, as shown in figure 2.12, where the and Taiwan, China, as one might expect. expected roles of self-support and the state The relative role of public transfers in East in financial support easily dominate expec- Asian and Pacific middle-income countries tations of support from adult children and is minor compared to other middle-income family. The results also indicate that expecta- countries such as Brazil, although famil- tions of the state are considerably higher in ial transfers are more significant except in the region’s middle-income countries than Japan. The results are a useful confirma- in its higher-income countries like Korea tion of the broad patterns seen in the analy- and Singapore where societal aging is more sis above, though some differences appear, advanced. For caregiving support, the expec- most likely driven by the time period of the tations of the state are also growing (and data. (For example, Thai NTA data are from are already higher in richer East Asian and T H E W E L FA R E A N D L A B O R MA R K E T B E H AV I O R O F O L D E R P E O P L E I N E A S T A S I A A N D PAC I F I C 61 FIGURE 2.11 The relative roles of public and private transfers and of savings and assets across the life cycle varies notably across East and Southeast Asian economies and between East and Southeast Asia and other regions Percentage of average life-cycle deficit for public and familial transfers and asset-based reallocations 130 120 110 100 90 % of average life-cycle deficit 80 70 60 50 40 30 20 10 0 –10 –20 –30 p. s es sia na an y a ico en nd il ne an di Re az at hi p ne ed pi In ila ex rm Br Ja St a, ,C ilip do Sw a M re Ge d Th an In ite Ko Ph iw Un Ta Public transfers Asset-based reallocations Familial transfers Source: National Transfer Accounts, http://www.ntaaccounts.org. Pacific economies), though the expectations A worrisome feature of early withdrawal of children and family remain significant. from the labor force in East Asia and Pacific is The persistent importance of labor income the likelihood of better-educated older people for older people in East and Southeast Asia is retiring early, in contrast to those in regions consistent with working until often very old such as Eastern Europe and Central Asia. As age, but population averages conceal signifi- figure 2.14 shows, those with college educa- cant differences between urban and rural areas tion are most likely to withdraw from work and by gender. As figure 2.13 shows, thinking early in China and countries in Southeast Asia. in terms of distinct urban and rural labor mar- Although the strength of the effect varies across kets for older people in the region makes sense. countries, those with the least education have Although on average people work until late in the longest working lives, often out of neces- life, many in urban areas withdraw from work sity because of low assets and savings and lim- relatively early. In cases like urban Chinese ited access to adequate old-age social security women, labor market withdrawal is very early, programs. At age 60, only 40 percent of men with only around one third still working at age with postsecondary education in China are 60. Nearly everywhere, urban women are the employed, but nearly 65 percent of men with least likely to work into old age, and gender only middle school education are still working. gaps in participation in both urban and rural Similar patterns hold for women, and for men areas are substantial (but see later discussion and women in Thailand. In Indonesia, gaps in on caring roles undertaken by older people). employment rates of middle school graduates 62 LIVE LONG AND PROSPER FIGURE 2.12 Attitudes on the preferred source of financial support in old age and on the expected source of personal care are shifting in East and Southeast Asian economies Percentage of adults reporting government, selves, or family as likely sources of support, selected economies a. Who, ideally, should be most responsible for providing income to retired people? 70 60 % of respondents 50 40 30 20 10 0 p. m na ia a sia re s nd ne in Re ys na po hi ne Ch ila pi ala et ,C a, ga do a ilip re Vi Th M an Sin In Ko Ph iw Ta b. Who, ideally, should be mostly responsible for providing personal care to retired people when they need help with everyday living or are sick or disabled? 70 60 50 % of respondents 40 30 20 10 0 p. m na ia a ia re s d ne in Re ys es an na po hi Ch pi ala on ail et ,C a, ga ilip re Vi d Th M an Sin In Ko Ph iw Ta Retirees themselves Government Grown children or other family members Source: Jackson and Peter 2015. and elderly with high school or college educa- people in work as they age. East and Southeast tion become evident at somewhat older ages, Asian countries commonly have self-employed although in Vietnam, the gaps across educa- rates of 90 percent for older rural men and tion groups are not as pronounced. These pat- women, most of whom work in agriculture or terns contrast with the EU13,11 where partici- related activities. In urban areas, an increase in pation rates of those with college education the self-employed share around age 60 reflects are 20 to 25 percentage points higher than the fact that self-employed urban workers, those with middle school education at age 60 often in commerce and trade, remain in the (Giles, Lei et al. 2015). workforce while employees from the formal As figure 2.15 shows, self-employment sector retire. In countries with a higher share of accounts for an increasing share of older urban residents in the informal self-employed T H E W E L FA R E A N D L A B O R MA R K E T B E H AV I O R O F O L D E R P E O P L E I N E A S T A S I A A N D PAC I F I C 63 FIGURE 2.13 Labor force participation rates vary by location and gender in various East and Southeast Asian countries Percentage in labor force by age, gender, and urban or rural location, selected countries a. Cambodia, 2011 b. China, 2011 100 100 90 90 80 80 % of cohort working % of cohort working 70 70 60 60 50 50 40 40 30 30 20 20 10 10 0 0 40 50 60 70 80 40 50 60 70 80 Age Age c. Indonesia, 2007 d. Republic of Korea, 2010 100 100 90 90 80 80 % of cohort working % of cohort working 70 70 60 60 50 50 40 40 30 30 20 20 10 10 0 0 40 50 60 70 80 40 50 60 70 80 Age Age e. Philippines, 2009 f. Thailand, 2011 100 100 90 90 80 80 % of cohort working % of cohort working 70 70 60 60 50 50 40 40 30 30 20 20 10 10 0 0 40 50 60 70 80 40 50 60 70 80 Age Age Urban, men Urban, women Rural, men Rural, women Sources: Giles, Hu, and Huang 2015 based on data from CHARLS 2011; IFLS 2007; KLoSA 2010; World Bank East Asia and Pacific Standardized Household Surveys, various years; and ThaiSES 2011. 64 LIVE LONG AND PROSPER FIGURE 2.14 Better-educated people tend to withdraw from the labor force earlier in East and Southeast Asia Employment rate by age, gender, and urban or rural location, selected countries a. China, men b. China, women 100 100 Employment rate (%) Employment rate (%) 80 80 60 60 40 40 20 20 0 0 40 50 60 70 80 40 50 60 70 80 Age (years) Age (years) c. Indonesia, men d. Indonesia, women 100 100 Employment rate (%) Employment rate (%) 80 80 60 60 40 40 20 20 0 0 40 50 60 70 80 40 50 60 70 80 Age (years) Age (years) e. Thailand, men f. Thailand, women 100 100 Employment rate (%) Employment rate (%) 80 80 60 60 40 40 20 20 0 0 40 50 60 70 80 40 50 60 70 80 Age (years) Age (years) g. Vietnam, men h. Vietnam, women 100 100 Employment rate (%) Employment rate (%) 80 80 60 60 40 40 20 20 0 0 40 50 60 70 80 40 50 60 70 80 Age (years) Age (years) Middle school or less High school College and above Sources: Giles, Hu, and Huang 2015 based on data from CHARLS 2011, IFLS 2007, ThaiSES 2011, and VHLSS 2012. T H E W E L FA R E A N D L A B O R MA R K E T B E H AV I O R O F O L D E R P E O P L E I N E A S T A S I A A N D PAC I F I C 65 FIGURE 2.15 Self-employment is an increasingly important source of work as people age in East Asia and Pacific Percentage of self-employed as a share of total in work, by age, gender, and rural or urban location, selected countries a. China b. Indonesia 100 100 90 90 who are self employed (%) who are self employed (%) 80 80 Share of population Share of population 70 70 60 60 50 50 40 40 30 30 20 20 10 10 0 0 40 50 60 70 80 40 50 60 70 80 Age (years) Age (years) c. Thailand d. Timor-Leste 100 100 90 90 who are self employed (%) who are self employed (%) 80 80 Share of population Share of population 70 70 60 60 50 50 40 40 30 30 20 20 10 10 0 0 40 50 60 70 80 40 50 60 70 80 Age (years) Age (years) Urban, men Urban, women Rural, men Rural, women Sources: Giles, Hu, and Huang 2015 based on CHARLS 2011; IFLS 2007; ThaiSES 2011; and World Bank East Asia and Pacific Standardized Household Survey for Timor-Leste, various years. sector—Timor-Leste is an extreme case—one 30–45 hours per week depending on the coun- does not see as big an increase in the self- try. The situation in rural areas is even more employed share of older urban residents. striking, with rural men continuing to work in For those who continue working at older most countries 30–40 hours per week even at ages in East Asia and Pacific, hours of work age 75 and women working between 20 and remain quite high. In developed countries, 35 hours per week at the same age. Given the retirement is often a gradual process: workers nature of most rural work, such long hours and may reduce work hours at current employers, high participation rates confirm the impression move to work arrangements requiring fewer of many rural people in the region having to hours per week, or even transition in and out “work till they drop.” of retirement. In contrast, in much of East Asia Differences exist across countries also. and Pacific, hours of work for older people con- In urban China and Vietnam, hours of work ditional on working remain quite high in both for elderly still working (primarily the self- rural and urban areas (figure 2.16). For people employed) tend to decline little and then drop working at age 65 in the six countries shown, off sharply at age 70 and age 65, respectively, urban men work on average 40 hours or more in contrast to the more gradual decline in per week, and all women work on average hours worked in other countries. In rural 66 LIVE LONG AND PROSPER FIGURE 2.16 Those who continue to be employed in older age work long hours until advanced ages in East and Southeast Asia Hours of work per week among those working, by age, gender, and rural or urban location, selected countries a. Urban men b. Urban women 60 60 Hours worked per week Hours worked per week 50 50 40 40 30 30 20 20 10 10 0 0 45 50 55 60 65 70 75 80 85 45 50 55 60 65 70 75 80 85 Age (years) Age (years) China Indonesia Japan Korea, Rep. Vietnam Philippines c. Rural men d. Rural women 70 70 60 60 Hours worked per week Hours worked per week 50 50 40 40 30 30 20 20 10 10 0 0 45 50 55 60 65 70 75 80 85 45 50 55 60 65 70 75 80 85 Age (years) Age (years) China Indonesia Korea, Rep. Vietnam Philippines Sources: Giles, Hu, and Huang 2015 based on data from CHARLS 2011, FIES 2009, IFLS 2007, JSTAR 2011, KLoSA 2010, and VHLSS 2012. Note: Employment is defined as those who have worked for more than one hour weekly on average. Vietnam, the decline in work hours exhibits person and his or her spouse, (c) demands for a steeper slope, and hours worked tend to be caring for grandchildren or older parents, and well below those of other countries. In con- (d) work status of the spouse. Each driver is trast, in China men and women still working discussed in turn. in agriculture put in long hours at relatively First, availability of a pension appears to old ages, raising the question whether the have a strong association with lower labor absence of migrant adult children contributes force participation when the pension pro- to a scarcity of labor and higher labor input vides adequate financial protection. For of rural elderly. example, after controlling for age, educa- What are the primary drivers of the diverse tion, health status of respondent and spouse, labor market behavior of older people in East wealth, and family demographic characteris- Asia and Pacific, both within and across coun- tics, we found that receiving a formal sector tries? Potential drivers include (a) availability pension is associated with 52 and 54 percent of a pension, (b) health status of the older reductions in the probability that men and T H E W E L FA R E A N D L A B O R MA R K E T B E H AV I O R O F O L D E R P E O P L E I N E A S T A S I A A N D PAC I F I C 67 women, respectively, are working in urban most likely because of the immaturity of the China. The behavioral response of older pension system and modesty of benefits for workers to mandatory retirement provisions the current elderly cohort;12 and in Japan, can be seen in the hazard rates in figure 2.17, because availability of disability and survivor which contrasts the smooth glide into with- benefits is likely to have larger effects. For drawal from work in rural areas with spikes the Japanese, however, receipt of a pension is in retirement at official pension ages in urban associated with 3.0 hours less work per week areas (ages 50 and 55 for blue-collar and for those who work, and receipt of disabil- white-collar women, respectively, and age ity benefits is associated with 6.0 hours less 60 for men). Although China has no explicit work per week for men and 3.3 hours less provisions against working beyond retire- for women. ment age, the relative generosity of the urban As would be expected, the generosity of employee pension combined with manda- pensions is another important factor in labor tory retirement makes them much less likely. market effects. The continued work after Similarly, one sees a pronounced (though retirement of residents receiving social pen- weaker) negative effect on employment of sions in China, the Philippines, and Thailand men in urban and rural Indonesia, with 19 and the high labor supply rates of workers and 25 percent reductions, respectively, in over age 55 in Japan and Korea suggest that probability of working. In contrast, in Japan retirement benefits are simply not generous and Korea, there is a negligible effect of enough to lead to exit from work. For exam- pension receipt for either gender: in Korea, ple, figure 2.18 shows that over 80 percent FIGURE 2.17 Pension availability and mandatory retirement rules have strong effects on labor force behavior of older Chinese workers Retirement hazard rates by gender, age, and urban or rural location, China, 2011 a. Retirement hazard rate of urban retirees 100 80 Percent 60 40 20 0 45 50 55 60 65 70 75 80 Retirement age (years) b. Retirement hazard rate of rural retirees 100 80 Percent 60 40 20 0 45 50 55 60 65 70 75 80 Retirement age (years) Male Female Sources: Giles, Lei et al. 2015 based on data from CHARLS 2011. 68 LIVE LONG AND PROSPER FIGURE 2.18 Labor force participation is influenced by pension receipt in select East and Southeast Asian countries Percentage of people in labor force by age, gender, and urban or rural location, and percentage eligible for pensions, selected countries a. China: Urban male b. China: Urban female 100 100 Share of population (%) Share of population (%) 80 80 60 60 40 40 20 20 0 0 45 50 55 60 65 70 75 80 85 45 50 55 60 65 70 75 80 85 Age (years) Age (years) c. China: Rural male d. China: Rural female 100 100 Share of population (%) 80 Share of population (%) 80 60 60 40 40 20 20 0 0 45 50 55 60 65 70 75 80 85 45 50 55 60 65 70 75 80 85 Age (years) Age (years) Pension eligible Working total e. Indonesia: Urban male f. Indonesia: Urban female 100 100 Share of population (%) Share of population (%) 80 80 60 60 40 40 20 20 0 0 45 50 55 60 65 70 75 80 85 45 50 55 60 65 70 75 80 85 Age (years) Age (years) g. Indonesia: Rural male h. Indonesia: Rural female 100 100 Share of population (%) Share of population (%) 80 80 60 60 40 40 20 20 0 0 45 50 55 60 65 70 75 80 85 45 50 55 60 65 70 75 80 85 Age (years) Age (years) Pension eligible Working total T H E W E L FA R E A N D L A B O R MA R K E T B E H AV I O R O F O L D E R P E O P L E I N E A S T A S I A A N D PAC I F I C 69 FIGURE 2.18 Labor force participation is influenced by pension receipt in select East and Southeast Asian countries (continued) Percentage of people in labor force by age, gender, and urban or rural location, and percentage eligible for pensions, selected countries i. Japan: Male j. Japan: Female 100 100 80 80 Share of population (%) Share of population (%) 60 60 40 40 20 20 0 0 50 55 60 65 70 75 80 50 55 60 65 70 75 80 Age (years) Age (years) Receiving other social assistance Receiving old age pension Working total k. Republic of Korea: Urban male l. Republic of Korea: Urban female 100 100 Share of population (%) Share of population (%) 80 80 60 60 40 40 20 20 0 0 45 50 55 60 65 70 75 80 85 45 50 55 60 65 70 75 80 85 Age (years) Age (years) m. Republic of Korea: Rural male n. Republic of Korea: Rural female 100 100 Share of population (%) Share of population (%) 80 80 60 60 40 40 20 20 0 0 45 50 55 60 65 70 75 80 85 45 50 55 60 65 70 75 80 85 Age (years) Age (years) Pension eligible Working total figure continues next page 70 LIVE LONG AND PROSPER FIGURE 2.18 Labor force participation is influenced by pension receipt in select East and Southeast Asian countries (continued) Percentage of people in labor force by age, gender, and urban or rural location, and percentage eligible for pensions, selected countries o. Thailand: Urban male p. Thailand: Urban female Share of population (%) Share of population (%) 100 100 80 80 60 60 40 40 20 20 0 0 45 50 55 60 65 70 75 80 45 50 55 60 65 70 75 80 Age (years) Age (years) q. Thailand: Rural male r. Thailand: Rural female Share of population (%) Share of population (%) 100 100 80 80 60 60 40 40 20 20 0 0 45 50 55 60 65 70 75 80 45 50 55 60 65 70 75 80 Age (years) Age (years) s. Vietnam: Urban male t. Vietnam: Urban female Share of population (%) Share of population (%) 100 100 80 80 60 60 40 40 20 20 0 0 + 9 9 4 4 9 4 9 9 9 4 4 9 4 9 + –5 –4 –5 –6 –6 –7 –7 –5 –4 –5 –6 –6 –7 –7 80 80 45 50 55 60 65 70 75 50 55 45 60 65 70 75 Age group Age group u. Vietnam: Rural male v. Vietnam: Rural female Share of population (%) Share of population (%) 100 100 80 80 60 60 40 40 20 20 0 0 9 4 59 4 9 4 9 + 49 4 9 64 9 4 9 + –4 –5 –6 –6 –7 –7 80 –5 –5 –6 –7 –7 80 – – – 45 50 55 60 65 70 75 45 50 55 60 65 70 75 Age group Age group Pension eligible Other social assistance Working total Sources: Giles, Hu, and Huang 2015 based on data from CHARLS 2011, IFLS 2007, JSTAR 2011, KLoSA 2010, ThaiSES 2011, and VHLSS 2012. T H E W E L FA R E A N D L A B O R MA R K E T B E H AV I O R O F O L D E R P E O P L E I N E A S T A S I A A N D PAC I F I C 71 of Thailand’s population receives a pension reductions in probability of employment for by age 61, but employment rates for rural urban men and women, respectively, and 3.9 men are nevertheless well above 50 percent and 2.2 percent reductions for rural men and for those as old as age 75. Even in Thailand’s women, respectively. Japan shows similar pat- urban areas, both men and women continue terns. Poor health status is also associated to work at much higher rates than their coun- with fewer hours of work for those elderly terparts in urban China, where formal sec- continuing to work in rural China and urban tor pensions are more generous. In Vietnam, Indonesia (Giles, Lei et al. 2015). an early exit of both men and women from Third, provision of care to children and work—corresponding to pension eligibility elderly also affects the retirement behavior of ages—is evident, even though rates of pension women in East Asia and Pacific. A range of receipt are not as high as in China. studies shows that provision of family-based Second, lower health status, as mea- care is associated with fewer working hours sured through ADLs and IADLs, is strongly and a higher probability of exit from the work- associated with reductions in employment force (Jacobs et al. 2014; Meng 2011; Van (table 2.3).13 According to measures of ADL Houtven, Coe, and Skira 2013). For China, and IADL, in China, an increase in one dis- responses to CHARLS suggest that women ability is associated with 5.2 and 6.0 percent over the age of 50 contribute significant reductions in probability of employment amounts of time to provision of both child for urban men and women, respectively. In care (largely of grandchildren) and elder care. China’s rural areas, the same increase in dis- The shares and hours of people in China ages ability results in 4.2 and 3.8 percent reduc- 45 and above who provide care to grandchil- tions in the probability of working for men dren and older people are shown in table 2.4. and women, respectively. Because most work Overall, 28 percent of individuals ages 45 and in rural areas is more physically demanding, above provide care to grandchildren, which the weaker effect of poor health on the prob- rises to more than 45 percent for women ages ability of working in rural areas is striking 55 to 59. (Nearly 14 percent of individuals and underscores the difficulty of retiring for ages 45 and above care for older parents and China’s rural residents. A similar pattern is parents-in-law, which increases to 16 percent also witnessed in Indonesia. An increase in dis- for women ages 55 to 59.) For people over ability is associated with 3.0 and 5.7 percent age 45 who provide some care, average time reductions in probability that urban men and caring for grandchildren is equivalent to a full- women, respectively, will be employed, and time job (43 hours for women and just under 5.0 and 5.5 percent reductions for rural men 39 hours for men). Time spent providing and women, respectively. In Korea, where care for older people is also not insignificant reported disabilities are considerably lower (equivalent to a 20-hour per week, part-time than in China or Indonesia, the contribu- job on average). Although a higher share of tion of lower health status to exit from work women provide care and work longer hours is also less pronounced. Decreases associ- on average, the share of men providing care ated with disability are 4.5 and 2.3 percent and their time intensity is not dramatically TABLE 2.3 Poor health results in lower probability of work at older ages in East and Southeast Asia Percentage of reduction in employment rate when in poor health, by gender and urban or rural location Country Urban men Urban women Rural men Rural women China −5.2 −6.0 −4.2 −3.8 Indonesia −3.0 −5.7 −5.0 −5.5 Korea, Rep. −4.5 −2.3 −3.9 −2.2 Sources: Giles, Hu, and Huang 2015 based on data from CHARLS 2011, IFLS 2007, and KLoSA 2010. Note: All results are significant at 1 percent level. 72 LIVE LONG AND PROSPER TABLE 2.4 Hours spent providing care for children and elderly among Chinese cohorts ages 45 and above Hours worked by care provider Age of care Women Men provider Share providing Recipient of care (years) some care (%) All Urban Rural All Urban Rural Parents and parents-in-law 45–49 29 18.3 17.0 20.4 17.5 15.1 20.4 50–54 25 16.3 17.5 14.5 14.7 13.3 16.0 55–59 16 16.6 16.5 16.6 15.1 17.2 12.6 60–64 8 20.4 20.9 20.0 26.5 21.1 32.2 65 and above 3 24.1 32.9 14.9 23.8 23.1 24.7 Grandchildren 45–49 15 43.4 42.6 44.1 36.1 45.1 26.5 50–54 29 43.2 47.1 39.8 35.4 29.9 39.2 55–59 38 45.2 43.1 47.6 38.9 36.6 41.2 60–64 36 43.7 43.0 44.5 40.8 38.3 43.1 65–69 32 38.4 33.6 42.7 40.8 38.8 43.2 70 and above 18 33.1 30.8 36.0 38.2 36.0 40.5 Sources: Giles, Hu, and Huang 2015 based on data from CHARLS 2011. lower, suggesting that couples frequently 40 and above globally are engaged in volun- engage in joint provision of care. teer work at different ages. However, shares The contribution of older people to child are lower in most Asian countries (with the care is also underscored in multivariate notable exceptions of Japan, Korea, and labor supply analysis. In China, presence the Philippines) and sharply lower in China of a child under age six in the household is and India. The deficit is particularly notable associated with 17.6 and 3.0 percent reduc- among those ages 70 and above, suggesting tions in the probability of employment for a broader challenge of social participation urban and rural women, respectively. The for the oldest elderly. Offsetting low average presence of children or elderly showed no participation somewhat, in most East Asian effect on male employment. In Indonesia, and Pacific economies shown, those who vol- the presence of children ages 6 to 12 is asso- unteer work at or above the global average ciated with 11.7 and 9.7 percent reductions number of hours (Leeson and Harper 2007). in the probability of employment for urban Participation in volunteer work is important and rural women, respectively. In Korea, not only for its direct contribution to societal by contrast, the presence of children ages 6 welfare, but also as a channel for sustaining to 12 has a more significant negative effect social participation of older people. on labor force participation of older urban A strong correlation also exists for the men. However, this effect could be driven by employment of older spouses, suggesting already low participation of urban women. that spouses have joint labor supply deci- In Japan, the probability that a woman will sions and may have a preference for retiring be employed is 8 percent lower in house- together (table 2.5). The most pronounced holds with an older resident over age 80 effect is in urban China, where employ- (Giles, Lei et al. 2015). ment of a spouse is associated with 40 and Finally, in addition to care for family 58 percent increases in probability that men members, older people in East and Southeast and women are working, respectively, and Asia may contribute to lower labor force in Japan, especially for women. In rural participation because of volunteer work, China, a working spouse is associated with although the scale of involvement is gener- 23 and 33 percent increases in probabilities ally lower than global averages. Figure 2.19 that older men and women, respectively, shows that around 20 percent of people ages are working.14 In Indonesia and Korea, the T H E W E L FA R E A N D L A B O R MA R K E T B E H AV I O R O F O L D E R P E O P L E I N E A S T A S I A A N D PAC I F I C 73 FIGURE 2.19 Older people in East and Southeast Asia are a significant source of volunteer work Percentage of population by age participating in volunteer work, selected economies 40 35 30 Share of population (%) 25 20 15 10 5 0 40s 50s 60s 70s Age cohort World India Philippines Malaysia China Taiwan, China Singapore Korea, Rep. Hong Kong SAR, China Japan Source: Leeson and Harper 2007. TABLE 2.5 Influence of employment of a spouse on associated with longer working hours for probability of own employment rural men and women in China, Indonesia, Percentage change in probability of own employment and Korea, and for all men in Japan. with working spouse Joint decisions by spouses on labor supply, Country Men Women or retirement, are of considerable policy relevance when one considers incentives China to increase working ages of men and women, Urban 48 58 Rural 23 33 particularly in countries where male and Indonesia female retirement ages differ significantly Urban 11 19 (as in China and Vietnam). Rural 9 12 Japan 24 56 Conclusion Korea, Rep. Urban 13 21 The diverse patterns of well-being, labor mar- Rural 14 28 ket behavior, and sources of support across and Sources: Giles, Hu, and Huang 2015 based on data from CHARLS 2011. within East Asian and Pacific countries mean that policies will need to take into account dif- association between own work status and a fering initial conditions. Policy makers will working spouse is also significant, but lacks need to pay close attention to various aspects the same effect on probability of employ- of elderly welfare, a task made more complex ment. In terms of hours worked for those by the diversity of situations across countries who continue to work, a working spouse is and between different groups of elderly within 74 LIVE LONG AND PROSPER countries—rich and poor, rural and urban, measures of ability to perform activities and healthy and frail or depressed, as well important for independent living, such as as men and women and formal and informal shopping, preparing meals, and managing the workers. For example, improving the well- household budget. 4. These percentages exist even though median being of older workers who have spent their self-rated happiness tends to be higher for lives working in agriculture or the informal those above age 50 than for adults between sector and facilitating an ability to retire could ages 19 and 49 (Deaton 2007). be addressed by broadening access to pensions 5. Subjective well-being was calculated using the and other public support in old age. However, Center for Epidemiologic Studies Depression these interventions will need to be balanced Scale (CES-D 10), which is constructed from against potential disincentives to work at 10 questions commonly used to study the older ages, especially for the urban elderly. incidence of depression. Thus, a key challenge for policy makers lies 6. One should be careful in assigning causality in extending support for the elderly at risk of to this relationship because having a happier poverty while also incentivizing later retire- disposition throughout one’s life may have contributed to raising successful (and ment among older urban and formal sector educated) children. workers ages 50 to 69. 7. Results for women in urban and rural Building on the microlevel discussion from areas are quite similar in all countries (2–5 this chapter, the following chapters examine percentage points different and lower in most the policy challenges with respect to the mac- cases) except Korea, where women’s labor roeconomy, labor policies, pensions, health income share is 11–15 percentage points care, and long-term care. Chapter 3 focuses lower than that of men. on the macroeconomic and fiscal implica- 8. Taking account of nonincome support is tions of rapid aging in East Asia and Pacific. important because older people living with As noted, behavioral and policy responses to adult children in China, Indonesia, and aging will be crucial in determining the nature Thailand receive lower new transfers from adult children but have regular nonfinancial and scale of effects on growth and the fiscal support (see Giles, Lei et al. 2015). positions of East Asian and Pacific countries. 9. Private transfers include all cash and in-kind The microlevel insights from this chapter are transfers provided to households with elderly useful for anticipating potential responses. members, and public transfers include all forms of poverty assistance and pensions. The approach assumes no labor behavioral Notes responses. 1. Because of data limitations, old age poverty 10. See Mason et al. (2006) for a description of is measured according to households with methodology and for subsequent publications elderly members, households with only elderly edited by Andrew Mason and Ronald Lee members, and others with mixed composition. for cross-country applications. NTA also To the extent that intrahousehold resource provided an important underpinning for the allocation is biased toward younger members, Asian Development Bank’s 2012 volume on the welfare of individual elderly persons may aging and growth in Asia (Park, Lee, and be lower. Mason 2012). 2. ADLs are basic self-care tasks, and IADLs 11. The 13 new members of the European Union are tasks that allow an individual to function (EU13) are Bulgaria, Croatia, Cyprus, the independently. Unlike ADLs, IADLs can be Czech Republic, Estonia, Hungary, Latvia, delegated to someone else and thus are not Lithuania, Malta, Poland, Romania, the uniformly carried out by everyone themselves. Slovak Republic, and Slovenia. In addition, some IADLs relate to functioning 12. Research conducted in Korea showed that within a community rather than only within over 92 percent of respondents expressed the home. a desire to take their pension after the 3. ADLs are measures of physical functioning, eligible age, in part to improve adequacy ranging from the ability to walk 1 kilometer (Korea National Pension Research Institute to the ability to get out of bed. IADLs are 2012). T H E W E L FA R E A N D L A B O R MA R K E T B E H AV I O R O F O L D E R P E O P L E I N E A S T A S I A A N D PAC I F I C 75 13. If one looks at spouse health status using an the Developing World.” Social Protection index of spouse ADLs and IADLs, there is and Labor Policy Note 17, World Bank, evidence of a relatively small added worker Washington, DC. effect of poor spouse health status on Eyres, Harry. 2014. “The Slow Lane: The employment of men and women in urban and Cruel Cost of the Crisis.” Financial Times, rural areas of China, but no apparent effect in July 25. http://www.ft.com/intl/cms/s/2 other countries. / b4cbef16-10c9-11e4-b116-00144feabdc0 14. This effect is seen in OECD countries also. .html#axzz3iMcpEjMf. Schirle (2008) finds that increases in wives’ FIES (Family Income and Expenditure Survey). participation in the labor force can explain 2009. Philippine Statistics Authority, Manila. one-fourth, one-half, and one-third of the http://psa.gov.ph/content /family- income-and increase in older married men’s labor force -expenditure-survey-fies. participation in the United States, Canada, Giles, John, Yuqing Hu, and Yang Huang. 2015. and the United Kingdom, respectively. “Understanding the Retirement Decision in Aging East Asia.” Background paper for the East Asia and Pacific regional report on aging, References World Bank, Washington, DC. Cai, Fang, John Giles, and Xin Meng. 2006. “How Giles, John, and Yang Huang. 2015. “Are the Well Do Children Insure Parents against Low Elderly Left behind in a Time of Rapid Retirement Income? An Analysis Using Survey Demographic and Economic Change? Data from Urban China.” Journal of Public A Comparative Study of the Poverty and Well- Economics 90 (12): 2229–55. doi:10.1016/j Being of East Asia’s Elderly.” Background .jpubeco.2006.03.004. paper for the East Asia and Pacific regional CHARLS (China Health and Retirement report on aging, World Bank, Washington, DC. Longitudinal Study). 2011. National School of Giles, John, Xiaoyan Lei, Yafeng Wang, and Development, Peking University. http://charls Yaohui Zhao. 2015. “One Country, Two .ccer.edu.cn/en. Systems: Evidence on Retirement Patterns in China Census. 1982. Integrated Public Use China.” World Bank, Washington, DC, and Microdata Series, International, series 6.3, Peking University, Beijing. Minnesota Population Center University of Giles, John, and Ren Mu. 2007. “Elderly Parent Minnesota, Minneapolis. Health and the Migration Decisions of Adult ———. 1990. China Population and Information Children: Evidence from Rural China.” Research Center 1% Sample of the 1990 China Demography 44 (2): 265–88. Population Census Data, China Archive, Texas IFLS (Indonesia Family Life Survey). 2007. RAND, A&M University, College Station. Santa Monica, CA. http://www.rand.org/labor ———. 2000. National Bureau of Statistics 2000 /FLS/IFLS.html. Fifth Population Census Data, China Statistics Jackson, Richard, and Tobias Peter. 2015. “From Press, Beijing. Challenge to Opportunity: Wave 2 of the CHNS (China Health and Nutrition Survey). East Asia Retirement Survey.” Global Aging Various years. http://www.cpc.unc.edu Institute, Alexandria, VA. /projects/china/data. Jacobs, Josephine, Audrey Laporte, Courtney Van Christensen, Kaare, Gabriele Doblhammer, Houtven, and Peter Coyte. 2014. “Caregiving Roland Rau, and James Vaupel. 2009. “Aging Intensity and Retirement Status in Canada.” Populations: The Challenges Ahead.” Lancet Social Science and Medicine 102: 74–82. 374 (9696): 1196–208. JSTAR (Japanese Study of Aging and Retirement). Costa, Dora L. 1998. The Evolution of Retirement: Various years. Research Institute of Economy, An American Economic History, 1880–1990. Trade, and Industry, Hitotsubashi University Chicago: University of Chicago. and University of Tokyo, Tokyo. http://www Deaton, Angus. 2007. “Income, Aging, Health, .rieti.go.jp/en/projects/jstar/. and Wellbeing around the World: Evidence KLoSA (Korean Longitudinal Study of Aging). from the Gallup World Poll.” NBER Working 2010. Korea Labor Institute, Seoul. http:// Paper 13317, National Bureau of Economic www.kli.re.kr/klosa/en/about/introduce.jsp. Research, Cambridge, MA. Knodel, John. 2014a. “Is Intergenerational Evans, Brooks, and Robert Palacios. 2015. “An Solidarity Really on the Decline? Cautionary Examination of Elderly Co-residence in Evidence from Thailand.” Asian Population 76 LIVE LONG AND PROSPER Studies 10 (2): 176–94. doi:10.1080/1744173 McDougall, Fiona A., Fiona E. Matthews, Kari 0.2014.902160. Kvaal, Michael E. Dewey, and Carol Brayne. ———. 2014b. “The Situation of Older Persons in 2007. “Prevalence and Symptomatology Myanmar: An Overview.” BOLD 24 (3): 9–16. of Depression in Older People Living in Knodel, John, Vipan Prachuabmoh, and Napaporn Institutions in England and Wales.” Age and Chayovan. 2013. The Changing Well-Being Ageing 36 (5): 562–68. doi:10.1093/ageing of Thai Elderly: An Update from the 2011 /afm111. Survey of Older Persons in Thailand. London: Meng, Annika. 2011. “Informal Caregiving and HelpAge International. the Retirement Decision.” German Economic Korea National Pension Research Institute. 2012. Review 13 (3): 307–30. “Middle- and Old-Aged Koreans’ Preparation Pang, Lihua, Alan de Brauw, and Scott Rozelle. Realities for and Expectations about Their Old 2004. “Working until You Drop: The Elderly Age.” Seoul. http://english.nps.or.kr/jsppage of Rural China.” China Journal 52: 73–96. /english/research/research_01.jsp. Park, Donhhyuan, Sang-Hyop Lee, and Andrew Lee, Chulhee. 2009. “Labor Force Participation Mason, eds. 2012. Aging, Economic Growth, of Older Males in Korea: 1955–2005.” NBER and Old-Age Security in Asia. Manila: Asian Working Paper 14800, National Bureau of Development Bank. Economic Research, Cambridge, MA. Qin, Huali. 2007. “Adding Life to Years: Predicting Leeson, George, and Sarah Harper. 2007. “The Subjective Quality of Life among Chinese Global Ageing Survey (GLAS): Ageing and Oldest-Old.” Gerontology thesis, Georgia Later Life—Hong Kong and Asia.” Research State University, Atlanta, May 4. http:// Paper 307, Oxford Institute of Ageing, scholarworks.gsu.edu/gerontology_theses/4. University of Oxford, Oxford, U.K. Reynolds, Charles F., and David J. Kupfer. 1999. Lei, Xiaoyan, John Strauss, Meng Tian, and “Depression and Aging: A Look to the Yaohui Zhao. 2012. “Living Arrangements Future.” Psychiatric Services 50 (9): 1167–72. of the Elderly in China: Evidence from doi:10.1176/ps.50.9.1167. CHARLS.” IZA Discussion Paper 6249, Schirle, Tammy. 2008. “Why Have the Labor Institute for the Study of Labor, Bonn, Force Participation Rates of Older Men Germany. http://papers.ssrn.com/sol3/papers Increased since the Mid-1990s?” Journal of .cfm?abstract_id=1981205. Labor Economics 26 (4): 549–94. Liang, Jersey, Xian Liu, and Shengzu Gu. 2001. Sunwoo, Duk. 2013. “Economic Costs and “Transitions in Functional Status among Opportunities for Ageing and Healthcare.” Older People in Wuhan, China: Socioeconomic Paper presented at the Global Forum on Differentials.” Journal of Clinical Epidemiology Innovation for Ageing Populations, Kobe, 54 (11): 1126–38. Japan, December 11. Lim, Wei-Yen, Stefan Ma, Derrick Heng, Vineta ThaiSES (Thai Household Social and Economic Bhalla, and Suok K. Chew. 2007. “Gender, Surveys). Various years. National Statistical Ethnicity, Health Behaviour, and Self-Rated Office, Bangkok. http://web.nso.go.th/survey Health in Singapore.” BMC Public Health /house_seco/socio.htm. 7 (1): 184. doi:10.1186/1471-2458-7-184. Van Houtven, Courtney, Norma Coe, and Meghan Liu, Xian, Jersey Liang, Naoko Muramatsu, and Skira. 2013. “The Effect of Informal Care Hidehiro Sugisawa. 1995. “Transitions in on Work and Wages.” Journal of Health Functional Status and Active Life Expectancy Economics 32 (2013): 240–52. among Older People in Japan.” Journals of Van Minh, Hoang, Peter Byass, and Stig Wall. Gerontology Series B: Psychological Sciences 2010. “Multilevel Analysis of Effects of and Social Sciences 50B (6): S383–94. Individual Characteristics and Household doi:10.1093/geronb/50B.6.S383. Factors on Self-Rated Health among Older Mason, Andrew, Ronald Lee, An-Chi Tung, Mun- Adults in Rural Vietnam.” Geriatrics and Sim Lai, and Tim Miller. 2006. “Population Gerontology International 10 (2): 209–15. Aging and Intergenerational Transfers: VHLSS (Vietnam Household Living Standards Introducing Age into National Accounts.” Survey). 2012. General Statistics Office of NBER Working Paper 12770, National Bureau Vietnam, Hanoi. http://www.gso.gov.vn of Economic Research, Cambridge, MA. /Default_en.aspx?tabid=491. T H E W E L FA R E A N D L A B O R MA R K E T B E H AV I O R O F O L D E R P E O P L E I N E A S T A S I A A N D PAC I F I C 77 WHO (World Health Organization). 2012. Pension Program: Income, Poverty, Expenditure, “Global Health and Aging.” WHO, Geneva. Subjective Wellbeing, and Labor Supply.” China Witoelar, Firman, John Strauss, and Bondan Sikoki. Economic Quarterly 14 (1): 203–30. 2012. “Socioeconomic Success and Health in Zimmer, Zachary, and Pattama Amornsirisomboon. Later Life: Evidence from the Indonesia Family 2001. “Socioeconomic Status and Health Life Survey.” In Aging in Asia: Findings from among Older Adults in Thailand: An Examina- New and Emerging Data Initiatives, edited tion Using Multiple Indicators.” Social Science by James P. Smith and Malay Majmundar, and Medicine, Comparative Studies of Com- 309–41. Washington, DC: National Academies petition Policy 52 (8): 1297–311. doi:10.1016 Press. /S0277-9536(00)00232-X. World Bank East Asia and Pacific Standardized Zimmer, Zachary, Xian Liu, Albert Hermalin, and Household Surveys. Various years. Internal Yi-Li Chuang. 1998. “Educational Attainment data set of World Bank, Washington, DC. and Transitions in Functional Status among Zhang, Chuanchuan, John Giles, and Yaohui Zhao. Older Taiwanese.” Demography 35 (3): 2014. “Policy Evaluation of China’s New Rural 361–75. doi:10.2307/3004043. Aging and the Economy II Aging and the Macroeconomy in East Asia and Pacific 3 Introduction some of the potentially negative effects of aging and a shrinking working-age popula- A common concern about aging is its poten- tion, such as improving workforce quality tial effect on economic growth and the fiscal and increasing the labor force participation position of rapidly aging countries. Although of different groups. Because of the important concerns about the effects of rapid aging on role played by domestic savings in the high growth and government budgets should be economic growth of East Asia and Pacific taken seriously, the channels through which to date, the third section looks at how aging such effects may occur, at what scale they could affect household saving in the region. may occur, and over what time frames they The fourth section then presents evidence on may take place in countries at different stages the projected fiscal effects of aging, empha- of demographic transition are equally impor- sizing developing economies in East Asia and tant considerations. Pacific with rapidly aging populations to try This chapter explores the possible chan- to understand broad orders of magnitude and nels through which aging in East Asia and sources of fiscal pressure. Pacific may affect countries’ growth and fiscal situations. The chapter begins with Aging and growth in East Asia a brief description of aging and economic and Pacific growth trends, using the broad typology of red, orange, and green economies introduced Developing economies of East Asia and in chapter 1. It does not attempt to provide Pacific are aging at historically low income quantitative estimates of possible growth levels compared with their richer neigh- effects, but rather it explores the broad direc- bors, and many economies are getting old tion and intensity of effects and their inter- before getting rich. As figure 3.1 shows for action. The second section analyzes channels 1980–2010, the orange and green economies for augmenting the labor force to mitigate have been aging—in some cases rapidly—at This chapter is based on a background paper by Thomas Flochel, Yuki Ikeda, Harry Moroz, and Nithin Umapathi (2014). 81 82 LIVE LONG AND PROSPER FIGURE 3.1 East Asia and Pacific’s developing economies are getting old before getting rich Elderly dependency ratio and GDP per capita (2005 PPP), various economies, 1980–2010 60,000 50,000 GDP per capita (US$2005 PPP) 40,000 30,000 20,000 10,000 0 0.05 0.10 0.15 0.20 0.25 0.30 0.35 0.40 Elderly dependency ratio Singapore Hong Kong SAR, China United States United Kingdom France Russian Federation Korea, Rep. Malaysia Cambodia Timor-Leste Thailand Vietnam Lao PDR Philippines China Poland Japan South Africa Mongolia Indonesia Mexico Brazil Sources: GDP data in 2005 US$PPP for 1980 to 2010 from the World Bank’s World Development Indicators database; elderly dependency ratios calculated as ratio of population over age 65 to population ages 15–64, using data from UN 2013. Note: PPP = purchasing power parity. much lower levels of gross domestic product generally migrated upward and to the right, (GDP) per capita than the red economies and but they have followed very different paths.1 those of countries in the Organisation for The orange group, where the turning point Economic Co-operation and Development for dependency is either just past or not far (OECD) at similar stages of old-age depen- away, is particularly heterogeneous. Malaysia dency. The green economies, which are just experienced high growth, nearly tripling its beginning to undergo a demographic transi- GDP per capita with a still reasonably low tion, have lingered in the bottom left-hand ratio of elderly dependency, while Thailand’s corner at low levels of per capita income grow th was accompanied by particu- and low elderly dependency. Early declines larly rapid aging, with demographic ratios in youth dependency have even shifted some approaching one elderly for seven working- of them toward the left as the working- age adults. China has grown fast but from age share of the population began to rise. an even lower base; its rate of elderly depen- Economies in the orange and red groups have dency is not far behind Thailand’s. What is AG I N G A N D T H E MAC R O E CO N O M Y I N E A S T A S I A A N D PAC I F I C 83 clear for this group is that they will not be affect income per capita. Labor productiv- able to reach the income levels of red econo- ity interacts with aging in several ways. As mies at similar points of demographic transi- the population ages, the net effect on GDP tion. For example, Indonesia in 2010 had an per capita is determined by labor productiv- elderly dependency ratio similar to that of the ity at each age. Evidence suggests that age Republic of Korea in 1980, but with purchas- and individual productivity are correlated, ing power parity per capita income around but the relationship is complex and depends that of Korea in 1976 and output per worker on the relative endowments of human and equivalent to Korea’s in 1969 (or around physical capital workers are equipped with. 15 percent of Korea’s 2010 level of output The current cohort of elderly in East Asia and per worker). At the high end of the spectrum, Pacific is far more productive than the previ- red economies have continued solid growth ous cohort at any given age, particularly in (though flattening in Japan) while doubling countries where education and health have their rates of elderly dependency. improved rapidly. More healthy and better The contribution of favorable demo- educated, the elderly are increasingly able to graphics to the region’s growth performance develop skills as they age. Labor participation in recent decades has been significant. rates and productivity at each age are also Literature on East Asia and Pacific finds that affected by the two drivers of aging—lower up to one-third of the high per capita income fertility and extended longevity—through growth in the region from 1960 to the 1990s effects on economic behavior and on physi- can be attributed to the demographic divi- cal and human capital accumulation. From a dend (Bloom, Canning, and Malaney 2000; public policy perspective, demographics may Bloom, Canning, and Sevilla 2003; Bloom be a given, but labor participation and pro- and Williamson 1998). 2 Other authors ductivity are not. The size of the economic have estimated that demographic transi- bonus or burden that results from population tion accounted for as much as 44 percent aging depends on how policy influences labor of growth in per capita incomes from 1960 force participation, saving, human capital to 1990 (Kelley and Schmidt 1995, 2005). accumulation, and total factor productivity. Part of this effect is driven by rapid falls in fertility, which result in a large working-age share of population relative to youth share, Channels to augment the labor thus raising per capita incomes. In the case of force many East Asian countries, the pure demo- Although the share of working-age popula- graphic effect was positively compounded tion will experience significant declines in by improvements in educational attainment, red and some orange economies over the capital investment, and other factors. coming decades, significant potential exists Although demographics have provided a to augment the labor force in these countries clear tailwind for East Asian growth in recent through different channels. The potential decades (and will continue to do so in lower- channels through which aging may affect income countries for some time), demograph- growth include labor force effects, impacts ics are only part of the story. The effect of on saving and investment, indirect impacts population aging on GDP per capita also of rising fiscal obligations, and productivity. depends on labor force participation and One body of literature seeks to unpick these productivity, which in turn are affected by disparate effects for East Asia and Pacific demographic forces. Figure 3.2 decomposes and other parts of the world. The most pes- the channels through which aging influences simistic estimates of the impact of aging per capita income and provides the frame- on growth in East Asia and Pacific (typi- work for this chapter: the confluence of cally using basic growth accounting models) demographic, employment, and productivity assume unchanged labor force participation factors determines how aging has and will rates, retirement behavior, population health, 84 LIVE LONG AND PROSPER FIGURE 3.2 Aging’s influence on GDP per capita is the product of the employment rate, the working-age share of the population, and GDP per employed worker a. GDP per employed b. Share of the c. Share of total d. Labor productivity worker working-age population population who are ages 15–64 who are 15–64 employed Singapore 53,160 73% 74% 98,792 Average Average Hong Kong SAR, China 42,556 67% 75% 85,129 Japan 31,409 78% 64% 63,435 Korea, Rep. 26,830 67% 73% 55,133 Malaysia 13,793 61% 67% 33,402 Thailand 7,982 81% 72% 13,800 China 6,692 76% 74% 12,058 Indonesia 3,864 68% 65% 8,741 Mongolia 3,686 60% 69% 8,823 Philippines 3,532 64% 61% 9,095 Vietnam 2,920 82% 70% 5,095 Lao PDR 2,189 81% 59% 4,525 Papua New Guinea 2,187 74% 58% 5,076 Cambodia 1,949 88% 63% 3,526 Timor-Leste 1,482 39% 50% 7,585 Myanmar (missing data) 81% 69% GDP per capita Percent Percent GDP for those working (PPP, 2005 dollars) (PPP, 2005 dollars) Sources: World Bank estimates using population data from UN 2013, GDP per capita in 2005 US$PPP from the World Bank’s World Development Indicators database, and employment data from ILO 2011. Note: PPP = purchasing power parity. and migration patterns as populations age. 3 participation as a mitigating measure in the Yet each of these factors depends on behav- face of rapid aging. First, as of 2010 a large ior that is likely to change—or that with a gap remained in male and female participa- supportive policy environment has the poten- tion levels in a number of the region’s coun- tial to change—in ways that mitigate the tries, which suggests potential for higher mechanical and largely unavoidable effect participation among women. Second, the of shrinking working-age population. Such figures demonstrate how men’s and women’s estimates may thus be considered to reflect a participation rates have been moving in dif- “no-change” scenario and in that sense are ferent directions over time. For most countries useful indicators of the importance of policy in the region, male labor force participa- reform to stimulate behavioral change. A tion rates have declined notably, with some countervailing strand of research is far more countries such as Japan showing a decline sanguine about the growth effects of aging close to 15 percentage points. In comparison, in East Asia and Pacific but is underpinned female labor force participation rates have by optimistic assumptions about the ease of increased despite aging, which implies that policy change and the degree of behavioral in several countries, such as Korea, Malaysia, response to policy. and Singapore, participation rates have been If we look at recent decades, important able to partially counteract the demographic gender differences in participation trends effect on female labor supply. This is not the indicate the past and potential future con- case in China and Japan, where age-specific tribution of raising female labor force participation rates for women did not AG I N G A N D T H E MAC R O E CO N O M Y I N E A S T A S I A A N D PAC I F I C 85 rise as much. The evidence suggests, however, because of the income effect. According to that medium-term convergence toward fuller figure 3.3, relative trends in labor force par- use of a given working-age population is pos- ticipation rates and gains in health-adjusted sible. Policies to increase female labor force life expectancy years suggest the increas- participation and outcomes to date in East ing potential for extending working lives in Asia and Pacific are discussed in chapter 4 of East Asia and Pacific, particularly among this report. women. For example, on average, Japanese A second channel for augmenting labor men and women at ages 55–59 in 2010 forces is extending productive working lives could expect to live 1.4 and 1.2 extra years, of older workers. As shown in chapter 2 of respectively, compared with 1990. This this report, people in East Asia and Pacific “longevity dividend” is not gender neutral: work until fairly advanced ages already, since 1990, improvements in health-adjusted with notable exceptions in urban areas of life expectancy in the region have increased some middle-income countries. Many rural participation by women in China, Japan , people continue to work well into older age and Indonesia. out of necessity, not infrequently even when Apart from effects on quality of the labor in poor health. Among richer economies of force (see below), rising educational attain- the region, Japan and Korea have the high- ment across East Asia and Pacific can be est effective retirement age among OECD expected to play a role in promoting higher countries where, on average, people retire labor force participation across ages and later than pensionable age (D’Addio, Keese, genders through two channels. First, higher and Whitehouse 2010). Presumably, there- educational attainment increases incentives fore, this channel for enhancing labor force to work in better-paid employment. Second, participation would seem more constrained studies find that more education is associated than in other regions. However, certain with better health outcomes and lower mor- countries—crucially China and Vietnam— tality, thereby improving the ability and will- have significant potential to increase pro- ingness of older people to participate in the ductive working lives of urban people and labor force (Burtless 2013; Cutler and Lleras- mitigate the impacts of aging. The risk also Muney 2008). According to this logic, the exists that growth, urbanization, and expan- regional education projections to 2040 imply sion of social security systems in middle- and that labor force participation, especially at lower-income countries of East Asia and middle and older ages, should be expected Pacific will see current rural workers increas- to expand. Overall, as long as an increasing ingly adopt the retirement behavior of urban share of the population is better educated and workers, which would result in significant successive cohorts supplant the less-educated downward shifts in average participation generations, labor participation rates are rates at older ages in most countries of the expected to grow—especially among the region. That downside risk is considerable older population for whom the participation in the absence of further reform of pension rate tends to be lower. systems. Chapter 4 of this report discusses In East Asian countries with aging pop- regional and global experience with policies ulations, immigration of younger work- to extend active working lives. ers from other countries can help lift labor The potential for higher employment at force participation rates. Migration can older ages is influenced by gains in life expec- provide a certain level of relief, especially tancy and, especially, by gains in healthy in the short term, in the labor markets of years (Eggleston and Fuchs 2012). Bloom, countries with aging populations suffering Canning, and Moore (2004) show that from labor shortages. Migrants, on average, theoretically healthier life expectancies will tend to be younger and exhibit higher rates extend working lives but that improvements of participation relative to natives. Two tend to increase less than proportionately features distinguish East Asia and Pacific. 86 LIVE LONG AND PROSPER FIGURE 3.3 Older women in East Asia and Pacific will be able to work later in life by virtue of rising health-adjusted longevity Percentage of women in the labor force and healthy years of life expectancy by age between 1990 and 2010, various countries a. China b. Indonesia 70 25 65 25 65 60 60 20 55 20 55 Percent Percent Years 50 Years 50 15 15 45 45 40 10 40 10 35 35 30 5 30 5 50–54 55–59 60–64 65+ (65–69 50–54 55–59 60–64 65+ (65–69 for HALE) for HALE) Age cohort (years) Age cohort (years) c. Japan d. Republic of Korea 75 25 65 25 70 60 65 20 20 55 60 Percent Percent 50 Years 55 Years 15 15 50 45 45 40 10 40 10 35 35 30 5 30 5 50–54 55–59 60–64 65+ (65–69 50–54 55–59 60–64 65+ (65–69 for HALE) for HALE) Age cohort (years) Age cohort (years) Health-adjusted life expectancy, 1990 (right axis) Health-adjusted life expectancy, 2010 (right axis) Female labor force participation rate, 1990 (left axis) Female labor force participation rate, 2010 (left axis) Sources: World Bank calculations based on IHME and World Bank 2013 and ILO 2011. Note: HALE = health-adjusted life expectancy. First, significant variation exists in terms about the long-run levels of the United of demographic transitions. Thus, migra- States and a number of western European tion of young workers from countries with countries). However, the low base effect a younger workforce would benefit aging also indicates sociocultural, linguistic, and host countries now and the lower-income political barriers to immigration that may sending countries in the future by easing be more difficult to overcome. In contrast, the aging challenge in both (“chronological Hong Kong SAR, China, and Singapore arbitrage”). Second, some of the intensely stand out with high international migrant aging countries in the region are marked by stocks that make up about 40 percent of the very low rates of immigration. For example, total population. in Japan and Korea only 1 to 2 percent of The policy changes required will be the population is international migrant politically challenging and require signifi- stock (Özden and Testaverde 2014; World cant behavioral change on the part of work- Bank 2013d). This low base effect pres- ers, employers, and society more broadly. ents a significant opportunity if inflows of Migration is an unusually sensitive issue in new migrants ages 25–35 can be raised to potential host countries in East Asia and 10 percent of the labor force that age (that is, Pacific, and the historical record is modest AG I N G A N D T H E MAC R O E CO N O M Y I N E A S T A S I A A N D PAC I F I C 87 in some regional countries most affected by T ABLE 3.1 The educational attainment of aging.4 Increasing participation of older future generations in East Asia and Pacific will be workers will require reforms of social secu- considerably higher than for the current working- rity systems and labor policies, which have age population (secondary and above) Percentage of population ages 20–64 with secondary or proven politically difficult in many coun- higher education, 1990, 2030, and 2040 tries, though recent reforms in some East Asian countries provide encouragement. And Economy 1990 2030 2040 increasing female labor force participation Hong Kong SAR, China 72.5 89.4 88.9 will require incremental investments in child Japan 99.9 100.0 100.0 care and other programs as well as overcom- Korea, Rep. 74.6 99.5 99.8 ing inertia and long-held social attitudes to Singapore 63.3 90.3 92.5 the role of women in society once they have China 48.5 83.7 89.7 had children. These issues are discussed in Indonesia 31.6 72.3 79.2 chapters 4 and 5 of this report. Malaysia 50.4 89.9 93.0 Mongolia 75.8 91.4 92.3 Thailand 20.8 65.6 75.7 Labor force quality and Vietnam 24.2 49.1 56.9 productivity Cambodia 16.2 45.2 52.7 Lao PDR 17.8 48.5 56.3 Just as important as mitigating quantity of Myanmar 27.6 58.2 65.9 workforce declines will be efforts to enhance Philippines 52.7 84.0 88.3 worker quality. Better-educated people are Source: World Bank calculations based on data from World Population more prepared for lifelong learning, health- Program, International Institute for Applied Systems Analysis, http://www ier, and more productive. On this front, .iiasa.ac.at/web/home/research/researchPrograms/WorldPopulation /Introduction.html. as shown in tables 3.1 and 3.2, East and Southeast Asia have a positive starting point in terms of the massive expansion in second- ary and higher education in recent decades, TABLE 3.2 The educational attainment of which will flow through into the labor force future generations in East Asia and Pacific will be in coming decades. At all income levels, considerably higher than for the current working- each generation of future East and Southeast age population (tertiary) Asian workers will be significantly more Share of population ages 20–64 years old with tertiary educated than the last. The second trend is education in 1990, 2030, and 2040 that the educational gap between older and Economy 1990 2030 2040 younger workers is expected to diminish con- siderably in China, Indonesia, Malaysia, the Hong Kong SAR, China 15.9 34.0 35.7 Philippines, and Thailand—and even disap- Japan 23.5 55.2 60.5 Korea, Rep. 15.6 49.7 55.6 pear in countries such as Korea. Although Singapore 15.1 50.2 55.2 productivity differences may persist because China 3.1 11.5 14.7 of age-specific factors, intergenerational Indonesia 4.2 15.3 18.7 inequality of educational attainment is going Malaysia 5.7 26.9 32.4 to diminish significantly. This outcome Mongolia 20.1 30.5 35.1 implies improving prospects for employment Thailand 6.8 27.6 33.9 at older ages. Vietnam 3.0 9.3 12.1 Several studies find that significant Cambodia 0.5 3.3 4.3 increases in human capital such as those Lao PDR 5.0 19.0 22.3 experienced in East Asia and Pacific may Myanmar 3.4 12.7 15.6 be sufficient to offset the effects of popula- Philippines 15.4 37.4 42.9 tion aging on aggregate productivity. Bloom, Source: World Bank calculations based on data from World Population Program, International Institute for Applied Systems Analysis, http://www Prettner, and Strulik (2013) show that under .iiasa.ac.at/web/home/research/researchPrograms/WorldPopulation plausible production function specifications, /Introduction.html. 88 LIVE LONG AND PROSPER the rise in education and health investments the share of students achieving basic literacy that is in part spurred by declines in fertil- has been estimated to translate into an annual ity rates raises workers’ productivity enough growth rate that is 0.3 percentage points to compensate for declines in labor supply. higher than it would otherwise be for that And according to Lee and Mason (2010), country (Hanushek and Woessmann 2009). the effect of spending on education is strong Although red economies are in a strong posi- enough to offset the adverse effects of popu- tion in terms of educational attainment and lation aging. Significant increases in educa- skills, some orange economies and many tion levels will be particularly important green economies are still far from par. OECD sources of growth in aging orange econo- (2012) PISA (Programme for International mies where the levels in 1990 were low. For Student Assessment) results show that parts red economies, where the educational gap of the East Asian region outperform the across cohorts of workers is narrower, other world. Hong Kong SAR and Shanghai, forms of lifelong education and retraining China; Japan; Korea; and Singapore are of older workers that are not reflected in above the OECD average, and Vietnam’s per- tables 3.1 and 3.2 deserve further attention formance—with 15-year-olds performing on (see box 3.1). a par with those in Germany and Austria—is Falling fertility rates are usually associated exceptional. But other orange economies, such with greater parental investments in educa- as Indonesia, Malaysia, and Thailand, lag the tion and health of children, but translating OECD average. Furthermore, according to these investments into productivity growth the Program for International Assessment of requires investments in the quality of edu- Adult Competencies, which measures capac- cation. Becker and Lewis (1973) and Willis ity for retraining and adaptation among (1973) first described the quality-quantity adults, Japan leads globally, with Korea trade-off that occurs as parents have fewer close to the OECD average. More alarming children and they invest more in the human is the state of adult skills and competencies capital of each child. A 10 percent increase in in green economies, such as the Lao People’s BOX 3.1 The scope for improving skills of an aging workforce A commonly held belief is that the ability to learn older workers should not be judged by the learning new skills declines with age. In particular, Baltes, abilities of the current elderly population. Second, Lindenberger, and Staudinger (2006) find that the because healthy life expectancy is increasing, cogni- speed at which an individual processes new infor- tive decline is found to start later in life. Skirbekk mation declines from 25 years of age onward. et al. (2013) conclude that the cognitive plasticity Cognitive plasticity, the ability to learn, appears curve is shifted not only up, but also outward: indi- to follow an inverted U shape over an individual’s viduals are more able to learn at each age, and decline lifetime. Should we then conclude that older work- occurs increasingly late in life. Third, scientific evi- ers cannot be retrained to renew their skills if theirs dence shows that the decline is not deterministic and become obsolete? the elderly’s ability to learn improves given the right Recent research nuances the conventional wisdom. circumstances. Physical exercise has been found to First, because successive cohorts are healthier and revitalize key parts of the brain that improve learning more educated, their brains function better. This phe- abilities (Christensen et al. 1997; Voelcker-Rehage, nomenon, known as the Flynn effect, suggests that Godde, and Staudinger 2011), and learning outcomes testing a cross-section of elderly people at one point among the elderly are particularly responsive to moti- in time ignores improvements across cohorts. Because vational factors and supportive attitudes in the work- cognitive plasticity rises with each cohort, future place (Kessler and Staudinger 2007). AG I N G A N D T H E MAC R O E CO N O M Y I N E A S T A S I A A N D PAC I F I C 89 Democratic Republic (World Bank 2013c) profiles that emerge from industry or firm-level and Timor-Leste (World Bank 2013b), which empirical research in Japan have consistently have scored very low on core literacy tests found productivity to follow an inverted U as part of the World Bank’s STEP (Skills shape with respect to age, peaking for employ- Toward Employability and Productivity) ees at about 20 years of experience, which cor- skills measurement program. responds to an age of 40 to 46 years (Ochiai Severe child health and nutrition problems 2008; Shirakawa 2010). But productivity in some green economies could seriously cur- does not seem to fall rapidly after that. In tail the demographic dividend. Several green fact, research from Japan finds that the labor economies have some of the highest rates of productivity of the 55–59 age group is virtu- stunting in the world, with up to 44 percent ally identical to that of the 40–44 age group of children below age 5 being stunted (Murphy and Welch 1990). Evidence from (figure 3.4). Children who are stunted are Europe points to similar findings. In Austria more likely to have lower cognitive abilities. and Sweden, for example, at worst, no evi- Key parts of the brain develop less in children dence of age-related productivity decline is who are severely malnourished, making these found (Göbel and Zwick 2013; Malmberg, children much less likely to attend school or Lindh, and Halvarsson 2008). more likely to drop out early. Compared with In red economies, past trends in the age nonstunted children, stunted children score distribution of the population have been par- 7 percent lower on math tests, are 19 percent ticularly propitious to workforce productiv- less likely to be able to read a simple sentence ity growth, but their reversing trends do not at age 8, are 12 percent less likely to be able necessarily imply dramatic reductions in pro- to write a simple sentence, and are 13 percent ductivity. A higher share of prime-age indi- less likely to be in the appropriate grade for viduals in the population, sometimes defined age at school. as the 30–54 age groups, has been found to be A second important question on which significantly associated with high productiv- evidence is encouraging but inconclusive is ity growth (Gómez and de Cos 2008).5 From how a falling share of “prime working-age” the mid-1960s to the mid-2000s, the share of population in red and a number of orange those ages 35–54 rose and peaked at a ratio economies could affect productivity. How individual productivity varies with age is a challenging empirical question that has FIGURE 3.4 Prevalence of stunting among children under age five important implications for aggregate pro- could undermine investments in education and affect cognitive skills in East Asian and Pacific economies of the green group ductivity, particularly in advanced econo- Percentage of children under five who are stunted and GNI per capita mies where the workforce is aging and (2005 dollars), selected countries shrinking. Population aging has implica- tions for growth not only through its direct 50 effect on workers’ labor productivity as they Ethiopia Share of children under age 5 (%) 45 Lao PDR age, but also through its effect on capital 40 Cambodia Nepal Indonesia accumulation, total factor productivity, 35 Uganda Philippines 30 Togo and innovation. Empirical evidence on this 25 question is mixed and limited in East Asia 20 Vietnam Sri Lanka and Pacific but, in general, provides some 15 Thailand grounds for encouragement. 10 China Empirical evidence from Japan and 5 other aging countries suggests that the 0 0 500 1,000 1,500 2,000 2,500 3,000 3,500 age-productivity profile of workers follows an GNI per capita (2005 dollars) inverted U shape, with a mild decline after the mid-40s and a significant dip only after age Source: World Health Organization, Global Database on Child Growth and Malnutrition. 65 (box 3.2). Estimates of age-productivity Note: GNI = gross national income. 90 LIVE LONG AND PROSPER BOX 3.2 A labor productivity paradox? Evidence on the relative productivity of older and verbal abilities matter more (“crystallized” workers that emerges from empirical research in abilities), less or no reduction in productivity occurs advanced economies is somewhat contradictory. among elderly workers (Skirbekk 2008). Findings While the productivity-age profile is robustly found also appear to differ depending on the measure of to follow an inverted U shape, with the peak age productivity used and even using same approach. somewhere between 35 and 54, macro estimates Thus, supervisor assessment studies have found no of the effect of age on aggregate productivity find clear effect of age on productivity (McEvoy and that older workforces are at worst as productive as Cascio’s [1989] review of 96 studies) while others younger workforces. using quantity and quality of output measures have Several possible explanations exist. First, indi- found lower productivity of older workers. Matched vidual worker productivity does not dramatically employer-employee surveys show mixed results decline after it peaks around age 40. The decline across studies and specific measures of productivity. is found to be very mild relative to the rise in pro- But some surprising results have emerged from ductivity in the first 20 years of work experience in the research. Older workers who remain in the various studies in Japan (Fukao et al. 2006; Shinada labor force despite technological changes that affect 2011). In a famous study of BMW, Börsch-Supan their job content are found to have adapted well, and Weiss (2008) also find that productivity does experiencing greater growth in tasks with intense not begin to decline until age 60. Among Organisa- use of cognitive abilities (Romeu Gordo and Skir- tion for Economic Co-operation and Development bekk 2013, following a line of research initiated by countries, the share of those ages 50–64 was found Autor 2003). Focusing on the distinction between to have a positive influence on growth, while a mental and physical productivity, Van Ours (2009) higher share of those 65 and older contributed nega- finds that although the latter declines after age 40, tively (Lindh and Malmberg 1999). Second, as noted the former does not. Similarly, perhaps counter to previously, some issues with the research method- popular opinion, studies have found lower absen- ology may bias the results. The decision to retire teeism (Martocchio 1989) and fewer job shifts or retain a worker as he or she ages likely depends among older workers than younger ones, perhaps on the worker’s productivity. It follows that when saving on training costs (AARP 2002). The most measuring the productivity of older workers, only recent studies provide a generally more upbeat those elderly workers with “high enough” produc- assessment of older-worker productivity (Bloom tivity can be surveyed, whereas this selection effect and Sousa-Poza 2013). is much less present among younger workers. Finally, productivity is not only an individual but Third, the link between age and productivity is also a team concept, and evidence indicates that skills also highly dependent on the context and task under of the old and the young may be complementary in consideration. In general, productivity reductions at some contexts. Grund and Westergaard-Nielsen older ages are strongest in job tasks where problem (2008) find that those companies with workers of solving, learning, and speed are important (“fluid” mixed ages are actually more profitable than those abilities), whereas for work tasks where experience with exclusively young or old workers. of one in three people on average in the red cross-sectional evidence ignores cohort effects economies before declining again (figure 3.5). that drive a worker age 60 today to be both Other things being equal, the workforce in more productive and more able to acquire these countries could be expected to experi- new skills than a worker of the same age three ence a steady decline in labor productivity as decades earlier. it ages rapidly in the next 30 years and prime In orange and green economies, the age groups are less represented. However, maturing of the population offers increas- all things will not be equal. To begin with, the ing opportunities for productivity growth, AG I N G A N D T H E MAC R O E CO N O M Y I N E A S T A S I A A N D PAC I F I C 91 FIGURE 3.5 Productivity depends on the age distribution of the population, but mitigation channels exist Share of population ages 35–54, 1950–2050 32% 32 Share of population ages 35–54 (%) 30 29% 28 26 24 25% 22 20 20% 18 17% 16 16% 1950 1960 1970 1980 1990 2000 2010 2020 2030 2040 2050 Years Hong Kong SAR, China; Japan; Republic of Korea; Singapore China; Indonesia; Malaysia; Mongolia; Thailand; Vietnam Cambodia; Lao PDR; Myanmar; Papua New Guinea; Philippines; Timor-Leste Source: World Bank calculations based on UN 2013 data. yet the pace of growth will depend on physi- entrepreneurship, which is of particular con- cal and human capital investments and how cern in countries seeking to promote innova- efficiently they are put to use. The next 15–20 tion. Figure 3.6 shows the share of workers years will see a rise in the average share of the by age who manage and own a business and prime-age population of orange economies, those who expect to launch a start-up busi- although a flatter rise than that experienced ness in the coming three years. Nonetheless, in the 30 years since 1985. Green economies China, Japan, and Korea have relatively should experience some positive growth in high ownership of businesses among older labor productivity from the maturing of workers, and Japan is distinctive in that this their workforce, peaking around 2045. Yet increases as a share with age. although the age-productivity profile is the most obvious link between population aging and labor productivity, aggregate productiv- Saving in East Asia and Pacific: ity growth in the future will be determined by Will aging play an influential role? investments in human capital and the quality The net effect of aging on labor produc- of education. Saving behavior and the types tivity depends in part on the capital stock of saving mechanisms available to households and how capital accumulation evolves with will also have a role, as will the capacity of population aging.6 The concerns over saving the economy to make efficient use of these in East Asian countries with aging popu- human and physical inputs. The next section lations reflect the important role played focuses on the saving channel and aging. by domestic savings in the high growth of One feature that appears to decline the region to date. Savings not only fund with age of the workforce globally is investment but also allow lower-income 92 LIVE LONG AND PROSPER FIGURE 3.6 Although in many countries, entrepreneurship seems to decline with worker age, business ownership in East Asian countries is significant among older workers Adults who own a business and engaged in start-up activities, by age, selected areas a. Manages and owns a business for more than 42 months 20 Share of adult respondents (%) 15 10 5 0 Japan China Malaysia Korea, Rep. United States ECA LAC b. Expects to start a new business in the next 3 years 60 Share of adult respondents (%) 40 20 0 Japan China Malaysia Korea, Rep. United States ECA LAC Ages 18–34 Ages 35–44 Ages 45–54 Age 55 and over Source: Global Entrepreneurship Monitor 2011. Note: ECA = Eastern Europe and Central Asia; LAC = Latin America and the Caribbean. countries to adopt more advanced technolo- 2011). It could possibly even accelerate gies, which matters enormously for growth. productivity growth and thus mitigate the Aging may affect supply and demand for effects of a shrinking labor force. However, household savings, potentially corporate as dependency ratios stabilize, the supply savings, and almost certainly government channels discussed below ultimately deter- savings because of the effects of age-related mine the magnitude of long-term productiv- public spending. On the demand side, one ity and growth rates. important channel through which chang- Household saving in much of East Asia is ing demographics affect growth is increases high and provides a robust base as the region in labor productivity. A decrease in the size ages. In East Asian economies, the average pri- of the labor force in aging economies leads vate savings rate between 1960 and 2000 was to an increase in capital intensity (or the 25 percent—exceeding those of most other capital-to-labor ratio), thus increasing labor regions. China has accounted for the predom- productivity. Such capital deepening and inant share of this impressive increase since consequent increase in productivity leads the 1960s (figure 3.7). That said, domestic to higher standards of living, even in the savings across the region vary substantially: absence of higher savings (Lee and Mason from 1966 to 2007, the savings rate ranged AG I N G A N D T H E MAC R O E CO N O M Y I N E A S T A S I A A N D PAC I F I C 93 FIGURE 3.7 Compared with household savings rates in other regions, the rate in East Asia and Pacific is relatively high and stable Private savings as a percentage of GDP by region, 1980–2010 a. East Asia and Pacific b. Europe and Central Asia c. High-income OECD countries 40 40 40 Private savings as a Private savings as a Private savings as a share of GDP (%) share of GDP (%) share of GDP (%) 30 30 30 20 20 20 10 10 10 1980 1990 2000 2010 1980 1990 2000 2010 1980 1990 2000 2010 Year Year Year d. Latin America and the Caribbean e. Middle East and North Africa f. South Asia 40 40 40 Private savings as a Private savings as a Private savings as a share of GDP (%) share of GDP (%) share of GDP (%) 30 30 30 20 20 20 10 10 10 1980 1990 2000 2010 1980 1990 2000 2010 1980 1990 2000 2010 Year Year Year Source: World Bank’s World Development Indicators database. Note: OECD = Organisation for Economic Co-operation and Development. from well over 30 percent in China; Hong expected duration of consumption in retire- Kong SAR, China; Malaysia; and Singapore ment. Predictions on which of these two to below 20 percent in the Philippines and effects will dominate in East Asia and Pacific Vietnam (Horioka and Terada-Hagiwara are mixed. Some authors suggest the compo- 2012). Although household saving has fluctu- sitional effect will dominate the behavioral ated considerably, corporate saving has been effect; others maintain that extensions in on a steady upward trajectory since 2001. longevity will positively affect savings rates The demographic effect on the savings in the future, while others find mixed effects rate is difficult to predict because the dif- (see below). On balance, the studies suggest ferent demographic transformations driv- less cause for concern in East Asia and Pacific ing population aging will have ambiguous than in other parts of the world. effects on the savings rate.7 Two offsetting Whether aging in East Asia and Pacific effects determine the net effect of longev- will adversely affect savings growth prospects ity on aggregate saving: a compositional depends on three regional characteristics. effect that leads to a reduced aggregate sav- First are the patterns of life-cycle profiles of ings rate because of an increase in the share savings, which affect the scale of the compo- of older people in the adult population; and sitional effect as the population age distribu- behavioral effects, such as the rise in the tion changes. Second are the adjustments to individual savings rate to finance a longer saving behaviors in response to lower fertility 94 LIVE LONG AND PROSPER and increased longevity. Third is improving on aggregate savings rates. An analysis of efficiency in saving associated with social profiles of life-cycle income and consump- security and the increasing quality of finan- tion flows using National Transfer Accounts cial markets. across China, Korea, Thailand, and Vietnam Predictions regarding the impact of aging also suggests that for the wider region, con- on household saving in East Asia and Pacific sumption profiles remain essentially flat from are mixed but on balance suggest less cause around age 20, whereas labor income contin- for concern that in other parts of the world. ues to be positive at older ages.11 China shows Under the life-cycle savings hypothesis (LCH) both a relatively flat saving profile by age in (Modigliani 1970), demographic factors are adulthood (indeed, a higher savings rate from important determinants of savings rates, the late 40s than for those in prime age) and though evidence from East Asia and Pacific an increase in saving at all ages over time. A is mixed. According to the LCH model of stronger bequest motive, incomplete cover- saving, people save when young and dissave age of public pensions, and shallow financial to finance consumption during retirement. protection from health insurance in East Asia From a macroeconomic perspective, the life- and Pacific appear important in reconciling cycle theory implies a decrease in the national this pattern. All in all, the compositional savings rate in an aging economy8 because effect of aging on saving to date appears to the share of retirees in the adult population have been more muted than in other regions. increases. Empirical evidence on saving is A number of behavioral factors can poten- broadly consistent with the main predic- tially mitigate the adverse impact of aging on tions of the life-cycle theory (Browning and household saving in East Asia and Pacific, Crossley 2001). Loyaza, Schmidt-Hebbel, thus making the potential impact on aggre- and Serven (2000) find that these factors gate saving more modest. As discussed, are important determinants of cross-country declining fertility is associated with higher differences in savings rates of developing female labor force participation, which countries while Bloom et al. (2007) report implies higher national saving because of the same finding for a cross-country panel the income growth of a larger female labor of data for 1960 to 2000. Attanasio and force. Increased life expectancy can also lead Székely (2000) find that household saving to increased working lives and increased is consistent with the LCH in Mexico, Peru, saving to fund retirement. Bloom, Canning, and Thailand, but only for the most educated and Sevilla (2003) incorporate longevity in a households. Evidence for Japan from both standard life-cycle model of saving and show macro- and microstudies is consistent with that, under certain assumptions, longevity the LCH.9 The retired elderly as well as the leads to higher savings rates at every age. working elderly dissave in Japan. Their model fits the evidence from East Asia Overall, however, the exceptional increase and Pacific well. However, the effect of lon- in savings rates in East Asia and Pacific gevity on aggregate savings is transitory and is difficult to reconcile with the LCH. dissipates as the population ages. They thus Demographic changes and income inequal- conclude that the positive effect of increases ity between cohorts can explain only a in life expectancy on saving in East Asia and small part of this trend. A singular feature Pacific will dissolve as the population ages. of the East Asian saving surge during this Kinugasa and Mason (2007) have a more period is an increase in the rate of saving optimistic view. Their model of saving incor- at every age.10 This finding is inconsistent porates declines in youth dependency as well with the standard life-cycle theory. Schultz as increases in life expectancy and fits the (2004), for instance, finds that changing age saving trends in East Asian countries. They structure across 16 Asian countries from find that the increase in life expectancy is 1952 to 1992 had an insignificant impact the more important determinant of savings, AG I N G A N D T H E MAC R O E CO N O M Y I N E A S T A S I A A N D PAC I F I C 95 accounting for three-fourths of the increase be too negative because the precautionary in Asia. Moreover, their analysis predicts that motive suggests efficiency gains can be made the behavioral effect of longevity increases by shifting to lower levels but higher quality outweighs the compositional effect, implying of saving: for example, by expanding social that population aging will not lower saving insurance. A precautionary motive also in Asia. Using cross-country panel data, Li, appears to explain the high corporate savings Zhang, and Zhang (2007) find that the effect rate in China. In the future, policies decreas- of increased longevity on aggregate sav- ing income volatility can also lessen the need ings is positive whereas that of an increased for inefficient precautionary saving by firms. old-age dependency rate is negative. World An increase in the variance of permanent Bank (2013a) also argues that the drag on income shocks implies not only a higher sav- saving from the compositional effect will be ings rate but also a change in the portfolio offset by the behavioral effect in East Asia allocation of savings toward safe assets and a and Pacific. decrease in investment (Cherif and Hasanov The concern about reduced household sav- 2012). Conversely, International Monetary ing induced by aging in East Asia and Pacific Fund (IMF) analysis of the impact of finan- may thus be overstated because broader cial liberalization on corporate saving glob- macroeconomic channels are likely to play ally reveals that financial reforms may have a dominating role. First, a precautionary a considerably larger effect in Asia (IMF motive appears to be driving high saving in 2009). This evidence suggests that financial East Asia and Pacific, suggesting important sector reforms are likely to reduce precau- inefficiencies despite high national rates of tionary saving by the corporate sector in saving. Second, further financial liberaliza- Asia. In fact, recent policy changes indicate tion may contribute to a decline in the sav- that many middle-income countries in East ings rate, which will be a sign of improved Asia and Pacific are already moving in this efficiency. Finally, improvements in financial direction. intermediation, inclusion, and social security Improvements in financial inclusion will coverage will play a crucial role in determin- play an important role for channeling house- ing levels of saving and in channeling savings hold savings into productive investment. into productive investment. Greater financial inclusion, either through A precautionary motive appears to be bank or nonbank financial institutions, driving high household and corporate sav- allows for efficient flow of savings and invest- ing in East Asia and Pacific, suggesting ments in the economy (Levine 2003; Levine, inefficiencies that, if reduced, would allow Loayza, and Beck 2000). At the household stronger economic effects from a given rate level, bank account penetration differs across of saving. Blanchard and Giavazzi (2005) countries in East Asia and Pacific. With a rate stress the importance of reducing precau- of 55 percent, East Asia and Pacific is ahead tionary saving to boost consumption and of most regions in bank account penetration rebalance growth. In China, market-oriented and only behind high-income economies, yet reforms led to large-scale corporate restruc- access to personal saving instruments and turing and downsizing of the public sec- credit varies significantly across countries. tor between 1995 and 2005, consequently According to Demirgüç-Kunt and Klapper weakening the public social safety net. (2012), in countries such as Cambodia more Households responded by building up sav- than 95 percent of adults do not have an ings to self-insure against future health and account at a formal financial institution. In old-age income needs (Chamon and Prasad sum, these regional stylized facts reflect that, 2010; Ma and Yi 2010). Therefore, even aging pressures notwithstanding, East Asian if savings rates decline in future, the aging and Pacific countries have varying abilities to effect means the impact on growth may not mobilize savings, allocate capital efficiently, 96 LIVE LONG AND PROSPER and diversify risks, which suggests that sig- cases and that will require strong policy nificant gains remain to be made. leadership to mitigate. As noted, the most obvious fiscal risks are posed by pension, health, and long-term care systems, all of Fiscal effects of aging which are affected by aging (see chapters 5 The biggest risk to aggregate savings rates to 7 of this report for detailed discussion in East Asia and Pacific will be through the and summaries of country-specific expendi- impact on government saving caused by ture projections). The fiscal risks are com- rising pension, health, and long-term care pounded in many East Asian and Pacific spending. This risk is of particular concern countries because governments are pur- in developing East Asian and Pacific coun- suing rapid coverage expansion of health tries where projected fiscal space is very insurance and pensions and exploring more limited (see Pradelli and Van Doorn 2015). active public financing for long-term care. Although reasons exist to be sanguine Various projections of pension spending about the potential effects on household in East Asia and Pacific reveal significant and corporate savings from aging, the same fiscal pressure from rising pension spending is not the case for government saving. Even over coming decades. Various cross-country assuming relatively benign effects of aging projections are available that include East on economic growth in the region, major Asia and Pacific, some country-specific and fiscal risks from aging remain that are some more stylized around regional coun- already manifesting themselves in several try groupings. Although estimated impacts cover a considerable range (and are impor- tant to supplement with more detailed FIGURE 3.8 Projected increases in pension spending to 2030 point country projections presented in chapter 5 to a significant increase in China and Malaysia, with more modest of this report), they all point to the fiscal increases for Indonesia, the Philippines, and Thailand risk presented by pension systems in the Projected increase in pension spending as a percentage of GDP, 2010–30 region. In the following, we examine a few estimates. Turkey IMF projections with simplified but Egypt, Arab Rep. standard assumptions across countries China Jordan (figure 3.8) point to a significant increase in Russian Federation pension spending in China and Malaysia, Saudi Arabia with more modest increases for Indonesia, Malaysia Ukraine the Philippines, and Thailand (although Romania the projection for Indonesia does not take Argentina into account the new pension scheme). The Brazil Mexico increase in China of over 3 percent of GDP Latvia is particularly notable. Not shown in the fig- Philippines ure, but available from the same source, are South Africa Lithuania estimates for Japan and Korea. Japan exhibits Thailand a slight fall in pension spending as a share of Indonesia Pakistan GDP, reflecting its already aged population India and cost-containment measures in its pension Bulgaria scheme (see chapter 5 in this report), whereas Average = 1.0 Chile Poland Korea experiences a significant increase in Estonia pension spending approaching 4.5 percent Hungary in the 2010–30 period, reflecting the matu- –4 –2 0 2 4 6 8 ration of its relatively young pension system % of GDP and the continuation of rapid aging. Simplified estimates of pension spending in Source: IMF Fiscal Affairs Department 2010. 2050 done for this report (which incorporate AG I N G A N D T H E MAC R O E CO N O M Y I N E A S T A S I A A N D PAC I F I C 97 FIGURE 3.9 Estimated increases in East and Southeast Asian pension spending for 2014 to 2050 find significant fiscal risks from pensions Projected increase in pension spending as a percentage of GDP, 2014–50, various economies 6 5 4 % of GDP 3 2 1 0 a DR ar ea s a sia nd m in , n ia . Ch AR ep ne di in pa ys nm na in ila ne oP Ch bo ,R a S pi ala Ja Gu et a ng do ya ilip a m La Th Vi re M M w Ko Ca In Ph Ko Ne ng ua Ho p Pa Source: Pradelli and Van Doorn 2015. Note: Estimates assume stable debt-to-GDP ratios from 2013 and nominal GDP growth of 7 percent. Malaysia estimate includes spending from its provident fund. debt dynamics) also find significant fiscal FIGURE 3.10 Pension spending projections by stylized groupings risks from pensions. China continues to have for APEC economies, 2010–70, illustrate rapid increase in pension a substantial fiscal burden from pensions spending in a no-reform scenario (similar in magnitude to the IMF estimate), Projected pension spending as a percentage of GDP in APEC economies (no- while the incremental costs in Malaysia, the reform scenario), 2009–75 Philippines, and Vietnam are also substantial (figure 3.9). 18 16 A third estimation exercise uses a stylized 14 approach to capture broad groupings of the 12 Asia-Pacific Economic Cooperation (APEC) % of GDP 10 economies and potential fiscal costs of future 8 pensions (figures 3.10 and 3.11) (Hinz and 6 Zviniene 2011).12 Notable about this projec- 4 tion, which uses the World Bank’s PROST 2 (Pension Reform Options Simulation Toolkit) 0 model, is that similar long-term outcomes in 2009 2019 2029 2039 2049 2059 2069 2079 overall pension spending are reached at dif- High-income economies, mature pension system ferent rates and through the interaction of Middle-income economies, maturing pension system different influences. The high-income mature High-income economies, maturing pension system systems experience increases earlier, add- Low-income economies, developing pension system ing more than 5 percent of GDP to pension spending costs in the first 20 years and nearly Source: Hinz and Zviniene 2011. Note: APEC = Asia-Pacific Economic Cooperation. 10 percent of GDP after 40 years. These addi- tions are primarily owing to the advanced stage of population aging, but assume and Pacific. The lower- and middle-income modest accrual rates that are characteristic cases experience lower spending increases in of mature high-income systems in East Asia the initial decades of only 2 to 3 percent of 98 LIVE LONG AND PROSPER FIGURE 3.11 Pension spending increases can be substantially GDP but then rapidly increase an additional reduced with key reforms 10 percent of GDP in spending between 2030 Projections by stylized groupings for APEC economies, 2010–70, assuming a and 2070. The rapid increase in the later gradual five-year increase in pension age and price indexation years is caused by the accelerating process of coverage expansion. The baseline is driven 20 18 by the assumption that these economies 16 continue to retain high annual accrual rates 14 12 that are characteristic of low- and middle- Percent 10 income countries. As the members of these 8 systems accrue more years of coverage and 6 4 participation moves deeper into the income 2 distribution, expenditures can be expected 0 2009 2019 2029 2039 2049 2059 2069 to escalate rapidly unless the benefit formu- High-income mature systems High-income maturing systems las for such a system are adjusted to compen- Middle-income maturing systems Low-income immature systems sate for the greater density of contributions. Figure 3.11 illustrates the major fiscal impact Source: Hinz and Zviniene 2011. of two reforms that have been under consid- Note: APEC = Asia-Pacific Economic Cooperation. eration in several developing East Asian and Pacific countries in recent years: gradually FIGURE 3.12 Projected increases in public spending on health caused increasing pensionable age by five years and by aging, though significant, are not as substantial as for pensions using price indexation of benefits. These two Projected increase in public spending on health care as a share of GDP, with reforms dramatically flatten the increase in aging-related and excess cost growth breakdown, various countries, 2010–30 pension spending and result in much lower (though still substantial) increments in pen- Poland sion spending. Hungary Although health care spending projec- Brazil Chile tions are much more difficult to make over Argentina extended time horizons, they point to two Lithuania broad trends: increasing public spending on Bulgaria health, and aging as a significant but not Turkey Romania dominant factor driving increased public Ukraine spending. Looking at various sources in turn, Thailand cross-country IMF estimates are presented Russian Federation in figure 3.12 from a simple model that South Africa Estonia also attempts to separate the aging or pure Mexico demographic effect on health spending from Saudi Arabia “excess cost growth” (driven by technology Latvia change and other non-demographic factors). China Malaysia Overall, for East Asia and Pacific, the pro- Indonesia jected increases in public health spending Philippines are not as substantial as for pensions, but India nonetheless are not insignificant, especially Pakistan proportional to current health spending. For 0 0.5 1.0 1.5 2.0 2.5 Japan (not shown) and Korea, the compa- % increase rable increments in health spending are just Aging Excess cost growth share under 1 percent of GDP and over 3 percent of Weighted average = 1.0 Unweighted average = 1.1 GDP (with around one-quarter of the effect from aging alone), respectively. Sources: IMF Fiscal Affairs Department 2010 based on OECD and WHO health data. Note: Excess cost growth is growth in public health spending after controlling for aging. Weighted OECD projections of health and long-term averages are based on purchasing power parity GDP. care spending are also instructive, though AG I N G A N D T H E MAC R O E CO N O M Y I N E A S T A S I A A N D PAC I F I C 99 available only for China, Indonesia, Japan, FIGURE 3.13 Projected increases in public health spending and Korea in East Asia and Pacific (De la reflect the importance of policy choices in sustainability of health Maisonneuve and Oliveira Martins 2013). All care systems East Asian and Pacific and OECD economies Percentage of GDP increase in public health care spending under cost pressure and cost containment scenarios, 2010–30 will experience significant increases in system costs over time. A few other pertinent points 3.0 emerge from figure 3.13. First, although the absolute projected increase in public health 2.5 spending is higher in red economies, the GDP increase (%) 2.0 relative contribution of aging is more signifi- cant in China and Indonesia and also higher 1.5 than other middle-income countries listed 1.0 (more than half in China and over a third in Indonesia for the cost-pressure scenario 0.5 and a higher share in the cost-containment scenario). This scenario assumes a stronger 0 n p. a sia il ico a n ey demographic effect than previous IMF esti- az in di pa tio Re rk ne ex Ch In Br Ja ra Tu a, do M de mates. Second, reflecting this assumption, re In Fe Ko ian the gap between cost containment (benign ss Ru policy) and cost pressure (pessimistic policy) Cost pressure Cost containment o/w demographic is significant and in Indonesia, for example, effect almost as large as the entire effect of aging. In short, aging matters and should be taken Source: De la Maisonneuve and Oliveira Martins 2013. Note: o/w = out of which. into account, but policy choices also matter a lot for the sustainability of health and long- term care systems. A final source to consider is country- FIGURE 3.14 Projected increases in public long-term care spending specific projections of health spending and indicate additional fiscal pressure in countries where the aging the impact of aging. Unfortunately, these effect is strong projections are limited for East Asian and Percentage of GDP increase in public spending on long-term care, Pacific economies, but the two available 2010–30, various countries actuarial studies that look at this question suggest effects from aging more aligned 0.9 with IMF estimates. For Malaysia, about 0.8 one-third of the total projected increase in 0.7 GDP increase (%) health spending between 2010 and 2030 is 0.6 attributable to aging (Rannan-Eliya et al. 0.5 2013), while for Hong Kong SAR, China, 0.4 projections until 2033 find a similar effect of 0.3 about a 1 percent of GDP increase in health 0.2 spending caused by aging in a total spending 0.1 increase of 3.3 to 3.5 percent of GDP (Leung, 0 Tin, and Chan 2007). n . a sia zil ico a n y ep e in di pa io a rk ne ex Ch ,R In at Br Ja Tu Although still nascent in East Asia and er do M ea d r In Fe Ko Pacific, a further fiscal risk to consider in the ian ss long term for countries with aging populations Ru is public funding for long-term care. As chap- Cost pressure Cost containment o/w demographic ter 7 of this report shows, publicly financed effect long-term care is present in richer economies of Source: De la Maisonneuve and Oliveira Martins 2013. the region such as Japan, Korea, and Taiwan, Note: o/w = of which. 100 LIVE LONG AND PROSPER China, but remains in its infancy in the region’s the World Bank do not incorporate such developing economies. However, long-term interactions between demographics and fis- care is an area in which governments in rapidly cal accounts. This decision is understandable aging countries such as China, Thailand, and in settings where the focus is on short-run Vietnam are showing increasing interest and macrofiscal situations, but such interactions already making policy commitments. OECD become relevant when looking at long-term projections for long-term care in the region debt sustainability. A highly stylized effort point to a further (though for the foreseeable at incorporating demographics into debt sus- future relatively modest) source of fiscal pres- tainability analysis follows here. sure where—as for pensions—the aging effect The model underlying the following is strong (figure 3.14). simulations is a standard fiscal sustainability Each estimation approach has its strengths model with additional links to demograph- and limitations, but all point toward ics and age-related expenditure. The analysis significant fiscal risks from increased age- uses a standard fiscal sustainability frame- related spending in East Asia and Pacific, work linking GDP, fiscal accounts, and debt particularly in the aging middle-income with long-term demographic projections countries (largely orange economies in the and simple models that link demographics report typology). Chapters 5, 6, and 7 of to production and to age-related expendi- this report look at current pension, health, ture. The model is highly stylized to make and long-term care systems in the region and sure that interactions and results are well discuss reforms that aim to strike a balance understood, so only the channels from demo- between managing cost escalation and fiscal graphics to GDP and from demographics to risk, ensuring basic adequacy of benefits and age-related expenditure have been modeled. services, and expanding coverage where it is Figure 3.15 shows a schematic overview of currently less than complete. the links between demographics and fis- A more complex (and uncertain) picture cal accounts that underpin the simulation of fiscal pressures from aging emerges if we results. try to incorporate debt dynamics into fiscal Using the characteristics of 16 developing projections. Although the preceding fiscal East Asian and Pacific countries, a number of projections are very useful for understanding stylized cases are presented on debt dynamics orders of magnitude of age-related spend- that reflect different combinations of three ing pressures, they fail to take into account dimensions: that countries may be running fiscal deficits and hence financing public spending with • Aging pattern. Broadly, two types of debt that incurs further interest obliga- countries exist: those with young and tions. To develop a more complete analysis slowly aging populations (green econo- of whether an economy will have the fiscal mies) and those with older and faster- resources to deal with the effects of aging, aging populations (orange economies such a few more factors need to be taken into as China, Indonesia, Malaysia, Thailand, account: national investment, external cur- and Vietnam). rent account, foreign direct investment ver- • Age-related expenditure levels. For each sus debt financing, and total external debt. group, the simulations analyze the impact Countries that already have a structural cur- of current per capita public spending on rent account deficit and high external debt health care, education, and pensions in will have a smaller envelope to work with the long run. Thus, two levels of current than those with the same demographics and aging spending are seen: low spending saving behavior but with a current account (spending on health care, education, and surplus and low external debt. However, the pensions at 1.5 percent of GDP each at standard debt sustainability analysis mod- end-2013) and high spending (spending at els used by agencies such as the IMF and 2 percent of GDP each). AG I N G A N D T H E MAC R O E CO N O M Y I N E A S T A S I A A N D PAC I F I C 101 FIGURE 3.15 Schematic overview of the links between demographics and fiscal accounts Demographics Fiscal accounts Y=A×L Labor force A: productivity (exogenous) GDP Participation rate L: labor force (exogenous) Revenue (fixed as % of Demographics GDP) (exogenous projections)/working-age Expenditure: population • Nonaging, noninterest (exogenous) • Pensions • Health • Education • Interest Interest Vector of age-related Fiscal balance payments expenditure parameters × Vector of demographics = Age- Financing related expenditure Debt accumulation Financing assumptions Debt at t = 0 (interest rate) Source: Pradelli and Van Doorn 2015. • Initial debt levels. The initial (end-2013) populations and low age-related spending public debt burden affects long-term sus- have the best fiscal sustainability outlook, tainability because it amplifies deficits as we might expect. Maintaining a primary driven by age-related expenditure: low fiscal surplus and a healthy growth rate debt means public debt at 20 percent of (higher than the interest rate paid on pub- GDP, and high debt means 60 percent of lic debt) are the key determinants of the fis- GDP.13 cal sustainability of aging. In cases 1 and 2 (aging countries), both high primary fiscal The simulations demonstrate for a stylized deficits and adverse interest rate growth set of East Asian and Pacific countries the dynamics lead to continuous debt accumula- major variability in public debt over time tion. Cases 3 and 4 have the same spending from different interactions of initial debt lev- in the beginning but more favorable demo- els, spending on age-related programs, and graphics, which leads to both a narrower demographic trends (figure 3.16). The initial primary deficit and higher growth. The last level of debt is almost a pure level effect. four cases start with low age-related spend- The results show that the case of an aging ing, which allows them to run primary country with a high level of age-related surpluses, at least for the first 15 years. spending and a high level of initial debt is The countries with younger populations still clearly unsustainable, followed closely by grow relatively fast and are able to reduce the equally unsustainable case with a lower debt. Although these are only stylized cases level of initial debt. Countries with younger (built on composites of real country initial 102 LIVE LONG AND PROSPER FIGURE 3.16 Initial debt and spending levels and future debt Addressing concerns about saving in the dynamics have important interactions with demographics that can region will require improved social secu- sharply increase the fiscal impacts of aging rity and the deepening of financial mar- Stylized debt dynamics and composition with East Asian and Pacific kets and the role of banks. Ultimately, the demographics, 2013–60, under different starting combinations of spending impact of aging on capital formation in levels, debt, and demographics East Asia and Pacific will depend on the interaction of a number of factors: the pace, a. Public debt evolution for each stylized country case magnitude, and intensity of demographic 300 changes; the consequent behavioral adjust- 250 ments in response to fertility and longevity changes; and the macroeconomic environ- Share of GDP (%) 200 ment. Numerous behavioral factors spurred 150 by lower fertility and longevity suggest that the compositional effect of aging may 100 be significantly offset. Furthermore, the 50 region is already marked by high savings rates and in the future will be subject to 0 macroeconomic forces, international capi- 8 3 8 3 8 3 8 3 8 0 –1 –2 –2 –3 –3 –4 –4 –5 –5 –6 tal flows, financial deepening, and expan- 14 19 24 29 34 39 44 49 54 59 20 20 20 20 20 20 20 20 20 20 Years sion of social security, which together may Case 1: Aging, high spending, high debt Case 5: Aging, low spending, high debt dominate the older age composition effect. Case 2: Aging, high spending, low debt Case 6: Aging, low spending, low debt All of these factors make predictions of Case 3: Young, high spending, high debt Case 7: Young, low spending, high debt Case 4: Young, high spending, low debt Case 8: Young, low spending, low debt aging impacts on aggregate savings tenta- tive but suggest that the demographic effect b. Cumulative contributions to increase in public debt 2013–60 may not be the primary factor determining 250 future saving behavior. 200 150 Conclusion 100 The preceding analysis suggests that even Percent 50 rapidly aging countries in East Asia and 0 Pacific have significant scope to mitigate the potentially negative growth effects of aging, –50 but such action will require sometimes chal- –100 lenging policy reforms. It will also necessitate 1 2 3 4 5 6 7 8 social and cultural change in many countries Cases Interest rate–growth differential effect of the region that at best will happen only Primary deficit effect gradually. Younger countries have the good Change in public debt fortune of time before aging takes hold, but they need to use that time as productively Source: World Bank estimates in Pradelli and Van Doorn 2015. as possible to prepare for the rapid aging Note: Case 1 = aging, high spending, high debt; case 2 = aging, high spending, low debt; that will come in due course. For others, case 3 = young, high spending, high debt; case 4 = young, high spending, low debt; case 5 = aging, low spending, high debt; case 6 = aging, low spending, low debt; case 7 = young, low spending, the demands are more pressing, although the high debt; case 8 = young, low spending, low debt. examples of red economies demonstrate that East Asian societies have shown themselves conditions in East Asia and Pacific), they do more willing than those in other parts of the point to the importance of thinking about world to make sacrifices in the interest of sus- debt dynamics when considering fiscal risks taining healthy economies and intergenera- of aging—and conversely, of thinking about tional equity. The labor market aspects of the demographic dynamics when looking at needed reforms are discussed in the following long-term debt dynamics. chapter. AG I N G A N D T H E MAC R O E CO N O M Y I N E A S T A S I A A N D PAC I F I C 103 The bigger risks from aging in the region account. The implication is that they will be are fiscal and require urgent policy attention aging faster. in rapidly aging countries. Distinguishing the 2. This literature generally excludes the Pacific. demographic and economic challenges fac- 3. See for example Park, Lee, and Mason (2012), ing different groups of East Asian and Pacific which provides useful growth accounting countries is important as they think about results, finding for a selection of East and South Asian economies that only in Korea and aging populations: Hong Kong SAR, China, does the negative • For red economies, reducing the structural growth effect of rising old-age dependency outweigh the positive contribution of falling labor force decline from aging will be the youth dependency from 2011 to 2020. most critical challenge. Although the fis- 4. Interestingly, however, surveys of global cal risks are significant, a number of red attitudes to migration find more positive economies have already adopted forward- attitudes to immigration in some potential looking policies in areas such as pensions host countries in East Asia and Pacific. and health that will help contain the fis- Korea, for example, had the lowest share cal impact of further aging, although areas of respondents globally agreeing that such as long-term care will continue to immigration should be further restricted (at pose challenges (see chapters 5–7 of this 25 percent), and Japanese respondents were report for details). almost evenly split. These findings contrast to • For orange economies, the challenges of those in Malaysia, where almost 90 percent the demographic transition require sus- of respondents wanted more restrictions on immigration (Pew Research Center 2007). taining high productivity growth. At the 5. For literature on the link between age same time, they will need to mitigate the composition and growth, see Bloom and labor supply and fiscal impacts of rapid Williamson (1998); Feyrer (2007, 2008); aging through ongoing pension and health Lindh and Malmberg (1999); Modigliani reforms and labor policies to extend the (1986); Modigliani and Cao (2004); Persson working lives of urban and formal sector (1999); and Sarel (1997). workers. 6. This section draws on a background paper by • For the younger green economies, the pri- Malathi Velamuri (2015). orities are to establish the conditions to 7. Studies that have shown strong links between realize maximum GDP growth from the national savings rates and age structure demographic dividend (that is, invest- include those by Bloom et al. (2007); Deaton ments to raise productivity and maximize and Paxson (1997); Fry and Mason (1982); youth employment) and to avoid adopt- Higgins (1998); Kelley and Schmidt (1995); ing policies in areas such as pensions and Lee, Mason, and Miller (2000); and Mason (1988). health that are affordable now but would 8. In theory, a stationary population has no rapidly become unsustainable once aging aggregate savings because in the absence of accelerates. a bequest motive, the dissaving of the old The main areas of services and entitlement and the saving of the young cancel out (Ando reform—pensions, health, and long-term and Modigliani 1963; Bloom, Canning, and Sevilla 2003). However, when the population care—are discussed in chapters 5–7 of this grows or the economy is growing rapidly and report. incomes of wage earners are high relative to the incomes of the retired (as in East Asia and Pacific), aggregate savings or dissavings may Notes occur. 1. Because only elderly dependency ratios are 9. See Horioka (2010) and the references therein. shown here, the patterns may not perfectly 10. See Deaton (1992) for Thailand, and Deaton match the patterns of aging across groups. and Paxson (1993) for Taiwan, China. Some countries, such as Singapore, appear 11. See the National Transfer Accounts website, much younger than they are when the rest http://www.ntaccounts.org/web/nta/show of the population age structure is taken into / NTA%20Data. 104 LIVE LONG AND PROSPER 12. A lt houg h t he e st i m ate s u s e st yl i z ed Change, Social S ecu rit y Systems, and composites from across APEC, two-thirds of Savings.” Journal of Monetary Economics 54 the economies are from East Asia and Pacific, (1): 92–114. and they particularly dominate the categories Bloom, David, David Canning, and Michael other than high-income mature. Moore. 2004. “The Effect of Improvements in 13. The model and simulations are discussed in Health and Longevity on Optimal Retirement Pradelli and van Doorn (2015). and Saving.” NBER Working Paper 10919, National Bureau of Economic Research, Cambridge, MA. References Bloom, David, David Canning, and Jaypee Sevilla. AARP (American Association of Retired Persons). 2003. The Demographic Dividend: A New 2002. Staying Ahead of the Curve: The Perspective on the Economic Consequences AARP Work and Career Study. Washington, of Population Change. Santa Monica, CA: DC: AARP. RAND. Ando, Albert, and Franco Modigliani. 1963. “The Bloom, David, Klaus Prettner, and Holger Strulik. ‘Life-Cycle’ Hypothesis of Saving: Aggregate 2013. “Declining Fertility and Economic Well- Implications and Tests.” American Economic Being: Do Education and Health Ride to the Review 53 (1): 55–84. Rescue?” Labor Economics 22 (C): 70–79. Attanasio, Orazio P., and Miguel Székely. 2000. Bloom, David, and Alfonso Sousa-Poza. 2013. “Household Saving in Developing Countries: “Ageing and Productivity: Introduction.” IZA Inequality, Demographics, and All That— Discussion Paper 7205, Institute for the Study How Different Are Latin America and South of Labor, Bonn, Germany. East Asia?” I DB Research Department Bloom, David, and Jeffrey G. Williamson. 1998. Wo r k i n g Pap e r 427, I nt e r - A m e r i c a n “Demographic Transitions and Economic Development Bank, Washington, DC. Miracles in Emerging Asia.” World Bank Autor, David H. 2003. “Outsourcing at Will: The Economic Review 12 (3): 419–55. Contribution of Unjust Dismissal Doctrine to Börsch-Supan, Axel, and Matthias Weiss. 2008. the Growth of Employment Outsourcing.” “Productivity and the Age Composition of Journal of Labor Economics 21 (1): 1–42. Work Teams: Evidence from the Assembly Baltes, Paul B., Ulman Lindenberger, and Ursula Line.” Universitätsbibliothek der Universität M. Staudinger. 2006. “Life Span Theory in Mannheim, Mannheim, Germany. Developmental Psychology.” In Handbook of Browning, Martin, and Thomas F. Crossley. 2001. Child Psychology, vol. 1, edited by William “The Life-Cycle Model of Consumption and Damon and Richard M. Lerner, 569–95. Saving.” Journal of Economic Perspectives Hoboken, NJ: John Wiley & Sons. 15 (3): 3–22. Becker, Gary, and H. Gregg Lewis. 1973. “On the Burtless, Gary. 2013. “The Impact of Population Interaction between Quantity and Quality of Aging and Delayed Retirement on Workforce Children.” Journal of Political Economy 81 Productivity.” CRR Working Paper 11-2013, (2): S279–88. Center for Retirement Research at Boston Blanchard, Olivier, and Francesco Giavazzi. College, Boston, MA. 2005. “Rebalancing Growth in China: A Chamon, Marcos D., and Eswar S. Prasad. 2010. Three-Handed Approach.” MIT Department “Why Are Saving Rates of Urban Households of E conom ic s Work i ng Pap er 05 -32 , in China Rising.” American Economic Massachusetts Institute of Technology, Journal: Macroeconomics 2 (1): 93–130. Cambridge, MA. Cherif, Reda, and Fuad Hasanov. 2012. “The Bloom, David, David Canning, Günther Fink, and Volatilit y Trap: Precautionar y Saving, Jocelyn E. Finlay. 2007. “Does Age Structure Investment, and Aggregate Risk.” I M F Forecast Economic Growth?” International Work i ng Pap er 12 /13 4, I nter nat iona l Journal of Forecasting 23 (4): 569–85. Monetary Fund, Washington, DC. Bloom, David, David Canning, and Pia N. Christensen, Helen, A. S. Henderson, Kathleen Malaney. 2000. “Population Dynamics and Griffiths, and Carmel Levings. 1997. “Does Economic Growth in Asia.” Population and Aging Inevitably Lead to Declines in Cognitive Development Review 26 (suppl.): 257–90. Performance? A Longitudinal Study of Elite Bloom David, David Canning, Rick Mansfield, Academics.” Personality and Individual and Michael Moore. 2007. “Demographic Differences 23 (1): 67–78. AG I N G A N D T H E MAC R O E CO N O M Y I N E A S T A S I A A N D PAC I F I C 105 Cutler, David, and Adriana Lleras-Muney. Fukao, Kyoji, Ryo Kambayashi, Daiji Kawaguchi, 2008. “Education and Health: Hyeog Ug Kwon, Young Gak Kim, and Izumi Evaluating Theories and Evidence.” In Yokoyama. 2006. “Deferred Compensation: Making Americans Healthier: Social and Evidence from Employer-Employee Matched Economic Policy as Health Policy, edited by Data from Japan.” Discussion Paper 187, Robert F. Schoeni, James S. House, George Institute of Economic Research, Hitotsubashi A. Kaplan, and Harold Pollack, 29–60. New University, Tokyo. York: Russell Sage Foundation. Global Entrepreneurship Monitor (database). D’Addio, Anna, Mark Keese, and Edward 2011. Global Entrepreneurship Research W hitehouse. 2010. “Population Ageing A ssociation, L ondon Business S chool, and Labour Markets.” Oxford Review of London, http://www.gemconsortium.org Economic Policy 26 (4): 613–35. / Data. De la Maisonneuve, Christine, and Joaquim Göbel, Christian, and Thomas Zwick. 2013. “Are Oliveira Martins. 2013. Public Spending on Personnel Measures Effective in Increasing Health and Long-Term Care: A New Set of Productivit y of Old Workers?” L abor Projections. Paris: Organisation for Economic Economics 22: 80–93. Co-operation and Development. Gómez, Rafael, and Pablo Hernández de Cos. Deaton, Angus. 1992. “Household Saving 2008. “Does Population Ageing Promote in LDCs: Credit Markets, Insurance and Faster Economic Growth?” Review of Income Welfare.” Scandinavian Journal of Economics and Wealth 54 (3): 350–72. 94 (2): 253–73. Grund, Christian, and Niels Westergaard-Nielsen. Deaton, Angus, and Christina H. Paxson. 1993. 2008. “Age Structure of the Workforce and “Intertemporal Choice and Inequality.” Firm Performance.” International Journal of NBER Working Paper 4328, National Bureau Manpower 29 (5): 410–22. of Economic Research, Cambridge, MA. Hanushek, Eric, and Ludger Woessmann. 2009. ———. 1997. “The Effects of Economic and “Do Better Schools Lead to More Growth? Population Growth on National Saving and Cognitive Skills, Economic Outcomes, and Inequality.” Demography 34 (1): 97–114. Causation.” NBER Working Paper 14633, Demirgüç-Kunt, Aslı, and Leora Klapper. 2012. National Bureau of Economic Research, “Measuring Financial Inclusion: The Global Cambridge, MA. Findex Database.” Policy Research Working Higgins, Matthew. 1998. “Demography, National Paper 6025, World Bank, Washington, DC. Savings, and International Capital Flows.” Eggleston, Karen, and Victor Fuchs. 2012. “The International Economic Review 39: 343–69. New Demographic Transition: Most Gains in Hinz, Richard, and Asta Zviniene. 2011. “The Life Expectancy Now Realized Late in Life.” Potential Fiscal Effects of Population Aging: Journal of Economic Perspectives 26 (3): Simulation of Aggregate Pension and Health 137–56. Expenditure for Four Stylized Countries at Feyrer, James. 20 07. “Demographics and Various Levels of Economic Development.” Productivity.” Review of Economics and World Bank, Social Protection and Labor Statistics 89 (1): 100–109. Practice, Washington, DC. ———. 2008. “Aggregate Evidence on the Link Horioka, Charles Yuji. 2010. “The (Dis)saving between Age Structure and Productivity.” Behavior of the Aged in Japan.” Japan and the Population and Development Review 34 World Economy 22 (3): 151–58. (suppl.): 78–99. Horioka, Charles Yuji, and Akiko Terada- Flochel, Thomas, Yuki Ikeda, Harry Moroz, and Hagiwara. 2012. “The Determinants and Nithin Umapathi. 2014. “Macroeconomic Long-Term Projections of Saving Rates in Implications of Aging: Demographics, Labor Developing Asia.” Japan and the World Markets, and Productivity in East Asia Economy 24 (2): 128–37. Pacific.” Background paper for the East Asia IHME (Institute for Health Metrics) and World and Pacific regional report on aging, World Bank. 2013. The Global Burden of Disease: Bank, Washington, DC. Generating Evidence, Guiding Policy—East Fry, Maxwell J., and Andrew Mason. 1982. Asia and Pacific Regional Edition. Seattle, “The Variable Rate-of-Growth Effect in the WA: IHME. Life-Cycle Saving Model.” Economic Inquiry ILO (International Labour Organization). 2011. 20 (3): 426–42. LABORSTA (database), 6th ed., ILO, Geneva. 106 LIVE LONG AND PROSPER IMF (International Monetary Fund). 2009. of Cross Countries.” Journal of Development “Corporate Savings and Rebalancing in Economics 84 (1): 138–54. Asia.” IMF Asia and Pacific Department, Lindh, Thomas, and Bo Malmberg. 1999. Regional Economic Outlook (October), “Age Structure Effects and Growth in the Washington, DC. OECD, 1950–1990.” Journal of Population IMF (International Monetary Fund) Fiscal Affairs Economics 12 (3): 431–49. Department. 2010. Macro-fiscal Implications Loayza, Norman, Klaus Schmidt-Hebbel, and of Health Care Reform in Advanced and Luis Serven. 2000. “Saving in Developing Emerging Economies . Washington, DC: Countries.” World Bank Economic Review International Monetary Fund. 14 (3): 393–414. Kelley, Allen C., and Robert M. Schmidt. 1995. Ma, Guonan, and Wang Yi. 2010. “China’s High “Aggregate Population and Economic Growth Saving Rate: Myth and Reality.” BIS Working Correlations: The Role of the Components of Paper 312, Bank for International Settlements, Demographic Change.” Demography 32 (4): Basel, Switzerland. http://www.bis.org/publ 543–55. /work312.htm. ———. 2005. “Evolution of Recent Economic- Malmberg, Bo, Thomas Lindh, and Max Demographic Modeling: A Synthesis.” Journal Halvarsson. 2008. “Productivity Consequences of Population Economics 18 (2): 275–300. of Workforce Aging: Stagnation or Horndal Kessler, Eva-Marie, and Ursula M. Staudinger. Effect?” Population and Development Review 2007. “Intergenerational Potential: Effects of 34: 238–56. Social Interaction between Older Adults and Martocchio, Joseph J. 1989. “Age-Related Adolescents.” Psychology and Aging 22 (4): Differences in Employee Absenteeism: A 690–704. Meta-Analytic Review.” Psychology and Kinugasa, Tomoko, and Andrew Mason. 2007. Aging 4 (4): 409–14. “Why Nations Become Wealthy: The Effects Mason, Andrew. 1988. “Saving Economic Growth of Adult Longevity on Saving.” World and Demographic Change.” Population and Development 35 (1): 1–23. Development Review 14 (1): 113–44. Lee, Ronald, and Andrew Mason. 2010. “Some McEvoy, Glenn M., and Wayne F. Cascio. Macroeconomic Aspects of Global Population 1 9 8 9. C u m u l a t i v e E v i d e n c e o f t h e Aging.” Demography 47 (1): S151–72. Relationship between Employee Age and Job ———. 2011. Po p ul atio n A g i ng a n d t he Performance. Journal of Applied Psychology Generational Economy: A Global Perspective. 74 (1): 11–17. Northampton, MA: Edward Elgar. Mod ig lian i, Franco. 1970. “T he Life Lee, Ronald, Andrew Mason, and Timothy Cycle Hypothesis of Saving and M iller. 20 0 0. “Life Cycle Saving and Intercountry Differences in the Saving Ratio.” Demog raphic Transition: T he Case of In Induction, Growth, and Trade: Essays in Taiwan.” Population and Development Honor of Sir Roy Harrod, edited by Walter A. Review 26: 194–222. Eltis, Maurice F. Scott, and James N. Wolfe, Leung, Gabriel M., Keith Y. Tin, and Wai-Sum 197–225. Oxford, U.K.: Clarendon Press. Chan. 2007. “Hong Kong’s Health Spending ———. 1986. “Life Cycle, Individual Thrift, and Projections through 2033.” Health Policy 81 the Wealth of Nations.” American Economic (1): 93–101. Review 76 (3): 297–313. Levine, Ross. 2003. “More on Finance and Modigliani, Franco, and Shi Larry Cao. 2004. Growth: More Finance, More Growth?” “The Chinese Saving Puzzle and the Life- Review of the Federal Reserve Bank of Cycle Hypothesis.” Journal of Economic St. Louis (July–August): 31–46. Literature 42 (1): 145–70. Levine, Ross, Norman Loayza, and Thorsten Murphy, Kevin M., and Finis Welch. 1990. Beck. 2000. “Financial Intermediation and “Empirical Age-Earnings Profiles.” Journal of Growth: Causality and Causes.” Journal of Labor Economics 8 (2): 202–29. Monetary Economics 46 (1): 31–77. Ochiai, Katsuaki. 2008. “Where Is the Key to Li, Hongbin, Jie Zhang, and Junsen Zhang. 2007. Improve Productivity? An Analysis on Gender, “Effects of Longevity and Dependency Rates Education, and Seniority [in Japanese].” on Saving and Growth: Evidence from a Panel In Human Capital for the Future in Japan AG I N G A N D T H E MAC R O E CO N O M Y I N E A S T A S I A A N D PAC I F I C 107 [in Japanese], 29–47. Tokyo: Japanese Center Paper 97/97, International Monetary Fund, for Economic Research. Washington, DC. OECD (Organisation for Economic Co-operation S c hu l t z , T. Pau l . 2 0 0 4 . “ D e m o g r ap h i c and Development). 2012. PISA 2012 Results: Determinants of Savings: Estimating and What Students Know and Can Do: Student Interpreting the Aggregate Association in Performance in Mathematics, Reading and Asia.” Working Paper 901, Economic Growth Science. Vol. 1. Paris: OECD. Center, Yale University, New Haven, CT. Özden, Çaglar, and Mauro Testaverde. 2014. Shinada, Naoki. 2011. “Quality of Labor, Capital, “International Migration in Aging Societies: and Productivity Growth in Japan: Effects Impacts and Benefits in East Asia and Pacific of Employee Age, Seniority, and Capital Countries.” Background paper for the East Vintage.” RIETI Discussion Paper 11-E-036, Asia and Pacific regional report on aging, Research Institute of Economy, Trade, and World Bank, Washington, DC. Industry, Tokyo. Park, Donghyun, Sang-Hyop Lee, and Andrew Shirakawa, H i romichi. 2010. “E mpirical Mason, eds. 2012. Aging, Economic Growth, Analysis of Relationship between Worker and Old-Age Security in Asia. Manila: Asian Age, Productivity, and Real Wages.” In Is the Development Bank; Cheltenham, U.K.: Aging of Society a Threat to Japan?, 48–71. Edward Elgar. Tokyo: National Institute for Research Persson, Joakim. 1999. “Demographic and Per Advancement. Capita Income Dynamics: A Convergence Skirbekk, Vegard. 2008. “Age and Productivity Study on Demographics, Human Capital, Capacity: Descriptions, Causes, and Policy a n d Pe r C ap i t a I n c o m e f o r t h e U S Options.” Ageing Horizons 8: 4–12. States.” FIEF Working Paper 1999/156, Skirbekk, Vegard, Marcin Stonawski, Eric Fackföreningsrörelsens Institut för Ekonomisk Bonsang, and Ursula M. Staudinger. 2013. Forskning, Stockholm. “The Flynn Effect and Population Aging.” Pew Research Center. 2007. Global Unease Intelligence 41 (3): 169–77. with Major World Powers: 47-Nation Pew UN (United Nations). 2013. “World Population Global Attitudes Survey. Washington, DC: Prospects: The 2012 Revision—Highlights Pew Research Center. http://www.pewglobal and Advance Tables.” Working Paper ESA/P/ .org/files/pdf/2007%20Pew%20Global%20 WP.228, Population Division, Department of Attitudes%20Report%20-%20June%2027 Economic and Social Affairs, United Nations, .pdf. New York. Pradelli, Juan, and Ralph Van Doorn. 2015. Van Ours, Jan C. 2009. “Will You Still Need “Fiscal Impact of Aging.” Background paper Me: When I’m 64?” De Economist 157 (4): for the East Asia and Pacific regional report 441–60. on aging, World Bank, Washington, DC. Velamuri, Malathi. 2015. “Aging and Savings in Rannan-Eliya, Ravi P., Chamara Anuranga, Adilius Asia: Current Trends and Future Prospects.” Manual, Ruwani Wickramasinghe, Prasadini Background paper for the East Asia and Perera, Shanaz Saleem, Jayalal Chandrasiri, Pacific regional report on aging, World Bank, Indra Pathmanathan, and Institute for Health Washington, DC. Systems Research Core Team. 2013. Malaysia Voelcker-Rehage, Claudia, Ben Godde, and Healthcare Demand Analysis: Inequalities in Ursula M. Staudinger. 2011. “Effects of Healthcare Demand and Simulation of Trends Different Exercise Interventions on Cognitive and Impact of Potential Changes in Healthcare Functioning in Older Adults.” Journal of Spending. Kuala Lumpur: Institute for Health Psychophysiology 25: 29. Systems Research. World Development Indicators (database). Romeu Gordo, Laura, and Vegard Skirbekk. Various years. World Bank, Washington, DC. 2013. “Skill Demand and the Comparative http://data.worldbank.org/data-catalog/world Advantage of Age: Jobs Tasks and Earnings -development-indicators. from the 1980s to the 2000s in Germany.” Willis, Robert. 1973. “A New Approach to the Labor Economics 22 (suppl.): 61–69. Economic Theory of Fertility Behaviour.” Sarel, Michael. 1997. “Growth and Productivity Journal of Political Economy 81 (2, pt. 2): i n A SE A N C ou nt ries.” I M F Work i ng S14–64. 108 LIVE LONG AND PROSPER World Bank. 2013a. “Global Saving in 2030.” In ———. 2013c. Skills for Quality Jobs and Capital for the Future: Saving and Investment Development in Lao PDR: A Technical in an Inte rde pe n de nt World , 61–112 . A s s e s s m e n t of t he C u r re n t C o n t e x t . Washington, DC: World Bank. Washington, DC: World Bank. ———. 2013b. “Labor Market Issues in Timor-Leste: ———. 2 0 1 3 d . W o r l d D e v e l o p m e n t Current State, Prospects, and Challenges.” I n d i c a t o r s 2 0 1 3 . Wa s h i n g t o n , D C : World Bank, Washington, DC. World Bank. Labor Market Policies to Address Aging in East Asia and Pacific 4 Introduction This chapter examines the experience with labor market policies to address aging Concerns over shrinking labor forces and out- in red economies of East Asia and Pacific put have prompted many policy initiatives in and elsewhere, drawing out policy implica- richer economies of East Asia and Pacific that tions for East Asian and Pacific countries may provide lessons for middle-income coun- going forward. The first section discusses tries experiencing rapid aging. Earlier chapters policies to stimulate fertility and female have illustrated the challenges of declining labor force participation as well as legacy labor forces in rapidly aging East Asian and policies that may counteract such initia- Pacific countries, albeit mitigated by improv- tives. The second section reviews efforts in ing worker quality and other factors. The the region and globally to extend produc- emerging experience of the region’s red econo- tive working lives with a primary focus on mies (continuing the typology introduced in the formal sector, given the already overlong chapter 1) in responding to these challenges working lives of those in the region’s rural provides some useful lessons for labor market and informal sectors. Migration policies are policy responses. One key theme is the need for discussed in the third section. The final sec- policy action across the life cycle in the context tion briefly discusses policies to stimulate of rapid aging. These strategies include efforts lifelong learning in the context of enhanc- to stimulate fertility in countries with aging ing both overall labor quality in the face populations, increase female labor force par- of declining labor supply and employment ticipation, encourage immigration to augment prospects of older workers. Compared with shrinking supplies of local labor, and extend other chapters, this chapter relies more heav- the productive lives of older workers. ily on examples from richer red economies This chapter draws from a background literature review for this report by Shonali Sen (2014), a back- ground paper on migration in East Asia and Pacific by Çaglar Özden and Mauro Testaverde (2014), inputs on labor and aging policies in the Republic of Korea by Dewen Wang, and a study visit to Japan by the task team. 109 110 LIVE LONG AND PROSPER and OECD countries, because they are more breaks and rebates for families with young advanced in trying to address the labor mar- children or large families; and generalized ket consequences of rapid aging. child allowances. In some cases, national initiatives have been supplemented by sub- national programs, as in Seoul. For several Fertility incentives East Asian and Pacific countries, such mea- In recent decades, multiple initiatives have sures have lagged the decline in fertility below been undertaken in East Asian and Pacific replacement, reflecting the legacy of earlier countries with aged populations to reverse very family planning policies that were successful low fertility rates and to increase the participa- in rapidly reducing previously high fertility tion of women in the labor force. As noted in rates. Korea, for example, introduced its first chapter 1, the richer red economies have expe- pronatalist strategy (First Basic Plan for Low rienced very low fertility rates for some time, Fertility and Aged Society) only in 2006, well and their earlier trajectories of fertility decline after fertility had declined to very low levels. are seemingly being followed by a number of The balance of evidence from East Asia middle-income countries in the region, includ- and Pacific on direct incentives to have more ing China, Thailand, and Vietnam. In addi- children suggests they have not been particu- tion, although East Asia and Pacific generally larly effective or cost-effective.1 Evaluation of has relatively high female labor force par- Japan’s history of pronatalist initiatives since ticipation by global standards, a number of the early 1990s found almost no effect on countries, including Fiji, Japan, Malaysia, and the likelihood of second and third children, the Republic of Korea, face major challenges although a slight recovery has occurred in getting women to reenter the workforce after women having at least one child (Ma 2010).2 childbirth (World Bank 2012). This difficulty The consistently low fertility rates across the has led to many initiatives to stimulate fertil- region’s red economies suggest that similar ity directly and to promote family-friendly conclusions are likely to be reached and that employment policies in countries such as a wider array of policies and changes in social Japan and Korea that are attempting to attitudes are likely to be needed. In a different encourage women to return to the labor force context, for example, surveys in urban China after childbirth. Coherence across fertility and found that only a minority of families who labor policies has not always been a strong are eligible to have second children wish to do point, however; for example, evidence from so (about 18 percent to 24 percent of eligible countries of the Organisation for Economic families in Beijing and less than 40 percent of Co-operation and Development (OECD) indi- eligible families in Jiangsu) (Jones 2009). cates that high child allowances reduce incen- tives for women to return to work (IMF 2012). Economies including Japan; Korea; Enhancing female labor force Singapore; Taiwan, China; and Vietnam have participation introduced different combinations of mea- Paid parental leave has been spreading sures to help stimulate fertility. Measures used across the region, although for developing in the region have included direct payments East Asian and Pacific countries, it is mostly to families for second, third, and fourth chil- restricted to the formal sector and thus dren (the “baby bonus,” for example, paid excludes the majority of families. A study for six years after birth in Singapore since of 13 East Asian and Pacific countries finds 2000, and the “mother hero” lifetime ben- maternity leave in all of them, most often efits in Mongolia for mothers having four or employer financed (World Bank 2012). For more children) and for any childbirth (as in example, maternity leave in Japan is for up Vietnam, where a lump sum of two months’ to one year and paid at 50 percent of salary, pay is provided within the social insurance which is around the OECD average. Over system); priority access to public housing; tax the years, take-up among eligible women L A B O R MA R K E T P O L I C I E S TO A D D R E S S AG I N G I N E A S T A S I A A N D PAC I F I C 111 has increased significantly, from just under Van Doorne-Huiskes 2007). Positive impacts 50 percent in 1996 to 84 percent by 2012 on female employment have been found also (MHLW 2013). In developing East Asian in middle-income countries such as urban and Pacific countries, however, maternity Brazil and Colombia (under its Hogares leave is typically only for those within for- Comunitarios program). In East Asia and mal social security schemes (a minority in Pacific, however, evidence from countries at nearly all cases) and tends to be shorter: 98 different income levels (including Fiji, Japan, days in China, three months in Thailand at Korea, and Malaysia) shows sustained drops 50 percent of salary, and up to six months in female labor force participation after in Vietnam.3 Leave for fathers is limited but giving birth, pointing to the challenges of also increasing across the region, although it balancing child-rearing responsibilities and remains modest, where available, in devel- work (World Bank 2012). oping East Asian and Pacific countries (for In response to the challenges of reentry example, five to seven days in Vietnam), and into the workforce after childbirth, several take-up tends to be low. In Japan, fewer than East Asian and Pacific countries have subsi- 2 percent of men took their parental leave in dized child care in different ways. Overall, 2012, similar to the 2 percent who avail them- East Asian and Pacific countries that have selves of such leave in France and Germany promoted child care tend to use subsidies but much lower than the 70 percent take- and incentives rather than mandates on up rate in Sweden (Steinberg and Nakane employers, an alternative approach taken in 2012).4 For OECD countries, parental leave countries such as Brazil, Chile, and India, after childbirth is positively, although mod- which mandate crèche provision by employ- estly, associated with female employment and ers over a certain size. Within East Asia and hours worked, although the positive impact Pacific, Korea has introduced subsidized day reverses when paid leave is longer than two care for children up to four years of age, with years (Thévenon and Solaz 2013). Evaluation variable subsidies according to family income. of the effects of paid parental leave in East Singapore provides a subsidy for children Asian and Pacific countries is limited, but evi- under seven years old in child care centers dence from Japan (Waldfogel, Higuchi, and when their mothers are working, with higher Abe 1999) indicates a much stronger impact subsidies for infants and low-income families. on likelihood of women’s return to work after Malaysia, in contrast, works through the tax childbirth than in the United Kingdom or the system, making child care facilities an allow- United States (Suruga and Cho 2003), and able expense for employers and making pay- employers providing parental leave report ments by workers tax exempt. However, only more women return to work. 166 centers—of which only 26 were in pri- Across the world, public subsidies for vate organizations—were established nation- child care affect female labor force par- ally in response to this policy. In Thailand, ticipation positively, although the degree the trade unions are involved in provision of of impact depends on country context and crèches and receive only very modest support alternative options for informal care (see from municipalities (for example, through chapter 2 of this report on family-provided milk subsidies). Vietnam has taken a differ- child care). A study of 18 OECD countries ent approach for informal sector women by (Thévenon 2013) found that the level of lengthening the school day, especially where public spending on child care services for attendance is usually a half day. China has parents of children under three years old taken a somewhat similar approach in its significantly affects employment of women efforts to promote mass coverage of early ages 25–54. Similarly, analysis from Europe child development programs. finds that public subsidization of child care In East Asian and Pacific countries where is associated with higher rates of women’s public subsidies for child care are significant, employment and fertility (Den Dulk and they have had some success in stimulating 112 LIVE LONG AND PROSPER women’s employment. However, disentan- FIGURE 4.1 The Kurumin mark is given to firms in gling the effect of child care alone is not Japan that pursue child-friendly policies always easy when such initiatives are bundled with other measures, as discussed below. Available assessments come from richer East Asian and Pacific economies and show posi- tive effects of child care provision in Japan (higher than cash allowances) and Korea (Kim 2007; Steinberg and Nakane 2012). When child care provision and subsidies were sharply reduced in Mongolia (World Bank 2013b; World Bank and ADB 2005) and China (Chi and Li 2008; Du and Dong 2010; Maurer-Fazio et al. 2011) in the context of state enterprise restructuring, female labor force participation fell in urban areas, which researchers attribute in part to the rising costs of child care. The emerging practice in richer East Asian and Pacific economies is to bundle packages of measures that aim to stimulate Korea’s Saeromaji Plan 2010 provided a female labor force participation, especially similar package of benefits, including lower after childbirth. In 2007, Singapore intro- taxes for families. duced a package of training programs, flex- Although OECD countries historically ible working arrangements, targeted active saw a trade-off between childbearing and labor programs and work fairs, and other increased labor force participation of women, measures under the auspices of its Tripartite recent evidence indicates this is mostly no Workgroup on Enhancing Employment longer the case. As shown in figure 4.2, the Choice for women. These programs com- relationship between women’s work and plemented a range of measures to stimulate fertility behavior has reversed over the past fertility. Taiwan, China, bundled paren- few decades in the OECD, and having more tal leave benefits, child care subsidies, and children is associated with higher rates of early childhood development programs employment among women ages 25–54. under its White Book of Population Policy However, while both Korea and Japan have (Ministry of Interior 2008). Japan has also seen rising female labor force participation, had a series of initiatives (including the they have not managed to combine this with Angel Plans in the 1990s and the Support increased fertility in the way that a number Plan of 2004) to help combine work and of other OECD countries have. The factors child rearing, including child care subsidies, behind such divergent patterns are discussed child allowances, and employer incentives to in box 4.1. become more family-friendly (Ma 2010).5 Notably, some East Asian and Pacific Japan’s target is to increase the percentage countries actually provide disincentives of women resuming their careers after child- to women’s work. A previously common birth from 38 percent in 2010 to 55 percent (although declining) feature of labor laws in by 2020. Firms judged to be taking proac- East Asia and Pacific and other parts of the tive measures to promote family-friendly world has been gender-specific restrictions workplaces are also given a Kurumin mark on the sectors, types of work, and hours of (figure 4.1), and women’s reemployment is work for women. Within the International supported through various public programs, Labour Organization, such policies are no such as the Mothers Hello Work program. longer considered to promote the interests of L A B O R MA R K E T P O L I C I E S TO A D D R E S S AG I N G I N E A S T A S I A A N D PAC I F I C 113 FIGURE 4.2 The relationship between women’s work and fertility has reversed sharply in recent decades in most member countries of the OECD Employment rates of women ages 25–54 and total fertility rates, 1980 and 2009, selected countries a. 1980 b. 2009 3.5 2.2 IRL NZL ISL MEX IRL USA 2.0 AUS FRA NOR 3.0 SWE KOR GBR FINDNK 1.8 BEL NLD Total fertility rate Total fertility rate 2.5 CAN GRC 1.6 LUX EST PRT SLV 2.0 NLZ FRA GBR GRC CZE CHE AUS SVK USA 1.4 R2 = 0.05 ITA ESP POL AUT ITA BEL JPN JPN DEU NOR SWE CAN 1.5 DEU DNK FIN HUN PRT NLD 1.2 LUX R2 = 0.16 KOR 1.0 1.0 30 40 50 60 70 80 50 60 70 80 90 Employment rate of women ages 25–54 Employment rate of women ages 25–54 Source: OECD 2011. Note: OECD = Organisation for Economic Co-operation and Development. Country name abbreviations are ISO 3166 standard. BOX 4.1 Fertility and public policy in East Asia and Pacific and the OECD Much of East Asia and Pacific already has below- and care commitments. The drop happened at a time replacement levels of fertility. A number of national of increased female participation in higher educa- governments in the region (Japan, the Republic of tion and the labor market. Increasing participation Korea, and Singapore) are especially concerned by in higher education, coupled with evolving societal the ultra-low total fertility rate (TFR) at or below attitudes toward gender roles, influenced career aspi- 1.3 children per woman, which is comparable to rations as well as partnership preferences, such as some of the lowest TFR countries, such as Germany, lower rates of marriage. These choices resulted in not Hungary, and Portugal. In comparison, the Organ- only postponement of childbearing but also increased isation for Economic Co-operation and Development propensity for childlessness. However, the relation- (OECD) average TFR is 1.74 children per woman. ship with employment has reversed in Europe since Although China is not considered to be one of the the 2000s (OECD 2011), as seen in many OECD “lowest low,” its official TFR is well below replace- countries, such as Greece, Hungary, Italy, Japan, ment level at around 1.6 children per woman (and Korea, Poland, and Spain, where low female employ- could be considerably lower, according to estimates ment rates coexist with low fertility. from a number of well-respected researchers). Since 2002, a small rebound has occurred in These fertility levels raise two important questions fertility in about half the OECD countries (OECD for East Asian and Pacific governments: What are 2011), though whether this trend will sustain over the main drivers of fertility decline and recovery, and time and can be replicated in low-fertility countries what public support is effective in enabling people of East Asia and Pacific is unclear. One part of the to realize their desired fertility plans? In Europe, fer- recovery is explained by a fertility catch-up, because tility rates declining from their 1970 levels generally births that occurred at young ages in the past are were a result of women’s inability to match career no longer taking place but recover as births take box continues next page 114 LIVE LONG AND PROSPER BOX 4.1 Fertility and public policy in East Asia and Pacific and the OECD (continued) place at older ages (Bongaarts and Sobotka 2012). countries, suggesting that fertility recovery may not be Another major influence behind the recovery is easy in the region. These features include less flexible improved public support. An OECD (2011) report labor markets, lack of public support, and traditional observes that although direct financial support had a attitudes toward gender roles related to care commit- temporary and limited impact on birth rates, invest- ments. High financial costs of child care and housing ments in formal child care services and flexible labor in Japan and Korea have also constrained people in markets that also involve a greater degree of per- realizing their fertility plans. Lack of stable employ- manent employment have been most effective. For ment for both parents seems to have a particularly instance, Nordic countries have successfully shifted adverse effect on the decision to have a second child perceptions about combining parenthood and career (Rendall et al. 2014) and is likely to be another fac- aspirations through universally accessible parental tor that challenges fertility recovery in East Asia and support, including paid and job-protected parental Pacific. In China and Vietnam, explicit fertility poli- leave, flexible work arrangements, and subsidized cies have played a more specific role. Although these child care and schooling. As a result, Nordic coun- policies clearly affected past fertility trends, to what tries have the highest fertility rates in the OECD. extent further relaxation of policies would result in Several structural features distinguish low-fertility a sustained increase in fertility is less clear, given the East Asian and Pacific countries from the OECD range of other factors at play in fertility decisions. women, and they seem even less appropriate men in child rearing. This issue is a chal- in countries experiencing rapid aging. More lenge in much of the world. Countries such specifically, incentives may exist in tax sys- as Singapore, through public information tems that encourage women to stay out of efforts, and Japan, through its 2005 Gender the workforce. Although promoted as part Equality Law, are aiming to address social of family-friendly policies, such measures attitudes toward the roles of men and women increasingly run counter to other policies to in family life. Japan is also targeting greater encourage return to work after childbirth. efforts at men with its Ikumen project, which The Japanese tax and pension systems are one recognizes individual men and groups of men example: dependent exemptions for heads of who take an active role in parenting.6 household (usually men) are provided as long as the dependent spouse’s income is under a fairly modest threshold, a measure considered Extending productive working to discourage full-time work among women. lives Similarly, the pension system has incentives As societies in East Asia and Pacific age, for nonworking women. These policies are urbanize, and formalize, the challenge of under review. More generally, analysis points ensuring long and productive working lives to the positive association between neutral becomes ever more pressing. As East Asia tax treatment of second earners, child care and Pacific continues its rapid urbanization, subsidies, and paid maternity leave (Steinberg coverage of pension systems grows, and ris- and Nakane 2012). ing incomes provide new opportunities for Although policy measures to stimulate people to retire before they are no longer female employment are welcome, realiz- physically capable of work, thereby creating a ing their full potential will require longer- significant risk that average lengths of work- term shifts in social attitudes in many East ing lives will fall. Although the potential trend Asian and Pacific countries toward more of falling retirement ages for rural and infor- shared responsibility between women and mal sector workers is in one sense a boon L A B O R MA R K E T P O L I C I E S TO A D D R E S S AG I N G I N E A S T A S I A A N D PAC I F I C 115 and an indicator of rising prosperity and wel- FIGURE 4.3 Japanese workers show a desire to work in old age fare, the experience of richer East Asian and well above OECD counterparts Pacific economies already points to a need for Percentage of Japanese workers desiring to work until indicated strengthening policies to increase the work- age, 2007 ing lives of existing urban and formal sector Until when do you want to work? workers and to ensure that any future fall in rural and informal working lives is not 2.4% unduly precipitous. The continuing increases in longevity and healthy years of life expec- tancy mean both that people will have to 9.7% 19.2% 23.0% 8.9% 36.8% work until older ages over time and that they will be capable of doing so. Surveys in richer East Asian and Pacific economies suggest that people recognize the need to work lon- ger and the desirability in terms of economic 0 10 20 30 40 50 60 70 80 90 100 benefits and strengthened social engagement Share of respondents ages 60 and over (%) (see figure 4.3 on willingness to work in old Until age 60 Until age 65 Until age 70 Until age 75 age in Japan). Ages 76 and over As long as I can work A common objection to extending work- Source: Cabinet Office, Government of Japan 2008. ing lives of older workers is the potential Note: OECD = Organisation for Economic Co-operation and Development. effect on employment among younger peo- ple, but evidence globally and from within the region indicates that such negative financial aid, depending on the number of impacts are not found in practice. The seem- employees covered (Phang 2011). In Korea, ingly intuitive concern that increasing the starting in 2011, a new system of subsidies participation of older workers will squeeze has been adopted in selected occupations, out jobs for younger people is referred to as replacing a previous employment grant the “lump of labor fallacy.” In fact, global for the elderly. The objective is to promote evidence shows that participation of older employment of workers ages 60 and older workers if anything has a modestly positive who will most likely be made unemployed impact on employment of younger people through a retirement-age extension subsidy, (see box 4.2). a retiree reemployment subsidy, and a new Richer East Asian and Pacific economies temporary subsidy for employment of peo- have introduced a range of financial incen- ple ages 60 and older (OECD 2012). Since tives for employers to promote longer work- 2011, employers in Singapore receive up ing lives that may offer lessons for developing to 8 percent of wages of workers over age East Asian and Pacific countries experiencing 50 under the Special Employment Credit rapid aging. Financial grants conditional on Scheme, which is expected to cover about employers boosting recruitment, retention, 350,000 older Singaporeans. This credit is and reemployment of older workers are being supplemented by a lump sum grant to pro- used in Japan, Korea, and Singapore, similar mote firm-level initiatives to recruit, retain, to practices in a number of European coun- or reemploy mature workers. tries and other countries such as Australia. An important question related to finan- Since 2003, employers in Japan have been cial incentives for hiring older workers is the given subsidies to hire older workers ages extent to which they should be targeted by 45–64 (OECD 2012; UNFPA and HelpAge age only or as part of wider schemes based on International 2012). Under another pro- characteristics such as unemployment or low gram, small or medium companies that pro- income. In Europe, over half of such measures vide their employees with the opportunity to target older workers specifically (for exam- work until age 70 are provided with variable ple, in Bulgaria, Denmark, Finland, France, 116 LIVE LONG AND PROSPER BOX 4.2 The lump of labor fallacy The idea that increased participation of older work- Economic Co-operation and Development countries ers will negatively affect employment for younger on the relationship between employment of people people is known as the lump of labor fallacy. This ages 55–64 and those ages 20–24. For all but one fear of displacement is grounded in the assumption country (Japan, although the negative impact is mod- of a zero-sum labor market in which every job occu- est), increases in employment of older workers had pied by an older worker is one less potential job for at worst a neutral association and in most cases a a younger person. positive association with employment rates of young In fact, higher participation of older workers tends workers. This relationship is confirmed in regression to stimulate the economy and in doing so creates analysis. Although more analysis is needed on the higher aggregate demand for labor, thereby benefiting question in developing East Asian and Pacific coun- young people as well. In addition, the types of jobs tries, studies to date for China confirms no impact among older and younger workers differ, so they can- on youth employment from higher employment rates not be considered substitutes in general. Figure B4.2.1 of older workers (Munnell and Wu 2013; Zhang and shows results of analysis for 12 Organisation for Zhao 2012). FIGURE B4.2.1 In OECD countries, higher employment rates of older workers are associated with modestly higher rates of youth employment 0.20 y = 0.257x + 0.023 0.15 R2 = 0.034 Change in employment, 0.10 ages 20–24 0.05 0 –0.05 –0.10 0 0.02 0.04 0.06 0.08 0.10 0.12 0.14 0.16 Change in employment, ages 55–64 Employed, ages 20–24 Linear (employed, ages 20–24) Source: Gruber, Milligan, and Wise 2010. Note: OECD = Organisation for Economic Co-operation and Development. Data are taken from beginning of the last economic upturn by country to latest available year. Germany, Hungary, Italy, and Romania), concludes that wage subsidies granted solely while the rest provide wage subsidies for on the basis of age may not be very effective employers recruiting older people as part compared to a subsidy that is targeted more of vulnerable groups such as the long-term narrowly at the older long-term unemployed unemployed, people threatened by redundan- or at low-income older workers. cies, and people with disabilities (for example, Wage subsidy schemes that are predomi- Austria, Denmark, Latvia, Lithuania, Malta, nantly targeted at older workers raise a num- Portugal, and Sweden) (Feifs et al. 2013). The ber of issues in terms of cost-effectiveness. only available evidence (from 2006) on impact First, because older workers are a very diverse L A B O R MA R K E T P O L I C I E S TO A D D R E S S AG I N G I N E A S T A S I A A N D PAC I F I C 117 group, wage subsidies targeted on age alone approach seems defensible in a period when risk being quite blunt instruments and may governments are trying to change the societal result in small net employment effects at con- mindset on the productive value of older siderable cost. They may involve substantial people, but close evaluation will be needed deadweight loss (that is, a large proportion of to assess the cost-effectiveness of different subsidized workers would have been employed approaches as the stock of older workers even without the subsidy) and substitution or grows. Incentives may need to become more displacement effects (that is, subsidized jobs targeted over time to those at highest risk of for eligible workers lower job opportunities early exit from the labor force or at highest for workers ineligible for the subsidy). Korea’s risk of failure to reenter the labor force. 2011 reforms of financial incentives for hir- Richer East Asian and Pacific economies ing older workers were driven by evaluations have increasingly sought to reinforce finan- that found substantial deadweight losses from cial incentives to extend working lives with previous efforts. Second, subsidies for older employment services targeted at older work- workers as a group may lead to stigmatiza- ers. As with financial incentives, a number of tion and reinforce negative attitudes among countries in East Asia and Pacific and beyond employers regarding hiring and retention of have instituted employment service programs older workers (OECD 2006). specifically targeted to older workers. In For East Asia and Pacific, the approach to the European Union, Germany provides an date has more commonly been to adopt poli- example with its Perspective 50 Plus program cies targeted at all people above a certain age of active labor policies and campaigns to pro- and to supplement them with targeted sup- mote attitudinal change among employers plements for particular groups among older and society (OECD 2012). In East Asia and people (for example, Singapore’s Employment Pacific, Japan has a set of measures from mid- Assistance Payments, which are targeted at dle age onward, as do Korea and Singapore workers otherwise unable to be reemployed, (see box 4.3). Evaluations of impact and cost- in addition to general programs to sup- effectiveness of these programs are generally port employment of older workers). This lacking. Evidence from Japan suggests small BOX 4.3 Employment services targeted to older workers in Japan and the Republic of Korea Japan has strengthened its public employment ser- society” philosophy and are complemented by the Act vices by establishing special offices that support older on Stabilization of Employment of Older Persons and job seekers and incumbent workers, as well as sup- Silver Human Resources Centers. porting the reemployment of older workers through Korea has established a talent bank for older job- the following programs: Career Exchange Plazas to seekers that has been connected to job centers since help middle-aged and older white-collar job seekers 2012. Moreover, some private agencies have been work in close cooperation with Talent Banks; Older designated as employment service centers to support Persons Vocational Experience Utilization Centers to middle-aged professional workers with a certain level provide free placement services and dispatching ser- of qualifications in their job search. Korea has also vices for persons ages 60 and older; Industry Employ- established special offices that support older job seek- ment Security Centers to provide matching services to ers as well as older incumbent workers through the middle-aged or older workers; Elderly Employment Aged Service Centers and Manpower Banks for Older Support Centers/Corners to establish an office in Workers. every prefecture to provide assistance to middle-aged and older workers to develop career life plans. These Sources: OECD 2006 and Ministry of Health, Labor, and Welfare presentations to services all fall under the umbrella of the “ageless World Bank staff for Japan; OECD 2012 for Korea. 118 LIVE LONG AND PROSPER improvements in placement of workers ages wage rates (IMF 2012). Various mechanisms 60–64 after the introduction of age-targeted are used in OECD and richer East Asian and employment services, but how much of the Pacific economies to take advantage of this effect was attributable to the programs can- phenomenon, including the following: not be determined (OECD 2006). Supplementing the preceding measures are • Providing incentives in tax and social programs to incentivize workplace adjust- contribution systems. These include spe- ments to accommodate older workers and cific income tax credits for older work- enhance their productivity. Given the onset ers (for example, Australia, Denmark, of physical limitations for older workers, the Netherlands, Sweden, and the United many countries have sought to induce adjust- Kingdom, with the tax credit increasing ments in the physical work environment in with the age of the worker in some cases); firms through grants, loans, and other tools. reductions in or exemptions from social This approach is also happening increasingly contributions for workers over a certain on a purely private basis, as firms realize age (for example, France, the Netherlands, the financial payoffs of often modest invest- and Spain); and tax credits that are gen- ments. The most famous example in private eral but for which older people are dis- enterprise is Germany’s BMW: in response to proportionately represented in the target concerns about productivity on an assembly groups (for example, in Korea, the earned line staffed by older workers, BMW spent income tax credit introduced in 2008 for €40,000 on 70 small changes to workplace households with no children and some design and equipment to accommodate the self-employed individuals). particular needs of older workers. Within • Reducing disincentives to work in social a year, the line’s productivity had improved security systems. Social security systems by 7 percent (matching that of others staffed may give disincentives to work at older by younger workers), and within two years, ages in a variety of ways. The first is treat- absenteeism caused by poor health had fallen ment of work income while in receipt of to 2 percent below the firm average—payoffs a pension. Historically, countries such as that much more than offset the initial invest- the former Soviet Union republics pro- ment (Loch et al. 2010). The positive outcome hibited receipt of both pension and work is mirrored in a larger study of productivity income, although these systems were and age profiles in response to different inter- reformed in most cases during transition ventions (Bussolo, Koettl, and Sinnott 2015; as the state ceased to be the dominant pro- Göbel and Zwick 2010). In East Asian and vider of both pensions and wages. Within Pacific red economies, governments provide East Asia and Pacific, Japan provides an grants and subsidized loans to firms to under- interesting case: it retains a social security take such adjustments. For example, Korea earnings test whereby pension benefits since 2011 has provided subsidized loans beyond the basic pension are withdrawn for installation of age-friendly facilities, and in proportion to labor income at vary- Singapore provides grants that may be used ing rates, according to age and income by firms for upgrading. Increasingly, firms in level. For those ages 60–64, benefits are Japan are also making such investments on a first withdrawn at a 50 percent rate and purely market basis. then suspended entirely for incomes A second broad set of policies com- over ¥460,000 (in 2012). For those ages monly used to stimulate work at older ages 65 and older, a 50 percent withdrawal is within tax and social security systems, and applies over the same threshold. Analysis, these methods are spreading in East Asia and not surprisingly, finds negative effects on Pacific. Overall, older workers globally have labor supply among workers ages 60–64 been found to be more sensitive than younger in Japan (Shimizutami, Fujii, and Oshio workers to financial incentives and after-tax 2012).7 Such an approach seems to run L A B O R MA R K E T P O L I C I E S TO A D D R E S S AG I N G I N E A S T A S I A A N D PAC I F I C 119 counter to other policies designed to stim- FIGURE 4.4 Tax force to retire is strongly correlated with labor ulate employment at older ages in Japan, force participation of older men in OECD countries and many OECD and transition countries have moved away from it. 0.7 United R2 = 0.81 France Belgium 0.6 Kingdom For developing East Asian and Pacific Netherlands Italy 0.5 Spain countries, the bigger work disincentive in Canada Germany Proportion pension systems lies in early retirement pro- 0.4 United visions, which in a number of countries States Sweden 0.3 (for example, China, the Philippines, and 0.2 Japan Vietnam) provide strong financial incentives to retire earlier than official pension ages, as 0.1 discussed in chapter 5 of this report. In con- 0 trast, Japan and Korea (along with a num- 0 1 2 3 4 5 6 7 8 9 10 ber of OECD countries, including Canada, Tax force to retire the United Kingdom, and the United States) Ages 55–65 not in labor force Log of ages 55–65 not in labor force structure accrual and other benefit rules to incentivize retirement after the official age Source: Gruber, Milligan, and Wise 2010. Note: OECD = Organisation for Economic Co-operation and Development. (Feifs et al. 2013). A further factor at play is retirement age, which is discussed below. East Asian and Pacific countries at all higher productivity of older workers, this income levels can take lessons from OECD system will make older workers less sought experience (including in East Asia and Pacific) after in the labor market and will constrain on the importance of minimizing “tax force efforts to extend their working lives. Evidence to retire” (that is, the implicit tax on work from OECD countries finds that seniority over a certain age). As shown in figure 4.4, wage systems have a strongly negative impact in OECD countries with the lowest tax force on hiring of workers ages 50–64, a find- to retire, fewer than 40 percent of men ages ing confirmed in firm-level data (Daniel and 55–65 are out of the labor force, whereas that Heywood 2007; OECD 2011). On the basis share rises to as high as two-thirds in coun- of an early 2000 survey of employers in 10 tries with the highest tax force to retire (for East Asian and Pacific and OECD economies, example, Belgium). Japan does particularly East Asia and Pacific showed higher-than- well on this measure, with very low tax force average reliance on seniority wage setting, to retire and only about 20 percent of men particularly in Japan and Taiwan, China ages 55–65 out of the labor force. The nega- (Lowe et al. 2002). Although the differences tive impacts of high tax force to retire can are not dramatic compared with other OECD already be seen in the formal sectors of some countries, employers in China and Korea, countries in developing East Asia and Pacific for example, were notable for thinking that (for example, China and Vietnam), so avoid- seniority should be more important in wage ing high tax force to retire will be important setting in the future. for all developing countries in the region as Richer East Asian and Pacific economies they reform their pension and tax systems. are trying to reduce the influence of seniority In contrast to various incentives in East in wage setting and increase that of perfor- Asia and Pacific to extend working lives, some mance in a variety of ways that offer lessons policies and practices hinder demand for older for other countries in the region. Box 4.4 dis- workers, such as seniority wage systems. As in cusses recent reforms of seniority wage systems many countries, East Asian and Pacific coun- in Korea and Singapore and the contrasting tries have inherited seniority-based wage sys- example of Japan. In the first two cases, sensi- tems that automatically increase pay with age, bly designed reforms are being adopted slowly independent of performance. Absent much by firms but show potential. Developing 120 LIVE LONG AND PROSPER BOX 4.4 Reforming seniority wage systems in the Republic of Korea, Singapore, and Japan The Republic of Korea, Singapore, and Japan offer of a broad set of incentives to firms to hire or retain examples of efforts to reduce the role of seniority in older workers. wage determination. Japan has taken a different route to the challenge Korea introduced a so-called wage-peak sys- of seniority wage systems. Japan has to date seen most tem from 2005, which incentivized firms to adopt experimentation in reduction of seniority wages hap- more performance-based wage systems and con- pen at firm level, though it is also making efforts in duct productivity-based collective bargaining. The the public sector to flatten wage-to-age schedules. government has disseminated industry-specific Since the 1990s, the importance of age-based pay performance-based compensation models and within base-pay setting has seen a gradual reduc- actively promoted the new system through media, tion, especially among managers, where emphasis awards, and the like. This strategy has been sup- on role-related pay (which within it has an element ported since 2006 by allowances for firms adopt- of performance-related pay) has increased (Conrad ing such measures, which were extended in 2011 to 2009). By 2006, about one-third of firms reported employers who extended their firm-level retirement relying entirely on role-related pay for their managers, age and adopted measures such as flexible work and a further third relied on it for setting 50 percent or and reemployment after retirement (OECD 2012). more of wages, while around three-quarters of firms Although adoption has been relatively slow (only reported role-related pay not being the primary driver about 12 percent of firms had adopted the wage-peak of nonmanagerial pay (Shakai Keizai Seisansei Honbu system by 2011), the pace of demographic transi- 2006). Despite this gradual dilution of seniority pay, tion and market forces are strengthening incentives the most common approach in response to amend- to adopt it. Several large employers have already ments to the Job Security Law has been for firms to adopted the system: Samsung, for example, reduced mandate retirement of workers, typically at age 60, worker pay by 10 percent annually after age 55 to and then to rehire them at usually half or less of prere- increase the relative attraction of older workers. The tirement wages and on a term contract as a nonregular policy is not uncontroversial, however. The system employee. About 85 percent of companies had estab- has been a subject of dispute between government, lished such rehiring systems (Phang 2011). However, organized labor, and employer representatives during initiatives exist at the firm level to reduce the wage- 2015. age link. One large Japanese firm interviewed for this Singapore has for many years aimed to increase report has changed the entire wage-age curve within the flexibility of wage setting and its links to produc- the firm, creating budget space for rehiring and ade- tivity growth and economic conditions, starting with quately compensating older workers by flattening the its “flexi-wage policy” from 1986. More recently, it age-wage trajectory for workers in their 30s and 40s. has introduced financial grants to firms that move away from seniority-based wage setting. This is part Source: Sen 2014. countries are also trying to increase the role availability and retirement. Evidence from of performance-based pay in many cases, for around the world, including developed and example in China, where teachers have had developing East Asia and Pacific, finds evi- an increasing element of performance pay in dence of significant impact of retirement age recent years. on withdrawal from work, effects which in A second significant policy factor in the East Asia and Pacific are exacerbated by gen- formal sectors of East Asian and Pacific coun- erous early retirement provisions in countries tries is retirement age policy. This is discussed such as China, the Philippines, and Vietnam. in chapter 5 of this report in terms of fiscal Richer East Asian and Pacific economies sustainability and pension adequacy, but it have pursued a distinctive approach to delay- also has a clear labor market dimension, as ing withdrawal from the labor force, which found in chapter 2 with respect to pension involves a nuanced interaction of mandates L A B O R MA R K E T P O L I C I E S TO A D D R E S S AG I N G I N E A S T A S I A A N D PAC I F I C 121 and incentives. Compared with many other and Singapore. This is also true in Indonesia, parts of the world, richer red economies in which has not had a mandated retirement age East Asia and Pacific have taken a creative (Giles, Wang, and Cai 2012). The second ele- approach to the issue of retirement age and ment of the approach is that firms are given stimulating longer working lives. The first flexibility in how they extend working lives element of this approach is the role of the over specified minimum ages, and in gen- state and firms. Although the state mandates eral they seem to prefer programs of con- minimum retirement ages (and imposes man- tinued employment rather than raising the datory retirement in the public sector), firms firm-level mandatory retirement age itself. have a greater role in setting firm-specific The approach, most elaborated to date in retirement in countries such as Japan, Korea, Japan, is outlined in box 4.5. It also provides BOX 4.5 East Asia and Pacific’s distinctive approach to retirement age and continued work Richer East Asian and Pacific economies have exhib- retirement age. The adoption of reemployment as a ited a distinctive but still evolving approach to the measure is inspired by labor practices in Japan and issues of mandatory retirement ages and promotion is believed to be more useful in raising the effec- of longer working lives. Japan is a key example of tive retirement age while enhancing mutual ben- the mixture of mandates with flexibility. The Act efits because it allows for more flexibility in work on Stabilization of Employment of Elderly Persons arrangements as well as wage and benefit adjust- 2012 obliges employers to do one of the following: ments (Phang 2011). (a) raise the mandatory retirement age from 60 to The Republic of Korea is an interesting example 65; (b) introduce a continued employment system of mandate and incentives shifting over time with to ensure that all employees who desire to work respect to retirement age. In Korea, retirement ages remain in the workforce until age 65; or (c) abol- historically have not been mandated by law, but ish the mandatory retirement age at firm level. The employment laws have allowed firms to set man- aim is to improve the opportunities for older work- datory ages. Government employees face manda- ers to continue their career until at least pensionable tory retirement. One feature of pension eligibility age (65). The contract for this work differs from traditionally has been an imperfect correspondence that covering work prior to retirement age. In addi- between mandatory retirement age and the age at tion, when a system for continuous employment is which pension-eligible retirees may start receiving introduced in an enterprise, the standard for treat- pension benefits. The national pension scheme does ment of older persons can be decided through a for- not begin paying benefits until age 60, yet a signifi- mal labor-management agreement. As of mid-2012, cant share of employees until recently faced manda- 82.5 percent of companies had adopted the contin- tory retirement at age 55 (Cho and Kim 2005). The ued employment system, and few have taken mea- government over time has introduced various mea- sures to raise or abolish the mandatory retirement sures to encourage firms to voluntarily extend their age system (MHLW 2013). mandatory retirement age. The Act on Prohibition The statutory retirement age in Singapore was of Age Discrimination in Employment and Aged age 60 at the introduction of the Retirement Age Employment Promotion was adjusted in 2008, so Act in 1993. In 1999, it was raised to 62, with the that the retirement age an employer sets should be long-term objective of increasing the retirement 60 or over. Firms with 300 or more workers must age progressively to 67. In an effort to encour- report their firm-specific mandatory retirement age employment of more older workers, however, age each year to the Ministry of Employment and in 2007 the Tripartite Committee recommended Labor, and the ministry may request employers to reemployment instead of urging a higher statutory raise the minimum if it is below 60. However, this box continues next page 122 LIVE LONG AND PROSPER BOX 4.5 East Asia and Pacific’s distinctive approach to retirement age and continued work (continued) monitoring-based approach was considered not to be to employ older people above the industry-specific achieving sufficiently rapid results, and in 2013 the employment quota for elderly workers and to sub- act was amended to require firms to set minimum mit a status report each year. The employment retirement ages of at least 60. The amendments will quota is set by the standard employment rate of come into effect from 2016 for large firms and from older workers, which varies between 4 percent and 2017 for firms with fewer than 300 employees and 42 percent by industry and occupation. The Ministry central and local governments. of Employment and Labor may request firms to for- Korea also uses age-specific quotas to induce mulate and submit plans for implementation of the longer working lives in firms having 300 perma- standard employment rate when they fall below that nent employees or more. Those firms are required standard. an interesting case of industrial relations in reduced average working hours and a subsi- parts of East Asia and Pacific, whereby trades dized program that allows older workers to unions, employer associations, firm man- combine paid work and voluntary activities agement, and government aim to develop at social enterprises and nongovernmental nuanced policies that reduce the social con- organizations (though take-up has remained flict often engendered by economywide low) (OECD 2006, 2012). retirement age increases in other parts of the world. Complementing retirement age policy Promoting immigration into have been initiatives to promote flexible aged East Asian and Pacific work and a “glide” into retirement, although economies these have had limited impact in East Asia The potential for migration to mitigate the and Pacific to date. As shown in chapter 2 of labor force impacts of aging across East Asia this report, when workers in East Asia and and Pacific is significant and would pres- Pacific (at all country income levels) continue ent a “win-win” for both sending and host work into older age, they often tend not to countries.8 As in other parts of the world, reduce hours significantly until well into their immigration policy is a sensitive issue and 70s. This finding contrasts with patterns one that requires careful brokering of con- in some OECD countries, where flexible sensus in countries potentially involved to work arrangements have been promoted to ensure its sustainability and good outcomes encourage greater participation of both older for both migrants and their hosts. Rapid workers and women. The United Kingdom, aging, however, may provide increased for example, has achieved some success in impetus for migration policies that can be extending working lives through this route, beneficial for both sending and host coun- but in Germany the impacts have been more tries. Opinion poll evidence from Japan and limited because of negative perceptions about Korea presented in chapter 3 of this report part-time work (EC DG-ESAEO 2012). suggests that populations in rapidly aging Japan and Korea have made efforts to reduce East Asian and Pacific societies increasingly the average working week, which remains realize the desirability of raising historically very high by OECD standards, though low immigration rates. China has benefited progress is challenging. In response, Korea from this economic phenomenon in a differ- in 2011 introduced subsidies for firms that ent way with the mass internal migration of L A B O R MA R K E T P O L I C I E S TO A D D R E S S AG I N G I N E A S T A S I A A N D PAC I F I C 123 young rural workers to higher-productivity TABLE 4.1 There is enormous variation in the importance of work in urban areas. The necessity of immi- emigration in East Asian and Pacific countries gration for East Asian and Pacific econo- Number and percent of total population emigrating in selected East Asian mies with older populations is reflected in and Pacific countries as of 2010 the words of the late Singaporean Prime Emigration rate Minister Lee Kuan Yew (2012), accord- Origin Emigrants (thousands) (% of population) ing to whom “in the future we will have to Indonesia 2,502 1.1 depend on immigrants to make up our num- Philippines 4,275 4.6 bers, for without them Singapore will face Samoa 120 67.3 the prospect of a shrinking workforce and a Tonga 47 45.4 stagnant economy.” Vietnam 2,226 2.5 In both host and sending countries in East Source: World Bank 2013a. Asia and Pacific, enormous variation exists in the scale of migration. In terms of emigra- tion, several countries are among the most important points of origin across the globe. TABLE 4.2 The share of immigrants ranges from very low to very high For example, more than 4.3 million Filipinos by global standards in East Asia and Pacific are working abroad, followed by 2.5 mil- Immigration to selected East Asian and Pacific economies lion Indonesians and 2.2 million Vietnamese. Immigrants in Immigration rate in 2010 Yet because of the large populations of those Destination 2010 (thousands) (% of population born abroad) countries, emigration accounts for only Australia 5,522 25.7 between 1.1 percent and 4.6 percent of their Hong Kong SAR, China 2,742 38.8 populations (table 4.1). In contrast, emigra- Japan 2,176 1.7 tion is critical for several Pacific Island coun- Korea, Rep. 535 1.1 tries, such Samoa and Tonga, where emigrants Malaysia 2,358 8.4 make up 67 percent and 45 percent of total New Zealand 962 22.4 population, respectively. Although a large por- Singapore 1,967 40.7 tion of these emigrants stay within the region, Thailand 1,157 1.7 a significant number go to other parts of the Source: World Bank 2013a. world, such as the Persian Gulf, Australia and New Zealand, and North America. In terms of immigration, several economies in East Asia and Pacific display high levels, labor market liberalization in the context of whereas others are among the lowest receiving wider economic integration among members countries globally (table 4.2). In Australia and of the Association of Southeast Asian Nations. New Zealand, historically important migrant- The fertility and household formation receiving economies (mostly permanent immi- behavior of migrants is a critically important grants), around a quarter of the population and rapidly growing research area.9 Evidence was born abroad. Singapore and Hong Kong shows that the speed with which migrants SAR, China, (often temporary immigrants) adapt to the social norms of the destina- have even higher levels at about 40 percent. tion society depends on personal charac- Another important case is Malaysia with teristics of the migrants as well as cultural significant temporary low-skilled migration proximity between origin and destination from countries such as Bangladesh, Indonesia, countries, such as attitudes on gender pref- and the Philippines. In contrast, Japan and erences, sexuality, and marriage. In aging Korea have minimal immigration, at less than societies, the critical issue is the number of 2 percent of their total populations. Overall, children migrants will have and its deter- large portions of migrants in East Asia and minants. Widespread evidence exists that Pacific originate from within the region, a migrants who arrive before adulthood have situation that is likely to be reinforced with fertility behavior that is indistinguishable 124 LIVE LONG AND PROSPER from natives. For those who arrive at a later to fill the vacancy in question. Admission age, Kahn (1988) shows that highly edu- is granted on both permanent and tempo- cated migrants have almost identical fertil- rary basis, but the conditions to “import” ity behavior as natives of the destination. a low-skilled worker are normally more Migrants from neighboring countries or with stringent and residency privileges are less similar cultural backgrounds, regardless of permanent than those for skilled labor. The age at migration or education, have almost overall intake is again controlled by quo- identical fertility patterns. All of the evidence tas or other instruments. Korea, Malaysia, shows that assimilation is completed, at the Singapore, Thailand, and Taiwan, China, latest, with the children born in the destina- are examples of such systems in East Asia tion, if not with the migrants themselves. and Pacific. The impact of migrants on the overall fer- The most common policy tool across the tility level of the destination seems to be, at region and the world in managing immigra- best, a medium-term phenomenon. tion flows is quotas, and less common but Although the impact of migration on rapidly expanding tools are levies or taxes. migrants’ own fertility behavior is a criti- Many key destination countries implement cal issue, migration may also affect the quantitative limits on the number of visas fertility behavior of natives when large issued for migrants. These quotas can be numbers of migrants are involved. Among specific to education levels, occupations, the main determinants of fertility behavior or origin countries. They can be modi- are the opportunity costs of having chil- fied frequently or be set at a certain level dren for women and how migration affects for years, regardless of changing needs and these costs. More specifically, availability conditions. Under the levy approach, either of migrant workers for household services the employees or the migrants are required as nannies and maids is known to affect to pay a specific per worker tax to the gov- the labor force participation and fertility ernment. This tax can be imposed annually decisions of women in destination coun- or collected at the issuance of the employ- tries (Cortés and Tessada 2011). This issue ment visa or upon entry. Again, the levies is especially relevant for Singapore and can vary depending on the education level, Taiwan, China, which host significant num- the occupation, or the sector of employ- bers of female migrant workers. ment of the migrant. In recent years, coun- Host economies in East Asia and Pacific tries such as Singapore have implemented exhibit a mix of supply- and demand- levies that are employee specific and depend driven migration systems. A supply-driven on the share of foreigners in the firm’s immigration system selects workers based total labor force. To discourage employ- on their human capital credentials (educa- ment of migrants, firms with higher shares tion and experience), and migrants do not pay higher per worker levies. The quota need prearranged employment contracts. and levy distinction closely resembles the Admission is normally granted on a tempo- quotas-versus-tariffs distinction in interna- rary basis but with high likelihood of even- tional trade. Over time, quotas have been tual permanent residency and citizenship. removed or replaced by tariffs as the latter Quotas set by central authorities limit the are observed to be more flexible and eco- total annual intake. Australia is an example nomically efficient, since they lead to col- of such a system in East Asia and Pacific. lection of taxes rather than to the surplus In a demand-driven system, migrants nor- being captured by the quota permit holder. mally enter the host country with prear- The same economic benefits would be real- ranged employment contracts. Employers ized in the migration context, but political are generally required to obtain permission and public opposition to migration usu- from the government beforehand and to ally prevents full removal of quantitative show that no native worker can or is willing restrictions. L A B O R MA R K E T P O L I C I E S TO A D D R E S S AG I N G I N E A S T A S I A A N D PAC I F I C 125 Origin countries can also play an protection of workers while they are abroad, important role in enhancing the benefits of helping them maximize the long-term eco- migration for all parties involved by imple- nomic and social benefits for both the menting effective and efficient policies. migrants and their families. Other countries Among the roles that these governments can implement similar programs with varying play are improving informational efficien- degrees of success (see Ahsan et al. 2014). cies in the system, enforcing contracts, and Encouraging migration that maximizes protecting migrants’ rights. Informational benefits for both host and sending coun- problems are prevalent in migration. Among tries requires policy action in a range of the key predeparture policies is supporting areas, and practice from within East Asia and overseeing the recruitment process. The and Pacific already offers lessons. A recent first step is providing accurate information World Bank report (Ahsan et al. 2014) has to potential migrant workers. Creating infor- analyzed challenges, market failures in need mation centers and having nongovernmental of public policy focus, and policy implica- organizations involved in the process can tions and options for stimulating migration generate significant benefits. Regulation of in East Asia and Pacific in greater detail (see recruitment agencies through effective polic- also Adams and Ahsan 2014). The summary ing and enforcement and increased competi- findings of Ahsan et al. (2014) are therefore tion are critical. reiterated in table 4.3 and provide a broad The main concern of origin govern- framework for action. Details on specific ments should be to protect the rights of recommendations and good practices are their citizens abroad while finding appro- provided in that report and in Özden and priate employment opportunities for them. Testaverde (2014). These dual aims require close coordina- tion of different government agencies, such as departments of education, labor, and Policy effects on the labor force foreign affairs. The Philippines Overseas Even if sensible policy responses raise par- Employment Administration (POEA) has ticipation rates, the question remains whether been relatively successful in designing poli- they will be sufficient to compensate for the cies that generate long-term benefits. In declining share of working-age population. 2010, around 1.5 million workers registered The following section examines this issue through POEA before leaving to work in using a simple simulation framework looking more than 150 different countries. One of at red and orange economies, for which the the key features of POEA is the licensing shrinking working-age population is a press- of more than 1,000 recruitment agencies, ing challenge (see box 4.6 for assumptions which enables the government to prevent underlying these simulations). abusive practices by both these agencies and For both the upper- and middle-income potential employers. POEA is able to pro- countries shown, the biggest impact on tect the rights of the workers from the first labor force participation is from increasing stages of the migration process until they female labor force supply; for China and return home. The sister organization, the Thailand, increasing the participation of Overseas Workers Welfare Administration, older people also has significant potential collects contributions from the migrants to (figure 4.5). For red economies, significant fund predeparture training programs and life scope exists to increase female labor force insurance schemes as well as to operate wel- supply through a gradual convergence fare centers abroad. Finally, the Department toward male participation rates, whereas of Foreign Affairs coordinates the activi- the potential for labor force growth from ties of migration-related agencies through a older workers is more modest because of network of labor offices attached to consul- their already high participation rates. For ates. Such coordinated action leads to better China, in particular, increasing both elderly 126 LIVE LONG AND PROSPER TABLE 4.3 Policies to maximize the benefits of migration in East Asia and Pacific for sending and receiving countries Issue or market failure Causes Policy implications and options Objective: Meet labor shortages in economy through well-targeted migration programs Labor shortages in certain sectors Higher education of native workers, Formulate a clear migration policy based on comprehensive and “dirty, dangerous, and higher reservation wages, low consultations with affected sectors and assessment of demeaning” jobs are caused by geographic mobility, and aging create labor shortages in industries, occupations, and regions. segmented labor markets. labor shortages and labor market Policy should recognize gains and the need to enlist segmentation. cooperation of labor-sending countries and employee firms through regional and bilateral labor agreements. Objective: Address possible adverse labor market impacts of migrant labor on native unskilled workers Social returns of migration are Higher profits made by firms are not Ensure that foreign-worker levy is charged to and borne by less than private return to firms. shared with the economy although employing firms, not by workers. migrant workers have a social presence. Ensure that minimum wage or minimum-wage-indexed Lower any possible social costs by adding thresholds and benefits for migrant workers are used a cushion against lower wages and to prevent excessive lowering of wages or a general increasing unemployment. deterioration in working conditions. Although firms profit, the economy can Provide job training, retraining, and education for native suffer from the use of less productive, workers to upgrade their skills. This will probably happen labor-intensive technologies. normally, but government assistance can accelerate Migrant workers use public facilities the process. Ensure labor market flexibility to facilitate and social benefits, crowding out mobility of native workers from less productive to more local citizens. productive occupations. Wage subsidies and hiring bonuses can be provided to that end. Improve climate for investment and technology innovation and retention of professional and skilled labor. Provide job training and technical upgrading for workers to match demand in more skilled occupations. Provide normal fiscal incentives to firms for innovation and investment in new technology. Enable migrants to join and contribute to social security; make benefits and public facilities accessible to documented workers who also pay taxes. Social and private returns are Poorly regulated recruitment agencies Ensure workers receive orientation about rights and lower because of regulatory and rent-seeking lead to defrauding obligations, certified by recruitment firms and employers. and information failures. of migrant workers and excessive Ensure minimum capital experience thresholds for importation of foreign workers. Workers recruitment firms. are not adequately informed about their Standardize, monitor, and enforce contracts between rights and working conditions. employers, recruitment agencies, and workers. Take advance deposits from recruitment agencies and employing firms to use as penalties against abuse of migration laws through defrauding workers or violating contracts. Source: Ahsan et al. 2014. and female labor force participation has working-age population in richer East Asian significant potential to stave off what is and Pacific economies. Figure 4.6 shows the otherwise a significant decline in labor impact of permanent and temporary migra- force size. Thailand also has potential from tion scenarios on the size of the labor force both channels, although less so for elderly in destination economies relative to the base- participation. line case and the enhanced female and older- Permanent or temporary migration also worker participation scenarios. The impact has potential to mitigate the decline in of increased female participation continues L A B O R MA R K E T P O L I C I E S TO A D D R E S S AG I N G I N E A S T A S I A A N D PAC I F I C 127 BOX 4.6 Assumptions underlying labor force scenarios The scenarios span the period 2010–40 (2010–50 to the maximum rate in 2050 is determined dif- for migration) and combine population projections ferently for the red and orange groups. For the from the 2012 revision of the United Nations World red group, the participation rate for older men Population Prospects (UN 2013) with labor force and women is assumed to gradually converge to participation data from the International Labour the current Japanese rates for workers ages 60 Organization’s LABORSTA database (ILO 2011). and older. For the second group of aging coun- The assumptions for the base case and alternative tries, the elderly participation rate will move to scenarios are as follows: current Indonesian rates for workers ages 60 and older. No scenarios were estimated for the group • Base-case scenario. The International Labour with the lowest share of older workers: the Lao Organization publishes male and female partici- People’s Democratic Republic, Myanmar, Papua pation rates for each five-year age group from age New Guinea, the Philippines, and Timor-Leste. 15 and average participation for individuals ages • Migration scenarios— permanent and tempo- 65 and older. Taking these age and sex-specific rary. With permanent migration, newly arriving participation rates as given at 2010 levels and immigrants are assumed to be ages 25–35 and to combining them with official population projec- constitute 10 percent of the labor force of destina- tions implies that the projected changes in the tion countries in that age group. The assumption national labor forces are entirely determined by is based on the current levels observed in many the population age structure projections for 2010 countries that accept mostly permanent migrants, to 2040. including the European Union and the United • Convergence for women. The convergence sce- States. Migrants are assumed only to arrive in nario assumes that female labor force participa- that age group and never to leave. In addition, tion grows at an annual rate of 1/40th of the 2010 migrants adopt the fertility, mortality, and labor difference between men and women. Although force participation patterns of natives. The tem- women’s rates are assumed to converge toward porary migration scenario assumes that new male levels in 2050, the gender gap in partici- migrants constitute 20 percent of the labor force pation will not be eliminated by the end of the between ages 25 and 35 but that migrants stay period in 2040. for only 10 years (with a 100 percent labor force • Increasing participation by older workers. These participation rate) and then return to home coun- projections assume increased participation for tries. After 10 years, a new cohort of the same populations ages 60 and older, where convergence ages 25–35 arrives. to be the most significant in the long run for However, within 20 years, the permanent the overall size of the labor force in destina- migration scenario dominates as migrants tion countries. At the same time, both types of settle in host countries. The cumulative nature migration increase the labor force size in host of permanent migration has a larger impact in countries considerably in the short run and the long term, almost twice that from tempo- are superior in this time frame to effects from rary migration. Over time, as the migrants age higher female and elderly labor force partici- and retire, the total labor force again declines, pation rates. Temporary migration leads to the even with migration. largest immediate increase in the labor force The impact of migration on the origin coun- in destination countries because the arriv- tries is quite different because those countries ing cohort is bigger and temporary migrants have younger and expanding populations. have higher labor force participation rates. The initial decline is very similar under both 128 LIVE LONG AND PROSPER FIGURE 4.5 The potential channels for mitigating labor force decline vary across aging countries in East and Southeast Asia Percentage changes in labor force size relative to 2010 baseline from increases in female and elderly labor force participation a. China b. Japan 106 105 104 100 % change relative to 2010 % change relative to 2010 102 100 95 98 96 90 94 85 92 90 80 2010 2016 2022 2028 2034 2040 2010 2016 2022 2028 2034 2040 c. Republic of Korea d. Singapore 120 140 135 115 130 % change relative to 2010 % change relative to 2010 125 110 120 105 115 110 100 105 100 95 95 90 90 2010 2016 2022 2028 2034 2040 2010 2016 2022 2028 2034 2040 e. Thailand 110 108 106 % change relative to 2010 104 102 100 98 96 94 92 90 2010 2016 2022 2028 2034 2040 Baseline Female Elderly Source: World Bank estimates in Flochel et al. 2014. L A B O R MA R K E T P O L I C I E S TO A D D R E S S AG I N G I N E A S T A S I A A N D PAC I F I C 129 scenarios for the first two decades, but the FIGUR E 4.6 Greater participation by women, older people, and gap widens over time (figure 4.7). Permanent migrants can mitigate the projected decline in the total labor migration leads to a decline of about 5 percent force in richer East and Southeast Asian economies in the labor force by 2050 while the impact Projected changes in labor force size of destination economies with changes in female, elderly, and migrant labor force participation of temporary migration is about half that. Given the growth in origin-country labor 130 forces during that period and the immediate benefits of remittances to the origin countries, Size of labor force in millions such effects appear manageable. Thus both 120 host and origin countries appear to enjoy clear gains from increased regional migration (Özden and Testaverde 2014). 110 In summary, significant potential exists for higher participation to boost the size of the 100 workforce in East Asian and Pacific coun- tries most affected by aging. The potential for increased employment can be realized by 90 weakening barriers to work for women, the 2010 2015 2020 2025 2030 2035 2040 2045 2050 elderly, and migrants, all of which add up Baseline Elderly increased participation to a countervailing force in dampening and, Female increased Permanent migration in some cases, even overturning the decline participation in the labor force predicted in the base-case Temporary migration scenario. Yet activation of higher labor force participation provides only a temporary Source: World Bank estimates in Özden and Testaverde 2014. “shot in the arm” that delays the inevitable Note: Destination economies are Hong Kong SAR, China; Japan; the Republic of Korea; Malaysia; and Singapore. Sending economies are Cambodia, Indonesia, Lao PDR, Mongolia, Myanmar, Papua decline in labor supply from aging. A more New Guinea, the Philippines, Thailand, Timor-Leste, and Vietnam. Green line shows increase in sustained response to aging societies can be labor force relative to baseline when women participate at same level as men. Red line shows increase given increased labor force participation by the elderly. Orange line shows increase given achieved only by improving growth rates of migration of 10 percent of labor force ages 25–35 (migrants stay permanently). Light blue line human and physical capital and total factor shows increase given temporary increase in migration of 20 percent of labor force ages 25–35 (migrants stay 10 years). Dark blue line shows the baseline. productivity. FIGURE 4.7 Migration leads to a 5 percent decline in the total labor Upgrading skills of older workers force size of the migrants’ origin countries and integrating strategies Aggregate labor force of Cambodia, Indonesia, Lao PDR, Mongolia, Myanmar, As noted in chapter 3 of this report, East Papua New Guinea, the Philippines, Thailand, Timor-Leste, and Vietnam Asia and Pacific is in a relatively solid posi- tion with respect to rapid improvements in 500 Size of labor force (millions) its human capital stock in recent decades, but 450 lifelong learning systems remain underdevel- oped. Despite the impressive improvements 400 in educational attainment, older workers in 350 many developing East Asian and Pacific coun- tries face significant challenges: for example, 300 those who had compromised education in childhood caused by conflict or sustained 250 civil disturbance. Evidence from the OECD 2010 2015 2020 2025 2030 2035 2040 2045 2050 Programme for the International Assessment Baseline Female participation rate Elderly participation rate Permanent migration of Adult Competencies in Japan and Korea Temporary migration also shows low rates of on-the-job training for middle-aged and older workers relative Source: World Bank estimates in Özden and Testaverde 2014. 130 LIVE LONG AND PROSPER to OECD averages and contrasts with coun- gradual increase in targeted training schemes tries such as the United States, which exhibits for older workers. Co-financing or providing more equal access across the work life cycle government subsidies for training programs is (figure 4.8). Apart from this transition chal- a popular measure used by Australia and sev- lenge, a broader challenge exists of building eral countries in Europe to incentivize training lifelong learning systems for continuous skill and skill building for older workers. Sweden upgrading of workers across the life cycle. provides an innovative example of a “learning Globally, education, skill, and labor systems account” to which both employers and are aiming to evolve from a traditional pat- employees can make tax-exempt deposits for tern of study, work, and retirement to one training. Such measures are driven by the need where skill upgrading and work lives blend to improve employability of older workers, more seamlessly and the work-retirement since schemes to improve training rates across divide becomes more blurred. the adult population have the lowest take-up Australia, high-income countries of East among older workers (EC DG-ESAEO 2007, Asia and Pacific, and Europe have adopted based on the European Commission’s Ageing a variety of measures that subsidize training and Employment: Identification of Good and lifelong learning activities for older work- Practice study; OECD 2006). ers to improve their productivity and employ- Richer East Asian and Pacific economies ability. Globally, countries face challenges have also instituted training programs for in uptake of training by older workers rela- older workers, both on a targeted basis and tive to others, and examples from East Asia within the context of broader programs. The and Pacific suggest that the challenge may be government of Singapore in Budget 2014 even more acute.10 This situation has led to a announced a S$500 million top-up to its FIGURE 4.8 Japan and the Republic of Korea show low rates of on-the-job training for middle-aged and older workers relative to countries such as the United States Index of learning at work by age and country, 2012, selected OECD countries 2.6 2.4 2.2 2.0 Index 1.8 1.6 1.4 1.2 1.0 24 or less 25–34 35–44 45–54 55 plus Age group Australia Canada Finland France Germany Japan Poland Korea, Rep. United States Russian Federation Source: Generated from the OECD 2013. Note: OECD = Organisation for Economic Co-operation and Development. The index is scored on a scale of 0 (lowest) to 4 (highest). Russian data are preliminary and exclude Moscow. L A B O R MA R K E T P O L I C I E S TO A D D R E S S AG I N G I N E A S T A S I A A N D PAC I F I C 131 Lifelong Learning Endowment Fund, taking Smyth 2012). In Korea, less than 10 percent the total size to S$4.6 billion with a lifelong of workers ages 50–64 participated in train- commitment to continuing education and ing of any form (and fewer older women than training to enhance employability, though men), less than half the rate for other work- to what extent older workers are accessing ers, and well below age-specific rates of about the scheme remains unclear (Government of 40 percent in the United States and Nordic Singapore 2014). In addition, ADVANTAGE countries (Phang 2011).12 Part of the issue is and Flexi-Works Policy grant financial assis- simple cost-benefit, with a justified percep- tance to employers who train older workers. tion that the shorter the remaining working To better reach out to low-wage workers, life the more returns to training fall, a factor the government introduced a three-year that has led James Heckman, for example, to Workforce Training Scheme that subsidizes suggest that wage subsidies may be a more 90 percent to 95 percent of absentee payroll efficient intervention as workers get older. and course fee outlay for employers who However, another strain of literature finds send their older low-wage workers for train- significant discrimination against older work- ing. Low-wage workers are encouraged to ers by employers in providing training and a go for training to upgrade their skills and feedback loop whereby some older workers improve their employability, upgrade to bet- internalize the effect and seek less training ter jobs, and earn more. Since the Workforce (OECD 2006). Training Scheme started in July 2010, more “Universities of the third age”—in which than 90,000 Singaporeans have benefited, fol- skilled older people provide education and lowing the guidance of an old Chinese phrase: skill training for other older people—have ㋀‫ܦ‬冾븽ᘕ‫ܦ‬冾 (“live till old, learn till old”) proliferated across East Asia and Pacific. (Government of Singapore 2013; Phang 2011; Initiated in France, this type of organization UNFPA and HelpAge International 2012). has spread to transition and OECD coun- In Korea, job training subsidies are paid to tries and has been taken up in East Asian employers who provide training opportuni- and Pacific countries such as Australia, ties to workers under the Occupational Skill China, Malaysia, and Singapore. These Development Scheme (Phang 2011). More mixed programs are oriented toward work- specifically, the scheme focuses on providing related skills as well as general intellectual training for older people who want to start enrichment and social interaction, with their own businesses (given the prevalence of an aim to provide stimulation and slow self-employment among older workers) or cognitive decline (UNFPA and HelpAge return to rural areas. International 2012). As in other parts of the world, participa- Although limited, the global evidence on tion in training among older workers tends effects of targeted training schemes on the to be relatively low, though how much this productivity of older workers is mixed, and is driven by weak demand, employer atti- more assessments of cost-effectiveness are tudes, or inflexible supply is difficult to dis- needed. Assessments from Germany find that entangle.11 This low participation rate can training of older workers did not appear to be seen throughout East Asia and Pacific. significantly affect productivity among those In Singapore, about 37 percent of workers already working (Göbel and Zwick 2010, ages 20–29 undertook training with support 2013). Similarly, in Sweden training for of the Skills Development Fund, whereas unemployed workers ages 50–64 resulted in the share was only 19 percent in the 50–64 regular job placement for only one in eight age group (Phang 2011). In China, the trainees (OECD 2006). In contrast, placement effect was found to be more pronounced, rates of older trainees in the Czech Republic with a 1 percent increase in age resulting in and Spain were encouraging, with 70 and a 27 percent to 35 percent decline in prob- 50 percent of trainees, respectively, in work ability of undertaking training (Mishra and 6–12 months after training (OECD 2006). 132 LIVE LONG AND PROSPER In East Asia and Pacific, evaluations of initia- under its Advantage! Scheme; the WorkPro tives in red economies would be beneficial. scheme of funding to help employers rede- The most notable feature of policies to sign jobs and work environments to adapt extend working lives in East Asia and Pacific to an older workforce; and the Special in recent years has been the adoption of mul- Employment Credit and Workfare Income tipronged strategies to promote productive Supplement, which aims to support employ- aging. Richer East Asian and Pacific econo- ers and employees to raise the employability mies have heightened their policy focus on of older, low-wage Singaporeans. Japan has employment of older workers through com- had several packages of measures over the prehensive programs combining different years under its Basic Policy of Employment types of interventions. Box 4.7 discusses the Measures for the Elderly and the 2012 example of Korea, which mixes incentives revised Act on Stabilization of Employment for enhancing productivity of older work- of Older Persons, including wage subsidies, ers, subsidies to employers to retain older financial support to employers with age- workers, and efforts to activate unemployed friendly workplaces, dedicated training cen- older workers or those approaching retire- ters (Silver Human Resources Centers), and ment or dismissal. Singapore has adopted employment services (sections for the elderly a similar range of programs, including in Hello Work job centers). These are part of BOX 4.7 Second Basic Employment Promotion Plan (2012–16) for the aged in the Republic of Korea 1. Strengthening support for intergenerational job 3. Strengthening support for retirement preparation sharing and skills development • Skilled middle-aged and old retirees will serve • Large firms will be required to provide a cer- as mentors and instructors for young employ- tain period of preretirement training to their ees at 300 small and medium enterprises, and middle-aged and older workers who are forced 1,600 retired experts will be used as “on-site to leave their jobs involuntarily. professors.” • Long-serving middle-aged and older workers • The job-sharing program, which provides a will be granted the right to ask for unpaid subsidy of W7.2 million per year to firms if educational leave of one year or less. they hire youth to fill vacancies that arise as 4. Expanding support for early reemployment and middle-aged and older employees work shorter jobs for the aged hours or receive training, will be promoted. • Opportunities to participate in the Employment 2. Expanding support for older workers to stay Success Package that links counseling, longer in their principal jobs vocational training, and job placement and to • Subsidies for the wage-peak system will be receive on-the-job training at small and increased. medium enterprises will be expanded. • Subsidies for extending the retirement age • Relevant training and funding will be provided or reemploying older workers will be differ- to older people who want to start up their entiated based on the length of the extended own business or return to farming or rural employment period to encourage firms to keep areas. older workers for a longer period. 5. Promoting social contribution and talent-sharing • T he coverage of workplace su r veys of activities retirement-age schemes will be extended from 6. Improving systems and infrastructure to cope those with 300 workers or more to those with with an aged society 100 workers or more, to prepare for discussions • The retirement pension system will be further on retirement-age reform. promoted to ensure old-age income security. L A B O R MA R K E T P O L I C I E S TO A D D R E S S AG I N G I N E A S T A S I A A N D PAC I F I C 133 a broader “ageless society” theme of public programs is needed to inform policy deci- policy in Japan. Such active labor programs sions. For many programs aiming to stim- along with reforms of retirement age and ulate enhanced labor force participation, continuous employment effectively combine very little is known about their employ- incentives with mandates and give firms ment impacts and cost-effectiveness. As considerable latitude for how they comply. discussed earlier in the chapter, one spe- The lead from public policy is increasingly cific issue is the extent to which policies being taken by large employers, with com- promoting higher labor force participa- panies such as JFE Steel introducing the tion are more cost-effective when targeted Senior Expert Program to retain and maxi- purely by demographic characteristics mize the experience of workers over age 60 such as age or gender or by more general (Shimbun 2015). economic characteristics such as unem- ployment or income level. • An appropriate balance must be struck Conclusion between mandates (on employers or work- ers) and incentives. Emerging experience Fortunately, the middle-income countries from East Asia and Pacific suggests that of East Asia and Pacific can benefit from neither approach in isolation is likely to the lessons of various labor market policies achieve the desired results. A related con- implemented in richer East Asian and Pacific sideration is the role of public policy and economies In responding to the challenges of market forces. shrinking labor forces and outputs, the expe- rience from East Asian and Pacific countries suggests a set of overarching issues in devel- Notes oping labor market policy responses: 1. Jones (2009) provides an excellent summary • As supported by the recommendations pre- of pronatalist policies in richer East Asian and sented in this chapter, an effective labor Pacific economies. See also Tsuya, Choe, and market policy response to aging in East Feng (2009). Asia and Pacific requires policy action 2. This finding is consistent with other analysis across the life cycle, ranging from efforts for Japan (Tsuya 2005, 2008; Tsuya, Choe, to stimulate fertility in countries with older and Feng 2009). 3. U.S. Social Security Administration (2015) populations to initiatives aimed at extend- provides a comprehensive summary of ing the productive lives of older workers. maternity and other insurance benefits in Asia. • Attention is needed to policy coherence 4. See Hegewisch and Gornick (2011) on slow and trade-offs in trying to mitigate the take-up of paternity leave in OECD countries. adverse labor market impacts of aging. For Japan, see Basic Survey of Gender Equality Although many interesting and potentially in Employment Management, MHLW (2013). useful initiatives are available to augment 5. Employer initiatives to promote family- the labor forces in aging East Asian and friendly workplaces are based on the 2003 Pacific countries, ensuring that labor mar- Act on Advancement of Measures to Support ket, social security, tax, and other policies Raising the Next Generation of Children all push in the same direction remains a and the Basic Act for Measures to Cope with a Society with a Declining Birthrate, the challenge in some cases. Thinking about former amended and strengthened in 2014. policy coherence raises potential trade-offs, 6. Presentation of Japanese MHLW to World for example between the economic impera- Bank team, November 2013. tives of sustaining labor forces and socio- 7. The earnings test on pensions in Japan was cultural norms around the role of women repealed in 1985 for workers ages 65–69 but in society. reinstated in 2002. • A better understanding of the cost- 8. This section draws from a background paper effectiveness of labor market policies and by Özden and Testaverde (2014) and the 134 LIVE LONG AND PROSPER World Bank regional migration report for Community Society of Elderly People” [in East Asia and Pacific (Ahsan et al. 2014). Japanese]. Government of Japan, Tokyo. http:// 9. See Adsera and Ferrer (2014) for an excellent www8.cao.go.jp/kourei/ishiki/h20 / sougou review. /zentai/. 10. Available evidence suggests that older Chi, Wei, and Bo Li. 2008. “Glass Ceiling or workers are much less likely to participate Sticky Floor? Examining the Gender Earnings in job training programs relative to younger Differential across the Earnings Distributions workers. An OECD (2006) report suggests in Urban China.” Journal of Comparative on-the-job training participation rates of Economics 36 (2): 243–63. workers ages 50–64 to be under 10 percent Cho, Joonmo, and Sunwoong Kim. 2005. “On in Korea and over 40 percent in Demark, Using Mandatory Retirement to Reduce Norway, and the United States. In terms Workforce in Korea.” International Economic of lifelong learning opportunities, whereas Journal 19 (2): 283–303. just over 10 percent of Koreans ages 50–64 Conrad, Harald. 2009. “From Seniority to participated in some form of lifelong learning Performance Principle: The Evolution of Pay activity during 2000, the corresponding Practices in Japanese Firms since the 1990s.” proportion was about 36 percent in the Social Science Japan Journal 13 (1): 115–35. United States and about 46 percent in Sweden Cortés, Patricia, and José Tessada. 2011. “Low- (Phang 2011). Skilled Immigration and the Labor Supply of 11. See Mayhew, Elliott, and Rijkers (2008) for a Highly Skilled Women.” American Economic useful discussion of training and employment Journal: Applied Economics 3 (3): 88–123. of older workers. Daniel, Kirsten, and John Heywood. 2007. “The 12. Japanese MHLW survey findings suggest Determinants of Hiring Older Workers: higher training among older men (36 percent U.K. Evidence.” Labour Economics 14 (1): in 2005), though the rate is less than half that 35–51. among older women. Den Dulk, Laura, and Anneke Van Doorne- Huiskes. 2007. “Social Policy in Europe: Its Impact on Families and Work.” In Women, References Men, Work, and Family in Europe, edited by Adams, Richard H., Jr., and Ahmad Ahsan, eds. Rosemary Crompton, Suzan Lewis, and Clare 2014. Managing International Migration for Lyonette, 51–70. Basingstoke, U.K.: Palgrave Development in East Asia. Washington, DC: Macmillan. World Bank. Du, Fenglian, and Xiao-yuan Dong. 2010. Adsera, Alicia, and Ana Ferrer. 2014. “Immigrants “Women’s Labor Force Participation and and Demography: Marriage, Divorce, and Childcare Choices in Urban China during Fertility.” IZA Discussion Paper 7982, Institute Economic Transition.” In Gender Equality and for the Study of Labor, Bonn, Germany. China’s Economic Transformation: Informal Ahsan, Ahmad, Manolo Abella, Andrew Beath, Employment and Care Provision, edited by Yukon Huang, Manjula Luthria, and Trang Xiao-yuan Dong and Sarah Cook, 173–91. Van Nguyen. 2014. International Migration Beijing: Economic Science Press. and Development in East Asia and the Pacific. EC DG-ESAEO (European Commission, Washington, DC: World Bank. Directorate-General for Employment, Social Bongaarts, John, and Tomáš Sobotka. 2012. Affairs, and Equal Opportunities). 2007. “A Demographic Explanation for the Recent Employment in Europe 2007 . Brussels: Rise in European Fertility.” Population European Commission. and Development Review 38 (1): 83–120. — — —. 2012. European Employment Observatory doi:10.1111/j.1728-4457.2012.00473.x. Review: Employment Policies to Promote Bussolo, Maurizio, Johannes Koettl, and Emily Active Ageing 2012. Luxembourg: Publications Sinnott. 2015. Golden Aging: Prospects for Office of the European Union. Healthy, Active, and Prosperous Aging in Feifs, Toms, Tina Weber, Oscar Vargas, Karel Fric, Europe and Central Asia. Washington, DC: Maurizio Curtarelli, and Irene Mandl. 2013. World Bank. Role of Governments and Social Partners in Cabinet Office, Government of Japan. 2008. Keeping Older Workers in the Labour Market. “Survey Concerning the Participation in the Dublin: European Foundation for the L A B O R MA R K E T P O L I C I E S TO A D D R E S S AG I N G I N E A S T A S I A A N D PAC I F I C 135 Improvement of Living and Working Kahn, Joan R. 1988. “Immigrant Selectivity and Conditions. Fertility Adaptation in the United States.” Giles, John, Dewen Wang, and Wei Cai. 2012. Social Forces 67 (1): 108–28. “The Labor Supply and Retirement Behavior Kim, Jeong-hee. 2007. “Child Care Support of China’s Older Workers and Elderly in Programs for Double Income Families.” Comparative Perspective.” In Aging in Asia: In Generational Change and Social Policy Findings from New and Emerging Data Challenges: Australia and South Korea, edited Initiatives , edited by James P. Smith and by Ruth Phillips, 140–163. Sydney, Australia: Malay Majmundar, 116–47. Washington, DC: Sydney University Press. National Academies Press. Loch, Christoph, Fabian J. Sting, Nikolaus Göbel, Christian, and Thomas Zwick. 2010. B a u e r, a n d H e l m u t M a u e r m a n n . “Which Personnel Measures Are Effective 2010. “How BMW Is Defusing the in Increasing Productivity of Old Workers?” D e m o g r a p h i c Ti m e B o m b . ” H a r v a r d ZEW Discussion Paper 10-069, Zentrum für Business Review 88 (3): 99–102. https:// Europäische Wirtschaftsforschung, Mannheim, h b r. o r g / 2 0 1 0 / 0 3 / t h e - g l o b e - h o w - b m w Germany. -is-defusing-the-demographic-time-bomb. ———. 2013. “Are Personnel Measures Effective Lowe, Kevin B., John Milliman, Helen De Cieri, in Increasing Productivity of Old Workers?” and Peter Dowling. 2002. “International Labour Economics 22: 80–93. Compensation Practices: A Ten-Country Government of Singapore. 2013. A Sustainable Comparative Analysis.” Human Resource Population for a Dynamic Singapore: Population Management 41 (1): 45–66. White Paper. Singapore: National Population Ma, Li. 2010. “Social Policy and Childbearing and Talent Division, Prime Minister’s Office. Behavior in Japan since the 1960s: An ———. 2014. “Budget 2014: Opportunities for Individual Level Perspective.” Paper presented the Future—Assurance for Our Seniors.” in Poster Session 1 at the European Population http://www.singaporebudget.gov.sg/data Conference 2010, Vienna, September 1–4. /budget_2014/download/FY2014_Budget_in Maurer-Fazio, Margaret, Rachel Connelly, Lan _Brief.pdf. Chen, and Lixin Tang. 2011. “Childcare, Gruber, Jonathan, Kevin Milligan, and David A. Eldercare, and Labor Force Participation of Wise. 2010. “Introduction and Summary.” Married Women in Urban China, 1982–2000.” In Social Security Programs and Retirement Journal of Human Resources 46 (2): 261–94. around the World: The Relationship to Youth Mayhew, Ken, Matt Elliott, and Bob Rijkers. Employment, edited by Jonathan Gruber and 2008. “Upskilling Older Workers.” Ageing David A. Wise, 1–45. Chicago: University of Horizons 8: 13–21. Chicago Press. MHLW (Ministry of Health, Labor and Welfare, Hegewisch, Ariane, and Janet C. Gornick. 2011. Japan). 2013. 2012 Basic Survey of Gender “The Impact of Work-Family Policies on Equality in Employment Management [in Women’s Employment: A Review of Research Japanese]. Tokyo: MHLW. http://www.mhlw in OECD Countries.” Community, Work, and .go.jp/toukei/itiran/eiyaku.html. Family 14 (2): 119–38. Ministry of Interior, Taiwan, China. 2008. ILO (International Labour Organization). 2011. Population Policy White Book . Taipei: LABORSTA (database), 6th ed., ILO, Geneva. Ministry of Interior. IMF (International Monetary Fund, Fiscal Mishra, Vinod, and Russell Smyth. 2012. “High Affairs Department). 2012. Fiscal Policy and Performance Work Practices and Workplace Employment in Advanced and Emerging Training in China: Evidence from Matched Economies. Washington, DC: IMF. Employee-Employer Data.” Monash Jones, Gavin W. 2009. “Recent Fertility Trends, Economics Working Paper 30-12, Monash Policy Responses, and Fertility Prospects in University, Department of Economics, Victoria, Low Fertility Countries of East and Southeast Australia. Asia.” Report for the United Nations Expert Munnell, Alicia, and April Wu. 2013. “Will Group Meeting on Recent and Future Trends in Delayed Retirement by the Baby Boomers Fertility, Population Division, United Nations Lead to Higher Unemployment among Department of Social and Economic Affairs, Younger Workers?” CRR Working Paper New York, December 2–4. 2012-22 (expanded version), Center for 136 LIVE LONG AND PROSPER Retirement Research at Boston College, Health Status, and Retirement Decisions: Chestnut Hill, MA. New Evidence from the Japanese Study on OECD (Organisation for Economic Co-operation Aging and Retirement (JSTAR).” RIETI and Development). 2006. Live Longer, Work Discussion Paper 12-E-050, Research Institute Longer: A Synthesis Report . Paris: OECD of Economy, Trade, and Industry, Ministry of Publishing. Economy, Trade, and Industry, Tokyo. ———. 2011. Doing Better for Families . Steinberg, Chad, and Masato Nakane. 2012. Paris: OECD Publishing. http://dx.doi.org “Can Women Save Japan?” IMF Working /10.1787/9789264098732-en. Paper 12/248, International Monetary Fund, ———. 2012. OECD Thematic Follow-Up Washington, DC. Review of Policies to Improve Labour Market Suruga, Terukazu, and Kenka Cho. 2003. “The Prospects for Older Workers . 21 separate Effect of Child Care Leave on Women’s Fertility country reports. http://www.oecd.org/els/emp and Work Continuation by the Use of Panel /ageingandemploymentpolicies.htm#country. Data” [in Japanese]. Journal of Household ———. 2013. OECD Skills Outlook 2013: First Economic Review 59: 56–63. Results from the Survey of Adult Skills. Paris: Thévenon, Olivier. 2013. “Drivers of Female OECD Publishing. http://www.oecd.org /site Labour Force Participation in the OECD.” /piaac/Skills%20volume%201%20(eng) OECD Social, Employment, and Migration --full%20v12--eBook%20(04%2011%20 Working Paper 145, Organisation for Economic 2013).pdf. Co-operation and Development, Paris. http:// Özden, Çaglar, and Mauro Testaverde. 2014. dx.doi.org/10.1787/5k46cvrgnms6-en. “International Migration in Aging Societies: Thévenon, Olivier, and Anne Solaz. 2013. Impacts and Benefits in East Asia and Pacific “Labour Market Effects of Parental Leave Countries.” Background paper for the East Policies in OECD Countries.” OECD Social, Asia and Pacific regional report on aging, Employment, and Migration Working Paper World Bank, Washington, DC. 141, Organisation for Economic Co-operation Phang, Hanam S., ed. 2011. Population Aging and and Development, Paris. Productivity in Asian Countries. Tokyo: Asian Tsuya, Noriko O. 2005. “Shoshika to Jyosei/ Productivity Organization. Gendaa Seisaku (Fertility decline and gender Rendall, Michael S., Alessandra De Rose, Ann Evans, policy in Japan).” In Shoshika no Seisaku-gaku Edith E. Gray, Doris Hanappi, Frauke Kreute, (Studies of policies relating to fertility decline Trude Lappegard, Lori Reeder, Marit Rønsen, to below-replacement levels), edited by Hiroshi Olivier Thévenon, and Laurent Toulemon. Ohbuchi and Makoto Atoh, 157–87. Tokyo: 2014. “Employment Impacts on Partnership Hara Shobo. and Parenthood Entry in Different Family- ———. 2008. “Japan’s Low Fertility: Causes, Policy Regimes.” Paper presented at the Consequences, and Policy Implications.” Paper European Population Conference 2014, presented at the 2008 Shanghai Forum, Fudan Budapest, June 25–28. University, Shanghai, China, May 25–27. Sen, Shonali. 2014. “Review of Policies to Improve Tsuya, Noriko O., Minja Kim Choe, and Wang Labor Market Prospects for Older Workers.” Feng. 2009. “Below-Replacement Fertility Background paper for the East Asia and in East Asia: Patterns, Factors, and Policy Pacific regional report on aging, World Bank, Implications.” Paper prepared for the 26th Washington, DC. International Population Conference of the Shakai Keizai Seisansei Honbu. 2006. Nihonteki International Union for the Scientific Study of Jinji Seido no Genjo to Kadai (Actual state Population, Marrakesh, Morocco, September and issues of the Japanese human resource 27–October 2. management system). Tokyo: Shakai Keizai UN (United Nations, Department of Economic Seisansei Honbu Seisansei Rodo Joho Senta. and Social Affairs, Population Division). Shimbun, Yomiuri. 2015. “Companies Turn to 2013. World Population Prospects: The 2012 Veteran Workers in Japan.” Japan News , Revision, Volume I: Comprehensive Tables. January 9. http://business.asiaone.com/news New York: United Nations. /companies-turn-veteran-workers-japan. UNFPA (United Nations Population Fund) and Shimizutami, Satoshi, Mayu Fujii, and Takashi HelpAge International. 2012. Ageing in the Oshio. 2012. “Option Value of Work, Twenty-First Century: A Celebration and L A B O R MA R K E T P O L I C I E S TO A D D R E S S AG I N G I N E A S T A S I A A N D PAC I F I C 137 a Challenge. New York: UNFPA; London: Bank. http://knoema.com/WBRIO2013 HelpAge International. /migration-and-remittances-factbook-2013. U.S. Social Security Administration. 2015. Social ———. 2013b. “Mongolia: Gender Disparities Security Programs throughout the World: in Labor Markets and Policy Suggestions.” Asia and the Pacific, 2014. Washington, DC. P o l i c y N o t e 7 7 5 0 8 , Wo r l d B a n k , http://www.ssa.gov/policy/docs/progdesc Washington, DC. /ssptw/2014-2015/asia. World Bank and ADB (Asian Development Bank). Waldfogel, Jane, Yoshio Higuchi, and Masahiro 2005. Mongolia: Country Gender Assessment. Abe. 1999. “Family Leave Policies and Women’s Manila: World Bank. Retention after Childbirth: Evidence from the Yew, Lee Kuan. 2012. “Warning Bell for United States, Britain, and Japan.” Journal of Developed Countries: Declining Birth Rates.” Population Economics 12 (4): 523–45. Forbes, May 7. World Bank. 2012. Toward Gender Equality in Zhang, Chuanchuan, and Yaohui Zhao. East Asia and the Pacific: A Companion to the 2012. “The Relationship between Elderly World Development Report. Washington, DC: Employment and Youth Employment: Evidence World Bank. from China.” MPRA Working Paper 37221, ———. 2013a. Migration and Remittances Munich Personal RePEc Archive, Munich Factbook 2013 . Washington, DC: World University Library, Munich, Germany. Preparing for an Aging East Asia and Pacific III Pension Systems and Aging: Reforms for Winning the Race 5 Introduction proved ineffective in this regard in developing countries. In terms of pension system design, East Asia This chapter describes the current state of and Pacific exhibits significant heterogeneity, pension provision in 15 economies of East as reflected in the policy challenges faced by Asia and Pacific and explores policy options different groups of countries. The pension for the three groups. The first section describes challenges facing East Asian and Pacific coun- the most important elements of these pension tries vary not only because of their disparate systems and puts them in an international demographic situations but also because of context. The second section then assesses past policy choices and the nature of their the systems’ performance based on three key labor and capital markets. The richer econo- criteria: pension coverage, spending and sus- mies with the largest populations ages 65 and tainability, and adequacy. It also discusses the above have achieved wide coverage and tend potential indirect impacts of the pension sys- to be more financially sustainable than most tem on the overall economy. Finally, the third high-income economies of other regions, but section explores the available policy options, they face challenges of pension adequacy and taking into account the very different condi- hence challenges of old-age financial protec- tions that characterize the region. tion. The middle-income economies are grap- pling with the dual challenges of improving the sustainability of their existing or legacy Overview of East Asian systems and trying to expand coverage to pension systems large, uncovered informal sectors. Both chal- This section looks at four elements of overall lenges are given urgency by the rapid pace of pension systems across East Asian and Pacific aging. For the poorer and largely young econ- economies: (a) contributory pension schemes omies, the key challenge is to find an appro- that are mandatory for private sector workers priate model that can achieve significant but may also cover public sector workers, (b) coverage over time, while recognizing that schemes for civil servants and government schemes financed by payroll taxes have employees, (c) supplementary private This chapter is based on a background paper by Robert Palacios (2015). 141 142 LIVE LONG AND PROSPER pensions, and (d) noncontributory pensions, global pension landscape. Most countries in or “social pensions.” As shown in table 5.1, East Asia and Pacific have adopted this all four elements are present in some but not approach, but how different countries have all East Asian and Pacific economies. done so varies widely. East Asian and Pacific economies can be broadly divided into three groups in terms of Government-mandated contributory the introduction of nationally mandated pen- pension schemes sion systems relative to demographic charac- The ultimate objectives of public provision of teristics: the globally average adopters, the pensions are twofold: to reduce poverty in old late adopters, and the global outliers that age and to smooth consumption over the life have yet to adopt a pension system. Figure 5.1 cycle. Globally, a small number of countries shows the year that different economies intro- (for example, New Zealand and South Africa) duced pension mandates that applied to most focus on only the first objective using budget- workers in the formal sector rather than to financed transfers to the elderly. However, just civil servants, compared with the share of most countries have mandated retirement the population over age 60 at the time. savings in an attempt to address both The globally average adopters—China objectives. Today, government mandates to (1951), Japan (1941), Malaysia (1951), save for retirement and usually to insure Mongolia (1950s), the Philippines (1954), against death and disability dominate the and Singapore (1953)—introduced national TABLE 5.1 Combined elements of pension systems in East Asia Mandated savings for pensions Social Target Civil Economy pensions Type of scheme benefit servants Voluntary pensions Cambodia None None n.a. Only China Significant Defined benefit and High Parallel Limited occupational pensions, defined contribution growing informal sector scheme Hong Kong SAR, Small Defined contribution Low Integrated China Indonesiaa None Defined contribution Low Parallel Fledgling occupational pensions Japan Small Defined benefit Low Integrated Significant occupational pensions Korea, Rep. Significant Defined benefit Low Parallel Significant occupational pensions Lao PDR None Defined benefit Low Parallel Malaysia Small Defined contribution Low Parallel Some voluntary savings within mandatory provident system plus limited occupational and personal pensions Mongolia Small Defined benefit Low Integrated Myanmar None None n.a. Only Philippines Small Defined benefit High Parallel Limited occupational and informal sector pensions Singapore None Defined contribution Low Integrated Some voluntary savings within mandatory provident system Thailand Significant Defined benefit Low Parallel Limited occupational pensions Timor-Leste Significant None n.a. Only Vietnam Small Defined benefit High Partly integrated Source: Palacios 2015. Note: n.a. = not applicable. a. The government of Indonesia issued Government Regulation No. 45/2015 on the implementation of its National Social Security System (Sistem Jaminan Sosial Nasional, or SJSN) pension plan as this report was being finalized, converting its pension scheme to a hybrid defined benefit and defined contribution scheme. PENSION SYSTEMS AND AGING: REFORMS FOR WINNING THE RACE 143 pension schemes relatively early. The timing FIGURE 5.1 East Asian and Pacific economies can be divided into of the introduction is significant because in globally average adopters, late adopters, and global outliers defined benefit schemes, generally the longer Timing of introduction of national mandated pension schemes the pension scheme has been operating, the higher the ratio of pensioners to contributors 16 and the greater the pressure on finances. The Hong Kong 14 Share of population over age 60 (%) late adopters—including Hong Kong SAR, SAR, China China (2001); Indonesia (1992); the Republic 12 of Korea (1988); the Lao People’s Democratic Republic (1999); and Thailand (1991)— 10 Korea, Rep. introduced their schemes fairly recently and Japan Thailand 8 relatively late in terms of their demographic China, Malaysia Vietnam transitions. Vietnam can be assigned to this Philippines Indonesia 6 group because it expanded its original scheme for public sector workers (1961) to the pri- 4 Lao vate sector only in the mid-1990s. The late Singapore PDR introduction of pensions in this group may be 2 consistent with a reluctance to expand the 0 role of the state. The global outliers—which 1871 1891 1911 1931 1951 1971 1991 2011 are not shown in figure 5.1 but include Cambodia, Myanmar, and Timor-Leste, and Sources: UN 2013; U.S. Social Security Administration 2012. which are also among the youngest countries in the region—have yet to move beyond civil gone the DB route. Workers in publicly man- service pensions. aged DC schemes receive only what they have A second major point of variation across put into their accounts plus investment earn- East Asia and Pacific is whether the mandated ings, and there is a risk that net returns will be scheme operates as a defined contribution too low to generate target benefit levels. In (DC) or defined benefit (DB) scheme. A num- contrast, DB scheme parameters tend not to ber of the region’s economies—including be as closely calibrated to match contribu- Indonesia, Malaysia, and Singapore—have tions with benefits, so their members often opted for DC schemes managed by the public receive an implicit return greater than prevail- sector that allow withdrawals for purposes ing interest rates, at least they have been dur- other than retirement, including housing. ing the first decades of the scheme. The major Variants of this “provident fund” model are risk for participants in DB schemes is that the also seen in several Pacific Island countries government will not be able to keep its prom- such as Fiji, Papua New Guinea, and the ises because of unsustainable parameters. As Solomon Islands. The policies resemble mid- evidenced by the recent wave of Organisation century policy developments in India, Nepal, for Economic Co-operation and Development and Sri Lanka, reflecting a shared influence (OECD) pension reforms, this risk increases from colonial history. Hong Kong SAR, as schemes mature. China, took the DC route half a century later In the late 1990s, the neat division between but opted for private, competitive manage- DC and DB countries was somewhat compli- ment (although confusingly perhaps, it con- cated by reforms in China and Mongolia. In tinued to call it a provident fund). The other China, a DC component was added to the DB East Asian and Pacific economies with man- scheme, effectively replacing part of the DB dated schemes have opted for DB schemes. scheme for workers above a certain age. The As a result of these choices, the provident intention was to accumulate funds to back the fund countries have not been accumulating new individual accounts. However, in its large unfunded pension liabilities,1 though the decentralized implementation, these funds opposite is true for the countries that have have not, in the large majority of cases, been 144 LIVE LONG AND PROSPER segregated as is normally done in a DC Unlike retirees in DB schemes, those in DC scheme.2 In most parts of the country, the schemes are not protected against longevity urban workers’ scheme seems in practice to be risk, that is, the risk that a person will outlive running on a pay-as-you-go basis. In his or her retirement savings. With the partial Mongolia, the concept of notional accounts3 exception of Singapore, the countries with or notional defined contributions (NDCs) was DC schemes allow workers to withdraw their introduced in 1999 and will affect cohorts funds at retirement in a lump sum. Recently, born after 1960. The intention was not to Singapore instituted a new policy for the pay- build a fund but rather to change the DB for- out stage: at retirement, members of the mula to reduce future pension liabilities. Central Provident Fund are required to pur- In developing East Asian and Pacific econ- chase one of two products that mimic annui- omies, indexation of DB pension payments is ties through a phased withdrawal (although often not automatic. In addition to different they are allowed to opt out of this default). target benefit levels, another important differ- This change in policy has helped Singapore ence is in the way benefits are paid out. The better address the issue of longevity risk. seven DB schemes all pay life annuities. In Overall, pension systems in East Asia and Japan and Korea these annuities are adjusted Pacific exhibit diverse philosophies with annually to compensate for changes in prices, respect to generosity, degree of redistribution, a common practice among OECD countries. and assumption of risk between citizens and In China and Vietnam, adjustments have been the state. Table 5.2 summarizes several char- made for changes in both prices and wages acteristics, including types of schemes, implied over the years, but on a discretionary basis.4 benefit targets, and the mode of benefit In Mongolia, the Philippines, and Thailand, payment. At one end of the spectrum, the role adjustments are discretionary or ad hoc, and of government in Hong Kong SAR, China, is real pension levels can and sometimes do fall, limited mostly to supervising the mandated exposing pensioners to inflation risk. In the scheme, which aims for modest benefits and Philippines, these adjustments have histori- is managed privately.5 The scheme has no cally tended to track price movements government financial commitments and has (see Mesa-Lago, Viajar, and Castillo 2011). liberal rules at the withdrawal stage. At the TABLE 5.2 Implications of pension scheme designs Degree of Country Benefit target redistribution Member risk Risk to government China High Large Low returns on defined contribution Large unfunded liability component Hong Kong SAR, Low Small Low returns on defined contribution; None China longevity risk Indonesia Low Small Low returns on defined contribution; None longevity risk Japan Moderate Moderate Sponsor risk or default Large unfunded liability Korea, Rep. Moderate Large Sponsor risk or default Moderate unfunded liability Malaysia Low Small Low returns on defined contribution; None longevity risk Mongolia Moderate Large Inflation risk; sponsor risk or default Large unfunded liability Philippines High Large Inflation risk; sponsor risk or default Large unfunded liability Singapore Low None Low returns on defined contribution None component Thailand Moderate Moderate Inflation risk; sponsor risk or default Moderate unfunded liability Vietnam High Small Sponsor risk or default Large unfunded liability Source: World Bank. Note: Sponsor risk refers to the possibility that a government may renege on defined benefit promises if the scheme becomes unsustainable. PENSION SYSTEMS AND AGING: REFORMS FOR WINNING THE RACE 145 other extreme, the Philippines and Vietnam Civil service schemes in East Asia and provide generous benefits through a DB Pacific tend to be more generous than those scheme in which government bears all of the for private sector workers, sometimes signifi- risk and allows a large unfunded liability or cantly so. The relationship between contribu- financing gap to grow over time.6 In the table, tions and benefits tends to be more imbalanced sponsor risk refers to the possibility that, as in from an actuarial perspective than national the case of the recent reforms in Japan, Korea, schemes in the same country. Benefit parame- and Mongolia, governments may renege on ters are more generous, and public sector pen- DB promises when schemes become sioners are likely to have greater life unsustainable. expectancy at retirement. In some cases, such as in China (prior to changes announced in Civil service pensions 2015) and in Myanmar, public sector workers have made no contributions, and pensions are As in other parts of the world, civil servants paid directly from the budget. Aside from fis- in East Asia and Pacific have usually been cal pressures caused by these maturing and covered by pension schemes before the rest of often generous civil service pensions, the the workforce (see Palacios and Whitehouse schemes demonstrate the apparent inequity of 2006). In every economy in the region, civil providing more generous pensions for public service pensions preceded mandates for pri- sector workers. Although a case may be made vate sector workers, often by decades. Japan’s for the government providing supplementary civil service pension scheme was introduced pensions as an employer, it is difficult to justify in the late 19th century (see Casey 2004). The different levels of minimum pensions or most recent example is Hong Kong SAR, implied rates of return on contributions or China, which, after inheriting the British even different eligibility ages. colonial pension scheme for civil servants, Administering separate schemes raises waited until 2001 to mandate retirement sav- costs to the extent that economies of scale ings for private sector workers. cannot be exploited, especially in countries Civil service pensions in East Asia and with a small number of covered workers (see Pacific fall into three distinct groups: Sluchynskyy 2015). For example, the admin- • Countries with separate pensions for civil istrative costs of the Philippine civil service servants include China,7 Indonesia, Korea, pension scheme were roughly the same as Lao PDR, Malaysia, the Philippines, and those for the social security system, despite Thailand.8 the fact that the latter had seven times as • Economies with integrated pensions for many contributors and beneficiaries. all formal sector workers include Hong Similarly, the cost per participant was six Kong SAR, China;9 Japan;10 Mongolia; times higher for Thailand’s civil service pen- Singapore; and Vietnam following its sion scheme compared with the 2014 reforms. Indonesia also plans to national scheme. Sluchynskyy (2015) pro- move gradually in this direction, although vides the strongest cross-country evidence of implementation has not yet started. economies of scale based on a sample of 104 • Countries with coverage only for civil public pension funds from 87 countries: servants include Cambodia, Myanmar, and administrative costs per beneficiary are found Timor-Leste, which passed its civil service to fall by half when moving from 100,000 to pension legislation in 2012 but has not fully 500,000 participants and by another 25 implemented it.11 Except for Timor-Leste, percent with 2 million participants. these separate civil service pension schemes Finally, parallel pension schemes may are all DB and are fully mature, with retir- inhibit labor mobility between the public and ing civil servants having earned full pen- private sectors, because pension rights typi- sions and with relatively high ratios of cally are not portable.13 Civil service schemes pensioners to active employees.12 are almost always DB (Hong Kong SAR, 146 LIVE LONG AND PROSPER China, is the exception), and the typical for- rate targets and contribution rates in their mula rewards the later years of participation mandated schemes, and the majority of disproportionately (known as backloading). workers and employers pay income taxes. Given the lack of portability arrangements in • In China and Thailand, the existing occu- national schemes that cover private sector pational arrangements need restructuring workers, a public sector worker who moves to encourage and incentivize additional to the private sector suffers a large loss in pen- voluntary coverage. In China, members sion wealth. of the “enterprise annuity” plan (which For all of these reasons, an increasing num- are DC accounts) have been given limited ber of countries in all regions have started to tax preferences. Membership is around integrate civil servants into their national pen- 18.5 million workers, consisting mostly of sion schemes. Recent reforms in China, public enterprise employees—a small frac- Japan, Vietnam, and eventually Indonesia are tion of the overall formal sector workforce. examples in East Asia and Pacific. Integration In Thailand, private pension coverage rose presents an important policy question for the to around 7 percent of the labor force in rest of the countries that continue to have 2014.16 Thailand has relatively low ben- separate civil service schemes,14 but integra- efit targets and contribution rates, but few tion is particularly urgent in the poorer coun- qualify for favorable tax treatment. tries that do not yet have national mandates • Malaysia and Singapore have attempted that apply to private sector workers. to increase voluntary savings through Relatively small numbers of formal sector their existing provident fund schemes. workers and low administrative capacity in The programs offered by Malaysia17 and these countries suggest that they should seri- Singapore draw few participants—less ously consider avoiding parallel systems. than 5 percent of those covered by their respective provident funds. This low participation rate is partly the result of Supplementary private pensions already high forced savings rates in those The role of private, voluntary pension countries, which leave little space for addi- schemes is relatively limited throughout most tional voluntary pensions. of the region. Participation in these schemes • In the rest of the region, private voluntary typically is driven by tax preferences, and cov- pensions currently play a negligible role or erage tends to overlap with workers already are nonexistent. In Cambodia, Indonesia, participating in mandated schemes (with Lao PDR, Myanmar, and Timor-Leste, the some exceptions among the self-employed). enabling conditions for private pensions are When mandated pension schemes have high not yet in place, and few workers would target replacement rates and when contribu- benefit from tax preferences. The focus tion rates are already high, little room remains should be on reforming the existing public for voluntary provision. Another important systems of these countries. Vietnam offers factor is the share of the population subject to scope for launching or developing a vol- the income tax, which is extremely limited untary occupational provision targeted at outside of the richer countries. small and medium enterprises as the replace- The situation with regard to voluntary ment rates in the public system decline. pensions in the region can be summarized as follows: Noncontributory (social) pensions • The only countries with significant private Globally, the role of social pensions—cash pension coverage are Japan and Korea transfers paid to the elderly, financed from the with around 23 and 30 percent of the labor general budget and not linked to prior force covered, respectively.15 These coun- contributions—has expanded dramatically in tries have relatively modest replacement the past two decades (see Palacios and PENSION SYSTEMS AND AGING: REFORMS FOR WINNING THE RACE 147 Knox-Vydmanov 2014; Rofman, Apella, and Assessment of the performance Vezza 2014). To a large extent, this growth of East Asia’s pension systems can be attributed to frustration with the slow expansion of contributory pension schemes in Challenges the face of aging populations. Pension systems in East Asia and Pacific In East Asia, reliance on social pensions exhibit diverse challenges in terms of sustain- has increased in recent years, although their ability, coverage, and adequacy (for example, relative importance in national pension sys- see ILO FACTS 2013; OECD 2013; Palacios tems varies significantly. Three groups of 2015). This section looks at all three aspects economies can be distinguished: those with and also reviews the indirect economic effects recently expanded social pensions, those that of pension systems: have introduced targeted transfers, and those that have no social pensions, including • Financial sustainability is a particularly Cambodia, Indonesia (except a pilot), Lao acute challenge in countries facing unprec- PDR, Myanmar, and Singapore. China, edented population aging. On the one Korea, Thailand, Timor-Leste, and several hand, financial sustainability can directly Pacific Island countries, including Kiribati, affect adequacy when financial short- Samoa, Tonga, and New Ireland Province in falls lead to reduced benefits and partial Papua New Guinea, have recently expanded defaults on pension promises. On the social pensions to cover the majority of the other hand, rising pension spending can elderly. Korea and Thailand aim to address threaten the overall fiscal picture. the coverage gap left by their recently intro- • The coverage gap is defined as the share of duced contributory pension schemes. In the workers who will not have any kind of pen- case of Timor-Leste, social pension coverage sion income from contributory or noncon- was made universal at age 60, and benefits tributory pension schemes. Consistent with were set at a high level relative to local the global picture, coverage of contributory incomes. In contrast, the expansion of social schemes in East Asia and Pacific is closely pensions in Thailand was more gradual, tak- correlated with country income levels, with ing place over roughly a five-year period until very low coverage in low-income and lower- a modest social pension achieved practically middle-income countries in the absence of universal coverage among the elderly not broad-based social pensions.18 China is a qualifying for a formal pension. China’s notable exception because of its unique recent expansion of coverage for informal hybrid scheme for informal workers. sector workers is unique in terms of design • Adequacy directly relates to coverage and dramatic in terms of the pace of change, because many workers participate in con- as described in more detail below. tributory pension schemes for only part Social pensions play a more limited role of their careers, which results in relatively in economies where targeted transfers to the low benefits. This challenge is particularly poor elderly have been introduced, in some true for low-income workers who shift cases with relatively high eligibility ages (for between the formal and informal sectors example, 77 years in the Philippines). This as well as for women who tend to spend a group includes Hong Kong SAR, China; greater share of their working age caring Japan; Malaysia; the Philippines; and for children and elderly relatives. Vietnam, where social pensions are universal for those ages 80 and over but are means tested for those ages 60–79. Mongolia also Pension spending and sustainability has a significant and growing number of social With two exceptions, East Asian and Pacific pension beneficiaries. As discussed in the next economies currently spend less on pensions section, benefit levels and eligibility conditions than other economies after taking into vary significantly across these countries. account their demographics (figure 5.2). 148 LIVE LONG AND PROSPER FIGURE 5.2 East Asian and Pacific economies currently spend less on pensions than other economies after taking into account their demographics Public pension spending in East Asian and Pacific economies and the world 20 y = 0.0037x2 + 0.2572x – 0.3111 R2 = 0.6907 18 16 Public pension spending (% of GDP) 14 12 10 Japan 8 Vietnam 6 Mongolia 4 Timor-Leste China 2 Philippines Thailand Hong Kong SAR, China Korea, Rep. 0 0 5 10 15 20 25 30 35 40 Share of population age 60 and over (%) Sources: World Bank pensions database, http://www.worldbank.org/pensions; United Nations population data for the latest available year. Several factors contribute to this difference • The low spending levels in Cambodia and in pension spending: Myanmar can be explained by the fact that no mandated scheme exists for the • DC schemes in Hong Kong SAR, China; private sector or for social pensions. Indonesia; Malaysia; and Singapore limit • Another factor may simply be culture, their pension spending, at least in terms of which influences the three factors above. current budgetary outlays. Indonesia and Much of East Asia and Pacific tradition- Malaysia spend only on DB civil service ally has placed a strong reliance on fam- pension schemes, whereas Hong Kong SAR, ily and informal support networks such as China, spends only on social pensions. wantoks in the Pacific. Policy makers con- • Scheme maturation plays a role in Korea, tinue to assume that traditional support Lao PDR, and Thailand, where mandates can substitute or at least complement pub- for DB schemes have been introduced lic pension support, although chapter 2 recently. Their expenditures will rise presented evidence questioning the future toward the fitted line seen in figure 5.2 validity of this assumption. in the coming years as a greater share of workers qualifies for full pensions. Even Despite the strong relationship between after they mature, however, their low aging and public pension spending across target benefit levels suggest modest spend- countries, policy choices have room to play a ing by global standards. role. Independent of the policy approaches PENSION SYSTEMS AND AGING: REFORMS FOR WINNING THE RACE 149 taken, for decades pension spending in Japan earmarked budgetary resources are used to has been lower than in other countries at the cover part of pension spending by design. same stage of demographic transition, and According to the 2014 actuarial report, projections for Korea suggest that, even after under all but the most pessimistic case, its scheme matures, it will also spend less than reserves are never exhausted through the other aging countries. projection period that runs to 2055. Among Mongolia and Timor-Leste are exceptions the other countries, Mongolia has no to the general pattern of low pension spending reserves, and the scheme is already running in the region. Timor-Leste introduced univer- deficits. Spending is expected to exceed con- sal pensions for individuals age 60 that pay tributions in Vietnam by 2021, according to around 44 percent of average, non-oil income International Labour Organization projec- per capita, including some disability benefits. tions (although this projection was prior to In Mongolia, an inherited Soviet-style system 2014 reforms, which have somewhat had much higher coverage because of the role improved sustainability), and somewhat of the state as employer. These two countries later in Korea and the Philippines. Without spend notably more than other countries at reforms, these cash flow deficits will lead to the same stage of demographic aging. higher general government deficits.19 Although the regional picture is incom- Two caveats apply in the case of China. plete, available estimates of pension system First, some provinces still have reserves, sustainability point to substantial fiscal risks whereas others run significant deficits, and in a number of national pension systems. there is a separate but relatively small national The results are not strictly comparable fund, the National Social Security Fund because they are generated with different (NSSF). Second, the figures refer only to the models and assumptions and use data of DB scheme. However, as noted, evidence varying degrees of quality. Nevertheless, gen- shows that most of the contributions into DC eral patterns are evident in table 5.3, which accounts in China are used for current spend- shows the projected year of cash flow defi- ing and are therefore not accumulating (that cits as well as the year in which reserves are is, empty accounts). Nationally, subsidies to exhausted. Cash flow deficits are expected to the pension system have been estimated to cost emerge in all East Asian and Pacific DB around 3.8 percent of gross domestic product schemes in coming decades. In Japan, (GDP) annually (Dorfman et al. 2013). TABLE 5.3 Defined benefit schemes in East and Southeast Asia face significant sustainability challenges Indicators of long-run financial sustainability of main public defined benefit schemes, selected East Asian and Pacific economies Deficit or surplus in 2040 Country (year) Source (% of GDP)a China (2010) Herd, Hu, and Koen 2010 −3.5 Ma, Zhang, and Li 2012 −3.3 Korea, Rep. (2013) Korea National Pension Service −1.4 Mongolia (2011) World Bank 2011 −4.5 Philippines (2011) Philippine Social Security Service; Mesa-Lago, Viajar, Negative from 2026 and Castillo 2011 Philippine Government Service Insurance System; Negative from 2029 Mesa-Lago, Viajar, and Castillo 2011 Thailand (2011) Yamabana 2011 Negative from 2041 Vietnam (2012) ILO FACTS 2013 −1.7 Sources: Herd, Hu, and Koen 2010; ILO FACTS 2013; Ma, Zhang, and Li 2012; Mesa-Lago, Viajar, and Castillo 2011; World Bank 2011; Yamabana 2011. Note: Japan finances basic pensions from general revenues per policy, so a shortfall is not technically a deficit. a. Refers to cash flow deficits, that is, current contributions minus current benefits. For China, the figure excludes the National Social Security Fund. Thailand excludes civil servants. 150 LIVE LONG AND PROSPER Only Timor-Leste spends a large amount on As noted earlier, no mandated scheme for pri- social pensions: more than 2.5 percent of GDP vate sector workers exists in Cambodia, (non-oil), suggesting significant sustainability Myanmar, and Timor-Leste, where only civil issues. By comparison, no other East Asian servants are covered. In the case of Indonesia, and Pacific countries spent more than 0.4 poor collection and enforcement of the man- percent of GDP on social pensions. Based on date are likely to explain some of the short- current parameters, the present value of spend- fall. In Malaysia, low labor force participation ing on this program in Timor-Leste through rates of women are part of the explanation, as 2040 as a share of today’s GDP is more than well as the exclusion of most migrant work- 100 percent. The same present value figure for ers. The explanation for the low contributory Thailand, where coverage is high but benefits coverage found in Indonesia, Lao PDR, and are much lower, is about 20 percent. Thailand is less obvious. Overall, if current international patterns are a guide, most of the Coverage region is likely to experience its demographic In terms of participation in contributory pen- aging process before coverage rates in man- sion schemes, East Asia and Pacific is at or dated schemes rise to meet the challenge. somewhat below par globally, with the excep- The pension schemes in the Philippines and tion of several low-income countries where Mongolia cover a slightly larger share of their coverage is minimal—a worrying picture in a labor forces than predicted by their income region where aging is exceptionally rapid. (figure 5.3). In the Philippines, the main social FIGURE 5.3 Coverage of contributory pension schemes is around par globally for income levels and demography Coverage of working-age population in contributory pension schemes and income per capita, most recent year 120 y = 4.2633x2 – 56.792x + 192.16 100 R2 = 0.8735 Share of working-age population covered (%) Singapore 80 Japan Korea, Rep. 60 China Hong Kong Mongolia SAR, China 40 Vietnam, Philippines Malaysia Cambodia, Lao PDR, 20 Myanmar Indonesia 0 5 6 7 8 9 10 11 12 Log of PPP adjusted GDP per capita Source: World Bank pensions database, http://www.worldbank.org/pensions. Note: PPP = purchasing power parity. Data are from most recent year and refer to mandated contributory coverage only. PENSION SYSTEMS AND AGING: REFORMS FOR WINNING THE RACE 151 security scheme is known for its proactive FIGURE 5.4 Elderly pension coverage is low in the bottom income stance on collecting from the self-employed deciles in China, Indonesia, and the Republic of Korea and its migrant workforce. The higher-than- Pension coverage among the elderly by income decile predicted coverage rates in Mongolia are almost certainly due to the legacy of the cen- 100 trally planned economy. As the country contin- 90 Share of population age 60 and over ues its transition toward a market economy, 80 coverage has been falling, as has been the case 70 receiving pension (%) in Central Asia (see Schwarz and Arias 2014).20 China’s recent success in expanding cover- 60 age, in part because of innovations in design 50 of its scheme, is unprecedented in global expe- 40 rience. The expansion of coverage of informal 30 rural and urban workers in China since 2010 represents the single biggest increase in con- 20 tributors to a government pension system in 10 history. The expansion is unique not only in 0 terms of its scale but also in the design of the 1 2 3 4 5 6 7 8 9 10 scheme, which links contributions both to eli- (poorest) (richest) gibility for a basic pension at age 60 and to Income decile immediate benefits for the elderly parents of Poland Hungary Japan contributors. This incentive is important Korea, Rep. China Indonesia because contributions for these informal sec- tor workers are voluntary. If these new con- Sources: CHARLS 2011 (China); IFLS 2007 (Indonesia); JSTAR 2011 (Japan); KLoSA 2010 (Republic of tributors are counted as part of the covered Korea); Palacios 2015 (Hungary and Poland). population, coverage would increase from about where China’s income level would have predicted (about 28 percent) to well over 60 percent in East Asia and Pacific, highlighting percent. Given its importance, China’s experi- the challenges of expanding coverage through ence is discussed again later in this chapter. a purely contributory route. Figure 5.5 shows Countrywide averages on pension receipt the difficulties of increasing contributory hide the fact that coverage is often closely pension coverage among the poorer segments related to income level within countries, espe- of populations in East Asia and Pacific, who cially where overall coverage is low. Figure 5.4 nearly always work in the informal sector. shows that coverage in Indonesia is less than The only countries with substantial coverage 10 percent for the bottom half of the distribu- among poorer groups are those with wide tion but rises to around one-third for the top social pension coverage, such as China, quintile. The pattern is even more dramatic in Mongolia, Thailand, and Timor-Leste. China, where the elderly in the top decile are (Timor-Leste and recent expansions of cover- as likely to receive pensions as the same group age in China and Thailand are not reflected in in Japan.21 Meanwhile, fewer than one in five figure 5.5.) elderly Chinese in the bottom half of the In East Asia and Pacific, on average, a man income distribution received a pension in age 60 or above is more likely to receive a 2011.22 In the Philippines, only 2.8 percent of pension than a woman (figure 5.6). This cal- people in the lowest income decile were culation takes into account survivor pensions, receiving a pension in 2010 (Mesa-Lago, suggesting that the gap is even higher to the Viajar, Castillo 2011). extent that women disproportionately receive With notable exceptions, the share of the survivor pensions, which typically have much bottom 40 percent of elderly receiving any lower value. The gap is narrowest in Japan, pension has struggled to rise beyond 20 where a number of allowances are made for 152 LIVE LONG AND PROSPER FIGURE 5.5 Increasing contributory pension coverage among the schemes. To compare schemes in terms of poorer segments of populations is difficult individual outcomes or target benefit levels, Share of households with elderly in bottom 40 percent of distribution the analysis can simulate the results for a receiving pensions hypothetical worker who contributes during his or her entire career.23 Figure 5.7, which Rwanda Zambia compares the results in terms of gross and net Afghanistan Nigeria replacement rates,24 shows a wide range of Malawi Bhutan design choices in the benefit targets for aver- Uganda Mozambique age wage workers. China, the Philippines, and Lao PDR Indonesia Vietnam have high target replacement rates Honduras Maldives compared with the rest of the region and Cambodia Pakistan OECD countries.25 Japan, Korea, Mongolia, Guatemala Dominican Republic and Thailand have moderate net replacement Slovenia Nicaragua rate targets at around 40 to 50 percent, Bangladesh Ghana whereas the four countries that rely on DC Paraguay Sri Lanka schemes have the lowest target benefit levels. Vietnam Bolivia Indonesia stands out—with replacement rate Ecuador Philippines targets implied by the design of its schemes— Peru Venezuela, RB at below 20 percent.26 Colombia Mexico The degree of redistribution between low- Panama Iraq and high-income workers also varies consid- Costa Rica Thailand erably in the region’s pension systems. Many Chile Kosovo pension schemes have minimum pensions Bosnia and Herzegovina Argentina and ceilings on the wages on which contribu- Turkey Brazil tions are made. As a result, target replace- Uruguay Montenegro ment rates are often higher for low-wage Albania workers and lower for high-wage workers. Croatia Serbia Figure 5.8 shows the difference in simulated Armenia Tajikstan net replacement rates for workers with half Bulgaria Moldova and twice the average wage, respectively. Romania Poland Except in Indonesia, Malaysia, Singapore (all Hungary Mongolia DC), and Vietnam, target replacement rates Kyrgyz Republic Belarus are significantly higher for lower-income Slovak Republic Mauritius workers. The biggest differences are seen in Ukraine Korea and the Philippines. 0 10 20 30 40 50 60 70 80 90 100 Share of households with elderly receiving pension (%) In contrast to most high-income countries, actual replacement rates in DB schemes in Other countries EAP countries East Asia and Pacific (except in Vietnam) Source: World Bank Aspire data. tend to be lower than would be expected given the design parameters and are modest by global standards. Figure 5.9 shows the housewives and women with low rates of actual average replacement rates for labor force participation, one of those being earnings-related DB schemes in the region. provision of a basic pension. The gap is high- Unlike target replacement rates, this indicator est in Korea, perhaps because of how cover- applies to actual pensions today rather than age has expanded since the National Pension simulated pension outcomes for those start- Service (NPS) was set up in 1988. ing to contribute today, and it is an average of all existing pensions divided by the average Adequacy wage for all workers covered. Thus, it The target replacement rate varies signifi- includes early retirement pensions, retirement cantly among East Asian and Pacific pension pensions for workers across the income PENSION SYSTEMS AND AGING: REFORMS FOR WINNING THE RACE 153 distribution, survivor and disability pensions, FIGURE 5.6 Elderly pension coverage is higher for men than for women and pensions granted many years ago. The low replacement rates in Korea and Thailand 90 reflect the immaturity of their schemes—as Share of population age 60 and over 80 the schemes mature, replacement rates will 70 receiving pension (%) rise to their steady-state values, which in the 60 case of Korea is around 2028. Already, work- ers with more than 20 years of contributions 50 have replacement rates twice the average. 40 As discussed earlier, civil service schemes in 30 East Asia and Pacific tend to have higher 20 replacement rates than those for private sector 10 workers. Wang, Béland, and Zhang (2014) show that civil service replacement rates in 0 China Korea, Rep. Indonesia Japan China were about 89 percent, compared with around 45 percent for workers outside the Women Men civil service. In Vietnam, different benefit for- Sources: CHARLS 2011 (China); IFLS 2007 (Indonesia); JSTAR 2011 (Japan); KLoSA 2010 (Republic of Korea). mulas applied to private and public sector workers until the 2014 reforms. This resulted FIGURE 5.7 A wide range of design choices in the pension benefit in huge differences in replacement rates. is available Thailand’s civil servants have a hybrid DB/DC Simulated gross and net replacement rates for hypothetical workers scheme that provides replacement rates at least double the modest DB scheme for private OECD (average) sector workers. Civil servants in Indonesia LAC (average) have a fairly generous DB scheme, although in China practice, a large share of workers’ total com- Vietnam pensation is not considered as part of the pen- Thailand sionable wage base, so actual replacements Mongolia are lower. Nevertheless, their counterparts in Korea, Rep. the private sector have a very small DC Singapore scheme that provides little income during Philippines retirement and no longevity insurance. A simi- Japan lar situation is observed in Malaysia, where Malaysia civil servants have a DB scheme and private Hong Kong SAR, China sector workers contribute to a DC scheme Indonesia with low median accumulations. Finally, in 0 10 20 30 40 50 60 70 80 90 the Philippines, the replacement rate for civil servants is slightly lower than for private sec- Share of gross/net wage (%) tor workers, although absolute benefits are Gross replacement rate Net replacement rate higher as a result of wage differentials. Sources: OECD 2011, 2013; Wiese 2006. DC schemes in East Asia and Pacific face Note: LAC = Latin America and the Caribbean. OECD average is for 34 Organisation for Economic Co- significant challenges in the low level of ade- operation and Development (OECD) countries based on 2012 rules. Simulated results are prospec- tive and refer to new labor market entrants employed and making contributions throughout their quacy of retirement income, driven in some careers. For defined contribution schemes, the net rate of return is assumed to be 1.5 percentage countries by ability of workers to make prere- points higher than wage growth. Korean figures do not include retirement allowances. In Indonesia, a new regulation issued by the government of Indonesia in July 2015 took effect that would substan- tirement withdrawals as noted earlier. In tially increase the net and gross replacement rates, bringing them closer to the regional average. Indonesia, Malaysia, and the Pacific Islands, certain preretirement withdrawals are allowed of members age 54 have less than RM 65,000 that lower the balances available for retire- in their account.27 Given the life expectancy ment, in many cases substantially. For exam- at age 55, this amount translates into an ple, according to Malaysia’s Employees annual payment between 55 and death equiv- Provident Fund (EPF) website, more than half alent to one-quarter of income per capita. 154 LIVE LONG AND PROSPER FIGURE 5.8 In most economies, target replacement rates are Singapore, where the nominal return has significantly higher for lower-income workers essentially been fixed since the scheme began. Simulated net replacement rates for hypothetical high- and low-income Second, in Hong Kong SAR, China, the invest- workers ment return is an average, and returns for individuals vary widely according to choice of China investment portfolio and fund (MPFA 2014). The most disappointing case in terms of Vietnam DC returns is that of China, where returns are Korea, Rep. linked to short-term bank deposits. As a result, Mongolia returns on funds in DC accounts have barely Philippines exceeded inflation and have lagged far behind wage growth since long before individual Japan accounts were introduced. Figure 5.11 shows Thailand that they continue to do so. To the extent that Singapore this lag continues, the replacement rates will be much lower than those shown earlier. Malaysia Critics have noted that replacement rates Hong Kong SAR, China experienced by cohorts retiring in the past few Indonesia decades in Singapore have been very low, but recent trends in wage growth and policy 0 20 40 60 80 100 120 changes may improve the situation (Asher Share of net wage (%) 2012; Asher and Nandy 2011).29 Its Central Net replacement rate, low-income male Provident Fund (CPF) payouts are substantial Net replacement rate, high-income male relative to lifetime income but not relative to current incomes. Now that Singapore is a Sources: OECD 2011, 2013; Wiese 2006. high-income country, wage growth is stabiliz- Note: Simulated results are prospective and refer to new male labor market entrants employed and making contributions throughout their careers. High and low income here refers to workers with ing at lower rates. If combined with an twice and half the average wage, respectively. For defined contribution schemes, net rate of return increase in the effective retirement age, this is assumed to be 1.5 percentage points higher than wage growth. stability should result in higher replacement rates in the future. However, cohorts that are Public awareness that the accumulated bal- currently retiring do not have adequate ances are not sufficient is increasing, and savings—thanks to their CPF housing invest- recently a number of remedial measures have ments and appreciation of property values, been proposed, including delayed withdrawal many are asset rich and cash poor—which has (see, for example, Mahalingam and Wong led to several recent adjustments to the pro- 2013). Balances are even lower relative to gram. A lease buyback program allows elderly member incomes in Indonesia largely because Singaporeans to sell their housing back to the of low contribution rates combined with lib- government, along the lines of a reverse mort- eral withdrawal rules.28 gage. Most recently, the government has In some cases, another driver of low ade- announced the Silver Support scheme to sup- quacy is the actual returns on DC accounts plement the pensions of those without suffi- relative to the growth of wages. The simula- cient accumulated savings (Rodan 2014). tions presented assumed that the net rate of For many people in East Asia and Pacific, return on DC schemes exceeds growth of social pensions are the key determinant of wages by 1.5 percentage points annually. overall pension system adequacy, and benefit Figure 5.10 shows that the differential has levels are typically modest. This situation is been slightly higher than assumed in Hong confirmed by figure 5.12, which compares the Kong SAR, China, but much lower in China ratio of social pensions to income per capita and Singapore. Two important caveats should in selected countries with the ratio of social be pointed out: first, returns in Hong Kong pension recipients to the population ages 65 SAR, China, are much more volatile than in and above. PENSION SYSTEMS AND AGING: REFORMS FOR WINNING THE RACE 155 The East Asian and Pacific economies with FIGURE 5.9 In most economies, actual replacement rates in DB social pension programs fall roughly into schemes tend to be lower than would be expected three groups:30 Actual average replacement rates in selected countries • Social pension benefits range from 6 to 70 24 percent of income per capita in Japan, Malaysia, Mongolia, the Philippines, and 60 Vietnam. However, the number of social pensioners is very small. 50 • Benefit levels are low—between 5 and 9 percent of income per capita, in China; 40 Percent Hong Kong SAR, China; Korea; and Thailand—but the ratio of social pension- 30 ers to the population ages 65 and above is quite high, ranging from 54 to 94 percent. 20 As noted earlier, Mongolia has increased its reliance on social pensions over the 10 past decade, with benefits totaling about 20 percent of income per capita (World 0 Bank 2011). Korea, Rep. Mongolia Thailand China Philippines Vietnam • The only East Asian and Pacific coun- Sources: ILO FACTS 2013; NPS 2012a, 2012b; Mesa-Lago, Viajar, and Castillo 2011; SSO 2009; Wang, tries shown with relatively high social Béland, and Zhang 2014 (for urban workers only). Note: DB = defined benefit. Percentages refer to average of old-age, disability, and survivor benefits pensions and high coverage are Kiribati divided by relevant average covered wage. and Timor-Leste. Their programs are significantly more generous than other FIGURE 5.10 Real returns on DC accounts and growth of real wages programs for which data are available Average annual real returns on defined contribution accounts and growth and even compared with countries that of real wages have high coverage levels, such as South Africa. Timor-Leste’s ratio is actually 16 greater than 100 percent because its eligi- 14 bility age is 60. 12 The role of social pensions should be % change per annum interpreted with caution in light of interac- 10 tions with contributory pension schemes and 8 social assistance programs. In Japan, Malaysia, and Mongolia, contributory pen- 6 sion coverage is high, and the schemes are 4 mature. Many elderly have recourse to the formal pension system, and social pensions 2 play only a supplementary role. Although 0 coverage is much lower in Indonesia and the China Hong Kong SAR, Malaysia Singapore Philippines, both countries have large pro- (1998–2013) China (2001–13) (1984–2010) (1987–2011) grams that provide cash, in-kind transfers, Return Wage growth and health insurance to a significant propor- Sources: Asher and Bali 2013; MPFA 2014; World Bank 2012a. tion of poor households. With very high co- Note: DC = defined contribution. Malaysia shows income per capita growth. residence rates in both countries, especially among poor households, a significant pro- Overall, the least adequate pension sys- portion of the poor elderly benefit from tems in East Asia and Pacific are in those these programs.31 Nevertheless, a large cov- countries where absolute poverty is high, erage gap clearly remains in these countries pension coverage is low, and neither a broad and Vietnam. social assistance program nor a significant 156 LIVE LONG AND PROSPER FIGURE 5.11 Returns on funds in DC accounts in China have barely social pension is in place. These countries exceeded inflation and have lagged far behind wage growth include Cambodia, Lao PDR, and Myanmar, The growth of wages versus interest rates and DC returns which have no income support for the elderly outside of contributory formal sector 3,000 schemes that cover less than 10 percent of the population. The Philippines and Vietnam 2,500 have somewhat higher coverage through their contributory pensions, but their social 2,000 pensions are very small and pay very low Index (1990 = 100) benefits. Although Indonesia and the 1,500 Philippines do provide some social assistance and health insurance, both countries have a large gap in terms of adequate income pro- 1,000 tection for the elderly. Despite the broad social pension coverage that has now 500 emerged in China and Thailand, the level of pensions in both absolute and relative terms 0 is too low to be considered adequate. 1990 1995 2000 2005 2010 2013 Among richer economies in East Asia and Inflation Wage growth Pacific, the rapid growth of wages in recent Growth of one-year term deposits decades has resulted in pensions that are high relative to the lifetime income of the Source: China Statistical Yearbook Database, various years. Note: DC = defined contribution. person but not relative to current incomes. Although such rates of income growth are not expected to continue, the increases will affect the situation in China; Hong Kong FIGURE 5.12 Social pension benefits are modest in East Asia SAR, China; and Korea, where no pension and Pacific scheme operated during much of the high Social pension indicators in East Asia and selected economies growth period. Without forced savings schemes, the incomes of the elderly in Hong 50 Kong SAR, China, and in Korea are tied to Benefit as a share of per capita income (%) Kiribati 45 Timor-Leste personal savings decisions and the ability to 40 work into old age. Korea, for example, is 35 Australia clearly an outlier among OECD countries in France Brazil 30 Chile terms of its elderly, with high poverty rates 25 Germany South Africa and reliance on labor income. Rapid wage Japan 20 Mongolia Canada Georgia growth also affects Malaysia and Singapore Samoa United States because of the relationship between wage 15 Bolivia Malaysia Hong Kong growth and returns on account balances. 10 Korea, Rep. Vietnam SAR, China The DC schemes used in these countries do 5 Philippines Mexico India China Thailand not allow for intergenerational transfers, and 0 unless investment returns exceed wage 0 20 40 60 80 100 120 growth, the approach will generate low Social pensioners as a share of population age 65 and older (%) replacement rates. Sources: Based on HelpAge International database, http://www.helpage.org/resources/ageing- data/; OECD 2013. Indirect effects of pension schemes on Note: Data are for years 2010–12. Values are limited to 100 percent, but they exceed this ratio in some countries because they are universal and the eligibility age is 60. the economy Aside from the effects on public finances, pen- sion systems can affect labor markets on PENSION SYSTEMS AND AGING: REFORMS FOR WINNING THE RACE 157 several fronts, starting with the effects on the FIGURE 5.13 Older economies have higher pension contribution demand and supply of labor. Two potential rates than do younger economies effects are important: higher rates of informal Total contribution rates for pensions in East Asia and Pacific compared with sector activity and lower labor force partici- selected economies pation of the elderly. (A third potential effect Honduras of the pension system is on individual behav- Rwanda ior, caused by the lack of benefit portability.) Indonesia Thailand The international evidence for the first effect Congo, Dem. Rep. is limited, and what evidence does exist is Côte d’lvoire mixed. Two studies in Latin America found Ireland that a reduction in social security contribu- tions led to an increase in formal Mauritania Korea, Rep. Lao PDR employment—in both cases, a 10 percent Trinidad and Tobago reduction led to a roughly 5 percent increase Cabo Verde in formal employment (Heckman and Pagés Hong Kong SAR, China 2004; Kugler and Kugler 2009). Two other Philippines empirical studies, however, found no evidence Madagascar of such an effect when payroll taxes were Iceland reduced (Cruces, Galiani, and Kidyba 2010; Palau Islands Gruber 1997). Betcherman and Pagés (2007) Solomon Islands found that reducing social security contribu- Senegal tions led to a significant increase in formal employment in Turkey, and Koettl and Weber Nigeria (2012) found that the marginal effective tax Qatar rate is associated with higher rates of infor- Luxembourg mal labor market activity in European Union Japan countries. In addition, high legal minimum Bulgaria wages, restrictions against firing, compliance with various regulations, and transaction Mongolia costs involved with collection and reporting Turkmenistan all combine to make operating in the gray Kazakhstan economy more attractive. These incentives are Vietnam Norway no different in East Asia (World Bank 2014). Pension contribution rates and their poten- Finland Malaysia Singapore tial impacts on formality and competitiveness Liechtenstein vary widely, both globally and within East Bosnia and Herzegovina Asia and Pacific. Figure 5.13 compares total Croatia contribution rates (employee plus employer) Lithuania for pensions in about 150 countries. Not sur- China prisingly, older countries with mature public Czech Republic DB schemes in Europe and some Latin American and Caribbean countries have the Brazil Ukraine highest contribution rates, whereas younger, 0 5 10 15 20 25 30 35 40 immature pension schemes in Sub-Saharan Employee plus employer contribution rate (%) Africa or Central America tend to have lower contribution rates. The range is as large Source: World Bank pensions database, http://www.worldbank.org/pensions . within East Asia and Pacific as it is globally, Note: Singapore’s rates are for workers up to age 50, after which they fall gradually until they reach 1 from 6 percent to 28 percent (excluding percent at age 65 and above. Cambodia and Myanmar, which have no mandate). 158 LIVE LONG AND PROSPER East Asia and Pacific can be divided the population that is covered by the pension roughly into two groups in terms of pension system. This results in highly divergent contribution rates.32 The first group—which retirement patterns, particularly between includes Hong Kong SAR, China; Indonesia; urban and rural workers, as shown in chap- Korea; Lao PDR; the Philippines; and ter 2. That discussion notes, for example, Thailand—has modest total contribution that the strongest determinant for exiting rates that place them in the bottom third of work at older ages among urban Chinese is the distribution at between 5.7 and 10.3 the receipt of a pension. Receipt of a pension percent. The second group—China, Japan,33 was also associated with exiting the labor Malaysia, Mongolia, Singapore, and force in Indonesia and (with a weaker rela- Vietnam—has contribution rates between 18 tionship) in Japan and Korea, which is con- and 28 percent. sistent with the design of their pension Pension contribution rates in the second schemes.34 group of countries are likely to have a signifi- Retirement ages in national pension cant impact on formality decisions, especially schemes are relatively low in East Asia and contributions by smaller firms and the self- Pacific compared with life expectancy at age employed. This impact is a particular concern of retirement, as shown in figure 5.14. in China and Vietnam, which are aging Indonesia, Malaysia, and Thailand are well quickly and where the race to extend cover- above the OECD average for life expectancy age is most urgent. High historical rates of at retirement, and only the Philippines is coverage in Mongolia are also not guaranteed below average. The situation is more dra- to continue as the economy evolves. matic for women: their life expectancy is Ultimately, the factors that lead to formaliza- more than 25 years at the normal retirement tion, including incentives, economic structure, age, except in Korea and the Philippines. and enforcement capacity, encompass much Moreover, in countries with lower coverage, more than the pension system. However, in these figures understate the extent to which the context of rapid aging, upward pressure life expectancy at retirement exceeds that in on pension and other social insurance contri- OECD countries, because the mortality rates bution rates will continue unless financing used for the calculation are based on national policies change. To some extent this increase averages. The covered population, which is already happening, as in the case of the tends to be urban and in the upper half of growing role of earmarked value-added taxes the income distribution, almost certainly in Japan. The policy options will be discussed lives longer than the national averages. in the next section, on reforms. For most East Asian countries, LFPRs are The impact of the pension system on indi- at or above levels found in countries whose vidual decisions regarding when to retire is elderly have similar health-adjusted life expec- much more direct and can be attributed to tancy. The clear exception is Mongolia, where the design of the pension scheme itself. LFPRs are much lower than in comparable Several studies that focused on OECD coun- countries. However, figure 5.15 shows that tries have shown that the timing of retire- ability to work at older ages is not closely ment is influenced by the incentives built linked to LFPRs of the elderly.35 into the pension and tax rules. Gruber and A much stronger relationship is seen Wise (1999) and Wise (2004) find a strong between pension spending and LFPRs of the relationship between the incentives built into elderly. Figure 5.16 compares the average the pension systems and actual retirement share of GDP that countries spent on pub- patterns across 12 OECD countries. Duval licly mandated pensions in 2010 and LFPRs (2003) finds similar results for a sample of of people ages 60 and above. The expected 22 OECD countries. For East Asia and negative correlation is evident, but its inter- Pacific, the effect on labor force participation pretation is not straightforward. Higher pen- rates (LFPRs) will be restricted to a subset of sion spending is associated with both higher PENSION SYSTEMS AND AGING: REFORMS FOR WINNING THE RACE 159 FIGURE 5.14 Retirement ages in national pension schemes are relatively low in East Asia and Pacific compared with life expectancy at age of retirement Life expectancy at retirement age a. Male b. Female Thailand Philippines Malaysia United States Indonesia Germany Singapore Korea, Rep. Vietnam Singapore Japan Mongolia China Sweden Sweden Indonesia Korea, Rep. Japan United States Malaysia Germany Thailand Mongolia Vietnam Philippines China 0 5 10 15 20 25 0 10 20 30 40 Life expectancy after retirement (years) Life expectancy after retirement (years) Source: OECD 2013. Note: Official retirement ages vary by country. average pensions and larger numbers of pen- FIGURE 5.15 Ability to work at older ages is not closely linked to sioners, both of which should contribute to labor force participation rates lower labor force participation of the elderly. Health-adjusted life expectancy and labor force participation rates for males As public pension spending rises, the ages 60 and older expected decline in LFPRs of older workers 100 is observed, and the effect is generally larger y = 0.2414x2 – 11.302x + 140.21 90 Male labor force participation rate for spending increases from a lower base. R2 = 0.1738 80 However, the effect may be overstated if age 60 and over (%) 70 other important factors are not taken into 60 account.36 The major outlier is Japan, which Indonesia 50 Philippines has much higher LFPRs than would be 40 Myanmar Vietnam expected, given its pension spending levels. Lao Korea, Rep. This analysis may provide evidence that the 30 PDR Japan measures Japan has introduced, such as par- 20 Singapore 10 Mongolia tial pensions with continued part-time work, may have helped keep rates higher than they 0 6 8 10 12 14 16 18 would have otherwise been. Among East Asian countries, Singapore’s Health-adjusted life expectancy at age 60 (years) LFPR stands out as having LFPRs that have Sources: For labor force participation rates, World Bank 2013; for health-adjusted life expectancy, risen more in percentage terms than those of WHO 2010. any other country in the past two decades. Although Singapore provides incentives to living standards (see Asher and Bali 2013). employers to employ older workers, at the In fact, Singapore has the highest LFPR for same time its CPF balances are generating people ages 60 and above of any country very low replacement rates, possibly forcing close to its income level. Recent changes that people to work longer to maintain their restrict withdrawals before a certain age are 160 LIVE LONG AND PROSPER FIGURE 5.16 Pension spending is closely linked to labor force increase in pension wealth. Only Japan and participation rates Korea provide a significant reward for delay- Pension spending and labor force participation rates of persons ages 60 ing retirement. and above In developing East Asian and Pacific econ- omies, in many cases incentives for early 20 Pension spending as a share of GDP (%) y = 4.302 ln(x) + 18.415 retirement are strong, even before the already R2 = 0.5499 low official retirement age. Table 5.5 shows 15 the incentives for male workers, which are implicit in the rules regarding early and late 10 Japan retirement. In Vietnam, early retirement actu- China Timor- Leste ally increases pension wealth because the pen- 5 Mongolia Singapore Philippines alty is too low to offset the extra benefits Indonesia received. The Philippines rewards early retire- 0 Korea, Rep. ment with a large increase in pension wealth -5 and penalizes late retirement at a rate of 0 10 20 30 40 50 60 70 80 90 100 17 percent per year. In contrast, Japan and Korea heavily penalize early retirement. Labor force participation rate for population age 60 and over in 2010 (%) Such incentives for early retirement can Sources: World Bank pensions database, http://www.worldbank.org/pensions; World Bank 2013. have negative outcomes and are undesirable in countries where the overall size of the labor likely to push LFPRs among those ages 55 to force is already shrinking or set to shrink in 64 even higher. the face of aging workers’ retirement and fer- In most East Asian and Pacific economies, tility decline. In China and Vietnam, for exam- official retirement ages are low—especially for ple, the average retirement ages among formal women—and have often been static for long urban workers are 3–4 years below official periods. Even for economies with larger aged retirement ages for both genders, despite those populations, such as China and Korea, official already being low, especially for women. ages are low (table 5.4). Although some move- Notably, increasing the retirement age has ment toward higher retirement ages has been consistently shown to have a more posi- occurred, such as Malaysia’s increase from 55 tive impact on economic growth than raising to 60 years, others such as Vietnam have failed contribution rates or cutting benefit levels to to pass even modest increases, despite the reduce pension scheme deficits (see Barrel, rapid aging of the population. The relatively Hurst, and Kirby 2009; Karam et al. 2010). low retirement ages lead to increased spending In East Asia and Pacific, the impact of in China and Mongolia, especially for urban pension systems on retirement behavior is workers, making pension schemes less sustain- likely to be significant only when they are able. Indonesia is a special case in that with- contributory pension systems. As shown in drawals are allowed at any age as long as five chapter 2, receipt of a contributory pension years of contribution have been made. has a strong association with likelihood of Most economies with DB schemes provide retirement, which in East Asia and Pacific little or no incentive for delaying retirement. draws those with higher levels of education Retiring later than the normal age is not and productivity out of the labor market pre- allowed in Indonesia and Vietnam, and it is maturely. In contrast, the impact of noncon- effectively penalized in China and the tributory pensions or social assistance Philippines, where an additional year of con- programs is likely to be negligible unless ben- tribution does not increase the pension. The efits are increased significantly. For example, Philippines also explicitly prohibits working Zhang, Giles, and Zhao (2014) find little evi- and receiving a pension, but only until age 65. dence that the rural pension reduces the labor In Thailand, where the retirement age is very supply of elderly in China, given the rela- low, later retirement results in only a small tively low benefit level. PENSION SYSTEMS AND AGING: REFORMS FOR WINNING THE RACE 161 TABLE 5.4 Retirement rules for East Asian contributory pension schemes Normal retirement age (men/women) Conditions for early retirement Conditions for late retirement China 60/55 Blue-collar manual workers can retire at Allowed, but pension is not increased. 55/50. Hong Kong 60/60 Allowed only in cases of terminal illness Allowed, with interest on the balance SAR, China or permanent migration. accruing for the additional period. Indonesia 55/55 for formal Withdrawal of balance allowed at any age Not allowed. sector; 58 for most subject to having contributed for five years civil servants; 56 and after one month of unemployment. proposed in SJSN Japan 65/65 Allowed at 60 with 6 percent reduction per Deferring retirement increases the year. pension benefit by 0.7 percent per Early retirement at a reduced benefit is month or 8.4 percent per year. possible in both basic and earnings-related schemes. The benefit is reduced by 0.5 percent per month of early retirement, or 6 percent per year. Korea, Rep. 60/60 Allowed at age 55, but the minimum age At 60, early old-age pension will be is rising to 60 by 2033. 70 percent of normal old-age pension. Benefit is increased by 6 percent every year, so a person who retires at age 64 will be entitled to 94 percent of the full old-age pension. People can earn extra pension from retiring late. The benefit is increased by 7.2 percent every year, and the maximum deferral is five years until age 70. Lao PDR 60/60 Allowed at 55 with reduction based on n.a. points system. Malaysia 60/60 Some withdrawal possible prior to age 55. Allowed; interest on balance accrues for the additional period. Mongolia 60/55 Allowed for workers in hazardous jobs. n.a. Philippines 65/65 Allowed at age 60 subject to minimum Allowed, but pension is not increased. contribution period. Working while receiving a pension is not allowed. Singapore 63/63 (rising to 65 Subject to having the minimum required Allowed; interest on balance accrues in 2018) balance, a lump sum can be taken for the additional period. before the standard age of withdrawal. Annuitized unless retiree opts out. Thailand 55/55 Not allowed. Allowed with additional 1.5 percent per year accrual. Vietnam 60/55 Workers in hazardous jobs and certain Not allowed. Working while receiving a areas can retire at 55/50; pension reduced pension is allowed. 1 percent per year before normal age. Source: Palacios 2015. Note: n.a. = not applicable. SJSN = National Social Security System. Indonesia’s new regulations, effective July 1, 2015, will gradually increase the retirement age from 56 to 65 by 2043. Pension systems also can have indirect An important point for fiscal accounting pur- effects on the economy through their role in poses is that these two ways of accumulating public and private savings behavior. pension assets are treated very differently. Prefunding of public pension liabilities gener- Contributions that lead to a surplus in a DB ally takes two forms: partial funding of DB scheme are considered part of general govern- schemes and full funding of DC schemes. ment revenues, and the accumulation of 162 LIVE LONG AND PROSPER TABLE 5.5 Retirement rules and incentives for male workers in negative, the DB schemes contribute directly selected countries to larger deficits, a situation now common in Early retirement Late retirement Europe and already a reality in Mongolia. Country Normal age penalty (% change) reward (% change) In contrast, evidence indicates that DC schemes are likely to have a neutral to slightly Japan 65 −23 23 Korea, Rep. 65 −22 15 positive effect on private savings.39 Most of Philippines 65 18 −17 the literature examines the changes in private Thailand 55 Not possible 1 savings when economies move from unfunded Vietnama 60 15 Not possible DB schemes to privately managed DC Source: World Bank calculations. schemes. The amount of savings depends on Note: Figures in the last two columns refer to the annual change in present value of pension how much of the transition costs are covered payments from retirement until death for early and late retirement. Figures use national mortality rates at relevant ages. by borrowing from the new DC scheme. In a. Amendments to Vietnam’s social insurance law in 2014 increase the early retirement penalty four Latin American and Caribbean countries, slightly but not enough to offset the increase in pension wealth. estimates of the overall increase in savings ranged from 1 percent to 2.5 percent of pension reserves in DB schemes can reduce GDP.40 However, no evidence shows that the explicit government deficits and debt in the Eastern European reforms of this type increase near term if surpluses are not borrowed for net savings. In East Asia and Pacific, at least additional consumption. However, a measure two reforms have involved moving from an of the true impact on public savings should unfunded DB to a funded DC model—namely, take into account the fact that unreported in China and Thailand41—but no studies have pension liabilities are growing (see Goebel, been conducted on the impact of these reforms van de Ven, and Zwijnenburg 2015; on savings. Research on the provident funds Holzmann, Palacios, and Zviniene 2004). In in Malaysia and Singapore generally shows a contrast, contributions to individual account positive effect on private savings, although the schemes are not counted as part of revenues results vary. Husain (1995) found that man- because they are considered the property of dated savings in the CPF in Singapore were their owners, the individual workers. Pension fully offset by a reduction in voluntary private savings in the form of individual accounts are savings. In contrast, Hopf (2009) found that considered private savings, and here too, the despite some offset through lower private vol- government may offset any positive effect by untary savings, Singapore still showed an borrowing more than it would otherwise. important positive net effect. Wickramanayake However, in this case, no hidden liability is (1998) also found that the CPF contributed to affecting the government’s long-term finances. national savings during the period 1970–94. The evidence suggests that mandated DB The International Monetary Fund found no schemes are likely to reduce savings. In the statistically significant effect of Malaysia’s case of unfunded or partially funded DB EPF on private saving, suggesting that the schemes, as individuals accumulate pension reduction in voluntary savings and withdraw- wealth, they may reduce their own private als offset the impact of the mandate (IMF savings for old age. Some evidence of this 1994). effect can be seen in economies of East Asia Prefunding can also affect capital markets. and Pacific, though mostly it occurs in richer Walker and Lefort (2002) outline the possible countries.37 Yamada, Yamada, and Liu (1992) channels for such an impact. Their empirical found a reduction of 40 percent in private sav- results for Latin America and the Caribbean ings in Japan for every increased unit of public suggest that the presence of these funds can pension wealth.38 Especially in DB schemes lead to “a reduction in the cost of capital; that ignore growth in unfunded liabilities dur- lower security price volatility; and higher ing the immature stage of the system, investing traded volumes” (Walker and Lefort 2002, 3). heavily in government bonds will tend to Catalán, Impavido, and Musalem (2000) note reduce public savings. When cash flows turn that the causality could run in the opposite PENSION SYSTEMS AND AGING: REFORMS FOR WINNING THE RACE 163 direction, namely from improvements in the FIGURE 5.17 Only higher-income economies have pension assets capital markets to the pension funds. They that are large relative to GDP found that in most wealthier countries, pen- Pension fund assets as a share of GDP sion funds led to greater stock market capital- ization and liquidity, although the results Singapore varied across countries. In Malaysia and Malaysia Singapore, little or no evidence demonstrated Korea, Rep. causality between institutions and markets, Japan Hong Kong SAR, China which the authors attribute to centralized Thailand (total) public management. In a recent study, Hu Philippines (total) (2012) used data from 10 Asian economies, Vietnam including China; Hong Kong SAR, China; China Korea; Malaysia; Singapore; and Thailand, to Indonesia explore the effects of the growth of pension 0 10 20 30 40 50 60 70 80 assets on financial markets. The study found a Public pension assets as share of GDP (%) significant impact on stock market capitaliza- tion and, for developing countries, on liquid- Sources: CPF Board 2013; Government of Korea 2012; Mesa-Lago, Viajar, and Castillo 2011; MPFA 2014; ity. 42 In addition, it found no impact on OECD 2013; SSO 2009; World Bank 2013. Note: Thailand includes Social Security Office and Government Pension Fund, plus the Social government bond markets but a positive Security System and the Government Service Insurance System. China refers to the National Social impact on corporate bond markets. The Security Fund only. effects are more pronounced in less developed economies, but they may also lack precondi- on both dimensions. In a Transparency and tions such as a good supervisory regime and a Governance Index constructed by Souto stable macroeconomic environment. and Musalem (2012),44 Korea, Malaysia, and Among partially funded DB schemes, 3 of Singapore are ranked in the top third of 83 the top 10 pension reserves in the world in public pension funds globally, with only absolute dollar terms are in East Asia, that is, Korea ranked among the top 20 economies. China, Japan, and Korea (OECD 2013). As a Thailand’s Government Pension Fund (GPF) share of GDP, Japan and Korea occupy the and China’s NSSF fall in the middle third, top and third spots globally at more than one- while Japan and the two pension funds in the quarter of GDP, but China ranks among the Philippines are in the bottom third. Although smallest, with only around 2 percent of GDP. not included in the survey, available informa- As shown in figure 5.17, only the higher- tion for Vietnam suggests that it would not income economies—Hong Kong SAR, China; score well, as little public information is Japan; Korea; Malaysia; and Singapore—have available and investment decisions are essen- pension assets that are large relative to GDP. tially in the hands of the minister of finance. Notably, the provident funds in Malaysia and A similar situation holds for the Social Singapore are mature, whereas DB schemes in Security Office (SSO) in Thailand, where Korea and Thailand and the DC scheme in information is scarce and the most recent Hong Kong SAR, China, are at an early stage annual report available on its website is for and will grow significantly in the next few 2009. In contrast, Hong Kong SAR, China, decades. By 2030, Korean reserves are pro- relies on private trusts to manage pension jected to reach 50 percent of GDP. The funds, and the role of the government is to reforms planned in Indonesia, Myanmar, and supervise them. China’s Mandatory Provident Timor-Leste are also likely to lead to signifi- Fund Authority (MPFA) is a well-staffed, cant growth in pension reserves.43 professional body that supervises 41 schemes The impacts of pension systems on capital offering more than 400 funds with a high markets and fund performance are dependent degree of transparency.45 on governance and investment policy, and The experience of the Philippine Social East Asia and Pacific has mixed performance Security System (SSS) is particularly 164 LIVE LONG AND PROSPER FIGURE 5.18 Pension agencies in East Asia and Pacific generally been seen in the provident funds of some have conservative investment practices Pacific Island countries, particularly with Portfolio composition of mandated pension funds in East Asia respect to investments in domestic real estate ventures.46 Hong Kong SAR, China Pension agencies in East Asia and Pacific generally have conservative investment Philippines (SSS) practices, with the exception of Hong Kong SAR, China, and, more recently, Malaysia. Philippines (total) Figure 5.18 shows that Hong Kong SAR, China, has by far the most investments in Philippines (GSIS) equity markets among these economies. Malaysia However, many public pension funds glob- ally have increased their holdings in private Korea, Rep. securities and other risky investments, par- ticularly during the past decade in Korea Thailand (GPF) and Malaysia. At the other end of the spec- trum, Vietnam’s and Thailand’s SSOs are Indonesia heavily concentrated in cash and govern- Japan ment- or state-owned enterprise bonds. In Vietnam, the portfolio consists completely Thailand (SSO) of government bonds, and returns have been negative in real terms in recent years Thailand (total) (IHME 2013). In the case of Thailand, the Singapore fact that the GPF is able to have a signifi- cantly more diversified portfolio than the Vietnam SSO suggests that the portfolio is the result of a policy choice. 0 10 20 30 40 50 60 70 80 90 100 The size of pension or provident funds Percent relative to domestic markets is a factor even Bank deposits/fixed income Equities Other in East Asian and Pacific economies with deep stock and bond markets. A clear case Sources: CPF 2013; Government of Korea 2012; Mesa-Lago, Viajar, and Castillo 2011; MPFA 2014; OECD has been made for diversifying pension and 2013; SSO 2009; World Bank 2013. provident fund investments internationally Note: GPF = Government Pension Fund; GSIS = Government Service Insurance System; SSO = Social Security Office; SSS = Social Security System. to obtain the best risk-adjusted returns. The accumulated balances in Malaysia and instructive for demonstrating the link between Singapore are equivalent to 40 to 45 percent governance and performance. Mesa-Lago, of the value of their respective stock markets. Viajar, and Castillo (2011) document how In Korea, the NPS already has shares worth political objectives have influenced the choice about 5 percent of the domestic stock mar- for board membership and cite examples in ket, owns 15 percent of domestic bonds which projects and personal loans were made traded, and is expected to grow rapidly for under pressure from senior politicians. The the next two decades (Palacios 2015). Earlier authors also cite politically motivated invest- studies have shown that larger pension funds ments in airlines, hotels, and commercial with long track records have decent perfor- banks that had negative repercussions in capi- mance relative to other public pension funds tal markets and led to losses for the fund. in developing countries, and this record has Between 2000 and 2010, the real rate of persisted and even improved in the case of return in the SSS and the Government Service Korea and Malaysia (see Palacios 2002 for a Insurance System (GSIS) was 3.3 and 5.1 cross-country comparison). Nevertheless, percent, respectively, even as GSIS returns these figures show room for improvement in were less volatile. Similar challenges have terms of diversification and its potentially PENSION SYSTEMS AND AGING: REFORMS FOR WINNING THE RACE 165 TABLE 5.6 Assessment of pension system performance and level of reform needed Performance Indirect impact Formal labor force 60+ labor force participation participation Savings and Economy Sustainability Coverage Adequacy rate rate capital market Cambodia Minor Major Major Minor Minor Minor China Significant Significant Major Major Major Major Hong Kong SAR, Minor Minor Significant Minor Minor Minor China Indonesia Minor Major Major Minor Minor Major Japan Minor Minor Minor Minor Minor Significant Korea, Rep. Significant Minor Significant Minor Significant Minor Lao PDR Minor Major Major Minor Minor Minor Malaysia Minor Minor Major Significant Major Minor Mongolia Major Significant Significant Major Major Major Myanmar Minor Major Major Minor Minor Minor Philippines Significant Major Major Minor Minor Major Singapore Minor Minor Major Significant Minor Minor Thailand Significant Significant Significant Minor Significant Major Timor-Leste Significant Minor Minor Minor Minor Minor Vietnam Major Major Major Major Major Major Source: Palacios 2015, based on data from economy pension system sources. positive effects on domestic capital markets, In summary: especially in Thailand and Vietnam. Having • Major reforms are needed in the matur- government-run pension funds become the ing pension schemes of three rapidly aging largest owners of bonds and shares has countries: China, Mongolia, and Vietnam. implications in terms of corporate gover- All three countries have large unfunded nance and potential conflicts of interest. pension liabilities and will start to run Over the past few years, Korea, Malaysia, deficits in the next two decades. Although and the GPF in Thailand have all consciously coverage in China has improved dramati- increased their foreign investment shares to cally with the expansion of informal sec- between 15 and 20 percent of total assets. tor pensions, the social pension element Hong Kong SAR, China, invests 40 percent to date remains very low and adequacy in global assets, mostly equities. China and remains a concern. Sustainability in the Japan also hold some foreign assets, but the Chinese system is more complex to assess. rest of the pension funds hold only domestic The post-1997 reforms have improved the securities. long-run sustainability of the urban work- ers’ scheme, but there remain significant Summary of performance assessment issues with legacy costs of the pre-reform Given the economic, demographic, and pen- system, empty or partially funded individ- sion system diversity across East Asia and ual accounts, low retirement ages relative Pacific, finding that national pension systems to life expectancy, low pension reserves, face significantly different challenges is not and rapid demographic aging (Jackson surprising. Table 5.6 summarizes by differenti- et al. 2013, Mercer and Australian Centre ating between economies that require substan- for Financial Studies 2015, IMF 2011). tial or moderate reforms in certain areas from Coverage in Mongolia is high among the those where reforms are less urgent or require elderly but is falling for the working-age only minor changes (the need for reform is population (although this may be offset major, significant, or minor, respectively). by expanding social pension coverage). 166 LIVE LONG AND PROSPER In Vietnam, low coverage is a major shaping their future pension model. Because problem in the context of rapid popu- of Thailand’s long retirement duration and lation aging. These countries also have low contribution rates, its main pension high contribution rates and incentives scheme is projected to run deficits in about to retire early. Prefunding is not likely to two decades. Through the massive expan- deliver benefits as designed in China and sion of the social pension, coverage has Mongolia and increasingly so in Vietnam, become close to universal, but benefit lev- where management of reserves needs to be els are modest and have a relatively modest improved. The Philippines also requires impact on poverty among the old (although significant reforms in these areas. a more significant impact on extreme pov- • The DC schemes in Hong Kong SAR, erty). In contrast, pensions of civil servants China; Malaysia; and Singapore mainly in Thailand are much higher than for pri- face problems related to adequacy, albeit vate sector workers. In terms of indirect for different reasons. Having started its effects, the private sector scheme provides mandate only recently, Hong Kong SAR, incentives to retire earlier, and the SSO’s China, is already demographically aged prefunding approach undermines poten- before its scheme has matured, and even tial saving and capital market impacts. The when it does, retirement income will be situation in Indonesia is similar in many low for the lifetime poor and the unlucky ways: the low contribution rate for the DC (in terms of rate of return). Malaysia and scheme ensures low benefits, and coverage Singapore have been the victims of their is very low. The major difference is the lack own success in that growth in wages has of any social pension, which makes both been a boon to lifetime incomes. However, coverage and adequacy urgent priorities. modest rates of return on individual The proposed new pension scheme, which accounts are resulting in low replace- would include a DB plan, fundamentally ment rates for those retiring recently or changes the picture but only for those who in the near future. This pattern is likely to will be covered. At the time of writing, change in the long run and could lead to Timor-Leste was also planning to introduce higher replacement rates, but it will not a national DB scheme. The current univer- prevent poverty for the lifetime poor. sal pension raises concerns over longer- • Japan and Korea have chosen to rely term sustainability. Cambodia, Lao PDR, on a combination of partial funding and Myanmar have very low coverage rates and modest benefit targets to ensure even for their income levels and have no sustainability. Japan appears somewhat social pensions or broad social assistance more sustainable after its shift to financing schemes to address elderly poverty. through its consumption tax (the goods and services tax) and other measures that will effectively prevent an increase in the Reforms to date and options pension bill as a share of GDP in the long looking forward run (see Kashiwase, Nozaki, and Tokuoka Although a number of East Asian and Pacific 2012).47 As in Hong Kong SAR, China, countries have undertaken major pension the immaturity of the scheme in Korea reform efforts since the East Asian financial means that cohorts retiring soon will have crisis of 1997, many challenges have still not low pension income. This transitional been addressed. This section first reviews phenomenon in Korea helps explain why pension reforms in East Asia and Pacific, it has among the highest elderly poverty some of which were touched upon earlier in rates in the OECD. the chapter, then outlines key reform recom- • The picture is mixed among countries mendations. These recommendations deal of Southeast Asian countries, with low- mainly with changes to existing parameters income countries facing major choices in and, in some cases, to the integration of the PENSION SYSTEMS AND AGING: REFORMS FOR WINNING THE RACE 167 separate systems for civil servants and private In both DB and DC systems, most reform sector workers. The last subsection looks at activity has taken place in the past 15 years, more fundamental measures to move away in the period following the 1997 East Asian from the traditional social insurance approach financial crisis. For example: that is starting to fail in some parts of the • Indonesia and Lao PDR passed major region even before countries have grown old. pension legislation affecting DB systems during this period. In Lao PDR, a man- date for private sector workers was intro- Reforms since 2000 duced in a 1999 law and has subsequently Over the past 15 years, pension reforms in been implemented. The DB scheme uses East Asia and Pacific have been diverse in a point system to determine benefits and terms of the relative focus on coverage, ade- has a number of well-considered param- quacy, and sustainability. Reforms tended ini- eters. The Indonesian law represented a tially to focus on sustainability and included fundamental departure from historical new approaches to funding as well as reduc- pension policy by introducing a DB com- tion in liabilities through benefit reductions. ponent with much higher benefit targets In contrast, efforts in the past 10 years have (along with significant modification of the been much more focused on expanding cover- DC element). However, implementation age and improving adequacy, at least in terms has been severely delayed, and 10 years of alleviating poverty in old age. Table 5.7 after the law was passed, key parameters lists the pension reforms that have taken place have only just been determined. The impli- in the region since 2000. cations for adequacy and sustainability, TABLE 5.7 Pension reforms in East Asia since 2000 Late 1990s–2005 2006–2010 2011–present • China. Introduction of new individual • China. Creation of rural pension • China. Massive expansion of rural pension program, account, creation of National Social program introduction of urban resident pension scheme Security Fund • Republic of Korea. Gradual • Indonesia. Passage of law on social security administrators, • Hong Kong SAR, China. Introduction of reduction (from 2009 to 2028) significant improvement in legal structure and governance, private defined contribution schemes in replacement rate of national issuing of new regulations (June 2015) • Indonesia. Passage of Social Security Law scheme, introduction of social • Japan. Expansion of welfare payments for low-income • Japan. In 2000—increase in retirement pension pensioners age, shift to price indexation, reduction • Singapore. Introduction of • Republic of Korea. Reduction in civil servant pension in benefits; in 2004—modification of “workfare income supplement” replacement rate indexing to automatically stabilize, for low-wage workers • Malaysia. Increase in contribution rate for employers increase in budget subsidy for basic • Thailand. First expansion of for lower-income workers, establishment of voluntary pension to cover one-half of benefit social pension individual defined contribution scheme with initial (as of 2014), gradual increase in • Timor-Leste. Introduction of matching contribution contribution social pension • Myanmar. Passage of Social Security Law • Republic of Korea. Implementation of • Vietnam. Introduction of • Philippines. Introduction of scheme for informal sector reforms passed in 1998 that include social pension workers in 2012 gradual increase in retirement age and • Singapore. Increase in retirement age and tightening of reduction in benefit levels withdrawal conditions; increase in contribution rate • Lao PDR. Implementation of 1999 law • Thailand. Second expansion of social pension to almost introducing mandate for private sector universal, establishment of National Saving Fund and Social • Mongolia. Implementation of 1999 law Security Office scheme for informal sector introducing notional accounts • Vietnam. Revision of social insurance law to introduce • Vietnam. Extension of mandate to private automatic indexation, reduce accrual rate, apply slightly sector workers (1995) higher penalty for early retirement, and narrow gap between public and private sector replacement Source: Palacios 2015. 168 LIVE LONG AND PROSPER as well as the indirect impact on labor and Service itself and the public debates that capital markets, remain to be assessed. took place. The result was an increase in the • China’s shift from a DC system to a hybrid retirement age and reduction in benefit lev- DB/DC system with individual accounts els, with further reduction in accrual rates implied a major policy change in terms in the 2007 reform. Very few countries have of prefunding (at least nominally) and the made such farsighted reforms at a stage nature of benefits. Mongolia’s law in 1999 when the scheme is immature and running also fundamentally changed how benefits large surpluses. would be calculated by moving to an NDC system. As discussed earlier, in neither case The period 2006–10 was notable for the have the reforms unfolded as planned. In spread of noncontributory or social pensions China the funds allocated to the new indi- in East Asia and Pacific (see Rofman, Apella, vidual accounts have not in practice been and Vezza 2014 on Latin America). This was segregated in most parts of the country a major element in the 2007 reform in Korea and are used to pay accrued benefits. In and resulted in a huge increase in coverage. As Mongolia, spending in 2012 reached close discussed earlier, China and Thailand also fol- to 4.9 percent of GDP, exceeding projec- lowed a policy of extending social pension tions made after the reform that spending coverage to the vast majority of those not eli- in 2012 would be around 3.2 percent of gible for contributory pensions. More recently, GDP. The introduction of the mandatory Japan significantly expanded social assistance DC scheme in Hong Kong SAR, China, in for pensioners. The most ambitious social 2001 was another important milestone for pension was introduced in Timor-Leste, which individual accounts in the region, which joined a substantial number of countries with contrasts sharply with the developments universal social pensions, including Fiji, in mainland China. Nowhere is the “one Mauritius, New Zealand, Samoa, and Tonga. country, two systems” approach more evi- The 2007 Workfare Income Supplement in dent than in pension policy, with Hong Singapore took a different approach, with the Kong SAR, China, amassing close to a government supplementing wages and retire- third of GDP in assets in just over a decade. ment savings for low-income workers, signifi- • The most comprehensive reforms to an cantly raising replacement rates for these existing system during this period took workers (see Choon and Tsui 2012). place in Japan. All major parameters were The most important pension reforms since adjusted: retirement ages were increased, 2011 have aimed at expanding coverage. indexation was changed, and benefits were Myanmar’s Social Security Law includes, for reduced. Most important among those were the first time, a mandated contribution by pri- (a) the introduction of an automatic stabili- vate sector workers and employers. Thailand zation mechanism that reduced future ben- introduced two schemes to bring informal efits to offset increases in dependency ratios sector workers into the pension system, both and (b) an increase in the consumption tax with significant subsidies to encourage take- that was earmarked to partly pay the basic up, and Timor-Leste introduced a civil service pension. Together, these measures appear to pension scheme. The most dramatic develop- have achieved long-term financial sustain- ment, however, has been the massive increase ability (Kashiwase, Nozaki, and Tokuoka in coverage in China in the past five years 2012). Reforms in Korea were passed in thanks to the reform that links voluntary con- the midst of the financial crisis, but only a tributions from informal sector workers to decade after the scheme started, the gov- provision of a social pension for both work- ernment acknowledged that the original ers and their parents. By the end of 2014, just parameters were not sustainable. What was under 500 million rural and urban informal impressive was the focus on long-term pro- workers contributed, and 140 million elderly jections produced by the National Pension were receiving basic pensions. PENSION SYSTEMS AND AGING: REFORMS FOR WINNING THE RACE 169 Recommended reforms FIGURE 5.19 Except in Japan, contribution rates required to pay for defined benefits are significantly higher than actual rates Parametric reforms Actual versus required contribution rates for public pension schemes Across East Asia and Pacific, ongoing reforms of existing contributory systems are needed to China make the systems sustainable and to create fis- Vietnam cal space for expanding pension system cover- Thailand age. Figure 5.19 shows that the contribution Philippines rate required to pay for the DB in each coun- Korea, Rep. try is significantly higher than the actual con- tribution rate, with the exception of Japan. Japan Because contribution rates have risen over 0 10 20 30 40 50 60 time in most of these countries, the gap was Share of wage (%) even larger during the early years of schemes. Actual contribution rate Breakeven contribution rate Payroll tax rates at these levels would encour- age evasion or unemployment, and most Source: OECD 2013. Note: The Japanese contribution rate refers to 2017. countries would either have to use other financing sources or reduce benefits. This was salaries are typically higher relative to lifetime the conclusion of Korea’s reformers in two salaries for high-income workers, giving them rounds of parametric reforms that ultimately a better rate of return than is provided to low- reduced the target benefit from 70 percent to income workers. Price indexation provides 40 percent over a roughly 30-year period. protection against inflation and eliminates the Some consensus exists over desirable para- arbitrary treatment of different cohorts that metric reforms for existing pension systems, can occur when adjustments are discretionary, which are particularly pressing priorities in as is current practice in several East Asian and the middle-income countries. The following Pacific pension systems. Almost all OECD seven parametric measures have all been countries have reformed their DB schemes in implemented in a range of OECD and Eastern this way, although reforms in civil service European and Central Asian countries over schemes typically lag those of national the past two decades, but much less widely in schemes.48 East Asia and Pacific: The last three measures are related to • Moving away from final salary and toward eligibility. Again, most OECD countries are lifetime, revalued average wages as the base using these kinds of parameters, at least in wage for calculating replacement rates their main national schemes. Lower retirement • Using linear accrual schedules ages for women increasingly is recognized as • Indexing pensions to changes in prices leading to lower pensions for them. It is also • Having minimal vesting requirements clear that life expectancy has increased signifi- • Linking retirement-age changes to life cantly since retirement ages were set decades expectancy ago and that people can work longer than they • Equalizing retirement ages for men and did earlier. The average retirement age in women OECD countries has been rising gradually • Providing actuarially fair (neutral) incen- over the past 15 years, and legislation already tives for early or late retirement passed will continue to increase normal retire- ment ages.49 Retirement ages of men and The first four measures relate to the benefit women also are converging (figure 5.20). One formula used to calculate the value of the of the perceived advantages of notional pension and how it is adjusted. Longer aver- defined contributions, the approach followed aging periods, linear accrual rates, and short in Italy, Latvia, Poland, and Sweden, is that vesting periods reduce distortions and favors the value of the pension is automatically toward certain workers. For example, final adjusted according to age of retirement and 170 LIVE LONG AND PROSPER takes life expectancy changes into account. East Asia and Pacific, but with the exception However, other mechanisms can achieve the of Japan, Korea, and Singapore, little prog- same result, as described in box 5.1. ress has been made. The arguments for East Asian and Pacific countries are at dif- increasing and equalizing retirement ages, ferent stages in implementing parametric linking them to future changes in life expec- reforms. These types of changes have been tancy, and rewarding voluntary later retire- thoroughly implemented in Japan and ment are most relevant for China, Mongolia, Korea, and Mongolia’s new NDC scheme Thailand, and Vietnam. Such reforms are addresses many of these parametric issues. In gradual, and in the OECD (including Japan all other countries, the seven parametric and Korea), they have been phased in over measures described above are relevant. years, even decades. It will also be an impor- Retirement age reforms have been recom- tant design choice for new schemes in mended for many years by many experts in Indonesia, Lao PDR, Myanmar, and Timor- Leste, where policy makers are in the midst of making decisions and should avoid the FIGURE 5.20 Retirement ages for men and women are converging temptation to legislate low retirement ages. in OECD countries Retirement ages, 1949–2029 Integration of civil service pension schemes In terms of labor market, equity, and fiscal 65 effects, deepening the integration of civil ser- Normal retirement age (years) vice pensions is also a priority in East Asia 64 and Pacific. Among the East Asian and Pacific 63 countries reviewed, seven have parallel civil 62 service schemes, and three have schemes for 61 only civil servants.50 A few countries, includ- 60 ing China and the Philippines, have multiple schemes for public sector workers. All these 59 1949 1969 1989 2009 2029 countries have scope for reforms that shift Men Women toward an integrated national pension plan. The arguments for such reforms include the Source: Chomik and Whitehouse 2010. potential for reducing administrative costs Note: OECD = Organisation for Economic Co-operation and Development. (especially in countries with low numbers BOX 5.1 Automatic adjustments for changes in life expectancy in defined benefit schemes In the past decade, at least 20 Organisation for Eco- France has linked the minimum contribution years nomic Co-operation and Development countries have required for retirement to life expectancy. introduced parametric reforms linking the scheme’s The adjustment in Japan as part of the 2004 parameters to changes in life expectancy. In Finland reform was quite different. A reduction of 0.9 percent and Portugal, benefits will be reduced by a factor that in the benefit at retirement is being applied each year is directly related to life expectancy. The link is less through 2023 based on the estimated increase in life direct in Germany, where benefit adjustment is based expectancy during this period. However, the reduc- on the ratio of pensioners to workers. Pensioners liv- tion is not adjusted if the estimate turns out to be ing longer translates into a lower value of initial pen- lower or higher than the actual change. sions. Italy and Greece will link their retirement age to life expectancy starting in 2015 and 2020, respectively. Source: Based on OECD 2011. PENSION SYSTEMS AND AGING: REFORMS FOR WINNING THE RACE 171 of contributors) and ensuring portability of FIGURE 5.21 Expanding coverage has become difficult in recent benefits between public and private sectors. years Another argument in favor of integration is Predicted labor force coverage by income level, 1994 versus 2014 based on equity grounds but also has implica- Share of labor force actively contributing tions for government spending when public 100 sector workers have substantially more gener- 90 to public pension plan (%) ous pensions than private sector peers. 80 Reforms have generally sought to reduce this 70 disparity by bringing public sector workers 60 into the main national scheme, which is usu- 50 40 ally more financially sustainable and can help 30 create fiscal space for initiatives designed to 20 reach informal sector workers. 10 Integration is more challenging where the 0 disparities between public and private pen- 6 7 8 9 10 11 sions are substantial, as in China and Thailand Income per capita (ln) currently. An intermediate step would be to Latest data Early 1990s data follow the example of Vietnam, where para- metric reforms to the civil service scheme will Sources: World Bank 1994; World Bank Pensions database, http://www.worldbank.org/pensions. Note: Data are from the early 1990s and most recent available. gradually result in equivalent benefit levels in the long run. A complementary approach that was proposed in China’s reforms announced Reforms to expand coverage in 2015 would be to integrate the main pen- The relationship between the coverage rate of sion scheme but allow for supplementary contribution-based pension schemes and the occupational schemes for civil servants, to income per capita of countries has remained avoid dramatic falls in generosity across strong for at least several decades, globally cohorts. Another approach would be to ensure and in East Asia and Pacific. With the excep- portability between public and private sector tion of the recent expansion of the pension schemes, as the Philippines has done. Whatever scheme for rural and informal urban sector the approach taken, civil service pension workers in China, no country has managed to reform should ideally take place in the context bring a significantly larger proportion of its of an overall compensation review, as was workers into a contributory scheme than done in Hong Kong SAR, China, when new would have been expected given its income civil servants were put into the DC system. level. Moreover, the data behind figure 5.21 Countries currently without national pen- suggest that doing so has become more diffi- sion schemes should avoid creating parallel cult over the past 20 years. schemes for the public and private sectors. Consistent with global experience, the Cambodia and Myanmar are considering increase in coverage of contributory pension establishing parallel schemes for private systems in East Asia and Pacific has been sector workers, but the number of workers very gradual and is highly unlikely to keep likely to contribute to each of these schemes is pace with population aging. Figure 5.22 less than 1 million. Lao PDR introduced a shows the percentage change in the share of new scheme for private sector workers in the labor force contributing to a pension 1999 that had fewer than 100,000 contribu- scheme from the early 1990s to around 2010 tors. A new civil service pension scheme is (the figure for China does not include the being implemented in Timor-Leste, but the new voluntary pension program). It shows government is considering a national scheme. that coverage has been stagnant in Indonesia In light of limited financial resources and and may have even fallen in Mongolia, and administrative capacity, countries’ creating the increases in other countries are less than parallel pension schemes is difficult to justify. one percentage point per year. The increase 172 LIVE LONG AND PROSPER FIGURE 5.22 The increase in coverage of contributory pension Since 1995, the government has provided systems in East Asia and Pacific is unlikely to keep pace with matching contributions to farmers and fisher- population aging man, increasing their participation signifi- Change in mandated contributory pension coverage from the early 1990s to cantly (Moon 2013). 51 The result of the around 2010 policy changes is shown in figure 5.23. From 1988 to 2012, labor force coverage rose by 60 about 50 percentage points. 50 Frustration with the pace of expanding % change in labor force coverage coverage has led to at least two types of gov- 40 ernment interventions in the region and glob- 30 ally, the first being the introduction or expansion of social pensions as discussed ear- 20 lier. Social pensions can help address elderly 10 poverty in the short run, but they present two major challenges: cost and incentives.52 As 0 described earlier, Timor-Leste has the most expensive social pensions in East Asia and –10 Pacific because benefits are relatively large lia sia s ia re a m p. ne in and eligibility is universal at age 60. In con- ys Re po na go ne Ch pi ala et ga a, on do ilip Vi re M trast, China; Hong Kong SAR, China; Korea; Sin M In Ph Ko and Thailand minimize costs by paying very Source: World Bank calculation in Palacios 2015. Note: China excludes rural and urban resident pension schemes. low benefits to a large share of the elderly, achieving high coverage but not ensuring adequacy. Other countries control costs by in Vietnam is partly related to the expansion providing transfers to a very small fraction of of coverage to private sector workers and the elderly (for example, the Philippines and therefore is a “one-off” event. Global experi- Vietnam) or by providing nothing (as in ence indicates that per capita income growth Indonesia and Myanmar). of 6 percent per year over the next 20 years In countries that already cover a large would result in income levels consistent with proportion of the elderly population with less than 40 percent coverage in Indonesia, social pensions, such as China and Thailand, the Philippines, and Vietnam. In other benefit levels could be increased at relatively words, to reach income levels consistent low cost. For example, doubling the social with full coverage would take roughly 50 pensions in Thailand would cost approxi- years, and even longer in the poorest East mately 0.4 percent of GDP per year. Lu, He, Asian and Pacific countries. With the excep- and Piggott (2014) calculate that raising the tion of Korea, the available evidence sug- rural social pension benefit in China to a level gests that without a fundamental change in of benefits equal to the poverty line would approach, population aging will have taken cost between 0.5 and 1.5 percent of GDP per place in most of the countries long before year over 40 years. Starting at a higher level to contributory pension schemes are able to address absolute poverty but indexing these provide broad-based old-age income pensions to prices would, over the long run, security. control the cost relative to budget resources in Korea’s dramatic coverage increase took a growing economy. China is already moving place thanks to legislative changes over the in this direction with the 27 percent increase past 20 years. Four years after it began, the in 2014 in the basic pension for informal NPS reduced the size of firms required to par- workers, to around US$11.30 per month. ticipate from 10 employees to 5 employees. A similar case could be made for expand- Seven years later, it reduced the size of firms ing coverage of social pension schemes in the to those with fewer than five workers and Philippines and Vietnam. However, these finally included all self-employed workers. countries also have social assistance programs PENSION SYSTEMS AND AGING: REFORMS FOR WINNING THE RACE 173 FIGURE 5.23 In the Republic of Korea, a dramatic change in coverage took place Number of workers covered by Korea’s National Pension Service, 1988–2012 25,000 20,000 20,329 Number of workers (thousands) 17,740 17,182 15,000 16,262 All businesses Firms with Firms with Urban fewer than 5 or more residents 5 workers workers 10,000 Firms with 10 or more workers 7,497 5,000 Match for 5,021 farmers and 4,433 fishermen 0 1988 1992 1995 1999 2003 2006 2012 Source: NPS 2012a, 2012b. that operate in parallel to the social pension. One risk of expanding social assistance or Another option may be to consolidate pro- social pensions in all but the poorest countries grams as the overall budget envelope is can be the increased incentive to stay in the increased and to use a common targeting informal sector.53 In Thailand, for example, approach and delivery platform. the fact that a person has any pension from a In lower-income countries such as formal sector scheme negates eligibility for Cambodia or Myanmar, which have neither the social pension. If the value of the social social pensions nor significant social assis- pension were to rise significantly, so would tance programs, the best approach may be to the effective tax on contributions to the pub- expand targeted social assistance programs. lic pension scheme. A better approach where Survey data suggest that most poor elderly capacity exists would be to withdraw the reside with families. Thus, a well-targeted transfer gradually at higher income levels. transfer will tend to reach the elderly poor, This was done in Chile as part of reforms in achieving the same objective as a social pen- 2008 that expanded the social pension to sion while addressing broader antipoverty cover the bottom 60 percent of the distribu- objectives. In such cases, a social pension is tion (see box 5.2). It reduced incentive prob- unlikely to reduce poverty more than a pro- lems significantly relative to imposing an gram that determines eligibility based on effective 100 percent marginal tax on formal poverty rather than age. Notably, transfers wages and pensions. made to the bottom segments of the income Another approach to expanding coverage is distribution are unlikely to reduce formal to subsidize contributions of low-income and sector activity in countries where contribu- informal sector workers. Some international tory pension scheme coverage does not evidence suggests that matching contributions exceed 10 percent of the labor force. These can increase coverage at the margin (see Hinz workers simply do not have the choice to et al. 2013). In Korea, for example, matching move in and out of the formal sector. contributions for fishermen and farmers were 174 LIVE LONG AND PROSPER BOX 5.2 Integrating noncontributory and contributory pensions: Chile’s solidarity pillar Chile introduced a noncontributory (“solidarity”) pension. As a result, low- to middle-income individ- pension in 2008 that would complement the existing uals fell into a “coverage gap.” contributory pension system. In 1981, it had replaced Chilean policy makers faced a trade-off between its traditional defined benefit pay-as-you-go pen- providing income protection and reducing incentives sion system with a defined contribution system with to save. The new solidarity pillar had to be designed individual accounts managed by specialized private to be compatible with incentives to contribute to indi- firms. Two types of noncontributory benefit existed: vidual accounts. Since neither the existing social assis- (a) a social assistance pension targeted to the low- tance pension nor the minimum pension guarantee est income quintile of the population, with eligibility provided adequate income support to alleviate pov- precluded if a person had any other source of pen- erty in old age, policy makers decided that the new pil- sion, and (b) a minimum pension guarantee for pen- lar would replace both programs. To maintain some sion fund contributors who had attained only low incentives to contribute, the benefit was designed not balances after at least 20 years of contributions. as an absolute minimum floor guarantee but as a min- By 2006, it was clear that although the system imum pension for individuals with no contributions had been very successful in obtaining high real rates plus a top-up for individuals with contributions. The of returns for the pension funds, pension levels amount of the top-up would be reduced depending on would not be adequate for a large portion of the the level of the contributory pension. The reduction population. Most members of the system did not was designed so that the total pension (the sum of the contribute frequently enough to build a large bal- contributory and noncontributory pensions) would ance. Although the proportion of the labor force always be increasing along with the balance accumu- that worked in the formal sector was relatively high lated in the individual account. at around 60 percent, only a minority held steady Operationally, the new solidarity pillar was com- jobs in the formal sector. Berstein et al. (2006) esti- posed of two benefits: (a) the basic solidarity pension mated that around 50 percent of members would for individuals with no contributions and (b) a soli- receive a pension lower than the minimum pension, darity pension supplement, which is the top-up for and many would not reach the 20 years of contri- individuals with some contributory pension. In other butions required for the pension guarantee. At the words, the new solidarity pillar was designed as a same time, many of these individuals would not minimum pension benefit with a clawback, as shown be poor enough to qualify for the social assistance in figure B5.2.1. FIGURE B5.2.1 Design of the Chilean pension system Total pension Basic solidarity Solidarity pension pension supplement Self-financed pension PENSION SYSTEMS AND AGING: REFORMS FOR WINNING THE RACE 175 found to have increased participation by pensions than might otherwise have been around 10 percentage points. Under the paid. In China more than in most countries, Malaysia Retirement Savings Scheme, estab- strong arguments can be made for a larger lished as an incentive scheme for the self- social pension, given the huge disparities in employed and for individuals without fixed pension levels across different schemes as well monthly income they could save for retire- as the economic effects of excess savings.55 ment, the government provides a matching In the poorest countries in the region, contribution of 10 percent of the contribution, matching contributions are unlikely to be the subject to a maximum. However, only just best use of scarce fiscal resources. In these over 73,500 individuals have participated countries, the trade-off is more acute when it since the scheme was introduced in 2010, con- involves spending more on social assistance sistent with the modesty of the match (see or social pensions or diverting funds into indi- Dorfman 2014; Samad and Mansor 2013). vidual accounts. This trade-off is particularly Thailand has introduced two pension pro- difficult if subsidies cannot be targeted effec- grams for informal workers, which are admin- tively, since higher-income informal sector istered by separate ministries (see World Bank workers will benefit most. Furthermore, the 2012b). One is an extension of the DB scheme lack of a financial and institutional infrastruc- run by Thailand’s SSO, and the other is a DC ture that can channel savings is more prob- scheme that provides matching contributions lematic in countries like Cambodia and ranging from 50 percent below age 30, to 80 Myanmar than in China. percent between ages 30 and 50, to 100 Matching contributions may work better percent at age 50 and above, with a required in countries where extensive, well-targeted contribution of less than US$2 per month. programs already exist and where the infra- Almost 1.4 million people had signed up for structure for a DC scheme is already function- the SSO scheme as of March 2013. ing. The conditional cash transfers program China’s experience in combining a social in the Philippines provides a database that pension with a matching contribution to could be used to target subsidized contribu- workers illustrates both the potential and lim- tions, and it has a sophisticated system for itations of matching contributions to expand delivering transfers that could be exploited to coverage. A flat amount, initially around collect contributions. However, the country US$9 per month (US$11.30 per month from lacks the infrastructure for administering and mid-2014), is paid as a basic pension to those regulating other elements of a DC scheme. ages 60 and above living in rural areas and to Legislation in Vietnam envisions a DC scheme informal urban workers with a local residence (with a match) for informal sector workers, permit (hukou). Eligibility for those already but it has not yet been implemented. The out- over 60 is contingent on having working-age comes will depend on the incentives provided spouses or children who contribute to their as well as on the administrative infrastructure own pension accounts (a process referred to available to collect, track, and invest these as family binding, though implementation funds. demands appear to have resulted in weak enforcement of this requirement) (Dorfman Options for changing the paradigm et al. 2013). The massive enrollment that has The attempt to address the coverage gap in taken place demonstrates the feasibility of col- East Asia and Pacific is part of a global pat- lecting small amounts from informal sector tern occurring over the past decade or so for workers; however, without better returns and policies that move away from a sole reliance more direct incentives (for example, a higher on payroll taxes for financing pensions. As match), many will remain outside the system discussed in the previous section, financing for much of their working lives, and benefits the expansion of social pensions and match- will be inadequate.54 At the same time, subsi- ing contributions for informal sector workers dizing individual accounts results in an oppor- from the budget is a response to the persistent tunity cost, notably in the form of lower social failure of contributory pension schemes to 176 LIVE LONG AND PROSPER close the coverage gap. Meanwhile, the use of ties or phased withdrawals based on individ- consumption taxes in Japan and budget sup- ual accounts. port for the contributory pension scheme in This approach could be tailored to the ini- China is partly driven by reluctance to raise tial conditions of each country. In a country payroll tax rates further. The trend is not uni- with low mandated contributions and benefit form, however, as evidenced by the introduc- targets (such as Thailand), individual accounts tion of a social insurance scheme in Lao PDR could supplement the contributory pension. and by legislation in Indonesia and Myanmar. Because the individual account–based pension Although these countries are among the last would be the same for all workers, the effect to have followed the social insurance–based on replacement rates would be greater for path to provide pensions, the experience of low-income workers. In a country with higher dozens of countries that once faced the same benefit targets and required contribution rates choice suggests that it may not be the appro- (such as Vietnam), the latter could be reduced priate path. proportionately as the new universal individ- A way forward in pension reform can ual account fills the gap, achieving the same instead be based on the premise that the replacement rates but with less reliance on redistributive element of the pension system payroll taxes. Similarly, in countries that oper- should be financed from general revenues ate DC schemes, this approach would simply and clearly distinguished from the insurance involve a reduction in employees’ contribu- or savings component (World Bank 1994). tion by the flat amount contributed by gov- The latter would ideally be financed by indi- ernment. This method would increase viduals on an actuarially fair basis. The sim- take-home pay and reduce the tax wedge plest approach would be to eschew the more for low-income workers by acting as a mandated scheme altogether and instead progressive wage supplement and encourag- have a social pension with broad coverage ing workers with low wages into the formal (for example, universal or affluence-tested sector, as done in Singapore. In the high-cov- coverage), complemented by a voluntary erage DB schemes in Japan and Korea, financ- retirement savings scheme. This approach is ing the basic pension with other taxes would the situation in New Zealand, which also achieve the same objective without the need features an opt-in default for a voluntary DC for individual accounts. Japan’s earmarking of scheme that has led to relatively high partici- consumption taxes to finance the basic pen- pation rates. However, this option is not sion is already a step in this direction. likely to be accepted in countries that are In some East Asian and Pacific economies, about to introduce mandates, and it may fiscal constraints or governance challenges face resistance in countries with existing may make prefunding difficult or impractical. mandates. A government contribution to a funded Another option, first proposed in the con- scheme for all adult citizens would eventually text of Mexico, is for the government to con- require an allocation of several percentage tribute a minimum amount to individual points of GDP. The resources would have to be accounts that would be opened for every identified and effectively earmarked. In adult citizen. The government’s contribution resource-rich economies where sovereign funds would be calibrated to produce an accumula- exist (such as Mongolia and Timor-Leste), tion of funds sufficient to pay an annuity financing could involve earmarking of reve- equivalent to the social pension, phased in nues for this purpose. Regardless of the source, gradually.56 Social pensions or broad social allocating scarce resources to this purpose assistance could provide for the current involves significant fiscal trade-offs. However, elderly and transition cohorts until the indi- good equity and efficiency arguments can be vidual account system matures. The result made for shifting from payroll taxes to other would be a gradual transition from an revenue sources, especially where the tax unfunded social pension to minimum annui- wedge is large and where the relative taxation PENSION SYSTEMS AND AGING: REFORMS FOR WINNING THE RACE 177 of factors of production is currently overly reli- but the same approach is evident in Korea and ant on labor (as in China and Vietnam). Thailand. The new scheme in China that links Important challenges related to governance social pensions and a government contribution and potential investments would also have to to individual accounts is the most important be addressed. If centrally managed, address- example of what may represent a paradigm ing those challenges would result in the accu- shift in pension provision, not only in East mulation of a fund that was very large relative Asia but also globally. to the economy. Many economies have lim- The reform agenda for the low- and ited capital markets and may find investing middle-income countries of East Asia includes abroad difficult. It is also difficult to avoid improving existing pension systems, such as conflicts of interest, ensuring that investments parametric changes and harmonization or are not motivated by reasons other than good integration of parallel schemes, as well as pension outcomes. The international experi- expanding coverage of the informal sector. ence with the management of large public The design of pension schemes and especially pension funds does not provide many positive their retirement incentives can significantly experiences in this regard (Iglesias and affect labor force participation rates and Palacios 2001). If the individual accounts should be reformed to encourage longer were managed by competing private manag- working lives as healthy life expectancy rises. ers, as in Hong Kong SAR, China, the govern- However, historical experience suggests that ment’s role as regulator and supervisor would these reforms can provide only part of the be crucial, and few economies in the region solution and that a comprehensive plan will have the required capacity. include programs that can reach the majority Governments have two alternatives to pre- of people who operate in the informal sector. funding. The first is the NDC approach, in This observation is especially true in those which accounts are credited with contribu- countries where the race between coverage tions but have no segregated assets.57 The sec- and aging is already well under way. ond would be for the government to simply issue government bonds that would be depos- Notes ited into individual accounts. These would be special, nontradable bonds and could be 1. However, the civil service pension schemes linked to growth of GDP or income per cap- of Indonesia and Malaysia have significant ita. The impact of the latter approach on unfunded liabilities. 2. Zuo (2013) calls these “empty individual national savings depends on the overall fiscal accounts” and estimates them at 90 percent of adjustment, which could range from complete recorded contributions. tax financing to complete debt financing. 3. These accounts mimic DC schemes but are not funded. Individual accounts are Conclusion credited with contributions that then earn a notional interest rate that is typically linked A paradigm shift is needed (and already under to the growth of wages or the wage bill. See way) in how pension systems in East Asia and Holzmann and Palmer (2006) for a review of Pacific are designed and financed. As discussed international experience. throughout this chapter, a growing number of 4. Amendments in 2014 in Vietnam introduce countries, frustrated with the traditional model automatic price indexation. 5. The Mandatory Provident Fund Authority of contributory pensions that dominated the is proactive in terms of financial literacy and last century, are finding ways to plug the cov- making improvements to the system, for erage and adequacy gaps with programs example, introducing life-cycle investment financed by general revenues. These programs options. break the link between formal employment 6. Conceptually, the unfunded liability is the and pension (and health insurance) coverage. present value of future pension obligations. Timor-Leste is the most dramatic example, The financing gap is the difference between 178 LIVE LONG AND PROSPER the present value of contributions and pension cites March 2014 figures of about 70,000 payments. Unfortunately, international registered members (about 0.5 percent of the standards for reporting these key measures of labor force) in 44 private funds. sustainability for public pension schemes do 18. A World Bank pension database covering not exist. 78 low-income countries revealed about 7. China has had a distinct scheme for workers 100 million contributors from a labor force in public sector units, although the State of almost 900 million (about 11 percent). Council in early 2015 announced its intention 19. Because of the way public DB pensions are to merge civil service, public sector units, and treated in the fiscal accounts, this challenge urban worker schemes. The system relies on of sustainability is true even during the period occupational supplements to avoid a dramatic when reserves are being drawn upon, at least fall in replacement rates for civil servants. to the extent that investments consist of 8. Although Vietnam has one pension scheme government bonds. When a public DB pension for all workers, a case is made for including scheme invests in government debt, this is seen it in this first group because of the practice as one part of government lending to another of calculating benefits of the two groups and is netted out of general government debt. of workers using a different methodology, To the extent that contribution revenues though 2014 reforms seek to reduce the exceed benefit spending, this reduces the disparity over time. budget deficit in that time period. Conversely, 9. Although civil servants participate in the when contributions are lower than spending, same scheme as private sector workers, the general government deficits increase even government contributes more than private where earmarked bonds are redeemed to pay sector employers are required to contribute. pension benefits (see Goebel, van de Ven, and 10. In Japan, the systems for civil servants and Zwijnenburg 2015). private school employees are to be merged 20. Mongolia also allowed workers to “buy with the system for private employees on back” contribution years during the 1990s. October 1, 2015. 21. Although pension coverage is almost universal 11. As of April 2015, a new social security law in Japan, average effective retirement ages are that would cover all formal sector workers higher than age 60. was still being discussed. 22. Data from 2011 for China are from CHARLS 12. Note that the median age of civil servants (2011). tends to be higher than private sector workers. 23. This approach has been used in OECD and 13. The Philippines is an exception in this regard, Latin America and the Caribbean. They use having passed a portability law in 1994 that prospective simulations for workers entering makes it possible to combine service periods the scheme today, not what is actually paid (Mesa-Lago, Viajar, and Castillo 2011). out today. The aim is to compare the implied 14. Mesa-Lago, Viajar, and Castillo (2011) also objectives of different designs. advocate harmonization and integration of 24. The difference between gross and net multiple schemes in the Philippines. replacement rates is that the net replacement 15. In addition, Korea’s mandated retirement rate takes taxation into account. A standardized benefit system provides a lump sum payment set of assumptions was used for wage growth, equivalent to 1/12th of annual salary per interest, and mortality rates. The figures are for year of employment. However, this can men entering the system today with average be withdrawn by the worker prior to wages throughout their careers and retiring retirement and is therefore not defined here at normal retirement age. For DC schemes, as part of the pension system. See the Korean simulations convert account balances into law at http://www.moleg.go.kr/english annuities based on country- and sex-specific /korLawEng?pstSeq=47471. mortality rates. 16. See the Thai Securities and Exchange 25. Because of the different life expectancies at Commission’s website at http://capital.sec. retirement age across East Asian and Pacific or.th/webapp/thaipvd/apps/pvdreports2 countries, the present value of the target /stat12en.php. benefit, referred to as pension wealth, can 17. See those programs at Malaysia’s country be calculated. This method would change profile at Wilmington PLC, http://www the rankings only slightly, with China .pensionfundsonline.co.uk/content /country- becoming the most generous, followed by the profiles/malaysia/102. Dorfman (2014) Philippines. Pension wealth for women would PENSION SYSTEMS AND AGING: REFORMS FOR WINNING THE RACE 179 also generally be higher in DB countries compensation to late retirees than do all but a because women live longer than men and may few OECD countries. Also see Giles, Hu, and have lower retirement ages. Huang (2015) regarding East Asia and Pacific. 26. In Indonesia and Mongolia, reforms are being 35. Bloom, Canning, and Sevilla (2004) also find implemented that represent a major departure that the expectation that longevity gains will from current rules. In Indonesia, new lead to longer working lives is not borne out parameters will certainly increase the overall by evidence. replacement rate target reflected in figure 5.7. 36. The analysis relies on national-level data, so In Mongolia, the replacement rates shown it is not possible to separate LFPRs for urban here are much lower than those for workers and rural elderly for this sample. Controlling born before 1960, with the change from one for the rural share in each country increases cohort to the next being so dramatic that it R-squared significantly (0.66) and reduces the has been cited as a potential problem. coefficient of the spending variable. 27. The guide to diversifying retirement income 37. There are, for example, various studies from states, “Your EPF savings alone are not likely Italy, the United Kingdom, and the United to be enough to support your retirement.” See States. The only study known to the authors KWSP (2012, 5). of this report that includes East Asia and 28. The conditions for withdrawal are five years of Pacific is Shome and Saito (1980). There is contribution and one month of unemployment. also evidence of crowding out when voluntary 29. The calculation of replacement rates in savings for retirement are provided tax Singapore has added complexity owing to a incentives. See Kohl and O’Brien (1998) for a second account that can be used for housing review of international studies. or retirement and a third account that is 38. The authors also estimated a positive savings used for medical services. Choon and Tsui effect for the period 1960–80 resulting from (2012) show that replacement rates would be additional savings for longer retirement substantially higher if housing were properly duration. The net effect was 25 to 30 percent taken into account. displacement of savings. 30. As noted earlier, countries without social 39. See Onder and Pestieau (2014) for theoretical pension programs include Cambodia, evidence. Indonesia, Lao PDR, and Myanmar. Aside 40. The four countries are Chile (Coronado 1997; from the subsidized food program in Fuentes 2013; Schmidt-Hebbel and Servén Indonesia, national social assistance programs 1998); Colombia (Villar et al. 2013); Mexico are almost nonexistent in these countries. (Villagómez and Antón 2013); and Peru 31. Indonesia provides free rice to more than (Secada, Cusato, and Zapata 2013). half of households (although the new 41. In Thailand, the Government Pension Fund administration has announced that the scheme was introduced in 1997 to partially replace will be rationalized) and has a conditional cash the DB scheme for civil servants. In the same transfer program for around 3 percent of the year, China introduced individual accounts population. It also subsidizes health insurance that were nominally funded but soon became for poor households. In the Philippines, more “empty” accounts. than 20 percent of households benefit from 42. A concentration in government bonds, a conditional cash transfer program, and especially nontradable bonds, can inhibit a larger proportion is covered by a targeted potential effects of pension funds on capital health insurance scheme. markets and savings. Both Malaysia and 32. These crude comparisons hide significant Singapore invest their provident funds in differences in the scope of the wage bill to special government bonds. In Singapore, which they are applied, including the low CPF purchases of nonmarketable bonds were earnings ceilings in Hong Kong SAR, China, issued specifically to the CPF. In contrast, and in Thailand. For the most part, however, Malaysia’s EPF holds state and corporate the parameters do not affect the relative bonds, equities, and property, with only about position of these economies. one-quarter in government bonds. 33. Japan’s current total contribution rate is 43. Preliminary projections based on the 16.7 percent but is scheduled to rise to 18.3 parameters being considered in Indonesia percent by 2017. suggest that reserves could exceed 10 percent 34. D’Addio, Keese, and Whitehouse (2013) of GDP in 2030 and 25 percent of GDP by find that Japan and Korea give higher 2050. 180 LIVE LONG AND PROSPER 44. The index focuses on the availability of Furthermore, although rules for portability reliable information required for independent have been issued, the highly decentralized monitoring as well as composition, system has no mechanism for ensuring this independence, and qualifications of Board portability, which poses a major challenge in members. the face of mass rural-to-urban migration. 45. Accountability is further promoted by 55. For example, Baldacci et al. (2010) estimated competition between funds. The MPFA is that an increase in government spending on especially concerned with the poor choice of pensions of 1 percent of GDP would raise investment options and the lack of awareness permanent income by 1.25 percent of GDP. about the impact of administrative fees, so it 56. A version of the proposal is still under has embarked on an ambitious educational consideration in Mexico. campaign in the past few years. 57. See Cai et al. (2012) for a discussion of this 46. For a recent example in Fiji, see Hofman and approach in the context of China’s rural Chand (2012). pension scheme. 47. A key measure is the automatic adjustment mechanism, which reduces pensions depending on deviation from baseline References projections. Asher, Mukul G. 2012. “Singapore.” In Pension 48. In addition, disability pension reform has also Systems in East and Southeast Asia: been important in some countries, such as Promoting Fairness and Sustainability, edited Poland. by Donghyun Park, 85–98. Manila: Asian 49. As discussed in chapter 4, resistance to Development Bank. increasing the retirement age often stems Asher, Mukul G., and Azad Singh Bali. 2013. from concern about the effect on employment “Fairness and Sustainability of Pension of younger workers (known as the “lump Arrangements in Singapore: An Assessment.” of labor fallacy”), but regional and global Research Paper 13-21, Lee Kuan Yew School evidence suggests that such concern is of Public Policy, Singapore. misplaced. See, for example, Zhang (2012) Asher, Mukul G., and Amarendu Nandy. 2011. for China. “Singapore: Pension System Overview and 50. Indonesia plans to integrate civil servants Reform Directions.” In Pension Systems into a hybrid DB/DC pension scheme. and Old-Age Income Support in East and However, the process may begin only with Southeast Asia, edited by Donghyun Park, civil servants hired after 2028. Observers see 152–75. Oxon, U.K.: Routledge. little justification for such a long transition Aterido, Reyes, Mary Hallward-Driemeier, and period. Carmen Pagés. 2011. “Does Expanding 51. Despite this impressive achievement, even Health Insurance beyond Formal-Sector after the coverage expansion, close to one- Workers Encourage Informality? Measuring third of the self-employed in Korea, mostly the Impact of Mexico’s Seguro Popular.” Policy low-income workers, do not regularly Research Working Paper 5785, World Bank, contribute to the pension system. Washington, DC. 52. See Barrientos (2012) for a thorough Baldacci, Emmanuele, Giovanni Callegari, David discussion. Coady, Ding Ding, Manmohan Kumar, 53. Aterido, Hallward-Driemeier, and Pagés Pietro Tommasino, and Jaejoon Woo. 2010. (2011) document similar effects on informal “Public Expenditures on Social Programs activity for noncontributory health insurance and Household Consumption in China.” coverage in Mexico. Davis and Henrekson IMF Working Paper 10/69, International (2004) present similar data from richer Monetary Fund, Washington, DC. http:// countries. www.imf.org / external/pubs/ft/wp/2010 54. The scheme suffers from some of the same /wp1069.pdf. implementation problems as individual Barrel, Ray, Ian Hurst, and Simon Kirby. 2009. accounts for formal sector workers. “How to Pay for the Crisis or Macroeconomic Specifically, rates of return are set at bank Implications of Pension Reform.” National deposit rates, resulting in poor incentive to Institute of Economic and Social Research, contribute and low balances at retirement. London. PENSION SYSTEMS AND AGING: REFORMS FOR WINNING THE RACE 181 Barrientos, Armando. 2012. “What Is the Role of DC. http://www.federalreserve.gov/pubs Social Pensions in Asia?” ADBI Working Paper /feds/1998/199812/199812pap.pdf. 351, Asian Development Bank Institute, Tokyo. CPF (Central Provident Fund) Board. 2013. Berstein, Solange, Guillermo Larrain, Francisco “Review of Operations.” CPF, Singapore. Pino, and Eduardo Morón. 2006. “Chilean Cruces, Guillermo, Sebastian Galiani, and Pension Reform: Coverage Facts and Policy Susana Kidyba. 2010. “Payroll Taxes, Wages, Alternatives.” Journal of the Latin American and Employment: Identification through and Caribbean Economic Association 6 (2): Policy Changes.” Labor Economics 17 (4): 227–79. 743–49. Betcherman, Gordon, and Carmen Pagés. 2007. D’Addio, Anna Cristina, Mark Keese, and Edward “Estimating the Impact of Labor Taxes on Whitehouse. 2013. “Population Ageing and Employment and the Balances of the Social Labour Markets.” Organisation for Economic Insurance Funds in Turkey.” Washington, DC: Co-operation and Development, Paris. World Bank. Davis, Steven J., and Magnus Henrekson. 2004. Bloom, David, David Canning, and Jaypee Sevilla. “Tax Effects on Work Activity, Industry Mix, 2004. “The Effect of Health on Economic and Shadow Economy Size: Evidence from Growth: A Production Function Approach.” Rich-Country Comparisons.” NBER Working World Development 32 (1): 1–13. Paper 10509, National Bureau of Economic Cai, Fang, John Giles, Philip O’Keefe, and Dewen Research, Cambridge, MA. Wang. 2012. The Elderly and Old Age Support Dorfman, Mark. 2014. “Malaysia: Savings, Social in Rural China. Washington, DC: World Bank. Insurance, and Elderly Protection Note.” Casey, Bernard H. 2004. “The Japanese World Bank, Washington, DC. Retirement Income System: A Special Case?” Dorfman, Mark, Robert Holzmann, Philip Discussion Paper PI-0407, Pensions Institute, O’Keefe, Dewen Wang, Yvonne Sin, and Cass Business School, City University, London. Richard Hinz. 2013. China’s Pension System: Catalán, Mario, Gregorio Impavido, and Alberto A Vision . Washington, DC: World Bank. R. Musalem. 2000. “Contractual Savings or doi:10.1596/978-0-8213-9540-0. Stock Market Development: Which Leads?” Duval, Romain. 2003. “The Retirement Effects Policy Research Working Paper 2421, World of Old-Age Pension and Early Retirement Bank, Washington, DC. Schemes in OECD Countries.” OECD CHARLS (China Health and Retirement Economics Department Working Paper 370, Longitudinal Study). 2011. http://charls.ccer Organisation for Economic Co-operation .edu.cn/en. and Development, Paris. http://dx.doi China Statistical Yearbook Database. Various .org/10.1787/308728704511. years. http://tongji.cnki.net/overseas/brief Fuentes, J. Rodrigo. 2013. “Evidencia para /result.aspx. Chile.” In Contribución del sistema privado Chomik, Rafal, and Edward R. Whitehouse. de pensiones al desarrollo económico de 2010. “Trends in Pension Eligibility Ages and Latinoamérica , edited by Rodrigo Acuña, Life Expectancy, 1950–2050.” OECD Social, 181–238. Medellín, Colombia: SURA Asset Employment, and Migration Working Paper Management. 105, Organisation for Economic Co-operation Giles, John, Yuqing Hu, and Yang Huang. 2015. and Development, Paris. “Understanding the Retirement Decision in Choon, Chia Ngee, and Albert Tsui. 2012. Aging East Asia.” Background paper for the “Adequacy of Singapore’s Central Provident East Asia and Pacific regional report on aging, Fund Payouts: Income Replacement Rates of World Bank, Washington, DC. Entrant Workers.” Department of Economics, Goebel, Paul, Peter van de Ven, and Jorrit National University of Singapore, Singapore. Zwijnenburg. 2015. “Long-Term Liabilities Coronado, Julia Lynn. 1997. “The Effects of for Social Security and Employment Related Social Security Privatization on Household Pension.” Prepared for the meeting of the Saving: Evidence from the Chilean International Monetary Fund Government Experience.” Finance and Economics Finance Statistics Advisory Committee, Discussion Paper 1998-12, Division of Washington, DC, March 9–10. https:// Research and Statistics and Monetary www.imf.org/external/pubs/ft/gfs/gfsac Affairs, Federal Reserve Board, Washington, /meetings/2015/pdf/1514a.pdf. 182 LIVE LONG AND PROSPER Government of Korea. 2012. “Fund Management Kenneth Bercuson and Robert Carling, 42–51. Report.” National Pension Service, Seoul. Washington, DC: International Monetary Fund. Gruber, Jonathan. 1997. “The Incidence of Payroll IFLS (Indonesia Family Life Survey). 2007. RAND Taxation: Evidence from Chile.” Journal of Corporation, Santa Monica, CA. http://www Labor Economics 15 (3, pt. 2): S72–101. .rand.org/labor/FLS/IFLS.html. Gruber, Jonathan, and David A. Wise. 1999. Iglesias, Augusto, and Robert Palacios. 2001. “Social Security Programs and Retirement “Managing Public Pension Reserves: Evidence around the World: Micro-estimation.” NBER from the International Experience.” In New Working Paper 9407, National Bureau of Ideas about Old-Age Security: Toward Economic Research, Cambridge, MA. Sustainable Pension Systems in the 21st Century, Heckman, James, and Carmen Pagés, eds. edited by Robert Holzmann and Joseph Stiglitz, 2004. Law and Employment: Lessons from 213–53. Washington, DC: World Bank. Latin America and the Caribbean. Chicago: IHME (Institute for Health Metrics and University of Chicago Press. Evaluation). 2013. The Global Burden of Herd, Richard, Hu-Wei Hu, and Vincent Koen. Disease: Generating Evidence, Guiding Policy. 2010. “Providing Greater Old-Age Security Seattle, WA: IHME. in China.” OECD Economics Department ILO FACTS (International Labour Organization Working Paper 750, Organisation for Economic Financial and Actuarial Services). 2013. Co-operation and Development, Paris. http:// “Actuarial Valuation of the Public Pension dx.doi.org/10.1787/5kmlh4x7pc7k-en. Scheme of the Vietnam Social Security Fund.” Hinz, Richard, Robert Holzmann, David Tuesta, Social Protection Department, International and Noriyuki Takayama, eds. 2013. Matching Labour Organization, Geneva. Defined Contributions for Pensions: A Review IMF (International Monetary Fund). 1994. of International Experience. Washington, DC: “Malaysia: Recent Economic Developments.” World Bank. IMF, Washington, DC. Hofman, Helene, and Satish Chand. 2012. “Fiji ———. 2011. “The Challenge of Public Pension National Provident Fund Writes Off Millions Reform in Advanced and Emerging in Member’s Funds.” Radio Australia , Economies.” Fiscal Affairs Department, IMF, February 15. http://www.radioaustralia.net Washington, DC. .au/international/radio/onairhighlights Jackson, Richard, Neil Howe and Tobias Peter. / fiji-national-provident-fund-writes-off 2013. The Global Aging Preparedness Index, -millions-in-members-funds. Second Edition. Washington, DC: Center for Holzmann, Robert, Robert Palacios, and Asta Strategic and International Studies. Zviniene. 2004. “Implicit Pension Debt: Issues, JSTAR (Japanese Study of Aging and Retirement). Measurement, and Scope in International 2011. http://www.rieti.go.jp/en/projects/jstar/. Perspective.” Social Protection Discussion Karam, Philippe, Dirk Muir, Joana Pereira, and Paper 0403, World Bank, Washington, DC. Anita Tuladhar. 2010. “Macroeconomic Holzmann, Robert, and Edward Palmer, eds. 2006. Effects of Public Pension Reforms.” IMF Pension Reform: Issues and Prospects for Working Paper 10/2, International Monetary Non-financial Defined Contribution (NDC) Fund, Washington, DC. http://www.imf.org Schemes. Washington, DC: World Bank. /external/pubs/ft/wp/2010/wp10297.pdf. H o p f , G r e g o r. 2 0 0 9 . “ Wa s S i n g a p o r e Kashiwase, Kenichiro, Masahiro Nozaki, and Kiichi Extraordinary? A Comparative View of Tokuoka. 2012. “Pension Reforms in Japan.” Singapore’s Saving Performance 1965–99.” IMF Working Paper 12/285, International HSBA Working Paper 02/2009, Hamburg Monetary Fund, Washington, DC. School of Business Administration, Hamburg, KLoSA (Korean Longitudinal Study of Aging). Germany. 2010. Korea Labor Institute, Seoul. http:// Hu, Yuwei. 2012. “Growth of Asian Pension www.kli.re.kr/klosa/en/about/introduce.jsp. Assets: Implications for Financial and Capital Koettl, Johannes, and Michael Weber. 2012. Markets.” ADBI Working Paper 360, Asian “Does Formal Work Pay? The Role of Labor Development Bank Institute, Tokyo. Taxation and Social Benefit Design in the New Husain, Aasim M. 1995. “Determinants of Private EU Member States.” IZA Discussion Paper Saving in Singapore.” In Singapore: A Case 6313, Institute for the Study of Labor, Bonn, Study in Rapid Development , edited by Germany. PENSION SYSTEMS AND AGING: REFORMS FOR WINNING THE RACE 183 Kohl, Richard, and Paul O’Brien. 1998. “The ———. 2013. “Pensions at a Glance.” OECD, Macroeconomics of Aging, Pensions, and Paris. Savings: A Survey.” OECD Economics Onder, Harun, and Pierre Pestieau. 2014. “Is Aging Department Working Paper 200, Organisation Bad for the Economy? Maybe.” Economic for Economic Co-operation and Development, Premise 144, World Bank, Washington, DC. Paris. Palacios, Robert. 2002. “Managing Public Kugler, Adriana, and Maurice Kugler. 2009. Pension Reserves: Part II—Lessons from Five “Labor Market Effects of Payroll Taxes Recent OECD Initiatives.” Social Protection in Developing Countries: Evidence from Discussion Paper 0219, World Bank, Colombia.” Economic Development and Washington, DC. Cultural Change 57 (2): 335–58. ———. 2015. “Pensions in an Ageing East Asia KWSP (Kumpulan Wang Simpanan Pekerja). and Pacific.” Background paper for the East 2012. “EPF Savings and Your Retirement.” Asia and Pacific regional report on aging, KWSP, Kuala Lumpur. World Bank, Washington, DC. Lu, Bei, Wenjiong He, and John Piggott. 2014. Palacios, Robert, and Charles Knox-Vydmanov. “Should China Introduce a Social Pension?” 2014. “The Growing Role of Social Pensions: Journal of the Economics of Ageing 4: History, Taxonomy and Key Performance 76–87. Indicators.” Public Administration and Ma, Jun, Xiaorong Zhang, and Zhiguo Li. 2012. Development 34 (4): 251–64. A Study of China’s National Balance Sheet. Palacios, Robert, and Edward Whitehouse. 2006. Beijing: Social Sciences Academic Press. “Civil Service Pension Schemes around the Mahalingam, Eugene, and Wei-Shen Wong. World.” Social Protection Discussion Paper 2013. “Malaysians Don’t Save Enough to 0602, World Bank, Washington, DC. See Them through Their Retirement Years.” Rodan, Garry. 2014. “Singapore’s Vaunted CPF Star, August 10. http://www.thestar.com.my under Fire.” Asia Sentinel , September 8. /Business/Business-News/2013/08/10 /Saving http://www.asiasentinel.com/econ-business -for-retirement-Malaysians-dont-save-enough /singapore-central-provident-fund-under-fire/. -to-see-them-through-their-retirement-years/. Rofman, Rafael, Ignacio Apella, and Evelyn Vezza, Mercer and Australian Centre for Financial Studies. eds. 2014. Beyond Contributory Pensions: 2015. Melbourne Mercer Global Pension Fourteen Experiences in Latin America . Index, 2015 Report. Melbourne. http://www Washington, DC: World Bank. .mercer.com/content/dam/mercer/attachments Samad, Suhaimi Abd, and Norma Mansor. 2013. /global/Retirement/Melbourne-Mercer-Global “Population Aging and Social Protection in -Pension-Index-2015/Report.pdf. Malaysia.” Malaysian Journal of Economic Mesa-Lago, Carmelo, Verna Dina Q. Viajar, and Studies 50 (2): 139–56. http://ssrc.um.edu Rolly Czar Joseph Castillo. 2011. Pensions in .my/images/ssrc/Books/Journal%20papers the Philippines: Challenges and Ways Forward. /Population%20ageing%20and%20 Manila: Friedrich Ebert Stiftung. social%20protection%20in%20Malaysia Moon, Hyungpyo. 2013. “Matching Contributions _suhaimi_norma.pdf. and Compliance in Korea’s National Pension Schmidt-Hebbel, Klaus, and Luis Servén. 1998. Scheme.” In Matching Contributions for Pensions, “World Saving: Trends and Theories.” edited by Richard Hinz, Robert Holzmann, Estudios de Economía 25 (2): 191–215. David Tuesta, and Noriyuki Takayama, 161–76. Schwarz, Anita M., and Omar S. Arias. 2014. The Washington, DC: World Bank. Inverting Pyramid: Pension Systems Facing MPFA (Mandatory Provident Fund Authority). Demographic Challenges in Europe and 2014. Annual Report . Hong Kong SAR, Central Asia. Washington, DC: World Bank. China: MPFA. Secada, Pablo, Antonio Cusato, and Verónica NPS (National Pension Service). 2012a. “Fund Zapata. 2013. “Efectos en Perú.” In Management Report.” NPS, Seoul. Contribución del sistema privado de pensiones ———. 2012b. National Pensions Statistics: Silver al desarrollo económico de Latinoamérica, Book. Seoul: NPS. edited by Rodrigo Acuña, 239–77. Medellín, OECD (Organisation for Economic Co-operation Colombia: SURA Asset Management. and Development). 2011. “Pensions at a Shome, Parthasarathi, and Katrine Saito. 1980. Glance: Asia and the Pacific.” OECD, Paris. “Creating Capital through Social Security 184 LIVE LONG AND PROSPER Institutions: The Asian Experience.” Domestic Wise, David A. 2004. “Social Security Provisions Finance Study 61, World Bank, Washington, DC. and the Labor Force Participation of Older Sluchynskyy, Oleksiy. 2015. “Defining, Workers.” Population and Development Measuring, and Benchmarking Administrative Review Aging, Health, and Public Policy 30 Expenditures of Mandatory Social Security (Suppl.): 176–205. Programs.” Social Protection Discussion Paper World Bank. 1994. Averting the Old Age Crisis: 1501, World Bank, Washington, DC. Policies to Protect the Old and Promote Souto, Pablo, and Alberto R. Musalem. 2012. Growth. Washington, DC: World Bank. “Assessing the Governance and Transparency ———. 2011. Mongolia: Policy Options for of National Public Pension Funds.” Review of Pension Reform . Washington, DC: World European Studies 4 (2): 148–67. Bank. SSO (Social Security Office), Thailand. 2009. ———. 2012a. “Malaysia: Elderly Protection Annual Report. Nonthaburi, Thailand: SSO. Study.” World Bank, Washington, DC. UN (United Nations). 2013. “World Population ———. 2012b. “Reducing Elderly Poverty in Prospects: The 2012 Revision, Highlights, Thailand: The Role of Thailand’s Pension and and Advance Tables.” Working Paper ESA/P/ Social Assistance Programs.” World Bank, WP228, Department of Economic and Social Bangkok. Affairs, Population Division, UN, New York. ———. 2013. International Income Distribution U.S. Social Security Administration. 2012. Social Database (I2D2). World Bank, Washington, Security Programs throughout the World: Asia DC. Edition. Washington, DC: U.S. Government ———. 2014. East Asia Pacific at Work: Printing Office. Employment, Enterprise, and Well-Being . Villagómez, Alejandro, and Arturo Antón. 2013. Washington, DC: World Bank. “El caso de México.” In Contribución del Yamabana, Hiroshi. 2011. “Actuarial Studies Sistema privado de pensiones al desarrollo of SSO Thailand.” Presentation of the económico de Latinoamérica , edited by International Labour Organization Financial Rodrigo Acuña, 105–80. Medellín, Colombia: and Actuarial Service, Bangkok, November 30. SURA Asset Management. Yamada, Tetsuji, Tadashi Yamada, and Guoen Villar, Leonardo, Jonathan Malagón, Julio César Liu. 1992. “Interdependency of Personal Vaca, and Carlos Ruiz. 2013. “Experiencia Savings and Labour Force Participation de Colombia.” In Contribución del Sistema of the Elderly, and Social Security Wealth: privado de pensiones al desarrollo económico A Time Series Analysis.” Applied Economics de Latinoamérica, edited by Rodrigo Acuña, 24 (4): 379–88. 55–104. Medellín, Colombia: SURA Asset Zhang, Chuanchuan. 2012. “The Relationship Management. between Elderly Employment and Youth Walker, Eduardo, and Fernando Lefort. 2002. Employment: Evidence from China.” MPRA “Pension Reform and Capital Markets: Are Paper 37221, Munich Personal RePEc Archive, There Any (Hard) Links?” Social Protection Munich, Germany. Discussion Paper 0201, World Bank, Zhang, Chuanchuan, John Giles, and Yaohui Washington, DC. Zhao. 2014. “Policy Evaluation of China’s Wang, Lijiang, Daniel Béland, and Sifeng Zhang. New Rural Pension Program: Income, Poverty, 2014. “Pension Fairness in China.” China Expenditure, Subjective Wellbeing and Labor Economic Review 28 (C): 25–36. Supply.” China Economic Quarterly 14 (1): WHO (World Health Organization). 2010. Global 203–30. Health Observatory Database. http://www.who Zuo, Xuejin. 2013. “Designing Fiscally Sustainable .int/gho/mortality_burden_disease/life_tables/en/. and Equitable Pension Systems in China.” Wickramanayake, Jayasinghe. 1998. “Savings and Paper presented at the International Monetary Development.” Central Provident Fund and Fund Office for Asia and the Pacific–Fiscal Savings in Singapore 22 (4): 435–54. Affairs Department Conference on Designing Wiese, Patrick. 2006. “An Evaluation of Pension Fiscally Sustainable and Equitable Pension Reform Options for Mongolia.” World Bank, Systems in Asia in the Post Crisis World, Tokyo Washington, DC. January 9–10. Population Aging and Health Services in East Asia and Pacific 6 Introduction population aging and health service delivery challenges in East Asia and Pacific. The second Population aging will have profound effects section briefly discusses Universal Health on the health sector throughout East Asia Coverage, which is likely to exert major pres- and Pacific. As discussed in chapter 1 of this sure on the region’s health care systems in the report, population aging has been accompa- context of population aging and NCDs. The nied by an increase in exposure to risk factors third section delves into population aging and that increase the likelihood of illness and dis- health expenditures in the region, examining ability in old age, specifically the prevalence of whether health expenditure growth is being noncommunicable diseases (NCDs). Health driven by aging itself and/or by other cost care delivery systems in the region have com- pressures. The fourth section provides recom- mon features that make them ill prepared for mendations for addressing these challenges the NCD epidemic, especially among popu- in a manner that is fiscally sustainable and lations with rising shares of older people. improves health care outcomes and quality for Such features include a range of cost drivers the elderly. that are likely to undermine fiscal sustain- ability, with or without aging in future years. Because age-specific health spending typically Population aging and health rises steeply toward the end of life, a com- service delivery challenges in mon assumption is that aging will generate East Asia and Pacific substantial increases in health care spend- Health care delivery systems in East Asia ing. However, the effects of aging on health and Pacific face common challenges in expenditure are much more complex. three key domains: delivery system design, This chapter examines the health system people-provider interactions, and organiza- challenges of responding to population aging. tional settings (figure 6.1). These weaknesses The first section provides an overview of are already contributing to a low-value health This chapter is based on a background paper prepared by Aparnaa Somanathan (2015) and a literature review conducted by Harry Moroz, Naoko Miake, and Ying Ho. 185 186 LIVE LONG AND PROSPER FIGURE 6.1 Health care in East Asia and Pacific faces major health system and service delivery challenges Low-Value Health Sysem Inequalities in access to services Cost escalation and financial burden Low quality; uncontrolled disease progression Challenges in the Service Delivery System External Challenges (Noncommunicable Delivery Top heavy, diseases, aging, system hospital-centric Fragmented service Weak primary care urbanization) design delivery system delivery People- Limited emphasis on provider Discontinuity of care patient self-management interaction Distorted incentives: Insufficient focus on Organization overservicing and population health and settings unnecessary care chronic care Challenges in the Policy, Institutional, and Financial Environment Cost inducing fee-for-service provider payments Deficient purchasing and service monitoring Deficiencies in human resource qualifications and management Source: World Bank 2014a. system where cost containment is a growing most economies (figure 6.2). For the majority problem. Health care costs are a major of people living in the region, the first point financial burden on both governments and of contact with the health system is usually households. Moreover, the delivery of health at a hospital (around 53 percent in China). services is characterized by inequalities, and Furthermore, throughout the region, rela- quality of health services is poor. This section tively simple health conditions such as hyper- discusses each of the three challenges in turn tension and uncomplicated diabetes result and the way their effects are likely to be exac- in hospitalization. In China, according to erbated in a rapidly aging population with a estimates, between 30 and 50 percent of rising burden of NCDs. admissions could have been handled at an ambulatory care setting (World Bank 2014a). Yet, tertiary hospitals have registered higher Weaknesses in delivery system design growth for both inpatient and outpatient Health systems in East Asia and Pacific are visits compared to primary and secondary hospital centric. Hospital costs account for 40 hospitals (figure 6.3). As shown in table 6.1, to 70 percent of total health expenditures in of all outpatient visits to central general P O P U L AT I O N AG I N G A N D H E A LT H S E R V I C E S I N E A S T A S I A A N D PAC I F I C 187 hospitals in Vietnam, 35 percent could have FIGURE 6.2 Hospital share of health expenditures is often greater been handled at provincial hospitals and than nonhospital share 20 percent at district hospitals. Of all out- Health expenditures by level of care, selected economies, 2012 patient visits to provincial general hospitals, 41 percent could have been handled at district Australia hospitals. High average lengths of stay further con- Bangladesh tribute to the hospital share of total spending. China As shown in figure 6.4, average lengths of stay in China, Japan, the Republic of Korea, Taiwan, China and Mongolia are higher than the average in Hong Kong SAR, China Organisation for Economic Co-operation and Development (OECD) countries. Japan The hospital bias is even stronger among Korea, Rep. the elderly. Health care utilization rates among the elderly are dominated by hospi- Malaysia tal inpatient care. Household survey data Mongolia from China, Indonesia, Japan, and Vietnam (figure 6.5) show that inpatient admission New Zealand rates increase with the mean age of house- Sri Lanka hold from about age 50 for both men and women, with a spike after age 65 in all Thailand countries. Analysis for China also shows 0 20 40 60 80 100 that inpatient admissions for older house- holds are far more likely to be in general and Share of health expenditures (%) township hospitals and that most outpatient Hospitals Providers of ambulatory care Other providers care is sought at hospitals. Outpatient utili- zation rates are much lower for the elderly Source: Jeong and Rannan-Eliya 2010. (with the exception of Japan) and do not vary much with age. When coverage is more FIGURE 6.3 Tertiary hospitals experienced the fastest growth of generous as in the Civil Servant Medical inpatient and outpatient visits in China Growth rate (percent) of inpatient admissions, 2008–11 Benefit Scheme in Thailand, the elderly make even greater use of inpatient care. The spike 30 in hospital utilization rates after age 65 and the high average lengths of hospital stays in 25 older East Asian and Pacific countries such 20 as China, Japan, and Korea suggest that Growth rate (%) acute care hospitals may be doubling up as 15 nursing homes for the elderly. A very inef- 10 ficient pattern of service delivery for the elderly emerges; expensive acute care hospi- 5 tals provide services such as outpatient care 0 and nursing home care that could be pro- vided more efficiently in other settings. –5 The increasing hospital focus of health sys- –10 tems in the region can be explained in part 2008 2009 2010 2011 by weaknesses in primary health care. By CHC, THC, and commune Secondary hospital Tertiary hospital definition, primary care is the professional health station health care response when patients make Source: World Bank 2014a. their first contact with the health care system. Note: CHC = community health center; THC = township health center. 188 LIVE LONG AND PROSPER TABLE 6.1 Hospital patients in Vietnam are being treated at higher levels than is necessary Treatment of patients in hospitals in Vietnam, 2010 (percent) Level where patients were actually treated Level where patients should have been Central Central Central Provincial Provincial District treated general hospital obstetrics hospital pediatrics hospital general hospital specialty hospital hospital Central level 39.1 6.5 5.8 2.7 0.9 0.4 Provincial level 35.4 43.3 35.5 43.3 42.4 7.3 District level 20.2 49.3 58.3 41.4 47.4 71.4 Commune health station 0.0 0.0 0.0 10.9 5.8 19.1 Other specialty hospital 5.3 0.9 0.4 1.7 3.5 1.8 Total 100.0 100.0 100.0 100.0 100.0 100.0 Source: HSPI 2010. FIGURE 6.4 Average length of stay in East Asian and Pacific acute The hospital-centric nature of care provi- care hospitals is higher than or comparable to OECD averages sion combined with weak and poor-quality Average length of stay (days) in acute care hospitals, selected economies primary care has resulted in fragmented health systems throughout East Asia and 20 18 Pacific. Few of the region’s health systems 16 have formalized links between hospitals and 14 ambulatory providers, including structured 12 referral systems, patient discharge and hando- Days 10 8 ver mechanisms, and patient outreach services 6 (McCollum et al. 2014). In almost all cases, 4 2 referral mechanisms exist on paper but not 0 in practice. Policies requiring patients to pay much higher fees for self-referrals have had C . ( 0) ew anm (20 ) ala ( 2 ) H o u n S i n b od ( 2 0 ) R , m 1) In Nep (20 ) i L (2 ) M CD ng a (2 0) Ch ( 2 ) pu M e t n (20 ) ng uss e ( 11) ala ( 0) lad ( 2 ) (2 ) in (20 ) (2 8) ne (2 ) B a a i l a ( 20 ) 8) SA ntr (20 ) ng ei D ga ia ( 09) Ca s t r a ( 2 0 ) a N y am 08 m lia 10 in 10 Sr e s h 0 0 5 R, ies 11 ng nd 10 Gu ar 03 ka 0 8 do a l 1 0 T h y s i a 011 Z e Fiji 0 0 8 Vi ina 009 ao ou ia 11 d 1 Re 0 1 SA ala 201 OE Mo hin 201 M si a 0 1 ea 0 00 A u n 01 Ch (20 K o a r p o r 20 an 0 a c c ol 0 a, (2 2 little effect (Somanathan et al. 2014). re a n a p Ko J a p w Ne Pa Br Weaknesses in people-provider interactions Source: OECD Health Data 2012 database, http://dx.doi.org/10.1787/888932723266. Note: OECD = Organisation for Economic Co-operation and Development. Data for the Republic of Korea refer to average length of stay for inpatient care (including long-term care beds in hospitals). Fragmented health care delivery systems merely exacerbate an overly acute and epi- sodic pattern of health care use by the elderly. However, quality of care in primary care Poor older people with minor ailments in facilities is poor, particularly with regard to Cambodia, Indonesia, and Vietnam reported staff skills and competencies. Clinicians at seeking care from a range of practitioners such the community level lack the skills to pro- as private health practitioners, local pharma- vide effective diagnosis and treatment because cists, and traditional practitioners instead of inadequate training and limited access to of government primary health care centers, continuing medical education (box 6.1). Staff which they found to be less convenient and turnover is high in primary care settings as a unreliable. Notably, in most East Asian and result of poor living conditions in rural areas, Pacific countries, annual health check-ups which further impedes quality of care. In a and screenings are infrequent. Countries such survey in Singapore, older people complained as China (Strauss et al. 2012) and Thailand they saw a different doctor on each visit. (Knodel, Prachuabmoh, and Chayovan 2013) P O P U L AT I O N AG I N G A N D H E A LT H S E R V I C E S I N E A S T A S I A A N D PAC I F I C 189 FIGURE 6.5 Inpatient admission rates increase with mean age of household Inpatient admission rates, by age and percent, selected countries a. China b. Indonesia 40 40 30 Inpatient rate (%) Inpatient rate (%) 30 20 20 10 10 0 0 25–40 41–50 51–60 61–70 71–80 81+ 25–40 41–50 51–60 61–70 71–80 81+ Age group Age group Urban men Urban women Rural men Rural women c. Japan d. Vietnam 20 40 18 16 30 Inpatient rate (%) Inpatient rate (%) 14 12 10 20 8 6 10 4 2 0 0 25–40 41–50 51–60 61–70 71–80 81+ 25–40 41–50 51–60 61–70 71–80 81+ Age group Age group Men Women Urban men Urban women Rural men Rural women Sources: CHARLS 2011; IFLS 2007; JSTAR 2011; VHLSS 2012. have made efforts to encourage annual physi- East Asian and Pacific countries, the chroni- cal exams and outpatient visits by ensuring cally ill tend to have multiple contacts with the they are covered by insurance. In Thailand, delivery system, such as diagnostic services, over half of those ages 60 and above reported specialists, emergency rooms, outpatient clin- having an annual physical exam, with little ics, and pharmacies. Moreover, as discussed variation by gender, age, or area of residence, in chapter 1 of this report, the elderly tend to a trend associated with high levels of outpa- have multiple chronic conditions and comor- tient care coverage provided by the Universal bidities, and the lack of integration of care Coverage Scheme.1 However, a study of pri- means that eventually, they seek care from mary health care in Cambodia, Indonesia, multiple providers for multiple conditions. Singapore, and Vietnam found that annual Without coordinated care approaches to help health exams were rare and that the elderly the elderly navigate the delivery system, fol- cited cost as the major reason for not seeking low the course of their diseases, and provide them (Dodge 2008). continuous management and follow-up, they Fragmentation is also directly associated often do not receive effective or timely care. with discontinuity of care and duplication of This inadequacy results in acute complica- services, particularly for those—such as the tions, repeat visits, or duplicated care—often elderly—with chronic conditions. In most for the same illness episode—and higher costs 190 LIVE LONG AND PROSPER BOX 6.1 Poor qualifications of staff to provide effective diagnosis and treatment China schools without proper credentialing is of particu- lar concern. A significant positive correlation was • Doctors with higher levels of education and more observed between adult curative care scores and qualifications were found to provide better- the receipt of training in diagnostic algorithm, quality processes and more correct diagnoses. antibiotics for respiratory disease, and so forth. However, 80 percent of clinicians staffing village health posts and 40 percent of those at town- Vietnam ship health centers reportedly had only voca- tional high school education or less (Sylvia et al. • The best doctors are concentrated in the cities: 2014). Lack of qualified staff was also reported only 5 percent of postgraduates and 45 percent for urban commune health stations (community of university graduates work at district and com- health centers). mune levels. • Only 18 percent of village doctors applied appro- • Clinical training is hospital based, with little priate quality processes (based on national guide- preparation provided for working at the primary lines) for diagnosis and treatment of diarrhea and care level. New medical graduates are not confi- angina. A low 26 percent of diagnoses were cor- dent enough to work independently at commune rect, and in 64 percent of the cases, the doctors health stations (CHSs) without close technical provided unnecessary or harmful prescriptions supervision and support. Newly graduated doc- (Sylvia et al. 2014). Clinicians in township health tors are weak in 16 of 38 specific competencies, centers fared only marginally better. including critical competencies for primary care such as performing medical techniques and treat- Indonesia ment and outpatient care (personal communica- • Based on survey vignettes (Rokx et al. 2010), only tion with World Bank staff in Vietnam). 56 percent of providers in public sector primary • Staff at the district and commune levels lack ade- care facilities (puskesmas and pustu) scored cor- quate skills for the diagnosis and treatment of rectly in diagnosis and treatment of adults with noncommunicable diseases and chronic diseases. respiratory infection. In one rural district, staff in only seven of 18 • Large regional variation exists: 59 percent of pub- CHSs had received training on hypertension pre- lic providers scored correctly in Java-Bali com- vention and management (Van Minh et al. 2014). pared to 48 percent in Sumatra. An evaluation of professional competency at the • Poor diagnostic and treatment ability is directly commune level found that doctors and assis- linked to the lagging quality of preservice and in- tant doctors responded incorrectly to more than service training in Indonesia and to the limited 50 percent of questions on cardiovascular and supply and poor quality of internships for medi- internal diseases (personal communication with cal students. The rapid growth of new medical World Bank staff in Vietnam). for individuals, insurers, and government efforts are unconnected. This issue is dis- (World Bank 2014a). cussed further in chapter 7. An added challenge specific to the elderly is integrating emerging long-term care arrange- Weaknesses in the organizational ments with the health delivery system, which setting can be institutional, community-based, or home-based care or care provided by fam- Distorted provider incentives leading ily members or formal aged care workers. In to overprovision most East Asian and Pacific countries, sepa- Providers in many East Asian and Pacific rate ministries are responsible for health and countries face strong incentives to overpro- nonhealth services for the elderly, and their vide services because fee-for-service (FFS) P O P U L AT I O N AG I N G A N D H E A LT H S E R V I C E S I N E A S T A S I A A N D PAC I F I C 191 payment methods are widely used without and prioritization at present (Leung, Tin, and adequate measures of volume and price con- Chan 2007). trol. Admittedly, FFS provider payment sys- The distortion of provider incentives tems are used in Japan; Korea; and Taiwan, and consequent overprovision of expen- China, and they have been successful in con- sive services is driving up costs, particularly trolling cost growth because FFS is used in because strategic purchasing capacity is lim- combination with global budgets and a uni- ited in most East Asian and Pacific econo- fied fee schedule. The significant inflationary mies. Strategic purchasing capacity refers to effects of FFS systems are well established the ability of the purchaser (insurer, ministry in Canada; the Czech Republic; parts of the of health, and so on) to incentivize provid- Russian Federation; and the United States ers to improve quality processes and results (Langenbrunner and Somanathan 2011). In while contributing to cost containment (that China and Vietnam, the FFS approach is used is, reducing readmissions and unnecessary in a context where health facilities essentially admissions, long stays, repeat procedures, operate as commercial enterprises, with an and adverse events). In most systems financed emphasis on self-financing of facilities (for by general revenues (Cambodia; Hong Kong example, sales or business income comprised SAR, China; Malaysia; and most Pacific over 60 percent of hospital revenue in both Island states), the purchaser and provider are countries in 2010). The link between hospi- essentially the same (the ministry of health), tal business revenue and physician income which leaves little room for the development has encouraged overservicing, unnecessary of strategic purchasing capacity. Even in care, and inefficiency in service produc- insurance-based systems (China, Indonesia, tion (Somanathan et al. 2014; World Bank Mongolia, the Philippines, and Vietnam), 2014a). social insurance agencies have yet to estab- In addition, the lack of appropriate mea- lish the informational, managerial, and finan- sures to prioritize among new and available cial accounting capacities to exercise robust, health interventions has resulted in rapid arm’s-length service purchasing and monitor- growth in health technologies and phar- ing of providers, whether public or private. In maceuticals. Distorted pricing for treat- many cases, the fragmentation of governance ments has given health care providers strong and administration to subnational levels has incentives to generate demand for profit- further weakened the purchasing power of able high-technology services and drugs in the insurance agency. place of less-profitable basic alternatives. In China and Vietnam, hospitals’—and there- Insufficient focus on chronic disease fore doctors’—revenues are enhanced by fees management and population health earned through prescriptions and medical Primary health care (PHC) facilities tend to tests and procedures and are supplemented focus on treatment rather than health promo- by kickbacks from pharmaceutical compa- tion and education. Health promotion and nies (Somanathan et al. 2014; Yip and Hsiao education interventions are crucial for the 2008). In China, doctors’ prescribing behav- management of NCDs, particularly among the ior has been found to vary with the insurance elderly. In Cambodia, Indonesia, Singapore, status of patients and the incentives offered and Vietnam, elderly survey respondents by pharmaceutical companies (Currie, Lin, reported that PHC providers do not spend and Meng 2014; Lu 2014). More generally in time discussing behavioral changes such as the region, few measures are in place to pri- healthy diets and exercise (Dodge 2008). oritize and restrict the diffusion of new tech- Moreover, the readiness of PHC facilities nologies and their incorporation into routine to manage chronic NCD conditions is limited. care. In many cases, the underlying health Box 6.2 describes the limited readiness of pub- care system is practically or philosophically lic sector primary care facilities (puskesmas) incompatible with mechanisms for evaluation in Indonesia to diagnose and treat three key 192 LIVE LONG AND PROSPER BOX 6.2 Limited supply-side readiness to manage and treat noncommunicable diseases at the primary health care level in Indonesia Indonesia faces several challenges in effectively Hypertension implementing its universal coverage policy, especially • Percentage of prevalence: estimated to be in addressing noncommunicable diseases that are 31.7 percent in 2007 (over 18 years of age) chronic and require careful disease management over • Percentage of puskesmas with functioning blood time. The results of a supply-side readiness study that pressure apparatus for diagnosis: over 95 percent drew upon a national facility-level survey are sum- nationally marized below. • Percentage of puskesmas with basic drugs to treat Diabetes mellitus hypertension: availability of captopril reported in 86 percent (urban) and 83 percent (rural) of areas; • Percentage of public sector primary care facili- 50 to 60 percent in some provinces such as West ties (puskesmas) that could conduct a basic blood Sulawesi and Papua glucose test for basic monitoring and diagnosis of diabetes mellitus: 70 percent in urban areas; Hypercholesterolemia 51 percent in rural areas; less than 20 percent • Percentage of puskesmas able to conduct cho- in some provinces such as Gorontalo, Papua, lesterol screening tests: 44 percent (urban) and Maluku, North Sulawesi, Southeast Sulawesi, and 31 percent (rural) West Papua • Percentage of puskesmas able to administer sec- • Percentage of puskesmas with basic drugs for ondary prevention: simvastatin reported in treatment of diabetes mellitus: 90 percent with 50 percent (urban) and 27 percent (rural) of areas glibenclamide in stock; 48 percent with metfor- min in stock Source: World Bank 2014b. NCD-related conditions (diabetes, hyperten- Philippines, and Vietnam, 10 to 20 percent sion, and hypercholesterolemia). In Vietnam, of households incur OOP payments that are only 15 percent of districts and 2 percent of catastrophic (exceeding 25 percent of house- communes are implementing hypertension hold expenditures). The likelihood of cata- management under the national targeted pro- strophic OOP payments is greater for NCDs gram, and 97 percent of commune health sys- and chronic conditions. tems lack the capacity to carry out basic tests In the region, OOP payments for health (Van Minh et al. 2014).2 In the Philippines, care are consistently higher for the elderly patient access to diabetes care and treatment than for the remaining population. Household was found to be impeded by the poor avail- survey data also show that the share of OOP ability of adequate equipment and insulin and payments in total household consumption other medicines at the PHC level (Higuchi is consistently higher for elderly households 2010). (average ages 60 and above) relative to non- elderly households (table 6.2). In addition, Failure of low-value health systems to evidence exists that OOP payments are poten- provide equitable and affordable access tially catastrophic for elderly households. In to services China, Indonesia, and Vietnam, OOP pay- In East Asia and Pacific, higher levels of cov- ments exceed 25 percent of total household erage have not translated into affordable cov- expenditures for 18 percent, 6 percent, and erage for the population. The out-of-pocket 3 percent of elderly households, respectively, (OOP) share of total expenditures and the and this rises to 34 percent, 17 percent, and incidence of catastrophic OOP payments 30 percent, respectively, when only nonfood remain high in most countries. In China, the expenditures are considered (table 6.3). P O P U L AT I O N AG I N G A N D H E A LT H S E R V I C E S I N E A S T A S I A A N D PAC I F I C 193 TABLE 6.2 Significant shares of elderly are experiencing catastrophic out-of-pocket payments for health care relative to total household spending Percentage of households by age groupings with health spending of 25 percent or more of total household expenditure a. China Household mean age 25–49 (%) 50–69 (%) 70 and above (%) Total (%) Sector Rural 9.36 12.48 18.57 11.23 Urban 8.21 10.13 13.55 9.38 Total 8.78 11.35 15.82 10.30 Gender Male 8.03 10.10 14.64 9.34 Female 9.44 12.53 16.78 11.16 Total 8.78 11.35 15.82 10.30 Sector and gender Rural men 8.75 11.87 17.91 10.61 Rural women 9.91 13.08 19.13 11.80 Urban men 7.33 8.14 11.83 8.06 Urban women 8.99 11.96 14.90 10.54 Total 8.78 11.35 15.82 10.30 b. Indonesia Household mean age 25–49 (%) 50–69 (%) 70 and above (%) Total (%) Sector Rural 1.60 2.86 3.56 1.85 Urban 1.82 3.58 4.16 2.06 Total 1.70 3.15 3.76 1.95 Gender Men 1.60 3.15 3.98 1.84 Women 1.80 3.15 3.65 2.04 Total 1.70 3.15 3.76 1.95 Sector and gender Rural men 1.57 2.79 3.64 1.79 Rural women 1.63 2.92 3.53 1.91 Urban men 1.63 3.67 4.80 1.90 Urban women 1.98 3.51 3.90 2.20 Total 1.70 3.15 3.76 1.95 c. Japan Respondent’s age Gender 50–69 (%) 70 and above (%) Total (%) Men 0.25 0.12 0.20 Women 0.25 0.48 0.33 Total 0.25 0.31 0.27 d. Thailand Household mean age Sector 25–49 (%) 50–69 (%) 70 and above (%) Total (%) Rural 1.48 2.52 3.68 1.88 Urban 1.55 2.92 4.61 2.04 Total 1.50 2.65 3.95 1.94 table continues on next page 194 LIVE LONG AND PROSPER TABLE 6.2 Significant shares of elderly are experiencing catastrophic out-of-pocket payments for health care relative to total household spending (continued) Percentage of households by age groupings with health spending of 25 percent or more of total household expenditure e. Vietnam Household mean age 25–49 (%) 50–69 (%) 70 and above (%) Total (%) Sector Rural 1.84 4.16 5.88 2.37 Urban 1.59 3.76 6.78 2.00 Total 1.75 4.06 6.09 2.25 Gender Men 1.64 4.10 6.22 2.10 Women 1.90 4.02 5.97 2.43 Total 1.75 4.06 6.09 2.25 Sector and gender Rural men 1.68 4.28 5.64 2.16 Rural women 2.05 4.04 6.08 2.63 Urban men 1.54 3.58 8.01 1.97 Urban women 1.64 3.94 5.60 2.03 Total 1.75 4.06 6.09 2.25 Sources: CHARLS 2011; IFLS 2007; JSTAR 2011; ThaiSES 2011; VHLSS 2012. TABLE 6.3 Significant shares of elderly are experiencing catastrophic out-of-pocket payments for health care relative to total nonfood household spending Percentage of households by age groupings with health spending of 25 percent or more of total nonfood expenditure a. China Household mean age Sector 50–69 (%) 70–79 (%) 80 and above (%) Total (%) Rural 33.41 33.16 40.83 34.00 Urban 29.62 32.78 25.70 30.57 Total 31.69 32.96 33.57 32.33 b. Indonesia Household mean age Sector 50–69 (%) 70–79 (%) 80 and above (%) Total (%) Rural 15.86 18.50 17.94 16.74 Urban 12.15 14.79 4.57 12.29 Total 14.44 17.24 12.83 15.09 c. Vietnam Household mean age Sector 50–69 (%) 70–79 (%) 80 and above (%) Total (%) Rural 27.99 29.98 35.20 29.88 Urban 26.70 25.57 53.01 29.91 Total 27.72 28.98 38.60 29.88 Sources: CHARLS 2001; IFLS 2007; VHLSS 2012. P O P U L AT I O N AG I N G A N D H E A LT H S E R V I C E S I N E A S T A S I A A N D PAC I F I C 195 Health care access for the elderly is also health system delivery costs and weak cost- characterized by large inequalities between containment measures. Despite increases in rich and poor elderly. Household surveys from central government subsidies of 1 to 2 percent China, Indonesia, Japan, and Vietnam show a of gross domestic product (GDP) during the two- to threefold difference between the health past decade or so, payers (insurance orga- care utilization rates of the richest 20 percent nizations and local governments) and pro- of the population compared to the poor- viders are increasingly under financial stress est 20 percent, with the exception of Japan. because benefits packages are overly gener- In Thailand, civil servants who have more ous and not fully financed in many cases. generous coverage use outpatient care two to Meanwhile, as discussed earlier, the com- three times as often and inpatient care four to bination of distorted provider incentives, five times as often as the remaining popula- lack of strategic purchasing capacity, and tion. Earlier analysis has shown that higher weaknesses in the design of the service deliv- household per capita expenditure is positively ery system make cost containment a major associated with use of inpatient and outpatient challenge. As costs rise, providers increas- care in China (Strauss et al. 2012). ingly resort to cost recovery from patients, Thus, effective health coverage—people such as the common practice of extra bill- who need health care receive it in a timely ing for services in the Philippines and Vietnam. and affordable manner and at an accept- With or without regard to aging, current able level of quality—remains elusive for the trends in service delivery patterns and costs elderly as much as or more than for the over- will not only increase the OOP burden on all population. The only exceptions are the households, but also undermine the sustain- high-income East Asian and Pacific countries. ability of public financing for health. Even However, affordability of care is not the only without population aging, political commit- issue; a growing body of evidence indicates ments toward universal health coverage and that higher levels of insurance coverage do the consequent expansion in health care cov- not always lead to use of appropriate, good- erage are expected to lead to quite substantial quality services (Currie, Lin, and Meng 2014; increases in the demand for health care. Such Somanathan et al. 2014). increases are likely to put additional pressure on health systems that are already supply con- strained. Moreover, NCDs pose a new set of Universal health coverage in challenges to health systems because of the the context of population aging different models of care delivery required to and NCDs manage chronic diseases. Coverage—defined in terms of enrollment Adding an aging population to this toxic in universal coverage programs and ben- mix of cost drivers is widely expected to lead to efit levels—has increased quite significantly, rapid cost escalation. Exactly how large will the accompanied by an expansion in the over- aging effect be relative to other drivers of costs? all supply of health services. The drive for The next section examines this in more detail. universal health coverage, backed by large increases in public spending on health, is the single most unifying feature of health Population aging and health policy in East Asia and Pacific today. Rapid expenditures in East Asia increases in enrollment coverage occurred in and Pacific five of the largest countries in the region since Overall health spending is relatively high the mid-1990s. Benefit packages have also in East Asia and Pacific and has risen quite been expanded to include chronic conditions sharply in recent years, fueling concerns and outpatient care. about fiscal sustainability, particularly in Despite increased coverage levels, OOP the context of aging. This section examines payments remain high because of rising whether aging itself or other cost pressures 196 LIVE LONG AND PROSPER FIGURE 6.6 Health share of GDP has been rising in East Asia and exhibit a J-shape: per capita spending is Pacific with income growth somewhat high at very young ages, declines Total and public expenditure on health as a percentage of GDP, 1995–2012 and remains stable for most of the prime-age period, and increases again from about age 7.0 55 for men and age 60 for women. Figure 6.7 6.5 illustrates this age-expenditure profile in the 6.0 case of Japan. Given this trend, the frequent Share of GDP (%) 5.5 assumption is that aging—by increasing the 5.0 share of the population in older age groups— 4.5 will generate substantial increases in health 4.0 care spending. 3.5 However, numerous studies have shown that the effects of aging on health expenditure 3.0 are much more complex than apparent from 95 96 97 98 99 00 01 02 03 04 05 06 07 08 09 10 11 12 looking at per capita health care costs by age 19 19 19 19 19 20 20 20 20 20 20 20 20 20 20 20 20 20 Total Public profile in a given year. The evidence shows that the age profile of health spending changes for Source: WHO health data. two main demographic reasons: (a) concen- tration of the bulk of an individual’s lifetime drive health expenditure growth in the region. health care costs at the end of life (so-called Health spending trends and demographic and “death-related” costs) and (b) the possible non-demographic drivers of health spending improvement in health status following the are examined in light of OECD evidence. The increase in life expectancy. Depending on the available empirical evidence on decomposi- death-related costs component and the pos- tion of health expenditures—retrospective sible changes in health status, an observation and forward looking—for OECD countries at any given point in time that the elderly con- and a selection of East Asian and Pacific sume more health care services than the young countries is also presented. does not suggest that the process of aging alone As a share of GDP, total health expendi- will inevitably lead to increases in health care tures in East Asia and Pacific are more or less spending (Raitano 2006). Each of the two fac- compared to other regions, but have been tors underlying the relationship between aging ramped up significantly in recent decades and health expenditure is examined below. due to a combination of policy shifts, social pressure, and to some extent demographic Proximity to death change (figure 6.6). The share of public health Recent empirical studies have found that spending tends to be higher than average in proximity to death, rather than age, is an East Asia and Pacific, although this is driven important driver of costs, at least for groups in part by Pacific Island countries, which ages 65 and above (Raitano 2006). Typically, have exceptionally high shares (more than illness and disability worsen sharply in the 70 percent) of public spending. few weeks or months before death, leading to rapid upward-spiraling of the individual’s health expenditures because of medical inter- The “aging” effect on health ventions, hospital stays, hospice care, and so expenditures on. Evidence from the OECD shows quite Health expenditures per capita, measured consistently that per capita health expendi- nationally or globally, rise at older ages. The tures spiral in the last few years of life, par- trend is supported by the data (Fuchs 1998; ticularly in the final year (Lubitz and Riley Hagist and Kotlikoff 2005; Meerding et al. 1993; Spillman and Lubitz 2000). 1998; Mendelsohn and Schwartz 1993). With the majority of health expenditures Age-specific profiles of expenditures typically concentrated near death, one cannot assume P O P U L AT I O N AG I N G A N D H E A LT H S E R V I C E S I N E A S T A S I A A N D PAC I F I C 197 that the age profile of health expenditures will FIGURE 6.7 Age-specific expenditure profile exhibits J-shape remain constant when life expectancy increases. for Japan Decreases in age-specific mortality rates may Health cost per capita by age relative to average index = 100, 2001, 2011, lead to decreases in age-specific costs, because and 2012 declining mortality rates reduce the proportion 500 of high-cost users (that is, those close to death). In figure 6.7, the rightward shift of the J-curve 450 Index (average cost per capita = 100) for Japan between 2001 and 2012 shows 400 that as longevity increased in Japan, the older 350 age groups still cost more than younger age groups, but at progressively older ages. 300 Therefore, examining estimations of how 250 large the spike in deaths will be in East Asian 200 and Pacific economies during the next 20 to 30 years is useful. For the very young green 150 countries in the region (Cambodia, the Lao 100 People’s Democratic Republic, Myanmar, 50 Papua New Guinea, the Philippines, and Timor-Leste), the increase in deaths will be 0 + 50 49 60 9 70 69 10 9 20 19 40 39 80 79 almost negligible (figure 6.8). The moder- 30 29 25 4 55 4 15 14 35 4 45 4 65 4 75 4 4 4 85 –5 5– –2 –5 –3 –4 –6 –7 0– –8 – – – – – – – ately old orange countries (China, Indonesia, Malaysia, Mongolia, Thailand, and Vietnam) Age group will have to absorb about 0.28 percent more 2001 2011 2012 of the population dying annually by 2040. Source: MHLW 2011. The very old red economies (Hong Kong SAR, China; Japan; Korea; and Singapore) face the biggest challenge, with 0.36 percent more of FIGURE 6.8 Spikes in death rates vary with the speed of aging across East Asia and Pacific the population dying annually by 2040. Projected percentage of deaths in population by country grouping, various years Health status and life expectancy If proximity to death is a key driver of costs, 0.60 then so is the progression of illness and dis- ability leading up to death. If the bulk of 0.50 health care costs is incurred within a few months or years of severe illness before death 0.40 and if this timespan shrinks and shifts over time to ever-higher ages (morbidity compres- 0.30 Percent sion), then population aging would not be the main factor driving up costs. Alternatively, if 0.20 increased longevity without improvement in health status (morbidity expansion) leads to 0.10 increased demand for services over a longer 0 period of time, aging would lead to increased total lifetime health care expenditures and –0.10 overall health care spending. 2010–20 2010–30 2010–40 2010–50 Data on years lived with disability and Green economies Orange economies Red economies risk factors produced by the Global Burden of Disease project indicate that in East Asia Source: World Bank calculations based on United Nations mortality data. Note: Red = Hong Kong SAR, China; Japan; the Republic of Korea; and Singapore. Orange = China, and Pacific, the middle-age cohort of today Indonesia, Malaysia, Mongolia, Thailand, and Vietnam. Green = Cambodia, the Lao People’s is less healthy compared with the middle-age Democratic Republic, Myanmar, Papua New Guinea, the Philippines, and Timor-Leste. 198 LIVE LONG AND PROSPER cohorts of 10 to 20 years ago (figure 6.9). The never been a major driver of health care middle-age cohort of today has higher rates spending (Oxley and MacFarlane 1994). of hypertension and diabetes and higher rates Growth in health spending has been driven of smoking and exposure to other risk fac- by other demand effects (per capita income tors. Therefore, one can reasonably expect growth and extension of coverage by morbidity to expand as the current middle- national health systems), but primarily by age cohorts reach older ages. Putting in place supply factors (technological progress, infla- health policies and strategies that take a tion, and supply-induced demand effects in whole-of-life approach to healthy aging will a sector where information asymmetries are help curb these trends in the future. pervasive). These non-demographic drivers are considered below. As national income or wealth increases, The main cost drivers: Non-demographic both expectations and health spending rise, effects on health expenditures regardless of changes in needs (Getzen 2000). The idea that aging has a significant effect East Asian and Pacific governments’ priori- on health expenditures is often referred to tization of universal health coverage goals as a “red herring” because it distracts from and willingness to accompany those policy more important underlying cost drivers commitments with large increases in public (Zweifel, Felder, and Meiers 1999). From spending in recent years, especially in China, a historical perspective, demography has Indonesia, the Philippines, and Vietnam, are evidence of this rise. In these countries, rapid economic growth during the 1990s and early FIGURE 6.9 Years lived with disability are increasing in East Asia 2000s was associated with rising expecta- and Pacific as life expectancy increases tions and demand for greater access to health Percentage change in years lived with disability for the cohort, by gender, care and with governments’ acknowledge- ages 55 to 59, 1990–2010 ment of the need to address those demands by increasing public spending on health, as evi- Japan denced by elasticities of spending of greater Singapore than or equal to one in almost all countries Korea, Rep. China (figure 6.10). Taiwan, China All other non-demographic drivers aside Cambodia from income are referred to as excess cost Indonesia Lao PDR growth—essentially, the excess of growth Malaysia in per capita health expenditures over the Myanmar growth in per capita GDP after controlling Philippines Sri Lanka for the effect of demographic change. This Thailand approach captures the effects of two cost Timor-Leste Vietnam drivers: technological change and policy Oceania choices regarding key levers of health financ- Micronesia, Fed. Sts. ing. The available evidence on the effect of Fiji Kiribati these two drivers and the significance for East Marshall Islands Asia and Pacific are discussed below. A third Papua New Guinea element of excess cost growth worth noting Samoa Solomon Islands is Baumol’s cost disease: low-productivity Vanuatu increases in medical care relative to other 0 0.2 0.4 0.6 0.8 1.0 1.2 1.4 1.6 less labor-intensive sectors shift the relative % change in years lived with a disability prices of medical care upward (Baumol 1996, Men Women 2012). Technological change has accounted for Source: IHME 2013. the bulk of medical care cost increases over P O P U L AT I O N AG I N G A N D H E A LT H S E R V I C E S I N E A S T A S I A A N D PAC I F I C 199 time (Newhouse 1992; Smith, Newhouse, FIGURE 6.10 Total and public health spending has increased and Freeland 2009). Recent estimates sug- more rapidly than GDP in much of East Asia and Pacific in the 1990s gest that medical technology explains 27 to and 2000s 48 percent of growth in health care spend- Elasticity of total and public health spending to GDP, selected East Asian and Pacific countries, 1995–2012 ing in OECD countries since 1960 (Smith, Newhouse, and Freeland 2009). Although technological change often contributes to Timor-Leste efficiency improvements, in recent years, new technologies have been dominated by costly Lao PDR product innovations to alleviate disease rather Myanmar than to improve efficiency or strengthen pre- ventive care. Expensive end-stage cancer Vietnam drugs that do little to cure the disease but may improve outcomes temporarily are one exam- Thailand ple. This trend has also contributed to the Papua New Guinea growth in death-related costs (Baumol 2012; Zweifel 2003). In countries where cost con- China trol mechanisms are weak and efforts to rein in technology growth are limited, technologi- Indonesia cal change can lead to rapid cost escalation Mongolia in association with aging. The contribution of the costly product innovations to per capita Philippines expenditures at different ages, rather than Malaysia aging itself, is the driver in health spending. Rising incomes and expectations will likely 0.0 0.5 1.0 1.5 2.0 add further pressure to technology growth. Elasticity As noted earlier in the chapter, the absence of appropriate measures to prioritize the dif- Total Government fusion of new technologies combined with Source: World Bank 2015. distorted FFS-driven incentives have already resulted in rapid growth in health technology spending in East Asia and Pacific. In China, not only the introduction of new mechanisms for instance, the medical device market grew for evaluation and prioritization (discussed in from US$8 billion in 2006 to US$20 billion the next section), but also the political will to in 2011—a compound annual growth rate withstand demand pressures for new technol- of 20.1 percent, exceeding even the pace of ogies. In many countries, the elderly are more total health expenditures (US$156 billion to vocal politically than others and may drive up US$357 billion, compound annual growth rate the adoption of new technologies. of 18.0 percent). Box 6.3 discusses the role of As shown by Hagist and Kotlikoff (2005), technological change and interactions with in addition to costly product innovations, the aging in the escalation of health care costs. contribution of benefit levels at different ages In East Asia and Pacific, efforts to rein can drive up spending. Because health care in the growth of technologies are critical for benefit levels (health care expenditures per controlling health expenditure growth, partic- beneficiary at a given age) are much higher ularly in the context of an aging population. for the elderly than for the young, benefit Health policy needs to focus on ensuring that levels increase as the proportion of older age health technologies are adopted on the basis groups in the population increases. However, of evidence and need and are diffused appro- considerable heterogeneity exists across priately and that the costs of adoption are countries in health spending at older ages. managed carefully. This approach will require As figure 6.11 shows, in Austria, Germany, 200 LIVE LONG AND PROSPER BOX 6.3 The role of technological change and interactions with aging in health care cost escalation Technological change—the adoption and utilization The combination of population aging and of new pharmaceuticals and medical technologies—is advances in medical technology provides the rationale not necessarily problematic. For instance, Cutler and to spend more, with more medical technology being McClellan (2001) found that the use of new technol- oriented toward the elderly (Dormont, Grignon, and ogy helped increase the average coronary patient’s Huber 2006; Jones 2002). A Dutch study that looked life expectancy by one year (valued at US$70,000 at utilization trends and the state of medical technol- per case) while treatment costs increased US$10,000 ogy (medical patents) found that large advances in per case (4.2 percent per year), for a net benefit of medical technology were significantly associated with US$60,000 per case. Whether a particular technolog- higher growth in the probability of hospitalization, ical development increases or decreases costs depends particularly for the older population (Wong et al. on its effect on unit cost, its level of use, and the like- 2011). lihood that the treatment complements or replaces existing methods. FIGURE 6.11 Age-specific expenditure profiles vary for OECD Spain, and Sweden, per capita expenditures countries for those ages 75 to 79 and 80 and above are Health spending by age relative to reference group 50–64 = 1, selected only twice the level of per capita expenditures countries, latest available year, early 2000s of the reference age group—ages 50 to 64. In contrast, the oldest-old (ages 80 and above) 14 in the United States receive benefits that aver- age 8 to12 times those of the reference group, Relative health spending, reference group 50–64 = 1 12 whereas Japan falls in the middle. The policy choices of health systems about 10 how the benefits package is defined, how providers are paid, how services are pur- chased, and how the delivery system itself is 8 structured are all determinants of the pace of expenditure growth. As discussed in the pre- 6 vious section, the current set of policy choices in most East Asian and Pacific countries is 4 inimical to cost control. Of course, there are some exceptions such as Japan. Given that Japan aged much more rapidly than the 2 United States during 1970–2002 (Hagist and Kotlikoff 2005), per capita health care expen- 0 ditures should have grown more rapidly in 0–14 15–19 20–49 50–64 65–69 70–74 75–79 80+ Japan. Instead, the reverse was true because Age group of the much higher rate of benefit growth Australia Austria Canada Germany recorded in the United States, where cost con- Japan Norway Spain Sweden trol is weaker than in Japan. United Kingdom United States Policy choices are critical with respect to the control of NCDs, given that they will Source: Hagist and Kotlikoff 2005. Note: OECD = Organisation for Economic Co-operation and Development. account for nearly two-thirds of all deaths. P O P U L AT I O N AG I N G A N D H E A LT H S E R V I C E S I N E A S T A S I A A N D PAC I F I C 201 Although three-quarters of NCD deaths overall health expenditure projections cover occur over age 60, the burden of NCDs is a broad range and are probably not directly not borne by the elderly alone. With the comparable because of methodological dif- onset of NCDs increasingly occurring at ferences, health expenditure projections for younger ages, much of the effect of NCD- Malaysia show the extent to which effi- related morbidity will be felt by the working- ciency savings can actually help control age population. In China, where over half of health expenditure growth (Rannan-Eliya the incidence of NCDs falls on the working- et al. 2013). age population, the projected cumulative Nonetheless, in low- and middle-income cost associated with each of the five domains countries in East Asia and Pacific, aging may (cardiovascular disease, cancer, chronic respi- account for a greater share of total health ratory disease, diabetes, and mental health) spending than that found for OECD coun- is US$27.84 trillion (Bloom et al. 2013). A tries to date. In these East Asian and Pacific range of policy options are available to pol- countries, the pace of urbanization and health icy makers for the prevention and control care reforms to expand coverage will accel- of NCDs, some of which are more likely to erate in the next two decades or so. Both of drive up costs than others as discussed in the these transitions will increase the demand next section. for health care generally, but expenditure by the elderly is likely to increase more than by other age groups. For instance, a recent study Decomposition of health expenditures from China found that although health care The main question for most policy makers spending of rural residents is lower than that in the region is whether health expenditures of urban residents in all age groups, the age– will reach unsustainable levels in the future, expenditure profile is much steeper in urban and the potential contribution of aging to areas than in rural areas (Feng, Lou, and Yu spending growth. Health expenditure pro- 2015). This finding suggests that the health jections or actuarial cost projections are expenditures of the elderly increase more dur- fairly data intensive. Economy-specific pro- ing the process of rural–urban integration. jections for East Asia and Pacific are avail- The study also found a cohort effect that able only for Hong Kong SAR, China, and causes the elderly to increase health expendi- for Malaysia, although health expenditure tures by more than the younger group even if projections produced by the OECD and the they experience common social or economic International Monetary Fund include several changes (for example, health care reform). countries from the region. Projections car- As health coverage expands with univer- ried out by the European Commission for the sal health coverage–motivated policies and EU-27 countries3 are also worth examining. reforms and as health insurance expands into Notably, all of these projections estimated the rural areas, pent-up demand for health care effect of demographic and non-demographic is likely. Significant health spending will be drivers. incurred as the elderly who did not previously Two key points emerge from recent use health care now begin gaining access to expenditure analysis, primarily from OECD services. countries: (a) aging itself typically accounts The expected increase in disease bur- for one-third or less of the increase in health den attributable to NCDs, combined with spending, even in the older countries (a find- inefficiencies in the financing and delivery ing that is similar to estimates for Malaysia), of health services, mean that escalation of and (b) excess cost growth, specifically that health care costs in East Asia and Pacific may driven by technological change and policy prove unsustainable even without the effects choices, emerges as the most critical factor of aging, although aging will no doubt be a in health expenditure growth. Although the contributing factor. As discussed earlier, the 202 LIVE LONG AND PROSPER distribution of the burden of NCDs across • Preparing communities and institutions the region reflects rapidly increasing exposure outside the health system for the emerging to lifestyle and disease risk factors, including needs of an increasingly older population smoking and conditions such as hypertension and diabetes that predispose younger adults Reducing exposure to risks to NCDs later in life. Meanwhile, most of the region’s health systems are ill equipped to Reducing exposure to risk early in life is part screen, monitor, and treat NCDs and other of taking a life-course approach to reduc- aging-related conditions such as frailty and ing illness and disability among the elderly. dementia efficiently and effectively. Table 6.4 summarizes key cost-effective strat- egies for reducing population-level exposure Health policies and strategies for to NCD-related risks. an aging population Although effective strategies to control smoking exist, smoking remains the lead- Without immediate and effective reforms to ing cause of preventable deaths globally provider payments and purchasing and with- (WHO 2013). Robust evidence shows that out improved efficiency in service delivery, tobacco control as outlined in the Framework health expenditure growth in East Asia and Convention on Tobacco Control is the most Pacific may well undermine fiscal sustainabil- cost-effective set of strategies (table 6.4) for ity. The experience of OECD countries indi- controlling NCDs and that improvements cates that there are alternative pathways to can happen quickly. This is confirmed by the growing old that involve reducing the sever- experience of the United States and Western ity of population disease from a moderately Europe (Cutler, Deaton, and Lleras-Muney severe level without excessive burden on gov- 2006). ernment and household budgets. This section Increasing taxes on tobacco control draws from experiences of OECD countries products is the most cost-effective of the in reducing mortality and morbidity resulting Framework Convention on Tobacco Control from NCDs and aging and applies them to interventions, but low- and middle-income the East Asian and Pacific context to present countries in East Asia and Pacific have tended a range of health policies and strategies for to lag on this intervention (table 6.5). Evidence consideration. These crucial policy reforms, from Australia, Canada, and the United which are discussed in turn below, are as States indicates that higher taxes induce some follows: smokers to quit, reduce consumption by • Reducing exposure to NCD-related risk those who continue to smoke, prevent oth- factors through stronger health promotion ers from starting, and reduce the number of and public health ex-smokers who resume (Sunley 2009; WHO • Managing risk factors more effectively 2014a). However, low- and middle-income through primary and secondary prevention East Asian and Pacific countries have lagged • Improving self-management and medica- in implementing higher taxes, partly because tion adherence of pressure from domestic tobacco produc- • Transforming health care delivery into a ers who tend to be important economically primary care–centered and integrated sys- (for example, China and the Philippines). In tem of care delivery 2012, the Philippines passed landmark tax • Managing acute care episodes more legislation that raised excise taxes on ciga- effectively rettes, indexed the rates to inflation (starting • Preparing for functional and cognitive in 2017), allocated a portion of the revenue to decline among the elderly compensate tobacco farmers, and earmarked • Getting better value for money through the remainder to finance universal coverage strengthened purchasing and prioritiza- programs. The experience of the Philippines tion of health interventions has spurred other countries in the region such P O P U L AT I O N AG I N G A N D H E A LT H S E R V I C E S I N E A S T A S I A A N D PAC I F I C 203 TABLE 6.4 Effective strategies to reduce exposure to key NCD-related risks in East Asia and Pacific Areas needing intervention Recommended strategy Smoking Control of tobacco use as outlined in the World Health Organization Framework Convention on Tobacco Control: • Increased taxes on tobacco products • Enforcement of smoke-free workplaces and packaging and labeling of tobacco products with comprehensive health warnings supported by public education • Comprehensive banning of tobacco advertising, promotion, and sponsorship Malnutrition Promotion of physical activity and dietary quality and efforts to reduce obesity: • Introduction of taxes for unhealthy food • Provision of subsidies for healthy food • Promotion of labeling • Administration of marketing restrictions • Community-based physical activity at schools, workplaces, and so on Reduction of populationwide salt consumption: • Voluntary reduction of salt levels in processed foods and food additives • Sustained public education to encourage change in food choices Reduction of undernutrition and low birthweight: • Targeted interventions to reduce stunting and wasting in poor populations • Multisectoral interventions such as food fortifications Alcohol use Reduction of populationwide harmful alcohol consumption: • Increased taxes on alcoholic beverages • Limited access to retail alcohol • Comprehensive banning of alcohol advertising, promotion, and sponsorship Vaccinations Universal hepatitis B vaccination for countries at high risk for hepatitis B and liver cancer Source: Adapted from Beaglehole et al. 2011. Note: NCD = Noncommunicable disease. TABLE 6.5 Tobacco prices and taxes in East Asia and Pacific are not high enough Tax as a percentage of price Price at official exchange rate Specific excise tax Total tax Economy (US$) (%) (%)a East Asian and Pacific economies Tuvalu 6.32 0.00 14.82 Cambodia 0.42 0.00 17.13 Marshall Islands 3.50 0.00 28.57 Philippinesb 0.36 18.13 28.85 Solomon Islands 3.26 20.00 30.00 Timor-Leste 1.25 30.40 34.60 Nepal 0.51 23.69 35.19 Papua New Guinea 7.23 28.26 37.35 China 0.79 1.20 40.76 Fiji 4.02 27.79 40.84 Vietnam 0.72 0.00 41.59 Kiribati 6.32 0.00 41.67 Lao PDR 0.75 8.33 43.05 Mongoliab 0.59 40.27 49.36 Myanmarb 0.92 0.00 50.00 Indonesiab 1.76 42.60 51.00 Cook Islands 9.68 46.57 53.51 Palau 3.50 0.00 57.14 Malaysiab 3.18 44.00 57.30 table continues on next page 204 LIVE LONG AND PROSPER TABLE 6.5 Tobacco prices and taxes in East Asia and Pacific are not high enough (continued) Tax as a percentage of price Price at official exchange rate Specific excise tax Total tax Economy (US$) (%) (%)a Vanuatub 7.58 45.71 58.49 Samoab 3.76 46.74 59.79 Tonga 4.60 50.00 63.04 Australiab 14.35 51.19 60.29 Korea, Rep. 2.21 52.90 61.99 Japanb 5.25 59.73 64.49 Micronesia, Fed. Sts. 2.00 0.00 65.00 Singapore 9.52 59.16 65.70 Niue 9.68 0.00 66.53 Thailand 1.84 3.21 70.22 New Zealandb 11.61 61.40 74.45 Comparator economies Russian Federationb 1.37 17.73 40.48 Indiab 1.76 24.37 43.07 South Africab 3.78 33.40 45.68 Brazilb 2.07 21.18 63.15 Sri Lankab 3.79 59.85 73.57 Source: WHO 2013. a. Total tax includes excise taxes, import duties, value added tax, and other taxes. b. Country has increased tobacco excise taxes overall since 2010. However, because of price variability, the effect is not necessarily apparent in tax indicators. as Indonesia, Tonga, and Vietnam to develop a well-known cause of hypertension, is rela- similar legislation on so-called “sin taxes.” tively high in East Asia and Pacific (Intersalt Implementation of other interventions Cooperative Research Group 1988). Alcohol of the Framework Convention on Tobacco abuse is high in countries such as Mongolia Control such as smoke-free public places, and is associated with higher rates of liver dis- health warnings, and advertising bans also ease in the country. Undernutrition and low remains relatively weak in East Asia and birthweight, often a consequence of maternal Pacific, with the exception of a few coun- undernutrition, are associated with higher tries such as Australia, Korea, Malaysia, NCD-related risks later in life. Singapore, and Thailand. Samoa, Tonga, and Greater progress is needed on implement- Vanuatu each scored 0 of 5 on key aspects of ing effective health promotion and public tobacco regulation policies. Tonga, for exam- health interventions aimed at addressing poor ple, has a tobacco outlet for every 29 house- dietary quality, alcohol abuse, and lack of holds (WHO 2011b). physical activity. Robust evidence points to The majority of the priority actions identi- effective strategies in these areas, as summa- fied for the prevention and control of NCDs rized in table 6.6. In recent years, the Pacific are related to diets and nutrition. Malnutrition Island countries have made numerous efforts includes both over- and undernutrition. to address dietary risks through exercise cam- Overnutrition, because of unhealthy diets paigns at the community level as well as bans (for example, high consumption of salt and on the imports of specific food items with fats) and inadequate physical activity, and high fat content (Anderson 2013). As is usu- its consequences (for example, obesity) pose ally the case with behavior-related interven- significant health risks, including higher rates tions, however, progress in this area remains of diabetes, cardiovascular disease (CVD), slow. Headway in reducing low birthweight and cancer. For instance, salt consumption, also remains slow, not only in low-income P O P U L AT I O N AG I N G A N D H E A LT H S E R V I C E S I N E A S T A S I A A N D PAC I F I C 205 TABLE 6.6 Effective strategies for the management of key risks such as cardiovascular disease Areas needing intervention Recommended strategy Screening for and • Improved diagnosis rates of hypertension and high cholesterol management of high • Treatment with an essential package of four readily available and off-patent medicines: blood pressure and high aspirin, beta blockers, other antihypertensives such as angiotensin-converting enzyme cholesterol inhibitors or calcium channel blockers, and statins Screening for and • Improved diagnosis rates of diabetes in the population management of diabetes • Lifestyle interventions focused on diet and exercise patterns, including group-based activities • Drug therapy (metformin) Screening for cancer risks • Screening for and treatment of precancerous lesions to prevent cervical cancer and early treatment • Raising of awareness among women, promotion of self-exams, and mammograms for those with elevated risks Self-management • Patient education delivered in small groups by health professionals • Peer education programs to improve diet and physical activity • Structured phone support and telemonitoring • Online self-management programs and resources Medication adherence • Reduced out-of-pocket costs of medications • Introduction of programs such as reminder or calendar packaging combined with education and reminder strategies (Mahtani et al. 2011); use of mobile phones to send out reminders (Ettenhofer et al. 2009); and provision of new pill bottlesa a. New technologies such as pill bottles that glow if pills are not taken are designed to increase adherence rates by combating inattention. A randomized control trial is underway to evaluate their effectiveness (U.S. National Institutes of Health, Clinical Trials Registry, http://www.clinicaltrials.gov/ct2/show /NCT01798784). countries such as Myanmar and Timor-Leste, Cardiovascular disease but also among poorer populations in coun- Effective diagnosis and pharmacological tries such as Indonesia and Vietnam. management of those at high risk of devel- oping CVD played an important role in the cardiovascular revolution in the West. That Managing risk factors more effectively experience is a classic example of secondary A range of highly cost-effective interventions prevention: effective treatment interventions can be used to help people avoid developing are used to avoid complications from condi- disease (primary prevention) and avoid com- tions that have already developed. The use of plications from the disease once it has devel- inexpensive and readily available aspirin and oped (secondary prevention), particularly off-patent drugs such as statins (cholesterol- with regard to CVD and cancer. Early detec- lowering drugs critical for pharmacological tion and effective management of risk factors management of CVD-related risks) have been such as hypertension and diabetes, which identified as one of the best buys for address- can greatly complicate the treatment of CVD ing NCDs (Beaglehole et al. 2011). As shown later in life, are inexpensive and cost-effective in table 6.6, an essential package consists ways of controlling future health care costs. of four medicines: aspirin, beta blockers, This strategy is part of taking a life-course other antihypertensives such as angiotensin- approach to addressing the causes of illness converting enzyme inhibitors or calcium and disability in old age. Similarly, many channel blockers, and statins for reducing types of cancers are preventable, and disease LDL (low-density lipoprotein) cholesterol. is more amenable to treatment if diagnosed Extensive use of these medicines in primary and treated early. Table 6.6 summarizes cost- and secondary prevention explains more than effective strategies proposed in this section for half the reduction of CVD mortality in devel- management of CVD and other risks. oped countries in recent years (OECD 2007). 206 LIVE LONG AND PROSPER It should be considered a necessary strategy affordable under existing insurance and tax- today in responding to NCDs, in addition to financed programs. Because of inadequate other preventive measures. Proactive screen- health financing coverage, many elderly may ing, prompt treatment, and regular moni- go without the needed drugs. toring by qualified medical professionals, Effective diagnosis and management of combined with patient self-management, are diabetes are limited in the region for simi- all important. lar reasons. The complications arising from Yet such practices remain limited in East untreated diabetes require acute medical Asia and Pacific because of inadequate screen- care interventions (for example, dialysis and ing and monitoring in the health system, amputations) and result in severe disabilities. limited efforts at self-management, and high The evidence suggests that effective screening OOP costs. The available evidence points to and monitoring, early diagnosis and manage- high proportions of all of these conditions ment of diabetes with lifestyle interventions, going undiagnosed in most countries in the and improved self-management can bring region, and the likelihood of being diagnosed enormous cost savings. Even in cases where correlates with socioeconomic status and drugs are required to control blood sugar lev- education. Among those who are diagnosed, els, the costs are higher but still significantly medication adherence is low. Moreover, most less than the costs of treating the disease when countries have limited supply-side capacity to it progresses further. As an example, the costs screen people and carry out such secondary of treating diabetes in Vanuatu are presented prevention activities effectively (Mendis et al. in table 6.8. 2012). Although aspirin, antihypertensives, and statins are included in the standard pack- Cancer age of services in many countries (table 6.7), As the proportion of elderly increases, the their high cost means that many patients must majority of East Asian and Pacific countries pay out of pocket. Such OOP payments occur face the prospect of rapid increases in new even though more efficient procurement cancer cases over the next 10 to 20 years, not of these drugs could make them far more least because too little focus on or investment TABLE 6.7 Status of availability of essential package of cardiovascular drugsa in selected East Asian and Pacific countries Are drugs included in the standard benefits Are drugs free of charge in the public Country package? sector? Indonesia Drugs are included in the benefits package for Yes, in principle, but drugs are not always National Health Insurance. available. Malaysia Yes Yes Mongolia Yes If obtained at primary care level, drugs are discounted through Health Insurance Fund. Philippines Department of Health provides treatment packs Yes, no copayments are charged. (ComPacks) for hypertension and diabetes mellitus, but when the new health insurance benefit package takes effect, ComPacks will be discontinued. Sri Lanka Yes Yes Thailand Yes Yes Tonga Yes Yes Vanuatu Yes Yes Vietnam Drugs are included in principle, but they not always No, co-payments are charged (≥ 5 percent available, particularly at primary care level. depending on type of insurance). a. The package includes four key drugs: aspirin, beta blockers, other antihypertensives such as angiotensin-converting enzyme inhibitors or calcium channel blockers, and statins for reducing LDL (low-density lipoprotein) cholesterol. P O P U L AT I O N AG I N G A N D H E A LT H S E R V I C E S I N E A S T A S I A A N D PAC I F I C 207 TABLE 6.8 Costs of treating diabetes in Vanuatu, 2012 Type of cost Cost per patient per year (US$) Treatment of diabetes for a newly diagnosed patient at primary care level (includes 347 nurse-attended outpatient clinic visits once a month, glucose testing strips, minimal-level laboratory tests, and metformin tablets to control blood sugar levels) − Cost of minimal but essential drugs (metformin) 33 Treatment required as disease progresses (includes cost of regular insulin, increased 831 frequency of blood glucose testing, increased laboratory tests, syringes, and dressings) − Cost of insulin vials 262 Source: Anderson 2013. in cancer prevention and screening has increasing age, female gender, number of occurred to date (figure 6.12). Cancer screen- diagnoses, recent hospitalization, and depres- ing and early diagnosis and treatment are sion (Elliott 2006). In countries where cover- particularly important in aging populations. age is not universal, OOP payments for drugs However, most countries in the region have pose the biggest threat to medication adher- not yet put effective strategies in place. ence; evidence shows that OOP payments, no matter how small, reduce adherence to medi- cation regimes (Sidorenko and Zaidi 2013). Improving self-management and Anecdotal evidence from Mongolia shows medication adherence that medication for hypertension is often Improving self-management of conditions is taken only for a high reading of blood pres- an integral part of managing the progression sure and not otherwise. A range of interven- of chronic conditions, particularly in older tions to improve medical adherence through populations. The ability to self-manage—that self-management is also summarized in is, the patient and his or her family take a more table 6.6. proactive role in contributing to the patient’s health outcome in partnership with the health Transforming the service delivery care providers—is affected by access to health system care, education, socioeconomic status, and especially health literacy.4 In the United States Transforming service delivery to better meet and Australia, 70 to 80 percent of older peo- population needs calls for stronger care coor- ple are estimated to be at risk for low health dination throughout the health system. The literacy, while in the United Kingdom, one in reformed model would anticipate and shape three have problems reading and understand- patterns of care according to the projected ing basic health information (Renehan et al. health and medical needs of the popula- 2012). Self-management interventions for tion while placing considerable emphasis on chronic disease that are targeted at the elderly strengthening the role and raising the quality have proved effective in developed countries of primary care. It would include significant and are summarized in table 6.6. strengthening of community-focused care and Efforts to improve medication adherence vertical and horizontal integration (physical are particularly important for the manage- layout, information, and communication) of ment of hypertension and high cholesterol. facilities to provide comprehensive services In aging populations, cognitive decline and along a continuum or chain of care. The polypharmacy (having to take four or more model also would encompass formal outreach medications) usually lead to poor medication to long-term care providers5 as well as atten- adherence. For instance, older Australians tion to acute care management and rehabilita- take nine medications on average. The pre- tion and palliative care needs for populations dictors of multiple medication use include with a growing proportion of elderly. 208 LIVE LONG AND PROSPER FIGURE 6.12 The rising tide of cancer in East Asia and Pacific Projected percentage of new cases of cancers in the elderly, ages 65 and above, selected East Asian and Pacific countries and country groupings, 2012–35 Cambodia Singapore Vietnam Micronesia, Fed. Sts. Malaysia Vanuatu Phillippines Korea, Rep. Mongolia China Myanmar Thailand Papua New Guinia Indonesia Lao PDR Fiji Samoa Solomon Islands New Zealand Australia Timore-Leste Japan East Asia (by WHO definition) High-income countries Upper-middle-income countries Lower-middle-income countries Low-income countries 0 50 100 150 200 250 300 Projected percentage Source: World Health Organization, GLOBOCAN 2012, http://globocan.iarc.fr/Pages/burden_sel.aspx. Note: WHO = World Health Organization. Strong primary care is the backbone of primary care would be the supervisor for an effective delivery system and is the plat- directing patients through the delivery system, form for active coordination of care. Clear as shown in figure 6.13. In OECD countries, evidence exists that primary care contributes the primary and secondary prevention inter- to better health outcomes, higher-quality ventions described earlier (pharmacological and more appropriate care, improved access, management of hypertension and high cho- better prevention and early management of lesterol and screening and management of diseases, and cost containment (Macinko et diabetes) are increasingly located at the pri- al. 2010; Macinko, Starfield, and Shi 2003; mary level. Some countries have fostered new Saltman, Rico, and Boerma 2006; WHO family medicine and general practice models 2008). In a reformed service delivery sys- to facilitate their central role in coordinated tem that emphasizes coordination of care, care programs. Others, such as Germany, P O P U L AT I O N AG I N G A N D H E A LT H S E R V I C E S I N E A S T A S I A A N D PAC I F I C 209 FIGURE 6.13 A health system is transformed from a pyramidal model into a primary care–driven coordinated care model Diagnosis Trauma Call center Tertiary Primary Secondary Hospital Hospital health center Nursing Primary Materials home Transport Source: World Bank 2013a. TABLE 6.9 Key design features of an effective coordinated care strategy Area needing intervention Key design feature Reference population Definition of a target population (such as all diabetic patients in a region) that will be covered by the program or intervention Prioritization of health needs Intervention design based on the health needs of the population within a defined territory, with a focus on risks or pathologies Central role of primary care Primary care as the essential delivery platform, with regulated and structured access to other providers (for example, specialists, hospitals, diagnostics, and pharmacies) Continuum-of-care strategy Integration or linking of care (including promotion, prevention, treatment, rehabilitation, and palliation) among different levels and providers, definition of roles for each provider within an interdisciplinary framework, and use of evidence-based guidelines or clinical pathways involving multiple providers Organized decision-support Possible involvement of data management and communication systems for patient systems records, provider–patient interaction, and monitoring of provider performance and patient experience Patient and family participation Involvement of patients and family members in care management planning and implementation Special financing arrangements For some programs, financial incentives to providers for participation Source: World Bank 2013a. incentivize patients to register with primary design features of the reformed service deliv- care providers. ery model advocated in this report. Coordination of care involves linking pro- Coordination of care emerged in the fessionals and organizations at all levels of the OECD countries in response to pressures health system, emphasizing patient-centered similar to those currently faced by many care integration, managing patient referral East Asian and Pacific countries. These through the delivery system, and promoting pressures included cost inflation, growing follow-up care and the continuity of long- burden of chronic conditions, aging popu- term service provision. Table 6.9 lists the key lations, technological change, and rising 210 LIVE LONG AND PROSPER population expectations (Curry and Ham healthy life expectancy, and age at exit from 2010; Hofmarcher, Oxley, and Rusticelli the workforce. 2007; OECD 2010, 2012, 2013). Box 6.4 The good news is that a range of interven- outlines five of the main coordinated care tions can help slow functional decline among strategies adopted in OECD countries. The the elderly and save costs (table 6.10). For strategies are aimed at containing costs by instance, despite increases in aging, vision has reducing unnecessary use, frequent hospital- improved over time in Spain, Sweden, and ization, inappropriate admissions, and avoid- the United States as a result of effective and able readmissions. Box 6.5 summarizes the timely interventions (Christensen et al. 2009). evidence to date on the effect of coordination Interventions to prevent falls among the of care on key health system outcomes. elderly have been shown to produce positive public health benefits such as reduced costs of care and hospitalization, and treatment of Managing acute care episodes more osteoarthritis is known to improve quality of effectively life and functional capacity (Renehan et al. Even as the emphasis is correctly placed on 2012; WHO 2014b). shifting the balance of care from treatment Functional decline and NCD-related dis- to prevention and management of disease, abilities will mean increased demand for reha- the need to better manage the consequences bilitation and longer-term follow-up care, of CVD should not be ignored. Treatment which makes effective coordination of care of the consequences of CVDs accounted for critical. In East Asia and Pacific at present, 40 to 50 percent of the reduction in mortal- the majority of these services are provided ity in OECD countries. Acute medical treat- in acute care hospital settings. This pattern ment of the consequences of CVD will always of care is not only inefficient but also ineq- be needed, particularly in aging populations. uitable, because hospitals are used mainly Efforts must be made to ensure adequate by those who can afford such care. Effective staffing; well-equipped facilities; and adop- coordination of care can provide for post- tion of new, cost-effective treatment technolo- discharge planning and follow-up care for gies such as clot-dissolving drugs that are part chronic disease patients with complex needs of high-quality and effective cardiovascular and reduce the likelihood of hospital readmis- care. sions. By ensuring that chronically ill elderly patients receive the follow-up care they need in a primary care setting or at home instead of Preparing for functional and cognitive at the hospital, countries achieve cost savings, decline among the elderly and the patient’s quality of life is improved. The demographic shift in East Asia and Likewise, demand for palliative or end-of- Pacific will give rise to many of the degen- life care for cancers and other painful chronic erative diseases of older age, highlighting the conditions will increase, requiring coordi- importance of introducing timely and effec- nation of care between acute care and pal- tive interventions to slow the progression of liative care settings. In high-income countries, these conditions and ameliorate their effects. aggressive therapy beyond the point of a rea- These diseases include the sensory organ sonable chance of recovery is often expensive disorders of old age such as worsening eye- and unnecessary and not always consistent sight (cataracts and macular degeneration with the wishes of the patients themselves. emerge around age 60) and hearing, as well Although high-cost aggressive therapies to as weakening of bones and joints leading to treat cancer and other chronic conditions osteoarthritis and osteoporosis. Although not at the end of life are not a feasible option affecting mortality directly, these conditions for most people in middle- and low-income can have important effects on disability levels, East Asian and Pacific countries, spending P O P U L AT I O N AG I N G A N D H E A LT H S E R V I C E S I N E A S T A S I A A N D PAC I F I C 211 BOX 6.4 Care coordination strategies in Organisation for Economic Co-operation and Development countries Five main coordinated care strategies observed multidisciplinary team, is responsible for locat- in Organisation for Economic Co-operation and ing, monitoring, and coordinating care based on Development member countries are outlined below. a care plan. Some overlap occurs in program features across cat- • Community-based day centers for frail elderly, egories, and the list of strategies and initiatives is not which refers to special programs for frail elderly exhaustive. Many countries have launched coordinated and patients with multiple and usually disabling care efforts, but documentation is limited in terms of conditions. One-stop shops provide a full range of describing the what and how of these programs. health and social services and operate somewhat like geriatric day wards. Services are provided by Shifting care out of the hospital to ambulatory set- a multidisciplinary team of nurses, physicians, tings for shared management of patients with chronic nutritionists, social workers, therapists, and so diseases on and follow a care plan. Programs usually work • Relocation of clinicians from hospitals to primary closely with home support and assisted living ini- care settings. tiatives that aim to keep the elderly living in their • Incorporation of specialists into multidisciplinary communities rather than in institutional care. primary care teams: in some countries, this Fostering transitional care practices method is referred to as a multidisciplinary team approach to care management. • Postdischarge planning and follow-up care for • Educational outreach in which hospital-based spe- chronic patients with complex needs. These pro- cialists provide instruction and advice on screen- grams center on patients with multiple conditions ing, referrals, and treatment. who are discharged from hospitals. The model provides comprehensive medication manage- Shifting care from hospital and ambulatory settings ment, step-down rehabilitation, self-management to the home education for patients and families, and timely • Telemedicine and communication outreach between outpatient follow-up with primary care provid- outpatient clinics or hospitals and households. ers. Hospital and primary care providers work Interventions involve e-consultations, e-prescriptions, together to address clinical and social issues that e-diagnostics and monitoring, and e-reminders (the can affect the patient’s recovery. The objective is United Kingdom and the United States). to reduce readmissions and emergency room visits. • Outreach home visits and hospital-at-home pro- Establishing accountable primary care organizations grams, which consist of home visits by nurses or community health advisers for both chronic dis- • Patient-centered medical homes, defined as “an ease management and health promotion. This enhanced model of primary care that offers whole- approach can be used in combination with tele- person, comprehensive, ongoing, and coordinated medicine and electronic outreach interventions patient and family-centered care” (Kaye, Buxbaum, (for example, Sweden, the United Kingdom, and and Takach 2011). Initiated in the United States, the United States) (Health Foundation 2011). medical homes consist of an organized group of ambulatory providers that furnish comprehensive Targeting specific diseases or cases care to individuals, including the elderly and those requiring complex care. Patients register with a • Disease management programs that focus on primary care practice or center, which is respon- disease-specific objectives, treatment goals, and sible for coordinating care, navigating the delivery referral criteria for a defined population (for system, and arranging care as needed. Providers example, diabetic patients). Programs generally are financially accountable for furnishing quality target high-cost and high-risk patients with one care while containing costs. They generally present chronic condition. Interventions are oriented to a comprehensive menu of services, including diag- the entire course of a specific disease and usu- nostics and pharmacotherapies. ally are founded on evidence-based guidelines. A single care coordinator, usually part of a Source: World Bank 2013a. 212 LIVE LONG AND PROSPER BOX 6.5 The effect of care coordination on key health system outcomes Coordination of care has the following effects on key was conducted in person rather than by phone health system outcomes. (Sochalski et al. 2009). • When compared with the National Health Service Positive effect on patient functional status and out- (NHS) in the United Kingdom, Kaiser Permanente comes, but ambiguous effect on mortality performs better: of 11 medical conditions studied, • The effect is positive, especially when mul- the United Kingdom’s NHS uses 3.5 times the tiple strategies such as case conferences, care number of bed days as Kaiser for those ages 65 plans, and funding were used to coordinate care and above (Pike and Mongan 2014). (Ouwens et al. 2005; Powell-Davies et al. 2008). Mixed or insignificant effect on costs • Kaiser Permanente members also experience more comprehensive and convenient primary care ser- • A review of systematic reviews and meta-analysis vices and much more rapid access to specialist of the effectiveness of care programs in the United services and hospital when required (Pike and States and in the United Kingdom and other parts Mongan 2014). of the European Union found mixed or insignifi- cant evidence of integrated care resulting in cost Positive effect on efficiency through reduced savings (OECD 2012; Ouwens et al. 2005). hospitalizations • Medicare randomized trials showed that costs • In Medicare randomized trials, programs in were generally higher for patients enrolled in which case managers had substantial direct inter- the coordination arm of the trial compared with action with physicians and significant in-person patients in the control arm, although these dif- interaction with patients were more likely to ferences were not statistically significant in most reduce hospital admissions than programs with- instances (Peikes et al. 2009). out those features, but no program was likely to • Overall, evidence suggests that the effect on costs reduce overall costs (Peikes et al. 2009). depends on the types of diseases being treated • Review of effect of chronic care management and whether these conditions are associated with found that patients with heart failure were less high-cost unnecessary treatment or with under- likely to be hospitalized when care was pro- treatment before the integrated care initiatives vided by a multidisciplinary team and when (RAND Europe and Ernst & Young 2012). communication between providers and patients Source: OECD 2012. TABLE 6.10 Effective strategies for slowing functional decline among the elderly Area needing intervention Recommended strategy Worsening eyesight • Regular eye examinations, including interventions to diagnose and treat glaucoma and macular degeneration • Provision of cataract surgery (often on an outpatient basis), which can minimize vision impairment with older age Worsening hearing • Regular hearing tests with follow-up hearing support mechanisms such as hearing aids to help reduce hearing-related disabilities; tests often included in benefits packages in Organisation for Economic Co-operation and Development countries but rarely in East Asian and Pacific countries (WHO 2014b) Weakening bones and joints • Treatment of osteoarthritis using nonsteroidal and other pharmacotherapies • Improved calcium intake and exercise throughout life • Access to joint replacement operations to improve functional ability and independence • Introduction of fall interventions focused on both personal and environmental factors • Inclusion of information about personal interventions such as strength and balance training in routine health education programs for the elderly P O P U L AT I O N AG I N G A N D H E A LT H S E R V I C E S I N E A S T A S I A A N D PAC I F I C 213 their last few days or months in hospital is Assessment instruments that create often the only option available because hos- person-focused, longitudinal records that can pice and palliative care services barely exist. be viewed, understood, and used across the Eventually, systems in the region will need to continuum of care are vital for ensuring con- consider expanded coverage for hospice and tinuity of care for the elderly. The use of the palliative care, as well as enhanced case man- same protocols to support care planning in agement and web-based tools and informa- sectors between which people may move (for tion for end-of-life care. Good palliative care example, acute, community, and residential would include the use of assessment tools (see care) allow for a collaborative approach to below) that allow patients and their families service provision that spans the continuum of to decide on the level of comfort versus cure care. The focus is on recording information at the end of life (Lancet 2014). about the individual and his or her chang- Delaying the onset of cognitive function ing strengths, abilities, and preferences (for disorders and managing their symptoms will example, palliative care preferences) rather also require attention through integrated than the information required only by a par- health and social networks. There are few ticular care setting where that person hap- effective treatments for dementia, although pens to be at any given time. Currently, the the reduction of CVD risk factors has been interRAI suite of assessment instruments is identified as having a protective effect (WHO the only example of this type. This integrated 2012). Current expenditures on dementia- system enables transfer of common informa- related costs already amount to 1 percent of tion across care settings, thereby facilitating GDP worldwide and 1.24 percent of GDP continuity of care across transitions. A num- in high-income countries (Wimo and Prince ber of OECD countries are now introduc- 2010). Much of the care is provided through ing the interRAI instruments in response to family and social networks, but these are growing pressures on care delivery services straining with economic development. As (OECD and European Commission 2013).6 a result, ensuring that interventions are put The transformation of service delivery in place to diagnose early, delay onset, and described above is likely to require substan- manage symptoms through integrated health tial additional resources, particularly with and social networks is even more important. regard to health workforce requirements. As Similarly for depression, limiting psychosocial discussed, the current primary care system stressors through support and long-term care in most East Asian and Pacific countries is will become imperative. far from adequate to carry out the kinds of Stronger coordination of care is critical for functions described and will need complete improving access to care and saving costs in rebuilding. Strengthening the capacity of relation to cognitive function disorders, as and adequately resourcing the primary care well. In mental health services, a number of level will be essential for fostering and sus- OECD countries have seen a recent shift from taining coordinated care approaches. Many hospital-based care to coordinated commu- countries in the region such as China and nity care (for example, Australia, the Nordic Vietnam have organizational roadmaps in countries, and the United Kingdom) (OECD place for strengthening primary care as well 2013). Coordinated mental health care and as reform pilots, but scaling up of small-scale integrated treatment have led to improved reforms is yet to begin. Staff skills will need service delivery and better clinical outcomes. to be improved, and in some cases, a whole Following this approach, fewer readmissions, new cadre of staff members such as general reduced use of intensive care services, and practitioners and geriatricians may need to be contact with community crisis teams have created. resulted in cost savings (Rosenheck, Resnsick, The health infrastructure may eventually and Morrissey 2003; Stewart et al. 2012; be transformed, too. For the service delivery Woods and McCollam 2002). model described earlier, smaller hospitals and 214 LIVE LONG AND PROSPER intermediate facilities (ambulatory surgical allocations to the health sector. Health system units, day hospitals, and specialty clinics) and inefficiencies are already exerting pressure larger primary care centers combined with on tax and social health insurance revenues. greater use of communication technologies Service delivery reforms are thus critical to for community, long-term care, and home ensure the sustainability of health systems outreach would be more appropriate. As is for aging populations and justify additional the case in several OECD countries, hospitals investment. themselves may need to adapt more flexible In the short to medium term, significant designs that respond to changing demand and efficiency savings could be gained in three technological change (Rechel et al. 2009). All areas: (a) paying providers, (b) purchasing of these changes will require substantially pharmaceuticals, and (c) prioritizing among higher spending on health. available interventions and technologies. Stronger purchasing capacity is a cross- cutting issue. The key strategies in these areas Getting better value for money from are summarized in table 6.11. health systems Strengthening provider payment methods Because mobilizing new resources to meet all to align providers’ incentives with deliver- of the additional requirements of service deliv- ing good-quality, effective care is essential in ery reforms would simply not be sustainable, East Asia and Pacific. Optimal systems do generating efficiency savings to gain better not exist, and provider payment systems and value for money from existing health sys- incentives need to be designed to address the tems is imperative in the region. Although the specific policy issues and objectives inherent reformed service delivery model will generate in a country’s health sector. However, regional significant efficiency savings over time, the and global experience provide a few key initial investment costs are likely to be sub- trends that the region’s countries can draw stantial. Meanwhile, other policy priorities— on, as described below. not least of which is the policy commitment For physician services, three provider to achieve universal health coverage—will payment models tend to dominate: salary, continue to place huge demands on East capitation, and FFS, or some combination Asian and Pacific governments’ budgetary of the three. The specific FFS model used in TABLE 6.11 Effective strategies for getting better value for money from health systems Area needing intervention Recommended strategy Paying providers • Movement away from line-item budgets and fee-for-service payments • Introduction of capitation payment systems for physicians and primary care providers and global budgets with case mix–adjusted payments for hospitals • Consideration of bundled payments to providers for specific medical conditions (for example, diabetes) Purchasing pharmaceuticals • Reform of drug procurement and movement toward smart strategic purchasing • Risk-sharing arrangements, such as price-volume arrangements, for high-cost drugs • Sole-source tendering in which the winning bidder is sole supplier for a fixed term • Therapeutic reference pricing Prioritizing among available • Institution of clear, transparent processes to make decisions about new technologies interventions and technologies and drugs • Use of the findings from more established health technology assessment (HTA) agencies to inform their own prioritization process (for example, de facto HTA) Ensuring strategic purchasing • Movement toward a single purchaser capacity • Building of strategic purchasing capacity • Investment in health information management systems P O P U L AT I O N AG I N G A N D H E A LT H S E R V I C E S I N E A S T A S I A A N D PAC I F I C 215 Japan; Korea; and Taiwan, China, has been payment categories within the hospital sec- successful in avoiding the cost inflation asso- tor. In Western Europe, most countries have ciated with FFS payment in many settings. moved to a performance-based approach using Capitation payment has emerged as more some combination of payment per admission effective in controlling costs, particularly based on case mix–adjusted diagnosis-related in relation to primary care (see box 6.6 for groups. Turkey adopted a global budget for Thailand’s experience with capitation). all ministry hospitals, which has helped con- For inpatient care services, countries trol spending growth (see box 6.6). have been moving toward some combina- High-income countries are now experi- tion of global budgets and case mix–adjusted menting with bundled payments to providers BOX 6.6 Cost containment through provider payment and purchasing reforms in Thailand and Turkey A single purchaser, virtually have continued to rise. The cost per member rose from B 1,201.40 in 2002 to B 2,693.50 in 2011, The National Health Security Office (NHSO) is the or from a total of less than B 60 billion in 2002 to largest purchaser in Thailand, purchasing services over B 120 billion in 2011. This is equivalent to a on behalf of 75 percent of the population (approxi- 70 percent increase in real terms over the period. The mately 50 million Universal Coverage Scheme mem- increase is attributed to rising remuneration of health bers). NHSO has adequate strategic purchasing care staff, particularly a rapid increase in extra incen- capacity and has used its bargaining power to reduce tive payments to keep highly skilled professionals in the price of medicines, medical products, and inter- the system. ventions over time. For instance, the price of hemo- dialysis decreased from US$67 to US$50 per cycle, Performance-based supplementary payment system leading to savings of US$170 million per year. During the 2000s, Turkey undertook comprehen- Capitation payment and diagnosis-related groups sive hospital reforms in the context of the Health with a global budget Transformation Program, which was geared toward The combination of mixed payment methods (capita- improving effectiveness, equity, and efficiency. A key tion for outpatient care and diagnosis-related groups reform was the introduction of a performance-based for inpatient care) with a global budget for inpatient supplementary payment system for employees of care has further contributed to cost containment in Ministry of Health hospitals beginning in 2004, Thailand. The closed-end payment system has pro- based on specified individual and institutional crite- vided strong signals to providers to contain costs by ria. The changes in hospital financing were facilitated prescribing generic medicines, dispensing medical by reforms that brought together all previous health technologies appropriately, and encouraging pre- insurance schemes under the umbrella of the Social ventive treatments. At the same time, to counteract Security Institution. Furthermore, expenditure caps underprovision of services, NHSO has established for ministry, private, and university hospitals have procedures such as (a) a 24-hour complaint manage- been introduced in recent years to ensure controlled ment mechanism operated by NHSO staff, (b) quality growth in hospital spending in line with gross domes- assurance through hospital accreditation processes, tic product. In 2007, a fixed global budget for all (c) routine auditing by random medical audit teams ministry hospitals was implemented, which has been with financial penalties, (d) utilization reviews to successful in containing further spending growth in monitor utilization rates, and (e) an annual poll sur- these hospitals. Since 2010, the government has also vey of consumers’ and providers’ satisfaction con- successfully maintained spending with agreed limits ducted by independent polling institutes. for university and private hospitals. Despite those measures, one should note that overall Universal Coverage Scheme expenditures Source: Hanvoravongchai 2013; World Bank 2013b, 2013c. 216 LIVE LONG AND PROSPER for specific procedures or medical conditions, pays for any volume over the agreed threshold an approach that holds potential for aging (for example, Korea); sole-source tendering in populations (Mandeville and Sinnott 2014; which the winning bidder is the sole supplier Painter, Burns, and Bailit 2012). Under this for a fixed term; and therapeutic reference method, fixed payments cover a set of services pricing in which the purchasers set fixed reim- for a defined time period, usually including bursement levels for the price of a drug by ref- both inpatient and outpatient care. Bundled erencing a base drug within that therapeutic payments for chronic conditions such as dia- class (Mandeville and Sinnott 2014). Box 6.7 betes have created incentives for primary and compares purchasing strategies in Korea, Sri secondary care providers to move care out of Lanka, and Thailand with the highly ineffi- the hospital. High-quality primary care has cient purchasing arrangements in Vietnam. prevented expensive inpatient stays for avoid- Strong purchasing capacity is as critical for able complications. undertaking smart purchasing of drugs as it Finally, the presence of a large purchaser is for undertaking provider payment reforms. with adequate strategic purchasing capacity When new technologies or drugs become has generally proved critical for cost contain- available, most countries—particularly ment. Being a large, preferably single pur- those with social health insurance–financed chaser gives the purchaser greater bargaining systems—grapple with decisions about power to negotiate prices with providers and including them in the benefits package and to control utilization and quality. at what level of patient cost sharing. As dis- Inefficiencies in purchasing pharmaceu- cussed, the differential adoption and dif- ticals result in high prices being paid for fusion of costly product innovations in the drugs, suggesting opportunity costs in terms health sector in OECD countries explains the of improving cost-effectiveness and reducing increase in per capita costs of elderly across co-payments. Table 6.12 compares the prices countries, and not aging itself. Costs have of common statins in selected countries. For been contained far more effectively in coun- example, for off-patent common statins (ator- tries such as Japan that have been more suc- vastatin and simvastatin), the price in Vietnam cessful in rationing the growth of technology is more than twice that in New Zealand. compared to the United States, for example. East Asian and Pacific countries can learn Ideally, decisions regarding new technologies from pharmaceutical purchasing strategies and drugs should be based on objective cri- that have been used in countries within and teria such as cost-effectiveness, efficacy, and outside the region to reduce drug prices. These potential effect on the status of budgets and strategies include risk-sharing arrangements health insurance funds and should be deter- for high-cost drugs such as price- volume mined through a transparent process. In arrangements, whereby the manufacturer practice, decision making is a priority-setting TABLE 6.12 Price comparisons of cholesterol-lowering drugs in selected countries Drug Atorvastatin Pravastatin Lovastatin Simvastatin Quantity 30 × 20 mg 30 × 20 mg 30 × 20 mg 20 × 20 mg United Kingdom, retail price (US$) 2.74 2.47 — 1.53 New Zealand, retail price (US$) 1.10 4.31 — 0.53 Sri Lanka, retail price (US$) 0.17 a 0.20 — Thailand, retail price (US$) 2.89 a a 0.93 Vietnam, retail price (US$) 2.41 — 1.27 2.52 Sources: Data from U.K. National Health Service (NHS); New Zealand Pharmaceutical Management Agency, Sri Lanka Ministry of Health, and Vietnam Social Security Agency. Note: mg = milligram; — = not available. Drug prices listed are those in public system in each country (for example, NHS in the United Kingdom, Ministry of Health and Indigenous Medicine in Sri Lanka, National Health Security Office in Thailand, and Vietnam Social Security in Vietnam). a. This drug is not included in the drug list. P O P U L AT I O N AG I N G A N D H E A LT H S E R V I C E S I N E A S T A S I A A N D PAC I F I C 217 BOX 6.7 The power of smart purchasing: Pharmaceutical procurement in the Republic of Korea, Sri Lanka, and Thailand versus Vietnam Smart purchasing monopsony power to negotiate prices with service providers and suppliers. For high-price pharma- • Republic of Korea . Since 20 06, the Korea ceuticals, a central price negotiation system is National Health Insurance Service has had a used to collectively bargain for the best-priced positive list of drugs in the benefits package and items. The Universal Coverage Scheme previously provides two distinct steps for addition of new also used the threat of compulsory licensing of drugs in the list. When a new drug is imported medicines to obtain cheaper prices for drugs that or produced, the Health Insurance Review and are under patent. The National Health Security Assessment Service makes a decision to list the Office is estimated to have saved B 12.5 billion drug based on cost-effectiveness, clinical effec- through cost-control measures, including savings tiveness, and budget effect and whether and at on pharmaceuticals. what price the drug is reimbursed in other coun- tries. The National Health Insurance Service then Not-so-smart purchasing negotiates the prices of the drug directly with pharmaceutical manufacturers using a price- • Vietnam. The current procurement system is volume arrangement. The price is negotiated on highly decentralized and complex, involving more the basis of expected sales as well as the substitu- than 1,000 entities. It results in wide differentials tion effect. If actual sales exceed expected sales in prices of medicines, often for the same type, during a specific period, the price is reduced pro- dosage, and formulation across hospitals and sup- portionately. The decision to include the drug in plies. The procurement system is also character- the positive list is made at the end of the process. ized by irregularities that contravene procurement • Sri Lanka. The Ministry of Health undertakes regulations, which mandate awarding of the con- centralized procurement for drugs on behalf of all tract to suppliers with the lowest prices. Members public sector hospitals and facilities, which gives of hospital procurement committees report that it strong monopsony power. In addition, procure- winning bids are often based on doctors’ requests ment is done globally. By seeking global markets and perceptions of quality and efficacy or on hos- instead of regional markets alone, the ministry pital management decisions. Doctors’ perceptions is able to secure the best prices. Procuring drugs are influenced by commissions from pharmaceuti- at exceptionally low prices has enabled it to pro- cal companies. An audit found that tender prices vide universal coverage of the essential package were 47 to 357 percent higher than the actual of drugs needed to manage cardiovascular disease winning bids and that prices of winning bids were risks (antihypertensives, statins, and so on) at no often higher than prices offered by unsuccessful cost to patients in the public sector since 2012. bids. In Hanoi, the prices of winning bids are • Thailand. The pharmaceutical benefits package is 130 to 245 percent higher than the import prices, based on the National Essential Drug List, which including cost, insurance, and freight costs. includes medicines that are selected on the basis Sources: Hanvoravongchai 2013; Lee 2013; personal communication in 2014 of effectiveness, safety, and cost-effectiveness. with Ravindra P. Rannan-Eliya, Institute for Health Policy, Sri Lanka; Somanathan The Universal Coverage Scheme also has strong et al. 2014. exercise that involves value judgments and is This approach is a systematic and transparent influenced by a mix of political, social, and appraisal and deliberation process for making economic factors. decisions on public reimbursement of medical Many OECD countries have introduced technologies, devices, and procedures. It can explicit prioritization systems—health technol- be very effective; for instance, New Zealand’s ogy assessments (HTAs)—that transparently PHARMAC (Pharmaceutical Management assess the value for money of new technologies. Agency) has been highly successful at keeping 218 LIVE LONG AND PROSPER FIGURE 6.14 PHARMAC has positively affected drug expenditure over time in New Zealand Actual drug spending versus projected drug spending at 2000 subsidy levels, 2000–2015 3,500 Forecast Drug cost (millions of New Zealand dollars) 3,067 3,000 2,895 2,635 2,500 2,336 2,000 1,788 1,590 1,436 1,500 1,248 1,099 898 993 1,000 844 671 723 517 586 694 706 777 784 804 822 500 640 653 517 516 504 510 536 567 566 602 0 00 01 02 03 04 05 06 07 08 09 10 11 12 13 14 15 20 20 20 20 20 20 20 20 20 20 20 20 20 20 20 20 Fiscal year (ending June 30) Actual expenditure Estimated expenditure at 2000 subsidies Source: PHARMAC 2012. Note: PHARMAC = Pharmaceutical Management Agency. costs down over a 12-year period by combin- services and setting co-payment rates for ing HTAs with strengthened pharmaceutical different types of services, providers, and procurement (figure 6.14) (Mandeville and patients. Thailand’s Universal Coverage Sinnott 2014). Scheme and Korea’s National Health In East Asia and Pacific, two strategies Insurance program provide good examples can be pursued to support the prioritization of this process (box 6.8). of new technologies and drugs: institute a transparent process for prioritization and use Use findings from more established HTA findings from more established HTA agen- agencies to inform prioritization processes cies to inform prioritization processes. These Kosovo, Romania, and Serbia looked to approaches can be followed even if HTA the experiences of other HTA agencies. For capacity has not yet been established because example, they used a de facto HTA method by developing national capacity to undertake drawing on recommendations of the United HTAs is costly and can take several years. Kingdom’s HTA agency (National Institute for Health and Clinical Excellence) to assess Institute a transparent process for the cost-effectiveness of their national for- prioritization mularies. Using this method, Serbia was able This strategy typically involves a committee to renegotiate prices on a range of drugs and consisting of, among others, providers and save millions of euros. researchers who are charged with making recommendations based on clear and agreed- Preparing communities and institutions upon criteria such as cost- effectiveness, outside of the health system medical necessity, and financial burden on patients. The committee and decision process Several interventions outside of the formal can be used not only for inclusion of new health care delivery system can help pre- interventions, but also for delisting existing pare society for the emerging needs of an P O P U L AT I O N AG I N G A N D H E A LT H S E R V I C E S I N E A S T A S I A A N D PAC I F I C 219 BOX 6.8 Good examples of decision-making processes for new technologies in East Asia and Pacific Thailand Republic of Korea The Thailand National Health Security Office’s The National Health Insurance program carries out (NHSO) Committee on Benefits Package is charged the following steps to add new services and technolo- with revising the benefits package and making rec- gies to the benefits package: ommendations to NHSO on the adoption of new • Consumer or provider groups make a request to drugs and technologies based on guidelines estab- include a new health technology or intervention. lished in 2010. NHSO makes requests to the Health • The request is reviewed by expert committees at Intervention and Technology Assessment Program the National Health Insurance Service and Health and the International Health Policy Program—two Insurance Review and Assessment Service on the technical agencies working on health technology basis of various criteria (clinical effectiveness, assessment and health system evaluation under cost-effectiveness, number of patients affected, the Ministry of Public Health—to supply evidence financial out-of-pocket burden, and budget effect). such as the effectiveness and cost-effectiveness of • The final decision is made by the Health Insurance various health interventions that will be considered Policy Committee (a tripartite committee consist- for expansion of the benefits package. Financial ing of consumers, providers, and the government feasibility, budgetary effect, and ethical consider- or public sector). ations are among the criteria applied. The sche- matic diagram of the decision process is provided Sources: Hanvoravongchai 2013; personal communication from Soonman in figure B6.8.1. Kwon, Seoul National University, 2014. FIGURE B6.8.1 National Health Security Office’s decision process for new technologies involves extensive criteria in Thailand Seven groups of stakeholders Topic • Number of people affected Submitted topics nomination • Disease/health problem severity Secretariat • Effectiveness of technologies Preliminary assessment • Variation in practice of each submitted topic • Financial impact to the Topic households Working group on health topic selection prioritization • Equity/ethical for assessment implications (affecting Prioritized topics the poor and for rare diseases) Health technology assessment researchers Technology • Cost-effectiveness Health technology assessment assessment • Budget impact results/preliminary recommendations Subcommittee on development of health benefit package and services system of the National Health Security Office Appraisal Recommendations National Health Decision Security Office board making 220 LIVE LONG AND PROSPER increasingly older population (table 6.13). For often have invested time and money and who instance, ensuring ready availability of assis- hold a detailed understanding of systems and tive devices and aids such as upper and lower practices—and attract others will have a com- limb prostheses, assistive devices for toilet petitive advantage. An important factor in use and bathing, walking sticks and frames, increasing the motivation and thus retention wheelchairs, lifting hoists (for lifting people of mature workers will be the work environ- in and out of bed, for example), and adap- ment itself. Even if workers have the ability tive transport are key items that can assist in and motivation to work, a poorly adapted mobility and increase independence of older workplace culture will discourage retention people with disabilities. In addition, installing of mature workers (Eide, Krause, and Rosas environmental features (for example, wheel- 2001; Seike, Biggs, and Sargent 2011). chair ramps; wheelchair-friendly transport, homes, offices, and buildings; ample seating; and safe footpaths) can enable older people Conclusion with disabilities to circulate within their com- The range of recommended policy measures munities (WHO 2014b). Finally, improve- could help East Asian and Pacific countries ments in public spaces could encourage address their health care challenges in a fis- walking and other physical activities, which cally sustainable manner and improve out- not only reduce the risk of chronic disease, comes for the elderly. As discussed earlier, but also exert protective effects by strengthen- even without the effects of aging, the expected ing the physiological systems of older adults increase in disease burden attributable to and reducing functional limitation such as NCDs and the inefficiencies in the financ- that from osteoarthritis (Beard et al. 2011). ing and delivery of health services are likely Equally important are changes to organi- to drive up health care costs and undermine zational culture in the workplace to accom- fiscal sustainability. The various measures modate an increasingly older workforce. This proposed in this chapter would help address includes the social environment within busi- NCD risks in a highly cost-effective manner nesses, institutions, and service agencies. As and improve the effectiveness and efficiency older workers form a larger proportion of of health care service delivery, leading to bet- a shrinking workforce, businesses that can ter health outcomes for both the elderly and keep these existing workers—in whom they the broader population. At the same time, TABLE 6.13 Effective strategies outside of the health system to prepare for functional and cognitive decline in old age Area needing intervention Recommended strategy Access to assistive devices • Ensure assistive devices are widely available and affordable, potentially at subsidized prices. Access to recreation and public • Improve public spaces: wheelchair-friendly transport systems and ramps, ample spaces seating, safer footpaths, welcoming neighborhood shopping areas, and easily accessible parks. Improvement of the workplace • Introduce health promotion programs when health benefits are linked to work environment employment (for example, supervised fitness programs, smoking cessation, nutritional and improved dietary intake, and return-to-work programs). • Target health promotion for mature workers, taking into account relevant issues such as gender (for example, wellness programs that account for the needs of mature women, such as osteoporosis prevention). • Consider issues related to occupational demands such as balance, flexibility, and sensory requirements (for example, improved lighting, larger signage, and volume- adjusted communication technologies) (Eide, Krause, and Rosas 2001). P O P U L AT I O N AG I N G A N D H E A LT H S E R V I C E S I N E A S T A S I A A N D PAC I F I C 221 pressure on the health care system could also ———. 2012. The Cost Disease: Why Computers be relieved by reducing the current default Get Cheaper and Health Care Doesn’t. New reliance on health systems for long-term care Haven, CT: Yale University Press. of the elderly, which is discussed in detail in Beaglehole, Robert, Ruth Bonita, Richard Horton, Cary Adams, George Alleyne, Perviz Asaria, chapter 7 of this report. Vanessa Baugh, Hank Bekedam, Nils Billo, Sally Casswell, Michele Cecchini, Ruth Colagiuri, Stephen Colagiuri, Tea Collins, Notes Shah Ebrahim, Michael Engelgau, Gauden 1. Interestingly, the percentage of elderly who Galea, Thomas Gaziano, Robert Geneau, reported having an exam was considerably Andy Haines, James Hospedales, Prabhat Jha, higher among the one-fourth of those ages Ann Keeling, Stephen Leeder, Paul Lincoln, 60 and older who participated in an elderly Martin McKee, Judith Mackay, Roger club during the past year, especially at more Magnusson, Rob Moodie, Modi Mwatsama, advanced ages. Sania Nishtar, Bo Norrving, David Patterson, 2. M. L. Sang provided information based on Peter Piot, Johanna Ralston, Manju Rani, K. research carried out for preparation of the Srinath Reddy, Franco Sassi, Nick Sheron, Vietnam Health Professional Education and David Stuckler, Il Suh, Julie Torode, Cherian Training project. Varghese, and Judith Watt. 2011. “Priority 3. Members of the EU-27 are Austria, Belgium, Actions for the Non-communicable Disease Bulgaria, Cyprus, the Czech Republic, Crisis.” Lancet 377 (9775): 1438–47. Denmark, Estonia, Finland, France, doi:1016/S0140-6736(11)60393-0. Germany, Greece, Hungary, Ireland, Italy, Beard, John R., Alexandra Kalache, Mario Latvia, Lithuania, Luxembourg, Malta, the Delgado, and Terry Hill. 2011. “Aging and Netherlands, Poland, Portugal, Romania, the Urbanization.” In Global Population Ageing: Slovak Republic, Slovenia, Spain, Sweden, Peril or Promise?, edited by John R. Beard, and the United Kingdom. Simon Biggs, David E. Bloom, Linda P. Fried, 4. Health literacy refers to an individual’s Paul Hogan, Alexandra Kalache, and S. Jay capacity to seek, understand, and utilize health Olshansky, 93–96. Geneva: World Economic information to make informed decisions Forum. about his or her own health (U.S. Department Bloom, David E., Elizabeth T. Cafiero, Mark of Health and Human Services 2000). E. McGovern, Klaus Prettner, Anderson 5. The service delivery reforms proposed here Stanciole, Jonathan Weiss, Samuel Bakkila, draw extensively from work on service and Larry Rosenberg. 2013. “The Economic delivery reform in China led by the World Impact of Non-communicable Disease in Bank from 2013 to 2015. The final report of China and India: Estimates, Projections, and this study will be published in 2016 by the Comparisons.” NBER Working Paper 19335, World Bank. National Bureau of Economic Research, 6. For more information, see the interRAI Cambridge, MA. website at http://interrai.org/index.php?id=7. CHARLS (China Health and Retirement Longitudinal Study). Various years. National School of Development, Peking University. References http://charls.ccer.edu.cn/en. Christensen, Kaare, Gabriele Doblhammer, Roland Anderson, Ian. 2013. “The Economic Costs of Rau, and James W. Vaupel. 2009. “Aging Non-communicable Diseases in the Pacific Populations: The Challenges Ahead.” Lancet Islands: A Rapid Stocktake of the Situation 374 (9696): 1196–208. in Samoa, Tonga, and Vanuatu.” Health, Currie, Janet, Wanchuan Lin, and Juanjuan Meng. Nutrition, and Population Discussion Paper, 2014. “Addressing Antibiotic Abuse in China: World Bank, Washington, DC. An Experimental Audit Study.” Journal of Baumol, William J. 1996. “Children of Performing Development Economics 110: 39–51. Arts, the Economic Dilemma: The Climbing Curry, Natasha, and Chris Ham. 2010. “Clinical Costs of Health Care and Education.” Journal and Service Integration: The Route to of Cultural Economics 20 (3): 183–206. Improved Outcomes.” King’s Fund, London. 222 LIVE LONG AND PROSPER Cutler, David M., Angus S. Deaton, and Adriana Hanvoravongchai, Piya. 2013. “Health Financing Lleras-Muney. 2006. “The Determinants of Reform in Thailand: Toward Universal Mortality.” NBER Working Paper 11963, Coverage under Fiscal Constraints.” UNICO National Bureau of Economic Research, Study 20, Universal Health Coverage, World Cambridge, MA. http://www.nber.org/papers Bank, Washington, DC. http://documents. /w11963. worldbank.org/curated/en/2013/01/17208145 Cutler, David M., and Mark McClellan. 2001. /thailand-health-financing-reform-thailand “Is Technological Change in Medicine Worth -toward-universal-coverage-under-fiscal It?” Health Affairs 20 (5): 11–29. doi:10.1377/ -constraints. hlthaff.20.5.11. Health Foundation. 2011. “Evidence: Getting Dodge, Brooks A. 2008. Primary Healthcare Out of Hospital? The Evidence for Shifting for Older People: A Participatory Study in Acute Inpatient and Day Case Services from Five Asian Countries. Chiang Mai, Thailand: Hospitals into the Community.” Health HelpAge International Asia and Pacific Foundation, London. Regional Development Centre. Higuchi, Michiyo. 2010. “Access to Diabetes Care Dormont, Brigitte, Michel Grignon, and Helene and Medicines in the Philippines.” Asia-Pacific Huber. 2006. “Health Expenditure Growth: Journal of Public Health 22 (3): 96S–102S. Reassessing the Threat of Aging.” Health Hofmarcher, Maria M., Howard Oxley, and Economics 15 (9): 947–63. Elena Rusticelli. 2007. “Improved Health Eide, Asbøjrn, Catarina Krause, and Allan Rosas. System Performance through Better Care 2001. Economic, Social, and Cultural Rights: Coordination.” OECD Health Working Paper A Textbook. 2nd ed. Dordrecht, Netherlands: 30, Organisation for Economic Co-operation Martinus Nijhoff. and Development, Paris. Elliott, Rohan A. 2006. “Problems with HSPI (Health Strategy and Policy Institute). 2010. Medication Use in the Elderly: An Australian “Hospital Overcrowding and Undercapacity: Perspective.” Journal of Pharmacy Practice Issues and Solutions.” Study report, Hanoi, and Research 36 (1): 58–66. Vietnam. Ettenhofer, Mark, Charles H. Hinkin, Steven A. IFLS (Indonesia Family Life Survey). 2007. RAND, Castellon, Ramani Durvasula, Jodi Ullman, Santa Monica, CA. http://www.rand.org/labor Mona Lam, Hector Myers, Matthew J. /FLS/IFLS.html. Wright, and Jessica Foley. 2009. “Aging, IHME (Institute for Health Metrics and Neurocognition, and Medication Adherence in Evaluation). 2013. The Global Burden of HIV Infection.” American Journal of Geriatric Disease: Generating Evidence, Guiding Policy. Psychiatry 17 (4): 281–90. Seattle, WA: IHME. Feng, Jin, Pingyi Lou, and Yangyang Yu. 2015. Intersalt Cooperative Research Group. 1988. “Health Care Expenditure over Life Cycle “Intersalt: An International Study of in the People’s Republic of China.” Asian Electrolyte Excretion and Blood Pressure: Development Review 32 (1): 167–95. Results for 24-Hour Urinary Sodium and Fuchs, Victor. 1998. “Provide, Provide: The Potassium Excretion.” British Medical Journal Economics of Aging.” NBER Working Paper 297 (6644): 319–28. 6642, National Bureau of Economic Research, Jeong, Hyoung-sun, and Ravi P. Rannan- Cambridge, MA. Eliya. 2010. “SHA-Based Health Accounts Getzen, Thomas E. 2000. “Health Care Is an in Twelve Asia-Pacific Economies: A Individual Necessity and a National Luxury: Comparative Analysis.” SHA Technical Paper Applying Multilevel Decision Models to 10, Health and Social Policy Programme, the Analysis of Health Care Expenditures.” Organisation for Economic Co-operation Journal of Health Economics 19 (2): and Development, Paris, and Korea Policy 259–70. Centre, Seoul. Hagist, Christian, and Laurence Kotlikoff. Jones, Charles I. 2002. “Why Have Health 2005. “Who’s Going Broke? Comparing Expenditures as a Share of GDP Risen So Growth in Healthcare Costs in Ten OECD Much?” NBER Working Paper 9325, National Countries.” NBER Working Paper 11833, Bureau of Economic Research, Cambridge, MA. National Bureau of Economic Research, JSTAR (Japanese Study of Aging and Retirement). Cambridge, MA. 2011. Research Institute of Economy, Trade, P O P U L AT I O N AG I N G A N D H E A LT H S E R V I C E S I N E A S T A S I A A N D PAC I F I C 223 and Industry, Hitotsubashi University and Packaging for Improving Adherence to Self- University of Tokyo, Tokyo. http://www.rieti Administered Long-Term Medications.” .go.jp/en/projects/jstar/. Cochrane Database of Systematic Reviews Kaye, Neva, Jason Buxbaum, and Mary Takach. 2011, 7 (9): CD005025. 2011. “Building Medical Homes: Lessons Mandeville, Kate, and Emily Sinnott. 2014. from Eight States with Emerging Programs.” “Healthier Lives in Aging Societies.” Commonwealth Fund, New York, and Background paper for the East Asia and National Academy for State Health Policy, Pacific regional report on aging, World Bank, Washington, DC. Washington, DC. Knodel, John, Vipan Prachuabmoh, and Napaporn McCollum, Rosalind, Lieping Chen, Tang Chayovan. 2013. The Changing Well-Being of Chenxiang, Xiaoyun Lui, Barbara Starfield, Thai Elderly: An Update from the 2011 Survey Zheng Jinhuan, and Rachel Tolhurst. 2014. of Older Persons in Thailand. Chiang Mai, “Experiences with Primary Healthcare in Thailand: HelpAge International. Fuzhou, Urban China, in the Context of Health Lancet. 2014. “Palliative Care: More than One Sector Reform: A Mixed Methods Study.” Chance to Get It Right.” 384 (9938): 103. International Journal of Health Planning and L a n g e n b r u n n e r, J o h n C . , a n d A p a r n a a Management 29 (2): e107–26. Somanathan. 2011. Financing Health Care Meerding, Willem J., Johan Polder, Luc Bonneux, in East Asia and the Pacific: Best Practices Marc Koopmanschap, and Paul van der Maas. and Remaining Challenges. Washington, DC: 1998. “Health-Care Costs of Ageing.” Lancet World Bank. 351 (9096): 140–41. Lee, T.-J. 2013. “Pharmaceutical Price Setting Mendelsohn, Daniel N., and William B. Schwartz. in Vietnam.” Presented at the workshop on 1993. “The Effects of Aging and Population Vietnamese Health Insurance Law Revision, Growth on Health Care Costs.” Health Affairs Vinh Phuc, Vietnam, June 17–18. 12 (1): 119–25. Leung, Gabriel, Keith Y. K. Tin, and Wai-Sum Mendis, Shanthi, Igbal Al Bashir, Lanka Chan. 2007. “Hong Kong’s Health Spending Dissanayake, Cherian Varghese, Ibitihal Projections through 2033.” Health Policy 81 Fadhil, Esha Marhe, Boureima Sambo, Firdosi (1): 93–101. Mehta, Hind Elsayad, Idrisa Sow, Maltie Lu, Fangwen. 2014. “Insurance Coverage and Algoe, Herbert Tennakoon, Lai Die Truong, Le Agency Problems in Doctor Prescriptions: Thi Tuyet Lan, Dismond Huiuinato, Neelamni Evidence from a Field Experiment in China.” Hewageegana, Naiema A. W. Fahal, Goitom Journal of Development Economics 106: Mebrhatu, Gado Tshering, and Oleg Chestnov. 156–67. 2012. “Gaps in Capacity in Primary Care in Lubitz, James D., and Gerald F. Riley. 1993. Low-Resource Settings for Implementation “Trends in Medicare Payments in the Last year of Essential Noncommunicable Disease of Life.” New England Journal of Medicine Interventions.” International Journal of 328 (15):1092–96. Hypertension 2012: article 584041. Macinko, James, Inês Dourado, Rosana Aquino, MHLW (Ministry of Health, Labor, and Welfare), Palmira de Fátima Bonolo, Maria F. Lima- Japan. 2011. “Estimates of National Medical Costa, Maria G. Medina, Eduardo Mota, Care Expenditure.” MHLW, Tokyo. Venexa B. de Oliveira, and Maria A. Turci. Newhouse, Joseph P. 1992. “Medical Care 2010. “Major Expansion of Primary Care Costs: How Much Welfare Loss?” Journal of in Brazil Linked to Decline in Unnecessary Economic Perspectives 6 (3): 3–21. Hospitalization.” Health Affairs 29 (12): OECD (Organisation for Economic Co-operation 2149–60. and Development). 2007. Health at a Glance Macinko, James, Barbara Starfield, and Leiyu 2007: OECD Indicators. Paris: OECD. Shi. 2003. “The Contribution of Primary ———. 2010. “Improving Co-ordination of Care Care Systems to Health Outcomes within for Chronic Diseases to Achieve Better Value Organization for Economic Cooperation and for Money.” In Value for Money in Health Development (OECD) Countries, 1970–1998.” Spending , 125–50. OECD Health Policy Health Services Research 38 (3): 831–65. Studies. Paris: OECD. Mahtani, Kamal R., Carl J. Heneghan, Paul B. ———. 2012. “Ageing, Non-communicable Glasziou, and Rafael Perera. 2011. “Reminder Diseases (NCDs), and Emerging Models 224 LIVE LONG AND PROSPER of Care.” Fast Track Paper, DELSA/ European Network of Economic Policy HEA(2012)37, OECD, Paris. Research Institutes, Brussels. ———. 2013. “Report on Integrated Service RAND Europe and Ernst & Young. 2012. Delivery for Vulnerable Groups.” DELSA/ “National Evaluation of the Department of ELSA/WP1(2013)8, Working Party on Social Health’s Integrated Care Pilots.” Prepared for Policy, OECD, Paris. the Department of Health, London, by RAND OECD (Organisation for Economic Co-operation Europe, Cambridge, U.K., and Ernst & Young, and Development) and European London. Commission. 2013. A Good Life in Old Rannan-Eliya, Ravi P., Chamara Anuranga, Adilius Age? Monitoring and Improving Quality in Manual, Ruwani Wickramasinghe, Prasadini Long-Term Care. Paris: OECD. http://dx.doi Perera, Shanaz Saleem, Jayalal Chandrasiri, .org/10.1787/9789264194564-en. Indra Pathmanathan, and Institute for Health Ouwens, Marielle, Hub Wollersheim, Rosella Systems Research Core Team. 2013. “Malaysia Hermens, Marlies Hulscher, and Richard Healthcare Demand Analysis: Inequalities in Grol. 2005. “Integrated Care Programmes for Healthcare Demand and Simulation of Trends Chronically Ill Patients: A Review of Systematic and Impact of Potential Changes in Healthcare Reviews.” International Journal for Quality in Spending.” Report prepared for the Ministry of Health Care 17 (2): 141–46. Health, Malaysia, by Health Policy Research Oxley, Howard, and Maitland MacFarlane. 1994. Associates and Institute for Health Systems “Health Care Reform: Controlling Spending Research, in association with the Institute for and Increasing Efficiency.” Economics Health Policy, Sri Lanka. Department Working Paper 149, Organisation Rechel, Bernd, Jonathan Erskine, Barrie for Economic Co-operation and Development, Dowdeswell, and Stephen Wright. 2009. Paris. Capital Investment for Health: Case Studies Painter, Michael, Megan E. Burns, and Michael from Europe. Observatory Study 18, H. Bailit. 2012. “Bundled Payment across Copenhagen: European Observatory on the U.S. Today: Status of Implementation Health Systems and Policies and World Health and Operational Findings.” HCI3 Improving Organization. Incentives Issue Brief, Health Care Incentives Renehan, Emma, Briony Dow, Xiaoping Lin, Irene Improvement Institute, January. Blackberry, Irja Haapala, Ellen Gaffy, Elizabeth Peikes, Deborah, Arnold Chen, Jennifer Schore, Cyarto, and Kathleen Brasher. 2012. “Healthy and Randall Brown. 2009. “Effects of Care Ageing Literature Review.” Report prepared Coordination on Hospitalization, Quality by the National Aging Research Institute and of Care, and Health Care Expenditures Council on the Ageing for the Department of among Medicare Beneficiaries.” Journal of Health, Victoria, Australia. the American Medical Association 301 (6): Rokx, Claudia, John Giles, Elan Satriawan, Puti 603–18. Marzoeki, Pandu Harimurti, and Elif Yavuz. PHARMAC (Pharmaceutical Management 2010. New Insights into the Provision of Health Agency). 2012. “Introduction to PHARMAC,” Services in Indonesia: A Health Workforce Information sheet, PHARMAC, Wellington. Study. Washington, DC: World Bank. h t t p : / / w w w. p h a r m a c . g o v t . n z / 2 0 1 1 Rosenheck, Robert A., Sandra G. Resnick, and /09/16/01INTRO.pdf. Joseph P. Morrissey. 2003. “Closing Service Pike, Brigid, and Deirdre Mongan. 2014. “The System Gaps for Homeless Clients with Dual Integration of Health and Social Care Diagnosis: Integrated Teams and Interagency Services.” Health Research Board, Dublin. Cooperation.” Journal of Mental Health Policy Powell Davies, Gawaine, Anna M. Williams, Karen and Economics 6 (2): 77–87. Larsen, David Perkins, Martin Roland, and Saltman, Richard B., Ana Rico, and Wienke G. W. Mark F. Harris. 2008. “Coordinating Primary Boerma, eds. 2006. Primary Care in the Health Care: An Analysis of the Outcomes Driver’s Seat? Organizational Reform in of a Systematic Review.” Medical Journal of European Primary Care. Berkshire, U.K.: Open Australia 188 (8 suppl.): S65–68. University Press. Raitano, Michele. 2006. “The Impact of Death- Seike, Atsushi, Simon Biggs, and Leisa Sargent. Related Costs on Health-Care Expenditure: 2011. “Organizational Adaptation and Human A Survey.” ENEPRI Research Report 17, Resource Needs for an Ageing Population.” In P O P U L AT I O N AG I N G A N D H E A LT H S E R V I C E S I N E A S T A S I A A N D PAC I F I C 225 Global Population Ageing: Peril or Promise?, Prepared for the Bloomberg Initiative to edited by John R. Beard, Simon Biggs, David E. Reduce Tobacco Use. http://global.tobacco Bloom, Linda P. Fried, Paul Hogan, Alexandre freekids.org / files/pdfs/en/Cigarette_tax Kalache, and S. Jay Olshansky, 46–50. Geneva: _BIcountries.pdf. World Economic Forum. Sylvia, Sean, Yaojing Shi, Hao Xue, Xin Tian, Sidorenko, Alexandre, and Asghar Zaidi. 2013. Huan Wang, Qingmei Liu, Alexis Medina, “Active Ageing in CIS Countries: Semantics, and Scott Rozelle. 2014. “Survey Using Challenges, and Responses.” Current Incognito Standardized Patients Shows Poor Gerontology and Geriatrics Research 2013: Quality Care in China’s Rural Clinics.” article 261819. Working Paper 264, Rural Education Action Smith, Sheila, Joseph P. Newhouse, and Mark Program, Stanford University, Stanford, CA. S. Freeland. 2009. “Income, Insurance, and https://reap.fsi.stanford.edu/sites/default/files Technology: Why Does Health Spending / Standardized_Patients_China_13Feb2014. Outpace Economic Growth?” Health Affairs pdf. 28 (5): 1276–84. ThaiSES (Thai Household Social and Economic Sochalski, Julie, Tiny Jaarsma, Harlan M. Surveys). 2011. National Statistical Office, Krumholz, Ann Laramee, John J. V. Bangkok. http://web.nso.go.th/survey/house McMurray, Mary D. Naylor, Michael W. Rich, _seco/socio.htm. Barbara Riegel, and Simon Stewart. 2009. U.S. Department of Health and Human Services. “What Works in Chronic Care Management: 2000. Healthy People 2010: Understanding The Case of Heart Failure.” Health Affairs 28 and Improving Health. 2nd ed. Washington, (1): 179–89. DC: U.S. Government Printing Office. Somanathan, Aparnaa. 2015. “Ageing and Van Minh, Hoang, Young K. Do, Mary Ann C. Health in the East Asia and Pacific Region.” Bautista, and Tran T. Anh. 2014. “Describing Background paper for the East Asia and the Primary Care System Capacity for Pacific regional report on aging, World Bank, the Prevention and Management of Non- Washington, DC. communicable Diseases in Rural Vietnam.” Somanathan, Aparnaa, Ajay Tandon, Huong L. International Journal of Health Planning and Dao, Kari L. Hurt, and Hernan Fuenzalida- Management 29 (2): e159–73. Puelma. 2014. Moving toward Universal VHLSS (Vietnam Household Living Standards Coverage of Social Health Insurance Survey). 2012. General Statistics Office of in Vietnam: Assessment and Options. Vietnam, Hanoi. http://www.gso.gov.vn Washington, DC: World Bank. /Default_en.aspx?tabid=491. Spillman, Brenda C., and James Lubitz. 2000. WHO (World Health Organization). 2008. World “The Effect of Longevity on Spending for Health Report 2008: Primary Health Care Acute and Long-Term Care.” New England Now More than Ever. Geneva: WHO. Journal of Medicine 342 (19): 1409–15. ———. 2011b. Noncommunicable Diseases Stewart, Malcolm V., Michael Wilson, Karla Country Profiles 2011. Geneva: WHO. Bergquist, and John Thorburn. 2012. “Care ———. 2012. “Global Health and Aging.” WHO, Coordinators: A Controlled Evaluation of an Geneva. Inpatient Mental Health Service Innovation.” ———. 2013. WHO Report on the Global International Journal of Mental Health Tobacco Epidemic 2013: Enforcing Bans Nursing 21 (1): 82–91. on Tobacco Advertising, Promotion, and Strauss, John, Hao Hong, Xiaoyan Lei, Lin Li, Sponsorship. Geneva: WHO. http://www.who Albert Park, Li Yang, and Yaohui Zhao. 2012. .int/tobacco/global_report/2013/en. “Healthcare and Insurance among the Elderly ———. 2014a. “Raising Tax on Tobacco: What in China: Evidence from the CHARLS Pilot.” You Need to Know.” WHO, Geneva. In Aging in Asia: Findings from New and ———. 2014b. The Health of Older People in Emerging Data Initiatives, edited by James P. Selected Countries of the Western Pacific Smith and Malay Majmundar, 342–70. Region. Geneva: WHO. Washington, DC: National Academies Press. Wimo, Anders, and Martin Prince. 2010. World Sunley, Emil M. 2009. “Taxation of Cigarettes in Alzheimer Report 2010: The Global Economic the Bloomberg Initiative Countries: Overview Impact of Dementia . London: Alzheimer’s of Policy Issues and Proposals for Reform.” Disease International. 226 LIVE LONG AND PROSPER Wong, Albert, Bram Wouterse, Laurentius C. J. ———. 2014a. “Fostering Primary Care Centered Slobbe, Hendriek C. Boshuizen, and Johan J. Integrated Care: Realizing High-Quality and Polder. 2011. “Medical Innovation and Age- Value-Based Service Delivery in China.” Study Specific Trends in Health Care Utilization: Concept Note, World Bank, Washington, DC. Findings and Implications.” Social Science and ———. 2014b. Supply-Side Readiness for Medicine 74 (2): 263–72. Universal Health Coverage: Assessing the Woods, Kevin, and Allyson McCollam. 2002. Depth of Coverage for Non-communicable “Progress in the Development of Integrated Diseases in Indonesia. Washington, DC: World Mental Health Care in Scotland.” International Bank Journal of Integrated Care 2: e21. ———. 2015. “Macrofiscal Implications of World Bank. 2013a. “Reshaping Health Service Achieving Universal Health Coverage in East Delivery through Strengthening Primary Care Asia and Pacific.” In Adjusting to a Changing and Service Coordination in China: Practical World, 61–77. Washington, DC: World Bank. Lessons and Models from National and doi:10.1596/978-1-4648-0618-6. International Experience.” Draft concept note Yip, Winnie, and William C. Hsiao. 2008. “The for the China Programmatic Health Analytical Chinese Health System at a Crossroads.” and Advisory Activities study, World Bank, Health Affairs 27 (2): 2460–68. doi:10.1377/ Washington, DC, October 22. hlthaff.27.2.460. ———. 2013b. “Thailand Case Study: Japan– Zweifel, Peter. 2003. “Medical Innovation: A World Bank Partnership Program for Challenge to Society and Insurance.” Geneva Universal Health Coverage.” World Bank, Papers on Risk and Insurance 28 (2): 194–202. Washington, DC. Zweifel, Peter, Stefan Felder, and Markus Meiers. ———. 2013c. “Turkey Case Study: Japan–World 1999. “Ageing of Population and Health Bank Partnership Program for Universal Health Care Expenditure: A Red Herring?” Health Coverage.” World Bank, Washington, DC. Economics 8 (6): 485–96. Long-Term Care in Aging East Asia and Pacific 7 Introduction living with disability (chapter 1), erosion of family support caused by shrinking family size Long-term care (LTC) refers to the broad and changes in the patterns of family-based range of services designed to support people care for the elderly (chapter 2), and pressure who are unable to perform activities of daily on health care and other systems of support living (ADLs) and instrumental activities of for those with disabilities (chapter 6). daily living (IADLs) independently because Formal LTC systems in developing coun- of the deterioration of physical or cognitive tries of East Asia and Pacific remain nascent, functions, or both. It encompasses family- but a growing number of aging countries are based care in the home and community as grappling with what is the appropriate and well as institutional care. It is quite distinct sustainable role of the state in an area that from health care in that while health care has traditionally been the domain of fami- services seek to change the health condition lies, communities, and the health system. (from unwell to well), LTC services seek to Rapid aging and social change have exposed make the current condition (frail or unwell) the limitations of traditional informal modes more bearable.1 Individuals may need LTC of LTC for frail, elderly people in East Asia because of disability, chronic condition, and Pacific, as shown in chapter 2 of this trauma, or illness that limits their ability to report. Part of the response has been default carry out basic self-care or personal tasks that reliance on health systems for LTC, but this must be performed each day. In modern soci- “solution” tends to be costly and complicates eties, concerns about LTC are a response to wider health reform efforts. Proactive policy three trends that were discussed in previous choices in the LTC domain are therefore chapters: increases in longevity and time spent important but require careful thought and This chapter is based on a background paper prepared by Aparnaa Somanathan (2015) and a literature review conducted by Harry Moroz and Naoko Miake. It is also based on a set of background papers pre- pared for China National Development and Reform Commission by Elena Glinskaya of the World Bank and Joshua Wiener, Zhanlian Feng, and Nan Tracy Zheng of Research Triangle International. It has also benefited from inputs from Professor John Campbell and Dewen Wang. 227 228 LIVE LONG AND PROSPER planning with respect to their interaction with has resulted in longer, healthier lives, thus informal care systems and existing formal postponing the need for LTC. health and welfare systems. With the proportion of the population This chapter outlines the key issues and ages 80 and above expected to double or options for LTC in developing East Asian and quadruple in some East Asian and Pacific Pacific countries. The first section describes countries, the demand for LTC is expected current trends and projections for demand for to increase significantly. In Japan and the LTC and LTC expenditures in the region. The Republic of Korea, the proportion of the second section provides an overview of LTC population ages 80 and older is expected provision within East Asia and Pacific and else- to increase three- to sixfold in the next four where, and the third section describes financing decades (figure 7.1). Disability rates tend to arrangements for LTC, including arrangements increase with age, generating demand for LTC for provision and types of benefits. Challenges services. Figure 7.2 shows projected ADL in coordinating health care and LTC and and IADL disabilities in China, holding cur- ensuring sufficient capacity for LTC provision rent disability rates constant. In East Asia and and quality of care are discussed in the fourth Pacific, particularly in lower-middle-income and fifth sections, respectively. Given the and middle-income countries, disabilities nascent state of formal LTC systems in most may not remain constant or even decline, as developing East Asian and Pacific economies was the case in the OECD. As noted in ear- the examples and policies cited in the chapter lier chapters, the population cohorts enter- rely heavily on experience from richer East ing their 70s and 80s in the next two decades Asian and Pacific economies and countries in have been more exposed to risks related to the Organisation for Economic Co-operation noncommunicable diseases in middle age than and Development (OECD). Furthermore, previous cohorts, making them more vulner- because the policy choices for LTC remain able to disabilities in old age. much more open at this early stage (compared The poor and rural elderly as well as the to, for example, pensions and health care, as female elderly are more likely than oth- discussed in previous chapters), the recommen- ers to need LTC services. The prevalence of dations are less prescriptive on what the most family-based care can lead to older people appropriate choices might be. with care needs being missed when they have no family nearby (or their family is unable or unwilling to help), no statutory or volun- LTC demand and expenditures: tary services can fill the gap, or older people Current trends and projections themselves cannot or choose not to ask for The proportion of the population surviving to help (Research Base 2012). A recent study very old ages (80 years and older) is a major by HelpAge International using data from driver of the demand for LTC, but it is not the World Health Surveys2 looked at unmet the only one. Demand for LTC is also greatly need for LTC among old people. On average, influenced by the prevalence of functional 55 percent of elderly men and 45 percent of and cognitive limitations among the elderly elderly women across the region faced care as measured by trends in ADLs and IADLs gaps. Figure 7.3 shows that the care gaps are and by the proportion of frail elderly in the greater for the rural elderly, the female elderly, population. As already noted in this report, and the poor in some of the poorer East Asian the extent to which aging is associated with and Pacific countries. functional and cognitive decline and frailty Predicting growth in demand for LTC in among older populations is not uniform East Asia and Pacific is complicated because across countries. In many OECD countries, the region’s LTC is largely informal. The the adoption of life-course approaches to level of unmet demand for LTC is presumed preventing the onset of noncommunicable to be very high. Where it exists, institutional diseases and associated chronic conditions care caters to a relatively small share of LO N G - T E R M C A R E I N AG I N G E A S T A S I A A N D PAC I F I C 229 the population. The relative underprovision FIGURE 7.1 Long-term care users in OECD countries are expected of LTC services at present, combined with the to increase with the aging population prospect of an unhealthy population bulge Proportion of total long-term care users by age, 2009 entering old age with greater dependency and disability than current elderly cohorts, has Hungary 31 42 27 created a sense of urgency among East Asian Poland 48 22 30 and Pacific policy makers about the need to Czech Republic 30 26 44 develop LTC systems. Austria 18 34 48 As with health care, LTC expenditures are driven by both demographic and non- Norway 32 19 49 demographic factors. The demographic driver Sweden 26 25 49 is related to the number of dependent people Netherlands 23 28 50 in the population, which depends on the evo- Luxembourg 23 28 50 lution of life expectancy and health expendi- ture. The non-demographic drivers are related Korea, Rep. 6 45 50 to income growth and changes in demand Germany 20 28 52 for publicly financed LTC services. Income Finland 21 25 54 has a direct effect through increases in liv- Switzerland 20 25 55 ing standards (gross domestic product [GDP] Australia 15 28 57 per capita) and an indirect effect through changes in relative productivity, or Baumol Denmark 11 28 61 effects. Because the LTC sector is highly labor New Zealand 3 36 62 intensive, room for productivity gains is more Japan 4 32 64 limited than elsewhere in the economy. With 0 10 20 30 40 50 60 70 80 90 100 equalization of wages across sectors, the rela- tive price of LTC vis-à-vis other goods and ser- Share of population in long-term care (%) vices in the economy tends to rise as aggregate 0–64 65–79 80+ productivity and GDP per capita increase. Sources: OECD Health Statistics 2012 data (http://dx.doi.org/10.1787/health-data-en), augmented With demand being price-inelastic, the share with additional Australian and Swedish data. of LTC expenditure in GDP would tend to Note: OECD = Organisation for Economic Co-operation and Development. increase over time. Given the importance of home production of LTC services, the demand for publicly financed LTC depends on devel- LTC insurance in 2008. This trend is expected opments in formal labor force participation to continue.3 (De la Maisonneuve and Martins 2013). OECD projections of LTC expenditures In 2010, OECD countries allocated an from 2010 to 2060 show that demographic average of 1.56 percent of GDP to public changes are not likely to be the most impor- spending on LTC, with private spending on tant determinants of future public LTC expen- LTC absorbing another 0.67 percent of GDP. ditures. The projections included all OECD Figure 7.4, panel a, shows total LTC expen- countries and Brazil, China, India, Indonesia, ditures in OECD countries. Although still and the Russian Federation. Because the cost relatively low, LTC expenditure—particularly of helping one person with ADL or IADL lim- public LTC spending—has shown a faster itations should be more or less the same, irre- upward trend in per capita terms than health spective of age, the pure age effect has only care spending, with an annual average growth a moderate impact on spending. The OECD of over 9 percent across 25 OECD countries, model also assumes “healthy aging,” which compared with 4 percent for total public further mitigates the age effect. As with pro- expenditure on health (figure 7.4, panel b). jections of public health expenditures, non- In Korea, the dramatic increase in public LTC demographic factors account for the largest expenditures followed the introduction of share of the increase. 230 LIVE LONG AND PROSPER FIGURE 7.2 In China, ADL and IADL disabilities may increase rather than decline as in OECD countries Projections of ADL and IADL disabilities by gender in China a. Female b. Male 20 20 15 15 Percent Percent 10 10 5 5 2010 2020 2030 2040 2050 2010 2020 2030 2040 2050 Year Year Activities of daily living Independent activities of daily living Sources: Giles et al. 2015; UN 2013. Note: ADL = activity of daily living; IADL = instrumental activity of daily living; OECD = Organisation for Economic Co-operation and Development. Share of (people have ADL or IADL disabilities) in (each gender group) × share of (people in each gender group in each year). FIGURE 7.3 Care gaps are greater for the rural elderly, the female elderly, and those living in poverty in East Asian and Pacific countries Percent of total population with unmet care needs by gender, rural/urban location, and poverty status, latest available year, 2000s a. Care gaps by gender b. Care gaps by rural/urban sector c. Care gaps by poverty level Brunei Brunei Brunei Darussalam Darussalam Darussalam Cambodia Cambodia Cambodia China China China Timor-Leste Timor-Leste Timor-Leste Share of population (%) Share of population (%) Indonesia Indonesia Indonesia Share of population (%) Japan Japan Japan Korea, Rep. Korea, Rep. Korea, Rep. Lao PDR Lao PDR Lao PDR Malaysia Malaysia Malaysia Myanmar Myanmar Myanmar Philippines Philippines Philippines Singapore Singapore Singapore Thailand Thailand Thailand Vietnam Vietnam Vietnam 0 25 50 75 100 0 25 50 75 100 0 25 50 75 100 Women Men Urban Rural Living in multidimensional poverty Other Source: Research Base 2012. Under a cost-pressure scenario that of GDP by 2060. For non-OECD countries, assumes little or no improvement in cost the increase in LTC spending is projected containment, the ratio of public LTC expen- to be, on average, broadly the same as that diture to GDP is projected to increase by experienced by OECD countries but starting 1.4 percentage points to reach 2.1 percent from a much lower level (0.1 percent in China LO N G - T E R M C A R E I N AG I N G E A S T A S I A A N D PAC I F I C 231 and Indonesia). Differences in the initial level FIGURE 7.4 Long-term care expenditures as a share of GDP are of female labor force participation rates as relatively low but growing faster than health expenditures well as policies in place also explain the dif- Public LTC expenditures as a share of GDP, 2010, and average annual real ferences across countries. growth in public LTC expenditure, 2000–10 The analysis of LTC expenditures and pro- a. LTC expenditures as a share of GDP, 2010 jections for OECD countries provides impor- Mexico tant lessons for the non-OECD countries in Slovak Republic East Asia and Pacific, where LTC remains Australia Portugal a gray economy that is largely unquanti- Estonia Czech Republic fied. First, reducing levels of severe disability Hungary (2008) combined with a scenario of healthy aging Poland United States can lead to better containment of LTC costs, Korea, Rep. Spain because older people remain at home for Luxembourg longer with less expensive care needs. Second, Slovenia (2008) Germany a degree of path dependence exists in how Japan Austria LTC systems develop, which can serve to Canada drive up costs well into the future or contain New Zealand Switzerland (2009) them. Responding to LTC needs with invest- OECD countries Iceland ment in LTC institutions rather than more France holistic systems based in home, community, Belgium Finland and institutional settings may lock countries Norway Denmark into a high-cost system that is hard to restruc- Sweden ture or shrink later. Countries such as Japan Netherlands provide useful lessons on this scenario. 0 1 2 3 4 Share of GDP (%) Providing LTC Health long-term care Social long-term care LTC is provided in three broadly defined b. Average annual real growth in public expenditure on settings—home, community, and institu- LTC (health), 2000–10 (or nearest year) Korea, Rep. tions—and includes both health and social Estonia care services. Social care services include Japan Netherlands assistance with performing ADLs and IADLs Hungary Switzerland as well as social support. Figure 7.5 illustrates OECD countries who requires LTC and what types of services Poland Portugal are available. East Asian and Pacific econo- Japan Australia mies that have already established formal Belgium LTC systems—or key elements—are China; Norway Czech Republic Japan; Korea; Singapore; Taiwan, China; Iceland Luxembourg and Thailand. In East Asia and Pacific more New Zealand broadly, LTC is delivered informally by family France Denmark members. The majority of informal caregiv- United States Slovenia ers, both within and outside the household, Austria Canada are women who are informally employed. Sweden Many informal caregivers are themselves Finland Germany older people. 0 10 20 30 40 50 60 Across OECD countries and elsewhere, Average annual growth rate in real terms (%) institutional care has tended to dominate formal (that is, paid) LTC provision, despite Sources: OECD 2013, using data from Colombo et al. 2011. research and polls invariably showing that Note: Based on the definition in the 2012 Joint Health Accounts Questionnaire of the System of Health Accounts, long-term care (LTC) expenditure comprises long-term (health) care and social the majority of older people with LTC needs services of long-term care. 232 LIVE LONG AND PROSPER FIGURE 7.5 Long-term care services include both social and health care provided in a variety of settings Types of LTC Assisted living Home care Frail elderly LTC Adult day care Unhealthy Nursing elderly homes Elderly Healthy elderly Hospice Sick elderly Health care care Note: LTC = long-term care. prefer services in their homes or community- (Campbell 2014). Of Japan’s elderly (ages based settings (Colombo et al. 2011; Keenan 65 and over) population, 4.7 percent live in 2010). Home- or community-based services publicly supported institutions, a higher pro- are also considerably more cost-effective. portion than in many developed nations. The These preferences appear to be universal, reason is historical and political: in the early regardless of cultural differences. 1970s, the Japanese government reduced co- In the United States, publicly financed LTC payments for medical care from 30 percent has been shifting away from nursing home or 50 percent to zero for those over 70 years care and toward home- and community- of age in response to political pressure. The based services over the past several decades, a result over several years was a flood of older process known as rebalancing (Feng, Fennell, people into hospitals, many newly built to et al. 2011; Wiener, Anderson, and Brown meet this demand. The problem of these 2009). Currently, home- and community- “social admissions” into hospitals, with little based services account for nearly half of total or no medical justification, has plagued gov- Medicaid LTC expenditures (including spend- ernments ever since: 24-hour institutional ing for people with intellectual and develop- care is perceived as attractive and normal, mental disabilities), up from just 20 percent a perception that once ingrained, is hard to in 1995 (Kaiser Family Foundation 2013). remove (Campbell 2014). Health reforms under the Affordable Care Similarly, the initial heavy reliance on Act of 2010 aim to further expand home- and LTC hospitals in Korea to meet the popula- community-based services and to accelerate tion’s LTC needs has become an entrenched the pace of rebalancing. Although progress pattern of care behavior now. As described has been made, the bulk of total spending is in box 7.1, LTC hospitals and LTC facili- still for institutional care. ties compete directly for LTC patients, even In Japan, the predominance of institution- though the LTC insurance scheme creates based LTC is at odds with family ideology, some (albeit weak) incentives to move LTC which is thought to make families reluc- patients to LTC facilities. In addition, the tant to put aging parents in an institution introduction of LTC insurance was followed LO N G - T E R M C A R E I N AG I N G E A S T A S I A A N D PAC I F I C 233 BOX 7.1 LTC hospitals and LTC facilities in the Republic of Korea: In direct competition In Korea, long-term care hospital (LTCH) services rehabilitative care services, whereas the LTCFs take are covered by national health insurance, and long- the less clinically demanding patients requiring sup- term care facilities (LTCFs) are covered by long-term port with physical and cognitive limitations. By com- care (LTC) insurance, which was introduced in 2010. parison, in Korea, no clear distinction exists between Before the introduction of LTC insurance, LTCHs patients in LTCHs and LTCFs. were the main source of institutional care for old or disabled patients, resulting in a high rate of “social admissions.” Once patients were discharged from FIGURE B7.1.1 Average length of stay in long-term LTCHs, they were no longer eligible for health insur- care hospitals in the Republic of Korea has declined ance coverage. only marginally Ideally, the introduction of LTC insurance should Length of stay in long-term care hospitals, 2010–12 have led to LTCFs taking over most of the LTC patients, leaving the more clinically complex patients 35 to the LTCHs. As table B7.1.1 shows, expenditures per person and the out-of-pocket share of financing 30 Share of elderly population (%) are lower at LTCFs. 25 In reality, LTCFs and LTCHs have ended up competing for the same patients because of the for- 20 profit nature of hospitals and a relative oversupply of LTCHs. As the first row of table B7.1.1 shows, the 15 share of the elderly population in LTCHs has con- 10 tinued to grow, as it has in LTCFs. As figure B7.1.1 illustrates, the frequency of very long average lengths 5 of stay (more than 180 days) at LTCHs has declined 0 only marginally. 2010 2011 2012 Clearly, the distribution of patients as measured (N = 106,739) (N = 166,887) (N = 197,597) by resource utilization groups can be rationalized further in Korea. Table B7.1.2 compares the share of patients by resource utilization groups in Korea and Less than 30 days 30–89 days 90–179 days Ontario, Canada, for comparable LTCHs and LTCFs. 180–269 days 270–359 days More than 360 days In Ontario, the LTCH equivalents take most of the clinically complex cases or those requiring specialist Source: Kim, Jung, and Kwon, forthcoming. TABLE B7.1.1 Comparison of LTCHs and LTCFs in the Republic of Korea LTCHs LTCFs Characteristic 2010 2011 2012 2010 2011 2012 Percent of elderly population (65 and older) in hospital 2.0 3.0 3.4 1.0 1.7 2.0 Number of days per person per year 170 158 155 258 254 222 Expenditure per person per day (W, thousands) 65.6 69.5 71.3 43.1 42.5 43.2 Expenditure per person per year (W, thousands) 11,190 10,824 11,005 11,174 10,905 9,721 Out-of-pocket payment (W, thousands) 2,171 2,101 2,155 1,412 1,493 1,363 Share of total expenditures (%) 19.4 19.4 19.6 12.6 13.7 14.0 Payment by public insurance (W, thousands) 9,019 8,723 8,850 9,762 9,413 8,358 Share of total expenditures (%) 80.6 80.6 80.4 87.3 86.3 86.0 Source: Kim, Jung, and Kwon, forthcoming. Note: LTCF = long-term care facility; LTCH = long-term care hospital. box continues next page 234 LIVE LONG AND PROSPER BOX 7.1 Long-term care hospitals and long-term care facilities in the Republic of Korea: In direct competition (continued) TABLE B7.1.2 Distribution of resource utilization groups in the Republic of Korea and Ontario by type of institution LTCH (%) LTCF (%) Korea, Rep. Ontario Korea, Rep. Ontario Resource utilization group (n = 1,364) (n = 14,600) (n = 1,472) (n = 90,115) Rehabilitation 44.20 53.90 19.77 15.00 Extensive special care 5.45 12.40 1.36 1.60 Special care 2.68 11.50 2.92 6.00 Clinically complex 13.35 16.70 11.28 16.50 Cognitive impairment 4.23 0.80 9.44 10.80 Behavior problem 1.97 0.10 4.62 2.80 Physical function 28.23 4.50 50.61 44.30 Sources: Kim, Jung, and Kwon, forthcoming; Hirdes et al. 2011. Note: LTCF = long-term care facility; LTCH = long-term care hospital. by a substantial increase in LTC hospitals more popular among governments and local and facilities. By 2012, Korea had 4,181 LTC authorities because building facilities and beds facilities and 10,759 ambulatory LTC provid- is easier and yields more tangible and visible ers. Over 60 percent of the LTC residential results than providing and managing services care facilities in Korea cater to 30 or fewer in homes and communities. Application of residents, making them too small to achieve appropriate policy instruments (including economies of scale (Kwon 2014). tax exemption, subsidies, and other financial Japan’s and Korea’s experiences are quite incentives) should be consistent with the high- pertinent for other countries in East Asia level policy decisions to build a three-team and Pacific, where a certain degree of pressure system of social services for the aged, with exists to expand institutional care despite fil- home-based family care as its bedrock, sup- ial piety–driven family ideologies and cultural ported by community-based care and under- preferences to remain at home. For instance, pinned by institutional care. Both preferences in China, older people prefer receiving LTC and high-level decisions clearly indicate that at home or in their communities (Shanghai as far as possible, LTC should be delivered in Daily 2013; Xinhua 2014). The government’s the least restrictive settings. overarching policy framework for a three- A range of community-based care mod- tiered LTC system aptly emphasizes home- els exists throughout the region and could and community-based services. However, be built upon as part of the rebalancing current policies and resource allocation pat- process (Research Base 2012). Singapore terns incentivize institutional care more than Programme for Integrated Care for the home- and community-based services (Feng Elderly (SPICE) centers operate integrated et al. 2012). services but also operate in home settings East Asian and Pacific policy makers need as required.4 Thailand has piloted a home to use policy instruments at their disposal health care scheme in 26 hospitals, targeted at to rebalance the mix of services as early as older people living at home. Services include possible and thus avoid a bias for an insti- health promotion, treatment, and rehabilita- tutional setting in the development of LTC tion. Thailand has also piloted service mod- systems. Institutional solutions to LTC are els that integrate health and social care—the LO N G - T E R M C A R E I N AG I N G E A S T A S I A A N D PAC I F I C 235 Bangkok 7 Model and the Community- about US$20 to community-based carers who Based Integrated Services of Health Care and receive minimal training in supporting others Social Welfare for Thai Older Persons—that in ADLs. Anecdotal evidence suggests that involve collaboration among local authori- carers are often younger elderly people who ties, volunteers, and older people. The ROK– have time and are likely to be more sensitized Association of Southeast Asian Nations to the needs of older elderly. This program (ASEAN) Home Care Programme, supported was nationally funded in the initial years, with by HelpAge International, builds on the suc- the intention to shift financing to subnational cess of a home care model involving volun- authorities over time. Perhaps even more teers in Korea and is the most prevalent one innovative is the growth of “time banks” in the region, having been adopted in all in China with support from local authori- ASEAN countries. Community groups and ties in cities such as Hangzhou. Under these networks and older people’s groups operate schemes, younger elderly provide home- and or are planned in Brunei Darussalam, Japan, community-based support to other elderly and Thailand, offering support networks, people in need of care and in return generate health information dissemination, and fit- a “time credit” that compensates the caregiv- ness activities. In all these initiatives, the role ers with equivalent care time when they reach of community-based organizations is vital, the stage of needing care. in many cases supported by older people’s In most OECD countries, provision of LTC associations as providers, trainers, facilita- services is largely the responsibility of the pri- tors, and advocates. In 2012, for example, vate sector, although the mix of for-profit and 410,000 older people’s associations existed nonprofit services varies. Notable exceptions across the region, according to the National to the dominance of private provision are the Commission on Aging, and more than 23,000 Nordic countries, where local municipalities exist in Thailand under the Senior Citizens provide a substantial share of services directly. Council of Thailand.5 In the United States, approximately 69 percent Examining pilots in several Chinese cities of all nursing homes are for profit, 25 percent to develop comprehensive community-based are nonprofit, and only 6 percent are govern- care services is also useful. The Chinese gov- ment run (CMS 2012). Among residential ernment is implementing pilots in 42 cities care facilities (for example, assisted-living (including Beijing, Chengdu, Hangzhou, and facilities and board-and-care homes) serving Shanghai) to provide comprehensive care ser- relatively less disabled elders than nursing vices with a focus on innovation and appli- homes, more than 82 percent are private and cability. Key elements of the pilot programs for profit. In England in the early 1980s, the are establishing or strengthening community majority of care-home providers were publicly care centers to provide comprehensive ser- owned, and community-based services were vices; conducting ability and needs analy- limited. Today, however, the overwhelming sis; purchasing services from the market or majority of adult social care service providers through service vouchers; using management are in the private sector (Malley et al. 2014). information systems and information com- Private but nonprofit providers play a larger munication technology for monitoring and role in some countries, such as Germany and evaluation, quality control, and management; Australia, with faith-based nonprofits playing and encouraging involvement of volunteers in a significant role in the latter. the provision of elderly care services. The mix of for-profit and nonprofit ser- A further distinctive feature of LTC pro- vices also varies across East Asia and Pacific. vision in East Asia and Pacific is the emer- Japan has a mixed model: for home- and gence of state-supported informal caring community-based services, for-profit compa- arrangements. The most notable example is nies compete with established social service the Thai scheme for informal carers in rural organizations and other nonprofits, while areas, which pays a monthly stipend equal to for-profit organizations are excluded from 236 LIVE LONG AND PROSPER the institutional sector that provides 24-hour has an important role to play in facilitating care. Under LTC insurance, institutions in and supporting self-care and family care, safe- principle compete for clients, but in reality, guarding individuals from abuse, and guaran- high demand plus government restrictions teeing quality and safety while encouraging on building new facilities has created wait- cost containment. Simply setting basic con- ing lists instead (Campbell 2014). Korea is sumer standards is often insufficient in a sec- an interesting intermediate case with respect tor where market forces do not ensure quality to LTC service provision. In principle, private and safety and where older people and fami- LTC providers are nonprofit, and the health lies are disempowered consumers. The role of law does not allow for-profit entities to own the state is therefore vital to safeguard qual- health care or LTC entities. However, non- ity and safety, not just through establishing profit providers, which are owned by health minimum standards for care but also through care professionals or nonprofit foundations, providing information and grievance mecha- often behave like for-profit entities in prac- nisms, training, setting up professional stan- tice, with the health professionals who own dards, carrying out inspections, and enforcing establishments as residual claimants on prof- standards. In Japan, instead of simply subsi- its. For example, Samsung and Hyundai have dizing or encouraging private provision, the hospitals, but they are owned by nonprofit government guides the program in the inter- foundations established by the conglomerates ests of national coverage, equity and fairness, (Jung et al. 2014). high quality, and efficiency in particular. As In developing East Asia and Pacific more has long been evident in Japan’s health care generally, formal or paid LTC services are field, a strong public role is crucial for keep- still at a very early stage of development. ing down expenditures (Campbell 2014). Traditionally, most countries have had a small number of social welfare institutions run by the state or local governments that served Financing LTC childless elders, orphans, and people with men- Informal LTC provided by family members tal illnesses. As the demand for LTC services constitutes the bulk of LTC consumed and increases, both state and private sector provi- is associated with significant uncompensated sion are growing rapidly. Chinese policy mak- costs in almost all countries in the world. In ers have recognized that the state is no longer the United States, 87 percent of the 12 mil- able to meet pressing needs without enlist- lion people who need LTC receive it from ing the private sector, and they have issued unpaid family caregivers, with the annual a series of national policy directives over the economic value of unpaid care estimated at past decade to speed up private sector devel- US$450 billion in 2009, which exceeded the opment of social services for the aged. These total reported LTC expenditure for paid care policy documents promote preferential policy of US$357 billion in 2011 (Kaiser Family treatments such as tax exemptions, govern- Foundation 2013). Although provision of ment subsidies for new and existing beds, land LTC by family members has significant nega- appropriation or leasing for new construction, tive impacts on their welfare (Van den Berg, and reduced rates for utilities. The policy doc- Fiebig, and Hall 2014), formal LTC services ument on aged care issued by the State Council are expensive and beyond the means of most in 2013 clearly signals that major reforms are families. Across the OECD, private LTC under way to further privatize the manage- insurance is rarely purchased voluntarily and ment of publicly built senior care facilities.6 is affordable only to a small number of fami- As service provision gravitates toward the lies (Colombo et al. 2011). private sector, concerns about quality of care Acknowledging the challenges in financ- will emerge, as happened in the OECD, which ing LTC, governments in many countries have calls for a stronger stewardship role by the stepped in to fill the void by developing pub- state. In provision of LTC services, the state licly supported LTC services to complement LO N G - T E R M C A R E I N AG I N G E A S T A S I A A N D PAC I F I C 237 informal care. In almost all developed coun- The philosophical premise in these systems tries, the government typically pays most or is that the government should take the lead a substantial share of total LTC costs for eli- in ensuring that all people with disabilities, gible recipients. For this reason, in most coun- regardless of financial status, should be eli- tries, much of the policy discussion revolves gible for the LTC services they need. Social around public financing of LTC services. In solidarity is highly valued, and the right to non-OECD East Asian and Pacific countries, LTC is viewed similarly to the right to medical public financing for LTC is minimal, but pol- care. This view is reflected in the public LTC icy makers are considering the introduction of insurance models in Germany, Japan, Korea, LTC financing schemes. Luxembourg, and the Netherlands, and in the Broadly, the approaches to public financ- personal care and nursing care through health ing of LTC can be classified according to the coverage in Belgium (Colombo et al. 2011; scope of entitlement to LTC benefits (universal Wiener 2011). In Germany, for example, vs. means tested), the structure of the service the primary source of financing is through delivery system at national and subnational a universal social insurance program for levels, and the types of services and benefits. LTC that provides nursing home and home These approaches vary greatly across OECD care benefits for people of all ages with dis- countries. This section assesses the different abilities (Campbell, Ikegami, and Gibson types of financing systems in OECD countries 2010; Cuellar and Wiener 2000; Gibson and to draw lessons for East Asia and Pacific. Redfoot 2007). The program is administered by sickness funds for 70 million Germans, and private health insurers cover an additional Scope of entitlement to LTC benefits 10 million, mostly upper-income individuals. In those countries characterized as having uni- Table 7.1 describes the key features of uni- versal coverage within a single program, LTC versal LTC insurance programs in Japan, coverage is provided through a single system. Korea, and Taiwan, China, which share TABLE 7.1 Features of long-term care insurance in Japan, the Republic of Korea, and Taiwan, China Feature Japan Korea, Rep. Taiwan, China Year introduced 2000 2008 2016 (planned) Premium contributors All people ages 40 and older All people All people Sources of financing Government subsidy 45% 20% 90% Premium 45% 60–65% Co-payment 10% 15–20% 10% (poor are exempt) Eligibility Ages 65 and older Ages 65 and older All ages with disability or age- Ages 40–64 with mental or All ages with age-related related conditions physical disability conditions Service benefits Home care Yes Yes Yes Community-based care Yes No Yes Nursing care No Yes No Institutional services Yes No Yes Cash benefits No Only exceptional cases Yes for family members and home help Management Municipal government but Under national health Under national health with uniform fee schedule insurance but financed insurance but financed independently independently Source: Lu 2014. 238 LIVE LONG AND PROSPER many similarities. In all three cases, financing Thus, a local approach can establish impor- is shared by governments and households. tant links between LTC and other services In those countries characterized as having often needed by people with disabilities. means-tested safety-net schemes, strict income Second, LTC is an intensely personal issue or asset tests are used to set financial thresh- involving decisions about how consumers olds for eligibility for publicly funded LTC want to live their lives. Thus, the planning services and benefits targeting the needy. Such and delivery of services can be influenced by schemes exist in England, New Zealand, the local circumstances, norms, and values as United States, and more recently Singapore.7 well as by the local preferences of the popu- The philosophical premise in these countries lation with disabilities, their caregivers, and is that the primary responsibility for care of providers. Finally, because subnational gov- older people and younger persons with dis- ernments are less driven to routinize their abilities rests with individuals and their fami- decision-making process and because indi- lies, and the government should act only as a vidual cases loom larger in the policy process, payer of last resort for those unable to provide locally administered programs are arguably for themselves. The means-tested programs less rigid and bureaucratic than centrally run thus limit public benefits to people who are programs. poor or who become poor because of the high At the other end of the continuum are costs of LTC and medical care. This system countries such as Germany and Japan, which results in inefficient use of personal and state have a more nationalized and centralized resources and inhibits personal savings. approach to LTC, although subnational gov- In between these two relatively clear-cut ernmental entities are often still involved. For schemes are mixed systems in other OECD example, under the LTC insurance program countries. In these systems, LTC coverage is in Japan, 2,895 municipal governments or provided through a mix of different universal alliances of municipalities are the insurers, programs and benefits or a mix of universal and they have a generalized responsibility and means-tested LTC entitlements. to provide adequate services (Campbell and Ikegami 2000, 2003). However, because almost all aspects of the program (eligibil- Structure of the service delivery system ity, most benefits, and reimbursement rates) A key issue in the design of LTC systems are fixed at the national level, the ability of is the level of government responsible for the municipalities to shape the program is financing and delivery. Many developed coun- strictly limited. Thus, although premiums are tries, including Canada, the Netherlands, set at the municipal level, almost all financing Sweden, the United Kingdom, and the United parameters are set at the national level. States, rely heavily on subnational govern- Two main arguments favor consolidation ments to design and administer their LTC at the national level (Wiener and Tilly 2003). systems, albeit often with substantial policy First, a uniform national program helps guar- guidance from the national government. antee horizontal equity across geographic For example, Sweden devolves almost all areas. In other words, national rules help responsibility for financing, organization, and ensure that similarly situated individuals in administration of LTC to municipalities, even different geographic areas are treated simi- though it is a small country with fewer than larly. In England, for example, which relies 10 million people (SALAR 2007). on subnational governmental units, beneficia- Advocates for devolution make three argu- ries often complain of a “postcode lottery” in ments in favor of assigning responsibility for which persons with similar needs and finan- LTC to smaller geographic governmental cial status are treated very differently because units (Wiener 1996). First, states, provinces, they are subject to different local authorities. and municipalities are heavily involved in a In countries with insurance approaches such variety of social services in many countries. as Germany’s, regional variations are thought LO N G - T E R M C A R E I N AG I N G E A S T A S I A A N D PAC I F I C 239 to be unfair, and efforts are made to eliminate few future prospects. Furthermore, even them (Cuellar and Wiener 2000; Wiener and if the costs of the allowances are low, the Cuellar 1999). Second, developing a single savings may be offset by the increased num- national program may involve less adminis- ber of users (Wiener 1996, 2011). Services trative expense because program rules and such as assistance with bathing are less systems need to be developed only once. Each attractive except when really needed, and subnational governmental unit need not rein- demand is likely to be limited, at least for sev- vent the wheel. eral years, in countries where traditional fam- ily caregiving is widely accepted as the norm (Campbell 2014). Types of services and benefits Public financing for in-kind services In most countries, publicly supported LTC is requires the development of infrastructure provided in the form of in-kind services rather for nursing homes and for community-based than cash benefits. Public payments and sub- services (for example, home help, day care, sidies for LTC typically go to service provid- respite care for family caregivers, and assis- ers rather than directly to care recipients. This tive devices). Nursing homes are expensive structure is primarily driven by concerns over to build and operate, but because the gov- the misuse of or lack of control over cash ben- ernment is likely to control capital expen- efits or by concerns over potential coercion of diture, it can control the pace of expansion. female family members to become LTC pro- Community care agencies do not require viders and forgo market opportunities. Japan, much new construction, but they do need for instance, operates a services-only model. trained staff and good management, which However, demand for cash allowances to require investment. However, they too can LTC recipients or family caregivers is increas- be developed gradually. A long-run benefit ing in a number of OECD countries (such is that the training and work experience will as Eastern European countries, England, raise the quality of the female workforce. Germany, Italy, and the Netherlands) to Private LTC insurance does not play a enhance consumer choice and flexibility. major role in financing LTC in any country In the United States, several recent policy (Colombo et al. 2011). In the United States, initiatives such as the Medicaid Cash and which has had an active market in private Counseling Program aim to promote partic- LTC insurance since the mid-1980s, only ipant-directed personal assistance services for about 12 percent of the population 65 years frail older adults with disabilities and other of age and older has any LTC insurance people with disabilities in the Medicaid pro- (almost all of which contains substantial grams. Case by case, it can be less expensive restrictions on coverage), and it accounts for than services: Germany offers a choice of cash only about 7 percent of total expenditures and services, and even though the value of the (Frank, Cohen, and Mahoney 2013; Kaiser services is about double the cash allowance at Family Foundation 2013). In France, which various levels of disability, most people take has significant market penetration, LTC the cash (Campbell 2014). The proposed LTC insurance is integrated with health insur- insurance in Taiwan, China, includes both in- ance, but the benefits are limited and private kind and cash benefits as a way of addressing insurance accounts for less than 1 percent problems of waning family support. of LTC expenditures (Nadash, Doty, and The disadvantages of cash allowances Racco 2013). Private insurance in Germany may outweigh the advantages, particularly is primarily offered as an alternative to the for middle-income countries. Budget con- mandatory statutory sickness funds for upper- straints mean the cash amounts would have income people and government officials. In to be low, insufficient to provide a decent short, private health insurance is not a viable income but perhaps enough to trap caregiv- option for financing LTC for the large major- ers (mostly women) into a situation with ity of the population. 240 LIVE LONG AND PROSPER Private out-of-pocket payments by recipi- care, which sets a floor on the quality of care ents are a feature of all public LTC coverage possible. Even though it is a means-tested schemes regardless of the specific financing system in which not everyone qualifies for mechanisms used (Colombo et al. 2011). government financial support, in the United Countries generally believe that individu- States, the government pays for approximately als should be financially responsible for the three-quarters of nursing home residents. In costs of “normal living,” such as housing addition, for private providers, threatening and meals. Even in universal coverage sys- to terminate participation in the government tems, services are often targeted to those with financing program is the principal enforce- the highest care needs, and not all countries ment mechanism for addressing poor quality provide full and comprehensive coverage of of care (AHCA 2013; Wiener et al. 2007). services. Individuals and families not infre- The degree of redistribution achieved quently pay a significant share of LTC costs by public financing schemes is a function out of pocket, as is the case in Switzerland of how funds are raised and how they are (60 percent), Portugal (45 percent), Spain spent. LTC financing systems redistribute (30 percent), and Korea (20 percent for insti- resources from people without disabilities to tutional care and 15 percent for home care). people with disabilities and potentially from In the United States, out-of-pocket and private younger (less likely to be disabled), working payments accounted for 22 percent (15 percent people to older, retired people. Depending on out of pocket, 7 percent private insurance) of the system, resources can also be redistrib- the total US$357 billion in LTC spending in uted from higher-income to lower-income 2011 (Kaiser Family Foundation 2013). people. Although disability affects people of Levels of public financing support and cost all ages and income or wealth levels, people sharing are determined on the basis of eligibil- with disabilities disproportionately have ity rules. Eligibility is based on objective and lower incomes and less wealth than people uniform assessments of individuals’ care needs, without disabilities. For example, in 2001 typically measured by the extent of physi- in the United States, the median income of cal and cognitive impairments (in addition to older people with no disabilities was about means testing, where it exists). Mechanisms twice that of older people with severe dis- for assessing eligibility for LTC services are abilities, while the median total household net discussed in a separate section below. worth for people with no disabilities was four The structure and details of public LTC times that of people with severe disabilities financing have large impacts on the bal- (Johnson and Wiener 2006). As a result, even ance between institutional and home- and universal coverage financing systems serve a community-based services, quality of care, disproportionately low-income population. supply and access to services, and cost con- Finally, cost containment goals can be tainment. Without adequate financing, the affected by the structure of financing. For need for LTC may not be translated into instance, the 2000 initiation of LTC insur- effective demand. That is, people may go ance in Japan was partly aimed at impos- without formal services (or rely on informal ing a more rational and efficient system for services) because they have no way to pay LTC. However, because the government had for them. This is already happening in China already been providing generous benefits to to some extent, with substantial vacancies in people who sometimes had relatively light some high-end residential care facilities. The needs, the government found backtracking on lack of sustainable and predictable, adequate the earlier generosity politically impossible. financing serves as a deterrent to private In contrast, when Germany started its quite sector investments into the industry. systematic program in 1995, it was providing For both publicly and privately provided rather low benefits to relatively few people, services, the financing system largely estab- so it could start out with stricter eligibility lishes the level of resources available to provide rules and lower benefits (Campbell 2002). LO N G - T E R M C A R E I N AG I N G E A S T A S I A A N D PAC I F I C 241 In most East Asian and Pacific countries, communities and relatively less on financial the policy challenge of providing broad-based support, also seem worth further exploration. and equitable access to LTC while ensuring financial sustainability looms large with the rapid aging of their populations. The experi- Coordinating LTC services ences of many OECD countries suggest that a The lack of integration of services across broad-based, universal LTC financing system different programs or settings, both within offers the most equitable coverage to people in LTC and across medical care and LTC, is a need of LTC. For East Asian and Pacific coun- common challenge across countries. Part of tries with rapidly aging populations (China, this problem is rooted in the common split Thailand, Vietnam), a targeted universalism between health care and social care. In most approach, which builds on the principle of countries, the LTC system is separate from universal coverage while targeting benefits for the health care system (a notable exception individuals with relatively higher care needs, is Belgium), even though the lines are not may be desirable. For younger countries, the always clear-cut. Bridging this gap has proved community-based approaches being piloted difficult because of the multiple provid- in countries such as Cambodia and Myanmar ers involved in service delivery, the complex with support from HelpAge International seem financing arrangements, and differences in promising in leveraging nonfinancial inputs working cultures (as discussed below). from communities and families that are crucial Many different actors and levels of gover- to achieving financial viability even with sup- nance are involved in service delivery, contrib- port from governments or development part- uting to fragmentation of services between the ners. Approaches such as the time banks in health and LTC sectors. Figure 7.6 illustrates some cities of China, which rely primarily on the problem of fragmentation in England, the convening power of government or local where multiple agencies are involved in the FIGURE 7.6 The health and long-term care systems in England are complex and riddled with perverse incentives Parliament Department of Department of Communities and Health Local Government (and social care) NHS England Care Quality Health improvement (national executive) Commission (quality (improvement agency and regulator) economic regulator) Local authorities Clinical commissioning (municipal groups (local service government) commissioners) Providers of health and care services (including public, private and voluntary sector providers) Citizens Funding Accountability Source: Ruthe Isden, personal communication, October 2015. 242 LIVE LONG AND PROSPER financing and delivery of health care and than social service employees. Differences LTC. The fragmentation is greater in those in interests, motivations, and behaviors countries where the acute care sector domi- at the provider level can lead to potential nates, the primary care sector is weak, and inefficiencies or failures in service delivery. community services are limited. Given that The workload of service providers is also a move toward greater coordination of care likely to increase, at least temporarily, as a and increased care of the elderly at home or result of integrating care services. Learning in communities is likely to result in decreased new referral and information systems and demand for hospital beds, a dominant acute adapting to new integrated ways of work- care sector can indeed be an impediment to ing can take time away from handling their reform (OECD 2013). existing workload (Maslin-Prothero and Moreover, the financing arrangements Bennion 2010). for health and LTC services are highly In East Asia and Pacific, health care and complex, with multiple sources of financ- LTC are completely separate from each ing associated with some indirect spillover other. In China, for instance, relatively few effects. In Germany, Japan, and Korea, pub- senior care facilities have medical or profes- lic LTC insurance operates separately from sional nursing staff available (Feng, Zhan, the health insurance system. As described et al. 2011). In most facilities, residents are in box 7.1, LTC hospitals and facilities in responsible for providing their own medica- Korea, financed by national health insur- tions (although facility staff often help with ance and LTC insurance, respectively, are in storing and administering them) and medical direct competition with each other. In the care, which may be covered by various health United States, the federal Medicare program insurance programs. In emergency situations (which provides health insurance—including that warrant hospitalization of a resident, a hospital insurance, medical insurance, and typical response by the facility is to call a fam- prescription drug coverage—for people ages ily member who will take the resident to the 65 and older as well as younger people with hospital, which could lead to problems in care certain disabilities) covers limited postacute transition and delays in transfer. The situa- care by nursing homes and home health agen- tion is compounded by the lack of portability cies, whereas Medicaid (a federal-state jointly of health insurance between regions in coun- funded, means-tested insurance program for tries such as China and Vietnam. low-income people) covers a broad array Bridging the gap between health care and of LTC services, although coverage varies LTC holds promise in two ways: by reducing by state (Grabowski 2007, 2012). In a set- the cost of acute and institutional care and ting with two or more bodies responsible for by improving service users’ satisfaction by funding a service, an incentive exists to pass enabling them to stay in their community on the costs to the other (Hofmarcher, Oxley, (OECD 2013). Integrated care for the elderly and Rusticelli 2007). It also increases the cost has been defined as “a coherent set of prod- to the individuals and to the public purse. ucts and services, delivered by collaborating In the United States, these concerns have led local and regional health care agencies through to the creation of a new Medicare-Medicaid securing liaison or links within and between Coordination Office under the Affordable the health and social care systems” (Hardy Care Act, with the express purpose of align- et al. 1999). It implies providing a compre- ing the two programs. hensive array of preventive, primary, acute, Differences in working cultures and ser- and LTC services, coordinated and provided vice providers’ roles in the two sectors have by an interdisciplinary team of professionals in also hindered coordination between health a community-based center and in participants’ care and LTC in the OECD. Health care homes, helping participants delay or avoid providers, particularly doctors, have tradi- long-term nursing home care. Coordinated tionally been seen as holding a higher status care experiments in Australia, Japan, and the LO N G - T E R M C A R E I N AG I N G E A S T A S I A A N D PAC I F I C 243 United States include efforts to align provid- Ensuring capacity and quality ers’ and users’ incentives for more efficient and for LTC provision quality-enhancing patterns of care. In time, they are likely to provide useful models that can be LTC is labor intensive, and the majority of the built upon in East Asian and Pacific countries. LTC workforce consists of front-line workers. In some cases, eligibility for LTC services In all countries, women—typically in their is determined by caseworkers employed middle ages—make up the vast majority of (or contracted) by municipal governments. the LTC workforce. The front-line workers These caseworkers evaluate the physical include certified nurses’ aides, home health and mental condition of the potential clients care aides, and home and personal care work- plus their income and assets and the extent ers who provide hands-on help for basic to which family care is available. The case- ADLs, such as eating, bathing, dressing, and worker draws up a personalized care plan using the toilet. In most countries, these are for services to be provided by the munici- low-skilled workers with minimal training pality, determining the amount of benefits requirements. In the United States, federal to be offered within the constraints of the law requires a minimum 75 hours of training municipal budget. An advantage of this or passing a certification exam for a certified system is that it allows the benefits to be nurses’ aide, although some states have addi- tailored flexibly to individual situations; tional requirements. moreover, if a means test for income or At the higher end of the skill set, the LTC assets is included, public resources do not workforce includes a group of licensed health go to people who could afford care on their professionals. Such professionals include own. A disadvantage of this system is that registered nurses, licensed practical and differences in the financial situations among vocational nurses, social workers, physical municipalities may result in regional varia- therapists, occupational therapists, physi- tions in eligibility criteria and benefit levels cian assistants and aides, and LTC facility that may be seen as unfair—or even as a administrators who often assume supervis- geographic lottery (Campbell 2014). ing or managerial responsibilities rather than The alternative is to decide on eligibility providing direct, hands-on care (Stone and and degree of disability through a national Harahan 2010). In only a few countries, such standardized instrument. The test usually as the Netherlands, are physicians directly measures how well applicants can perform involved in the provision of LTC. the standard ADLs. Japan uses a 74-item Many countries face a chronic shortage of questionnaire that is administered by a mod- LTC workers. Recruiting and retaining direct- erately trained public servant in a home visit care workers in LTC settings is particularly of an hour or so. The numerical results are challenging because of a multiplicity of fac- analyzed with a statistical algorithm to sort tors, such as low pay, low job prestige, few people into eight categories: independent; fringe benefits, and lack of career paths. As two grades of “needs support” at a relatively a result, these positions are often viewed as low level; and five grades of “needs care.” dead-end jobs characterized by high turnover, The computer’s decision is reviewed by a low retention, and job dissatisfaction. These local committee, which also looks at a brief workforce challenges are common across report on the applicant’s medical condition OECD countries, and policy efforts aimed to from the family physician and may move address them abound (Colombo et al. 2011). the grade up or down one level. Regardless Unfortunately, few successes exist thus far of which process is adopted, the key is to that can be readily shared and replicated in ensure that no one involved in making deci- different countries. sions about eligibility stands to gain from In this situation, high-income countries it and that the system has legitimacy in the such as Australia, Austria, Italy, and eyes of the public. Singapore are increasingly relying on migrant 244 LIVE LONG AND PROSPER workers from less-developed nations to other aspects of geriatric care. Governments provide care. In most cases, the immigrant should play a more active role in strength- workers are untrained women, often with ening workforce training programs in higher dubious immigration status, who live in the education or vocational schools through ear- household to provide full-time care for a frail marked funding and other forms of financial elder and to cook and clean for the family inducements (for example, fellowships, schol- as well. As their numbers grow, challenges arships, or targeted grants) similar to those emerge: quality of care, issues of fairness and provided to encourage the development of exploitation, stunting of the development of care facilities for seniors. Strengthening the more professional LTC, and demands for domestic LTC workforce may prove futile heavy and growing government subsidies if the same workers are attracted by better of the practice. Some high-income nations LTC work opportunities in higher-income have tried to regularize recruitment, train- East Asian and Pacific countries, so coordi- ing, and employment conditions of migrant nation between these programs is desirable. caregivers but without much success so far. To help ensure quality and standards in Japan is unusual in the extent to which it LTC, another priority is to address the lack relies on trained care workers employed by of effective regulatory oversight over the agencies in providing LTC (Campbell 2014). booming private senior care sector as well as At present, a program is being developed in home- and community-based providers. In Japan to bring and train LTC workers from most East Asian and Pacific countries includ- the Philippines and other Southeast Asian ing China, little effective regulatory oversight countries. exists for private sector LTC, partly caused by The lack of a qualified and professional a government focus on building more services workforce in LTC is a pressing issue across quickly to fill the supply gap, as described East Asia and Pacific. In China, for instance, earlier. Existing regulations are limited, focus the majority of direct-care workers in senior only on structure, and do not explicitly con- care facilities are inadequately trained and sider quality of care, let alone quality-of- poorly paid, with few having received any life issues. Furthermore, the various types professional training (Feng et al. 2014). In the of home- and community-based providers, emerging home- and community-based ser- such as assisted living facilities, board-and- vice sector, the workforce problems of recruit- care homes, and adult day care services, are ment and training are even worse. Inadequate subject only to some basic licensure require- training for direct-care workers is not the ments and loose regulations within individual only impediment to high-quality services: pro- states (Mollica, Sims-Kastelein, and O’Keeffe fessional clinical and management staff are 2007). The experiences of OECD countries also needed to ensure a transition to a mod- suggest a commonality in regulatory over- ern, information-based LTC delivery system. sight and quality assurance in LTC: they are Programs have been introduced in Taiwan, more visible, rigorous, and effective in insti- China, to train informal caregivers to cope tutional care settings compared to in-home with the shortage of skilled care workers. and community-based care settings.8 Quality East Asian and Pacific policy makers— assurance for home- and community-based particularly in countries such as China, care is particularly challenging where public Thailand, and Vietnam where LTC needs are financing of such care is limited and therefore growing rapidly—should prioritize education the government has little stake or incentive to and training initiatives to develop a profes- strengthen regulatory oversight. sionalized LTC workforce. Currently, few Information technology provides immense educational programs exist with curricula opportunities for East Asian and Pacific coun- focused on gerontological nursing, social tries to advance rapidly in the area of regu- work, and professional skills related to all latory oversight and quality assurance even LO N G - T E R M C A R E I N AG I N G E A S T A S I A A N D PAC I F I C 245 in their fledgling LTC systems. The experi- enormously challenging. Investing early on ences in other countries demonstrate that in a three-tiered formal LTC system that puts effective oversight entails building an infor- the emphasis on home- and community-based mation system to facilitate evidence-based care can avoid many of these challenges, and policy making, quality assurance, and regula- in the East Asian and Pacific context, it is also tory enforcement (Mor, Leone, and Maresso likely to be more culturally acceptable. An 2014). In the United States, all publicly certi- added advantage of the early establishment of fied nursing homes must report both facility- a formal LTC system in middle-income nations and resident-level data electronically, using a is that training and certification programs can uniform annual facility inspection survey and be built up gradually as demand expands. standardized resident assessment (Stevenson Whatever directions are chosen, piloting and and Bramson 2014). Building an informa- careful evaluation seem appropriate. tion infrastructure in East Asian and Pacific Even in younger countries, which do not countries may take time and resources, but yet have a large frail, elderly population, ini- the Internet and availability of cloud-based tiating a good LTC system early is fiscally systems mean that countries such as China prudent and makes political sense. Even at an can move more quickly than was possible in early stage in population aging, the burden of the United States two decades ago (Feng et al. caring for the elderly is acutely felt by more 2012). This process can be accelerated because and more ordinary citizens who are concerned the government is increasingly at risk—with about what will happen to their parents and significant investments of public resources in ultimately themselves. The earlier an LTC sys- the rapidly growing senior care sector—and tem is started, the lower the cost. When the better positioned to wield its growing “buyer number of qualifying older people is relatively power” to make regulations work. low, and traditional family supports are still working fairly well, the demand for public pro- grams will not be very high. Moreover, when Conclusion no public services had previously been avail- Middle-income and lower-middle-income able, even modest benefits will be welcomed. countries in East Asia and Pacific should The government can decide later on whether consider early adoption of systematic LTC coverage or benefits should be upgraded. programs, ideally before the frail, elderly Initiating a good LTC system early is also population becomes too large. Israel and good public policy. As the experience of high- Korea provide two examples of how early income nations demonstrates, a well-designed establishment of LTC systems can keep costs LTC system delivers good benefits at moder- down.9 If a comprehensive system is estab- ate cost—a key yardstick for all public poli- lished before other potential solutions to the cies. Good LTC will lead to real savings in problem have evolved much, serious difficul- health care provision because frail older ties can be avoided (Campbell 2014). As dis- people who get good day-to-day care are less cussed, in the face of a rapidly growing frail, likely to get sick and less likely to relapse after elderly population with inadequate family being hospitalized, and LTC is inherently less support, local governments or charities and costly than medical care because it uses staff other welfare groups generally resort to build- with lower wages and does not lead to open- ing lots of nursing homes to house those indi- ended benefits. Moreover, if care for frail viduals. Once institutional care becomes the older people can be handled by a dedicated norm, and if an increasing number of public system, the health care system can focus on or quasi-public nursing homes—or worse still, acute care and prevention while other social hospital owners—band together to protect programs can concentrate on pressing needs their interests politically, changing course and like poverty, urban housing, and helping adopting more rational programs becomes young people to succeed. 246 LIVE LONG AND PROSPER Notes receive tax exemption and other favorable policy treatments. In actual operation, 1. The health component of LTC spending however, they tend to behave more like for- relates to health and nursing care for patients profit facilities. who need assistance on a continuing basis 7. In Singapore, the government recently because of chronic impairments and reduced introduced subsidies to lower- and middle- degree of independence and ADLs. The income individuals to help with LTC costs. following items are included in the LTC health Subsidies are means tested, so that families component: palliative care; long-term nursing with fewer resources receive more than care; personal care services (assistance with families that earn more. See the Ministry ADL restrictions); and services in support of of Health’s MediShieldLife website at informal (family) care. The social component https://www.moh.gov.sg/content/moh_web of LTC includes home help (help with IADLs); /medishield-life/resources---faqs/healthcare residential (care) services; and other social -financing-in-singapore.html. The State services provided in an LTC context. Council document of 2013 is titled “Opinions 2. Data for unmet care needs were extracted of the State Council on Accelerating the from questions asked in the World Health Development of Services for the Aged” Survey (WHS), a multicountry study, to (unofficial translation, dated September 6, generate comparable information on the 2013). health of adults and health systems. One 8. In the United States, for instance, nursing strength of the WHS is the large sample sizes homes are among the most tightly regulated in participating countries. Survey findings sectors and must comply with a set of were applied to population estimates from minimum standards mandated by the federal the UN World Population Prospects data set government (Mor 2005; Stevenson and to create estimates of care at the national Bramson 2014). level. Eight countries in the region (Cambodia, 9. As described in Asiskovitch (2013), Israel, Indonesia, Japan, Korea, Singapore, Thailand, which established the world’s first LTC and Timor-Leste) were not included in system in 1988 when less than 10 percent the WHS. Estimates for these countries of its population was ages 65 and older, has were derived using data from countries in succeeded in keeping costs down and the the region with similar income levels. For number in institutions very low. As described example, Cambodia’s unmet care need was in Duk (2012), Korea established a system estimated by averaging the unmet care needs largely based on the Japanese model in 2008 identified in the Lao People’s Democratic when only 9 percent of its population was Republic, Myanmar, the Philippines, and ages 65 or older, and the costs have been quite Vietnam. Therefore, estimates of prevalence moderate. and burden of care in these countries should be interpreted with caution, because the underlying data do not come from the country References itself (personal communication from Meredith AHCA (American Health Care Association). Wyse, HelpAge International, 2015). 2 0 1 3 . “ LT C S t a t s : N u r s i n g F a c i l i t y 3. For more information, see Colombo et al. Patient Characteristics Report.” AHCA, (2011); OECD Health Statistics 2012, http:// Washington, DC. dx.doi.org/10.1787/health-data-en; and Asiskovitch, Sharon. 2013. “The Long-Term Care OECD and European Commission (2013). Insurance Program in Israel: Solidarity with the 4. Singapore Programme for Integrated Care for Elderly in a Changing Society.” Israel Journal the Elderly (SPICE), http://www.aic.sg/page of Health Policy Research 2 (3). http://www .aspx?id=782 (April 30, 2013). .ijhpr.org/content/2/1/3. 5. See HelpAge International (2015) on Campbell, John Creighton. 2002. “How Policies community-based social care in East and Differ: Long-Term-Care Insurance in Japan Southeast Asia. and Germany.” In Aging and Social Policy: 6. Currently, almost all private sector senior care A German-Japanese Comparison, edited by facilities in China are registered as nonprofit, Harald Conrad and Ralph Lützeler, 157–87. nonenterprise entities, a status required to Munich, Germany: Iudicium. LO N G - T E R M C A R E I N AG I N G E A S T A S I A A N D PAC I F I C 247 ———. 2014. “Japan’s Long-Term Care Insurance Feng, Zhanlian, Chang Liu, Xinping Guan, and Program as a Model for Middle-Income Vincent Mor. 2012. “China’s Rapidly Aging Nations.” In Universal Health Coverage Population Creates Policy Challenges in for Inclusive and Sustainable Development: Shaping a Viable Long-Term Care System.” Lessons from Japan, edited by Naoki Ikegami, Health Affairs 31 (12): 2764–73. doi:10.1377/ 57–68. Washington, DC: World Bank. hlthaff.2012.0535. Campbell, John Creighton, and Naoki Ikegami. Feng, Zhanlian, Heying J. Zhan, Xiaotian Feng, 2000. “Long-Term Care Insurance Comes to Chang Liu, Mingyue Sun, and Vincent Mor. Japan.” Health Affairs 19 (3): 26–39. 2011. “An Industry in the Making: The ———. 2003. “Japan’s Radical Reform of Long- Emergence of Institutional Elder Care in Urban Term Care.” Social Policy and Administration China.” Journal of the American Geriatrics 37 (1): 21–34. Society 59 (4): 738–44. Campbell, John Creighton, Naoki Ikegami, and Frank, Richard G., Marc Cohen, and Neale Mary Jo Gibson. 2010. “Lessons from Public Mahoney. 2013. “Making Progress: Expanding Long-Term Care Insurance in Germany and Risk Protection for Long-Term Services Japan.” Health Affairs 29 (1): 87–95. and Supports through Private Long-Term CMS (Centers for Medicare and Medicaid Services). Care Insurance.” Long Beach, CA: SCAN 2012. Nursing Home Data Compendium, Foundation. http://www.thescanfoundation 2012 Edition. Baltimore, MD: CMS. .org/sites/thescanfoundation.org/files/tsf_ltc Colombo, Francesca, Ana Llena-Nozal, Jérôme -financing_private-options_frank_3-20-13. Mercier, and Frits Tjadens. 2011. Help pdf. Wanted? Providing and Paying for Long- Gibson, Mary Jo, and Donald L. Redfoot. 2007. Term Care . Paris: OECD Publishing. “Comparing Long-Term Care in Germany doi:10.1787/9789264097759-en. and the United States: What Can We Learn Cuellar, Alison Evans, and Joshua M. Wiener. from Each Other?” Research Report 2007- 2000. “Can Social Insurance for Long-Term 19, AARP Public Policy Institute, Washington, Care Work? The Case of Germany.” Health DC. http://assets.aarp.org/rgcenter/il/2007_19 Affairs 19 (3): 8–25. _usgerman_ltc.pdf. De la Maisonneuve, Christine, and Joaquim Giles, John, Xinxin Chan, Yuqing Hu, and Yaohui Oliveira Martins. 2013. “Public Spending on Zhao. 2015. “What are the Current and Future Health and Long-Term Care: A New Set of Long-Term Care Needs of China’s Aging Projections.” OECD Economic Policy Paper 6, Population?” Background paper, World Bank, Organisation for Economic Co-operation and Washington, DC. Development, Paris. Grabowski, David C. 2007. “Medicare and Duk, Sunwoo. 2012. “The Present Situation Medicaid: Conflicting Incentives for Long- and Problems of the Long-Term Care Insurance Term Care.” Milbank Quarterly 85 (4): in South Korea: From Comparative Perspectives 579–610. between South Korea and Japan.” Japanese ———. 2012. “Care Coordination for Dually Journal of Social Security Policy 9 (1): 49–60. Eligible Medicare-Medicaid Beneficiaries under Feng, Zhanlian, Mary L. Fennell, Denise A. the Affordable Care Act.” Journal of Aging Tyler, Melissa Clark, and Vincent Mor. 2011. and Social Policy 24 (2): 221–32. “Growth of Racial and Ethnic Minorities in Hardy, Brian, Ingrid Mur-Veemanu, Marijke U.S. Nursing Homes Driven by Demographics Steenbergen, and Gerald Wistow. 1999. and Possible Disparities in Options.” Health “Inter-agency Services in England and the Affairs 30 (7): 1358–65. Netherlands: A Comparative Study of Feng, Zhanlian, Xinping Guan, Xiaotian Feng, Integrated Care Development and Delivery.” Chang Liu, Heying J. Zhan, and Vincent Health Policy 48 (2): 87–105. Mor. 2014. “Long-Term Care in China: HelpAge International. 2015. “Community-Based Reining in Market Forces through Regulatory Social Care in East and Southeast Asia.” Oversight.” In Regulating Long-Term Care Briefing Note, HelpAge International East Quality: An International Comparison , Asia/Pacific Regional Office, Chiang Mai, edited by Vincent Mor, Tiziana Leone, and Thailand. Anna Maresso, 409–443. Cambridge, U.K.: Hirdes, John P., Lori Mitchell, Colleen J. Maxwell, Cambridge University Press. and Nancy White. 2011. “Beyond the ‘Iron 248 LIVE LONG AND PROSPER Lungs of Gerontology’: Using Evidence Leone, and Anna Maresso, 180–210. to Shape the Future of Nursing Homes in Cambridge, U.K.: Cambridge University Press. Canada. Canadian Journal on Aging 30 (3): Maslin-Prothero, Sian E., and Amy E. Bennion. 371–90. 2010. “Integrated Team Working: A Literature Hofmarcher, Maria M., Howard Oxley, and Review.” International Journal of Integrated Elena Rusticelli. 2007. “Improved Health Care (10): 1–11. System Performance through Better Care Mollica, Robert, Kristin Sims-Kastelein, and Coordination.” OECD Health Working Paper Janet O’Keeffe. 2007. Residential Care and 30, Organisation for Economic Co-operation Assisted Living Compendium: 2007. Prepared and Development, Paris. for the U.S. Department of Health and Johnson, Richard W., and Joshua M. Wiener. 2006. Human Services, Office of Disability, Aging, A Profile of Frail Older Americans and Their and Long-Term Care Policy and Research Caregivers. Washington, DC: Urban Institute. Triangle Institute. RTI International: Research http://www.urban.org/publications/311284 Triangle Park, NC. http://aspe.hhs.gov/basic .html. -report/residential-care-and-assisted-living Jung, Hye-Young, Soong-Nang Jang, Jae Eun -compendium-2007. Seok, and Soonman Kwon. 2014. “Quality Mor, Vincent. 2005. “Improving the Quality of Monitoring of Long-Term Care in the Republic Long-Term Care with Better Information.” of Korea.” In Regulating Long-Term Care Milbank Quarterly 83 (3): 333–64. Quality: An International Comparison , Mor, Vincent, Tiziana Leone, and Anna Maresso, edited by Vincent Mor, Tiziana Leone, and eds. 2014. Regulating Long-Term Care Anna Maresso, 385–408. Cambridge, U.K.: Quality: An International Comparison . Cambridge University Press. Cambridge, U.K.: Cambridge University Press. Kaiser Family Foundation. 2013. “A Short Look Nadash, Pamela, Pamela Doty, and Nathalie at Long-Term Care for Seniors.” Journal of the Racco. 2013. “The Private Long-Term Care American Medical Association 310 (8): 786–87. Insurance Market in France: Implications Keenan, Teresa A. 2010. “Home and Community for the United States.” Presented at the Preferences of the 45+ Population.” annual meeting of the Gerontological Society Washington, DC: AARP. of America, New Orleans, LA, November Kim, Hongsoo, Yong-Il Jung, and Soonman 20–24. Kwon. Forthcoming. Delivery of Institutional OECD (Organisation for Economic Co-operation Long-Term Care under Two Social Insurances: and Development). 2013. Report on Integrated Lessons from the Korean Experience. Health Service Delivery for Vulnerable Groups . Policy. doi:10.1016/j.healthpol.2015.07.009. DELSA/ELSA/WP1(2013)8. Paris: OECD. Kwon, Soonman. 2014. “Long-Term Care OECD and European Commission. 2013. A Good Insurance and Care Coordination in Korea.” Life in Old Age? Monitoring and Improving Presented at the conference on Challenges Quality in Long-Term Care. Paris: OECD. of Population Aging organized by the Hong Research Base. 2012. “Care in Old Age in Kong University of Science and Technology Southeast Asia and China: A Situational and the World Bank, Hong Kong SAR, China, Analysis.” Commissioned by HelpAge April 14–17. International, Chiang Mai, Thailand. Lu, Jui-fen Rachel. 2014. “Reflections on LTC SALAR (Swedish Association of Local Authorities Insurance: What Can Be Learned from the and Regions). 2007. Care of the Elderly in Regional Experiences.” Presented at the Sweden Today, 2006. Stockholm: SALAR. conference on Challenges of Population Aging Shanghai Daily. 2013. “Elderly Prefer to Be Cared organized by the Hong Kong University of for at Home.” September 23. http://english Science and Technology and the World Bank, .eastday.com/e/130923/u1a7675125.html. Hong Kong SAR, China, April 14–17. Somanathan, Aparnaa. 2015. “Long-Term Malley, Juliette, Jacquetta Holder, Rachael Care in Aging East Asia and Pacific.” Dodgson, and Samantha Booth. 2014. Background paper for the East Asia and “Regulating the Quality and Safety of Long- Pacific regional report on aging, World Bank, Term Care in England.” In Regulating Washington, DC. Long-Term Care Quality: An International Stevenson, David, and Jeffrey Bramson. 2014. Comparison, edited by Vincent Mor, Tiziana “Regulation of Long-Term Care in the LO N G - T E R M C A R E I N AG I N G E A S T A S I A A N D PAC I F I C 249 United States.” In Regulating Long-Term by Robert H. Binstock and Linda K. George, Care Quality: An International Comparison, 309–22. London: Elsevier. edited by Vincent Mor, Tiziana Leone, and Wiener, Joshua M., Wayne L. Anderson, and Anna Maresso, 289–323. Cambridge, U.K.: David Brown. 2009. “Why Are Nursing Home Cambridge University Press. Utilization Rates Declining?” Prepared for Stone, Robyn, and Mary F. Harahan. 2010. Centers for Medicare and Medicaid Services, “Improving the Long-Term Care Workforce Baltimore, MD. Serving Older Adults.” Health Affairs 29 (1): Wiener, Joshua M., and Alison Evans Cuellar. 109–15. 1999. “Public and Private Responsibilities: UN (United Nations). 2013. “World Population Home and Community-Based Services in the Prospects: The 2012 Revision, Volume I: United Kingdom and Germany.” Journal of Comprehensive Tables.” Working Paper Aging and Health 11 (3): 417–44. ESA/P/WP.228, United Nations, Department Wiener, Joshua M., and Jane Tilly. 2003. “Long- of Economic and Social Affairs, Population Term Care: Can States Be the Engine of Division, New York. Reform?” In Federalism and Health Policy, Van den Berg, Bernard, Denzil G. Fiebig, and Jane edited by John Holahan, Alan Weil, and Hall. 2014. “Well-Being Losses Due to Care- Joshua M. Wiener, 249–92. Washington, DC: Giving.” Journal of Health Economics 35: Urban Institute Press. 123–31. Wiener, Joshua M., Jane Tilly, Anna Howe, Colleen Wiener, Joshua M. 1996. “Long-Term Care Doyle, Alison Evans Cuellar, John Campbell, Reform: An International Perspective.” In and Naoki Ikegami. 2007. Quality Assurance OECD, Health Care Reform: The Will to for Long-Term Care: The Experiences of Change , 67–79. Paris: Organisation for England, Australia, Germany, and Japan . Economic Co-operation and Development. Washington, DC: AARP. ———. 2011. “Long-Term Care Financing, Xinhua. 2014. “In-Home Nursing Gains Service Delivery and Quality Assurance: The Popularity in Aging China.” China Daily , International Experience.” In Handbook of March 13. http://www.chinadaily.com.cn Aging and the Social Sciences, 7th ed., edited /business/2014-03/14/content_17346869.htm. Looking Ahead IV To Live Long and Prosper: Looking Ahead 8 Introduction the pace of aging, the lower income levels at which it is happening, and the differences in A great deal still remains to be understood culture all point to a need for more research about aging in East Asia and Pacific. Given and experimentation specific to the region to the rapid pace of aging in the region, it is not guide policy development. surprising that large knowledge gaps remain This chapter outlines an agenda of aging- in understanding the process and drivers of related issues that will require close atten- aging, the emerging policy responses, the tion in East Asia and Pacific. Mirroring the effects of aging on societies and economies, issues addressed in earlier chapters, this and the channels through which those effects agenda ranges from better understanding of occur. Around the world, aging is a phe- demographic trends to the macroeconomic nomenon that remains inadequately studied, effects of aging, to policy change and societal although the explosion of research and policy responses, to the impacts at the household innovation in countries in the Organisation and individual levels. It is not comprehensive for Economic Co-operation and Development but rather identifies within the areas covered (OECD) and in Eastern Europe and Central in the report where more research is needed to Asia has begun to shed more light on the inform public policy development. A host of complex issues involved. In East Asia and other relevant issues beyond the scope of this Pacific, with notable exceptions such as Japan report will also require attention, including and Singapore, the knowledge gaps are even adjustments in urban environments to aging more acute. Although insights from OECD populations, the drivers of cognitive resilience experience and research on aging are valu- and decline, the role of technology in promot- able, questions remain on what aspects of that ing healthy and productive aging, the social experience are transferrable to the very dif- participation of older people, and the effects ferent setting of East Asia and Pacific, where of natural disasters on older people. 253 254 LIVE LONG AND PROSPER Demographics of these and other factors through careful sociological and quantitative research will be As discussed in chapter 1 of this report, crucial. One potentially important piece of demographic dynamics are inherently diffi- evidence in this regard from OECD countries cult to predict accurately, and deepening the is that fertility behavior in those countries capacity within governments and the regional in recent years is positively associated with research community to track actual trends women’s labor force participation, a reversal will be vital. This approach will require much of the situation in the 1980s (OECD 2011). more frequent updating of population projec- However, this reversal has not been seen in tions and modeling of different longevity and Japan and the Republic of Korea. fertility scenarios (as the United Nations itself Whatever the evolution of demograph- does with its multiple fertility scenario projec- ics, adopting alternative dependency mea- tions). East Asia and Pacific has witnessed a sures to better understand the real world fairly consistent underestimation of the rate effects of demographic, epidemiological, and of fertility decline in recent decades. Apart behavioral change would be useful. Future from the historical imprecision of population dependency in East Asia and Pacific is likely estimates, developments in medicine and tech- to continue longer in the early stages of life nology suggest that accelerated extensions in and shrink at the later stages (known as the life expectancy have the potential to radically Ben-Porath effect). In East Asian and Pacific alter the notion of “old” within our lifetimes, societies where informal sector workers—and in terms of both length of life and cognitive even formal workers in richer economies in and physical functioning. On a more modest the region—already work well beyond age 65 note, incorporating the rapidly changing edu- and where healthy life expectancy has been cational profiles of populations in the region steadily increasing, the traditional definition is an important factor that has led to lower of “working age” as finishing at age 65 needs fertility and longer life expectancy projections to be reexamined. Similarly, in a region where for many countries. This is an approach taken a growing number of countries have high by, for example, the Wittgenstein Centre for senior secondary and rising tertiary education Demography and Global Human Capital.1 completion rates, assuming that the working More work is needed to understand the age begins at age 15 is becoming less defensi- drivers of fertility behavior in East Asia and ble. New dependency measures that take into Pacific, including how such behavior may dif- account healthy life expectancy, years until fer from that in other parts of the world. In death, and observed work behavior will be contrast to some OECD countries, the experi- vital to inform public policy. They may also ence of richer East Asian and Pacific econo- help mitigate negative attitudes about older mies suggests that important differences exist people as unproductive or burdensome on in fertility behavior that raise questions about society. the likelihood of future reversals in fertility. Understanding the roots of these apparent differences is vital to shaping public policy responses and predicting future demographic Macro and fiscal issues trends. A range of hypotheses has developed Assessing the effects of aging on economic in this regard, including the persistence of growth and income distribution is a high traditional gender roles and responses of priority for policy makers in the region and younger women to them, the costs and pres- an issue with repercussions for the global sures of child rearing in some East Asian and economy. This is a research agenda that has Pacific societies, the degree of job security and already attracted much attention. However, it flexibility in employment terms and condi- has sometimes divergent findings and remain- tions, and more limited immigration in some ing knowledge gaps that suggest a few areas countries. Analyzing the relative importance for further work. TO LIVE LONG AND PROSPER: LOOKING AHEAD 255 An overarching research need is to move divergent predictions. Some analysts predict beyond simple growth accounting projec- inflationary pressures from aging as the share tions of the interaction of aging and growth of people producing relative to those consum- toward analysis that better reflects policy and ing shrinks, whereas other researchers argue behavioral change in response to aging. More that the net effect will ultimately depend on dynamic estimations of the growth effects of the ability of a particular cohort to influence aging are needed, as well as a better empirical monetary policy. Older and younger cohorts understanding of the channels through which may have competing demands for inflation- growth may be affected by aging (see discus- ary or deflationary policies, which may play sion below). More fundamentally, broader a key redistributive role in aging economies. definitions of societal welfare may be of par- If older generations have more influence on ticular relevance in aging societies. For exam- policy makers, then monetary policy will ple, factors such as improved healthy life favor a low or even negative rate of inflation expectancy and lower morbidity, falling crime to protect money holdings. This area will be rates, better working conditions for older important to track in practice, more so given people, and other welfare improvements may the higher reliance of older people on fixed lead to improvements in quality of life even in incomes from pensions, savings, or other cases where growth in gross domestic product assets. A related question is the evolution of per capita is slower. relative returns to factors of production in A key factor in understanding growth the face of aging. If labor forces shrink, this dynamics in aging societies is savings behav- may affect the relative price of labor. There ior across the life cycle and the way it evolves may also be potential effects on the returns in response to rapid aging and policy change. to different investment instruments as forced This is an area where existing research sug- savings through pension and old age savings gests that East Asia and Pacific may be dis- vehicles influence market demand (for exam- tinctive, but where findings are not always ple, for long maturity bonds) and, over time, consistent across countries, research methods, potentially affect the relative price of different time, and cohorts. Even where empirical find- investment instruments. ings concur, the drivers of observed savings In addition, understanding the potential behavior may differ, from cultural elements distributional effects of aging is an emerging such as variable bequest motives to rational priority in a region with growing concerns financial planning in the face of low social about inequality. The potential effects of insurance coverage and shallow financial aging on inequality may be divided into two protection. Whatever the historical patterns, broad groups: how aging plays out within changes in social insurance coverage, social population cohorts as they age and how it attitudes to support across generations, the may affect the distribution of income across relative emphasis of macroeconomic policy generations. Inequality will be driven not only between consumption and savings, and access by changes in the relative returns to labor, to financial services and capital markets all capital, and skill with population aging, but suggest substantial possibilities for shifts in also by the pension, health insurance, and savings behavior. A related research agenda other age-related policies in place. Although addresses the relationship between savings rising returns to labor may benefit the young behavior and intermediation of household in an aging society, redistribution through savings into capital investment. pensions and access to health care may sug- Divergent research findings also suggest a gest that older residents are better off. The need to understand better the effects of aging contribution of these divergent factors to on other macroeconomic variables, includ- overall inequality will vary according to ini- ing inflation and the relative returns to fac- tial conditions, in particular the relative living tors of production. With respect to inflation, standards of elderly and working-age people. research on the effect of aging has produced The role of population aging in inequality 256 LIVE LONG AND PROSPER dynamics is thus a topic that would benefit policy choice, health technology, and aging from increased attention as aging accelerates itself to cost pressures in health and long- across the region. term care systems, particularly in light of With regard to fiscal issues, regular projec- research on China that suggests that the tions of age-related spending scenarios and respective contributions seen in OECD improved approaches to disentangling the countries may vary in rapidly urbanizing contribution of aging and other factors to and aging countries. public spending are needed. Although pro- jecting the effects of pension, health, long- term care, and other age-related spending Labor markets has matured in the richer economies of the The importance of the labor market as a chan- region, it remains in its infancy in the devel- nel through which the socioeconomic effects oping countries. There is an ongoing agenda of aging are felt is well demonstrated. Thus, of actuarial capacity building in this regard, gaining a deeper understanding of evolving including education in actuarial science, labor market behavior across the life cycle development of examination and licensing and of the most effective policies to mitigate procedures, codes of conduct and profes- the effects of aging on the labor force is vital. sional bodies, and standards of practice. One area needing better evidence is the Strengthening actuarial capacity in relative productivity of workers across the government and the private sector in devel- life cycle, as well as the factors that can affect oping East Asian and Pacific countries is a worker productivity at older ages. In devel- priority, as is integrating such analysis sys- oping East Asian and Pacific countries, lim- tematically into public policy making (for ited information exists on productivity across example, by mandating regular updates of the life cycle, including how the picture may actuarial projections of pension and health be changing as more educated cohorts of insurance programs and by requiring actu- workers begin to pass through middle age. arial results be included in proposals for Assessing the productivity profile at the firm reform of social insurance programs). Many level based on the workforce age mix within OECD countries have a legal requirement firms is important in determining how older of periodic publication of actuarial results and younger workers can best combine their of their social security funds—for exam- relative strengths. This approach would help ple, an annual actuarial valuation in the inform further reforms of seniority wage sys- U.S. Social Security Administration (over a tems and, more generally, optimal combina- 75-year time horizon) and five yearly (over tions of workers of different ages in the work 90 years) and three yearly actuarial reports place. Related to this is assessment of work- in Japan and Canada, respectively (Hoskins place adjustments to enhance the productivity 2010). More broadly, Australia is obliged of older workers, an area still in need of evi- every five years to publish an intergenera- dence on the most cost-effective adjustments. tional report (latest edition 2015), which Few studies document workplace adjustments forecasts the effects of demographics on and organizational changes designed to facili- growth, labor force and productivity, and tate and encourage longer working lives in public finances, assuming unchanged pub- East Asia and Pacific. This is an area where lic policies (Commonwealth of Australia studies could be a catalyst to wider adoption Treasury 2015). It has proven to be an of such adjustments and changes and would effective vehicle for generating public warrant public research support. debate on the effects of fertility, longevity Understanding effective practices with and aging, and migration on the economy respect to development and retention of and fiscal position. At a basic research adult skills in middle age and beyond will level, more Asia-specific work is needed be vital. This includes both on-the-job train- to understand the relative contributions of ing and more structured adult learning and TO LIVE LONG AND PROSPER: LOOKING AHEAD 257 would help identify market failures in the As countries in the region continue to adult training market where public interven- reform their social security systems, having tion may be warranted, as suggested by evi- better insights on the labor market effects dence from Japan and Korea in chapter 4 of of incentives and disincentives to retire in this report. With regard to life-long learning social security and tax systems is impor- outside the workplace, this is an area where tant. Evidence from OECD countries is clear advocacy sometimes runs ahead of evidence that retirement incentives in social security with regard to workers in mid-career and systems substantially affect the behavior of beyond. In countries such as China, signifi- older workers. In addition, current retire- cant expansion has occurred in short-course ment rules and early retirement incentives adult training in the TVET (technical and in middle-income East Asian and Pacific vocational education and training) system in countries appear to lead urban formal sec- response particularly to the needs of migrant tor workers to retire earlier than will be sus- workers, though they are usually in their tainable in an aging society. However, more 20s and 30s. Such experience may be adapt- systematic analysis is needed of the observed able for those in mid-career and beyond. effects in the region’s developing countries of Continuing evaluations of the experience and how strongly pension rules alter labor force cost-benefit of older worker training in richer behavior, how formal sector work and self- East Asian and Pacific economies would also employment after withdrawal from wage be important. employment interact, and what levels of gen- In light of the proliferation of initiatives erosity of social pensions produce significant to stimulate female labor force participation effects on the work of older people. These across the region’s richer countries in recent questions are more challenging in environ- years, evaluating the relative effects of dif- ments with high informality, but are impor- ferent approaches will be important. Certain tant to analyze further. public interventions such as increasing the An important dynamic in labor markets supply of affordable child care and elder care of developing East Asian and Pacific coun- appear to have positive effects on working tries that needs to be better understood is the women in East Asia and Pacific and glob- interaction of the labor force participation of ally. The effects of other interventions such working parents of young children, the work as paid parental leave have less demonstrated and caring responsibilities of grandparents, effect in the labor market (though it may have and the child care and long-term care mar- other positive effects on children and family kets. The care provided by grandparents life). A more rigorous cost-benefit analysis to across the region appears to facilitate the inform future policy prioritization, including work of their adult children, and also par- the potential interaction effects of various tially fills an important gap in long-term care incentives and benefits, would be valuable. markets. One must understand the extent To date in richer East Asian and Pacific to which caring duties push older people, economies, little rigorous assessment has particularly women and especially those in been done of the cost-effectiveness of various urban areas, out of the labor force earlier financial incentives to promote extension of than they may otherwise choose and the for- active working lives. An important question mal child and long-term care arrangements related to financial incentives for hiring older that are needed to offset this effect if it is workers is the extent to which they should occurring. be targeted by age only or as part of broader A final and inadequately studied issue schemes based on characteristics such as in the East Asian and Pacific context is unemployment or low income. In addition, the degree to which extending the work- a better understanding of what influences ing lives of older people has any effects employer attitudes toward hiring or retaining on the employment prospects of younger older workers is needed. people. Evidence from OECD countries is 258 LIVE LONG AND PROSPER very clear that extending the working lives older people and their families across a range of older people does not harm, and may of dimensions. For countries with ongoing mildly help, the employment prospects of HRS-style surveys, the challenge is to sustain younger people. However, the phenomenon funding for future rounds and, ideally, to is not well studied in East Asia and Pacific, bring the surveys into the mainstream of offi- though studies in China find a similar effect. cial statistical systems. Scope also exists to Nevertheless, in economies where public further improve the aging surveys, through employment accounts for a higher share wider sampling of working-age people and than in OECD countries (for example, the the eldest elderly; introduction of a panel Lao People’s Democratic Republic, the element to the samples lacking that element; Pacific Island countries, and Vietnam) and and inclusion of more detailed questions fiscal constraints are tight, assessing the on issues such as asset wealth, physical and extent to which OECD findings hold would cognitive functioning, and caring arrange- be important. This approach is part of a ments. 3 Countries without an HRS-style global need to assess the lump-of-labor fal- survey should consider initiating one, poten- lacy in non-OECD settings. tially with co-financing from global research funding sources such as those that provided The situation of older people support for CHARLS (China Health and and households in East Asia and Retirement Longitudinal Study) in China Pacific (from the National Institute on Aging at the National Institutes of Health, U.S. Understanding the financial well-being, Department of Health and Human Services). health, family relations, and consumption In cases where funding or capacity is lim- and savings behavior of different age groups ited, inclusion of aging modules in regular requires better data than are currently avail- household living standards surveys would able from most household living standards be useful. For example, Thailand provides surveys in East Asia and Pacific. Current aging-related modules in its general house- national household surveys of income and hold survey every four to five years. consumption tend to be poorly geared toward Areas needing a better understanding of studying aging, because most use the house- the dynamics of elderly and household wel- hold as a unit of analysis and are thus not fare include the following: well suited to understanding intrahousehold well-being and behavior. • Total wealth of older people beyond cur- The region has benefited in recent years rent income and expenditure. The mixed from the spread of surveys dedicated to evidence from income and expenditure the situation of older people, but scope to surveys on elderly welfare highlights the expand geographic coverage and enhance importance of including wider aspects of the aging focus of regular surveys remains. household wealth in welfare and inequal- The expansion of aging-related surveys in the ity analysis, including housing and land region includes health and retirement liv- ownership, savings, and other assets. ing survey (HRS)–style surveys in China, • The extent that the labor force behavior Indonesia, Japan, Korea, and Thailand— of older people changes as societies urban- and also India in South Asia. 2 Countries ize and other factors such as health and such as Myanmar, Thailand, and Vietnam longevity change. The very distinct nature also have conducted useful surveys of older of rural and urban labor markets for populations, and initiatives such as the older people in East Asia and Pacific sug- World Health Organization’s SAGE (Study gests potential shifts in work and retire- on Global Ageing and Adult Health) survey ment behavior as urbanization deepens. include China. These surveys allow for more However, the behavior of urban older detailed analysis of the living conditions of people is likely to change over time in the TO LIVE LONG AND PROSPER: LOOKING AHEAD 259 face of demographic, policy, and behav- Health and long-term care ioral changes that affect return to work at older ages. A first crucial building block for informing • The way the family unit is changing in the health, long-term care, and other areas of region in terms of structure and behavior public policy is the continuation of moni- and the resulting implications for aging. toring the evolution of life expectancy rela- A key aspect of this area is the way co- tive to healthy years of life expectancy. residence patterns continue to change Understanding whether East Asia and Pacific in the face of changing fertility, migra- is experiencing compression or expansion of tion, urbanization, and other factors, and morbidity, and to what extent, need further the way mutual support within families analysis, particularly in a region where aging responds to changes in residence patterns. is occurring in populations that may have Research from China suggests that analy- had higher exposure to risks (smoking, poor sis needs to go beyond simple co-residence diets, and so on) at younger ages than those of parents and adult children to analogous in Western Europe and North America—the situations where adult children live nearby source of much of the morbidity compression and provide care and other support. In evidence. addition, an understanding of how chil- In addition, better analysis is needed of dren may split support responsibilities health expenditures in developing countries (for example, in-kind care from proximate in the region, including both disaggregation children and remittances from migrant of past expenditure trends (to analyze the children) is needed. relative contribution of aging and nonag- • The way sources of support for older ing effects in health expenditure) and more people evolve. Although the current pri- regular and robust expenditure projections mary role of own labor in elderly support for more countries. Health spending projec- in most East Asian and Pacific countries tions are inherently more challenging than is clear, the situation may change as coun- pensions because of the multiplicity of influ- tries get richer and social security cover- encing factors on the demand and supply age spreads. Of particular interest will be sides, particularly dramatic improvements the interaction between public and private in the availability and effectiveness of medi- transfers to older people, the degree of cal technology. Existing projections in East family support, and labor force behavior, Asia and Pacific tend to rely on the morbid- especially for rural people. ity compression hypothesis, which may not be a sound basis for projecting spending in the There is a significant agenda to develop the region. Existing projections also do not fac- monitoring mechanisms that will allow pol- tor in adequately the large-scale expansions in icy makers to assess the individual and soci- coverage and urbanization, which could dra- etal welfare effects of rapid aging for older matically increase health care use and expen- people and their families. Apart from the ditures. Furthermore, aside from a few studies need for better household surveys, East Asian from the United States and OECD, research is and Pacific countries would benefit from lacking about the extent to which technology more systematic monitoring of a core set of interacts with aging to drive up expenditures. age-related indicators to assess trends in the Although aging increases the urgency of welfare of older people. A useful example of greater emphasis on primary care with coor- such a framework can be found in the Active dination across levels of care, emerging expe- Aging Index, which measures indicators rience on how best and most affordably to related to work, social activities and partici- achieve such a transformation in lower- and pation, independent living, and the enabling middle-income countries needs to be shared. environment for healthy and productive In Thailand, the drive for universal health aging (box 8.1). coverage was accompanied by significant 260 LIVE LONG AND PROSPER BOX 8.1 The Active Aging Index: A tool for monitoring the situation and potential of older people The Active Aging Index (AAI) is a tool developed by 2. Social activity and participation of older people the European Commission and the United Nations a. Voluntary activities Economic Commission for Europe to measure both b. Care to children and grandchildren the current living conditions and the untapped c. Care to older adults potential of older people for active and healthy d. Political participation aging across 28 European countries. It measures the level at which older people live independent 3. Independent and autonomous living of older lives and participate in paid employment, social persons activities, and civic life, as well as their capacity to a. Physical exercise actively age. b. Access to health services The index is constructed from 22 individual indi- c. Independent living cators that are grouped into four domains: employ- d. Financial security (three indicators) ment, social participation, independent living, and e. Physical safety capacity for active aging. The results of the AAI are f. Lifelong learning presented in a country ranking for the overall AAI 4. Capacity and enabling environment for active and for each of the domains. AAI also offers a break- aging down of results by gender. It was developed in part in a. Remaining life expectancy at age 55 response to the review and appraisal of implementa- b. Share of healthy life expectancy at age 55 tion of the Madrid International Plan of Action on c. Mental well-being Ageing and its Regional Implementation Strategy. d. Use of information and com munication The domains and their indicators are outlined as technology follows: e. Social connectedness 1. Employment of older workers f. Educational attainment a. Employment rate ages 55–59 b. Employment rate ages 60–64 For details of the questionnaire that underlies the c. Employment rate ages 65–69 reporting and of the methodology, see Zaidi et al. d. Employment rate ages 70–74 (2013). scaling up of primary care service provision. fiscal implications. Better understanding of Brazil’s Family Health Program provides use- such factors will be important to assess what ful lessons for reorienting provision away lessons are most adaptable to the East Asian from a hospital-centric model and improving and Pacific context. In addition to changes health outcomes through multiprofessional throughout the health delivery systems, par- teams delivering primary health care. The ticular attention will be needed to develop program emphasizes the continuity of care, effective and sustainable models for manage- an integrated service package, a first point ment of age-related conditions such as demen- of access and coordination within the health tia and to balance the trade-offs between system, and a focus on the family. Turkey has quality of care, health outcomes, and costs in recently embarked on health system reforms end-of-life care. that take a primary care–centered approach. With respect to long-term care, although Yet, not enough is known about why these greater reliance on community- and home- three countries have been successful in pri- based care for most people is a priority, the mary care–centered reforms, the political models that are most viable in low- and middle- economy factors that drove them, and the income settings remain to be demonstrated. TO LIVE LONG AND PROSPER: LOOKING AHEAD 261 The emerging regional experience in countries significant voluntary participation of infor- such as China and Thailand and the experi- mal workers. It also raises questions about ence with community-based models being feasible, low-cost mechanisms to collect con- piloted in middle- and low-income countries tributions from informal workers. No clear in the region promise to contribute lessons on answer exists to the question of what level of balancing access, cost, quality, and cultural subsidy is enough globally, though success- appropriateness. Furthermore, government ful examples of coverage expansion within systems for monitoring hospital performance East Asia and Pacific have entailed substan- should be developed to give a clearer view of tial subsidies (for example, China and Korea those hospitals currently being used as default at both collection and pay-out stages). With long-term care arrangements. More broadly, respect to collections, evidence from devel- the experience of OECD countries suggests oping countries points to the importance of that assessing and promoting decent quality intermediary aggregators who are able to act long-term care will be vital, with a particular as collection channels for groups of work- emphasis on the trade-off between quality and ers (for example, farmer cooperatives, trade cost. This approach will help clarify where the unions, and microfinance organizations, as in market failures in provision lie and what seg- parts of India) and the potential for mobile ments of the market are most in need of incre- technology to facilitate collections and pay- mental public subsidies. ments (for example, Kenya’s Mbao pension plan) (Kwena and Turner 2013). Piloting and experimentation will be needed if the route Pensions and social security of matching contributory schemes combined In the region’s low- and middle-income with a more flexible and efficient collection countries, the adequacy and sustainability infrastructure is to be expanded. of pension systems and their effect on labor Yet no contribution-based program will markets are less of an immediate concern yield results in the short term, and a dual than ensuring wide coverage. This is espe- strategy is needed to combine rapid cover- cially true in countries such as China and age expansion with long-term adequacy of Vietnam where populations are aging at benefits. The current and soon-to-be elderly unprecedented speed. Because all contribu- population can be lifted from poverty only tory pension schemes—defined benefit or by social assistance programs, which may defined contribution, funded or unfunded— include social pensions (that is, categorical take decades to produce significant pen- cash transfers aimed at the elderly). Several sion incomes, the inability of the traditional East Asian and Pacific countries, and many social insurance model to provide a timely more globally, demonstrate the potential of solution to the pension coverage gap is social pensions to achieve rapid and wide- becoming increasingly clear. This conclusion spread coverage. In some, the level of the pen- is strongly supported by the international sion and its coverage are both high, raising experience. challenges of long-term sustainability as the If coverage expansion is an urgent pri- population ages. In others, the benefits are so ority, what are the options? In the long low as to render the programs ineffective in term, with time to accumulate pension preventing old-age poverty. A holistic social wealth, potential may exist for subsidized assistance policy must address any trade-offs contributory schemes where, as in many that may exist between programs targeted to health insurance schemes, the premium for the elderly and those targeted at poor house- the poor is paid by the government and the holds. In low-income countries, the elderly nonpoor (or near-poor) informal sector has poor tend to live with their families, so a pro- subsidized or matched contributions. This gram targeted at poor households will tend to approach raises questions about the required reach the elderly poor as well. In essence, only level of public premium subsidy to generate social pensions can address short-term elderly 262 LIVE LONG AND PROSPER poverty, whereas expansion of contributory focal agencies for a national aging strategy pension schemes, including those with subsi- or action plan, and in some cases, they also dies for the informal sector, can help address help monitor performance under specific laws the situation in the long term. Combining on the welfare of the elderly. Both the agen- these two policy instruments in a dynamic cies and the strategies perform a useful func- plan should be part of the overall strategy for tion, though the primary policy development a social protection system. function continues to rest with line ministries A third set of questions relates to the effect such as labor, social security, health, and wel- of pension systems on the broader economy. fare. Among those ministries, as in developed In light of the common objectives of East countries, uncertainties about the boundaries Asian and Pacific countries of sustaining eco- of sectoral leadership in areas such as the con- nomic competitiveness and gradually increas- tinuum of care from aged care to long-term ing formalization of the labor market, a need care services, or the interaction of labor and exists to understand better the responses of social security policies with respect to older firms and workers to different social contri- workers, sometimes exist. These are ongoing bution rates. Do high rates deter labor mar- challenges even in countries that have had ket formalization, as analysis from some aged populations for long periods. However, Latin American and Caribbean countries, the there are several examples of so-called super- European Union, and Turkey suggests? If so, ministries in OECD countries, which bundle at what level of social contribution are nega- the key policy and service delivery functions tive effects on labor market formality likely around aging populations (health and long- to be felt? What segments of the labor market term care, pensions and social security, aged are most likely to be affected, and what are care, and sometimes employment) into a sin- the options to reduce the effect? gle public agency. A second area in which sharing experience across and within countries would be valu- Cross-cutting issues able is the building of social consensus around A challenging aspect of aging societies is insti- policy and behavioral change needed to man- tutional change and the way public, market, age rapid aging. The experience of richer East and private institutions need to change in the Asian and Pacific economies suggests that face of rapid aging. In this respect, “institu- there may be a distinctive political economy tions” can have several meanings: organiza- to regional efforts to navigate societal aging. tions or stakeholders and their configuration; To date, the region’s countries have demon- public policies that set the institutional frame- strated less conflictual political economy work in specific areas (for example, labor dynamics across generations as they have regulations); or more broadly, sociocultural sought to respond to the rising needs of grow- norms with respect to the extended fam- ing elderly populations. At the same time, ily or gender relations. With respect to pub- older people do not yet appear to have mobi- lic institutions, developing countries in East lized as openly as in other parts of the world Asia and Pacific are still coming to terms to promote their interests, and the voice of with institutional arrangements to respond older people’s associations is still emerging. to rapid aging. Several countries, including How much such trends reflect strong resid- China, Indonesia, Thailand, and Vietnam, ual respect for older generations or different have national commissions, councils, and political systems and dynamics is unclear, as is committees on aging that act as the highest- how such dynamics may shift over time. level bodies to raise the profile of aging issues, perform some coordination functions, and coordinate reporting on international obliga- Conclusion tions such as the Madrid International Plan of Aging will fundamentally reshape societies Action on Ageing. Such bodies are often the and economies in East Asia and Pacific, but TO LIVE LONG AND PROSPER: LOOKING AHEAD 263 the shape of the future is very much in the and IFLS (Indonesia Family Life Survey) in hands of policy makers, communities, and Indonesia include a range of cognitive tests. citizens. Demography is a powerful force in 4. This position and a debunking of age-related development, but it is not destiny.4 Policy myths is efficiently summarized in Börsch- Supan (2013). makers in the region have the potential to shape responses that increase the chances for healthy and productive aging and promote References societies in which the compact between gen- Börsch-Supan, Axel. 2013. “Myths, Scientific erations is fair and realizes people’s potential Evidence, and Economic Policy in an Aging at all ages. World.” Journal of the Economics of Ageing In this process, East Asian and Pacific 1–2: 3–15. economies can learn from the experiences— Commonwealth of Australia Treasury. 2015. 2015 good and bad—of richer and older coun- Intergenerational Report: Australia in 2055. tries, as well as from one another. Crafting Canberra: Commonwealth of Australia. http:// appropriate policies will inevitably involve www.treasury.gov.au/PublicationsAndMedia experimentation and course correction. This /Publications/2015/2015-Intergenerational approach will also require strong leadership -Report. in the face of inevitable reluctance to embrace Hoskins, Dalmer D. 2010. “U.S. Social Security at 75 Years: An International Perspective.” Social change. Fortunately, the region’s strong eco- Security Bulletin 70 (3): 79–87. nomic and cultural traditions place it in a Kwena, Rose Musonye, and John A. Turner. favorable position to create environments in 2013. “Extending Pension and Savings Scheme which citizens will live long and prosper. Coverage to the Informal Sector: Kenya’s Mbao Pension Plan.” International Social Security Review 66 (2): 79–99. Notes McDonald, Peter. 2014. “Demographic Change in 1. For more information, see the Wittgenstein the Asian Century: Implications for Australia C e n t r e D a t a E x p l o r e r, h t t p : / / w w w and the Region.” EABER Working Paper 95, .wittgensteincentre.org/dataexplorer. See East Asian Bureau of Economic Research, also McDonald (2014) for an analysis of Australian National University, Canberra, demographics using the Wittgenstein Centre prepared for the Australian National projections for a number of East Asian and University conference on Prosperity in Asia: Pacific countries. The Intergenerational Dimensions, Canberra, 2. Originating with the U.S. Health and April 16–17, 2015. Retirement Living Survey, HRS-style surveys OECD (Organisation for Economic Co-operation are also conducted in a number of OECD and Development). 2011. Doing Better for countries, most notably ELSA (English Families. Paris: OECD. Longitudinal Study of Ageing) in the United Smith, James, and Malay Majmundar, eds. Kingdom and SHARE (Survey of Health, 2012. Aging in Asia: Findings from New and Ageing, and Retirement in Europe) in the Emerging Data Initiatives. Washington, DC: European Union, which in its latest wave National Academies Press. included 18 European countries (SHARE, Z a i d i , A s g h a r, K a t r i n G a s i o r, M a r i a http://www.share-project.org). See Smith and M. Hofmarcher, Orsolya Lelkes, Bernd Majmundar (2012) for a detailed review of Marin, Ricardo Rodrigues, Andrea Schmidt, aging-related data initiatives in Asia. Pieter Vanhuysse, and Eszter Zolyomi. 3. Future waves of CHARLS (China Health and 2013. “Active Aging Index 2012: Concept, Retirement Longitudinal Study) in China, LASI Methodology and Final Results.” European (Longitudinal Aging Study in India) in India, Centre, Vienna. ECO-AUDIT Environmental Benefits Statement The World Bank Group is committed to reducing its environmental footprint. In support of this commitment, the Publishing and Knowl- edge Division leverages electronic publishing options and print-on- demand technology, which is located in regional hubs worldwide. Together, these initiatives enable print runs to be lowered and shipping distances decreased, resulting in reduced paper consumption, chemical use, greenhouse gas emissions, and waste. The Publishing and Knowledge Division follows the recommended standards for paper use set by the Green Press Initiative. The majority of our books are printed on Forest Stewardship Council (FSC)–certified paper, with nearly all containing 50–100 percent recycled content. The recycled fiber in our book paper is either unbleached or bleached using totally chlorine-free (TCF), processed chlorine-free (PCF), or enhanced elemental chlorine-free (EECF) processes. More information about the Bank’s environmental philosophy can be found at http://www.worldbank.org/en/about/what-we-do/crinfo.