52277 Stress-Testing Households in Europe and Central Asia THE CRISIS HITS HOME THE CRISIS HITS HOME Stress-Testing Households in Europe and Central Asia THE CRISIS HITS HOME Stress-Testing Households in Europe and Central Asia Erwin R. Tiongson, Naotaka Sugawara, Victor Sulla, Ashley Taylor, Anna I. Gueorguieva, Victoria Levin, and Kalanidhi Subbarao Washington, D.C. ©2010 The International Bank for Reconstruction and Development / The World Bank 1818 H Street NW Washington DC 20433 Telephone: 202-473-1000 Internet: www.worldbank.org E-mail: feedback@worldbank.org All rights reserved. 1 2 3 4 12 11 10 09 This volume is a product of the staff of the International Bank for Reconstruction and Development / The World Bank. The fi ndings, interpretations, and conclusions expressed in this volume do not necessarily reflect the views of the Executive Directors of The World Bank 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 judgement on the part of The World Bank concerning the legal status of any territory or the endorsement or acceptance of such boundaries. Rights and Permissions The material in this publication is copyrighted. Copying and/or transmitting portions or all of this work without permission may be a violation of applicable law. The International Bank for Reconstruction and Development / The World Bank encourages dissemination of its work and will normally grant permission to reproduce portions of the work promptly. For permission to photocopy or reprint any part of this work, please send a request with complete informa- tion to the Copyright Clearance Center Inc., 222 Rosewood Drive, Danvers, MA 01923, USA; telephone: 978-750-8400; fax: 978-750-4470; Internet: www.copyright.com. All other queries on rights and licenses, including subsidiary rights, should be addressed to the Office of the Publisher, The World Bank, 1818 H Street NW, Washington, DC 20433, USA; fax: 202-522-2422; e-mail: pubrights@worldbank.org. ISBN: 978-0-8213-8222-6 e-ISBN: 978-0-8213-8223-3 DOI: 10.1596/978-0-8213-8222-6 Library of Congress Cataloging-in-Publication data has been requested. Contents Acknowledgments ix Abbreviations xi Executive Summary xiii A. Objectives of the Study xiii B. Main Findings xv Chapter 1 Macroeconomic Shocks 1 A. Introduction 1 B. Macro Shocks and Household Welfare: Framework 1 C. External Shocks and Transmission Channels 3 D. Context: Macroeconomic Strengths and Vulnerabilities 6 E. Shocks to Household Welfare 11 Chapter 2 Household Vulnerabilities 21 A. Introduction 21 B. Context: Poverty and Vulnerability in the Pre-Crisis Period 22 C. Shocks to Household Welfare: Empirical Strategy 23 D. Households and Credit Market Shocks 24 E. Households and External Price Shocks 38 F. Households and Income Shocks 46 Chapter 3 Coping with the Crisis 53 A. Introduction 53 B. Household Responses: Lessons from the ECA Experience 53 C. Context: Policy Response, Government Resources, and Constraints 55 D. Immediate Policy Responses: Some Illustrations 57 E. Longer-Term Policy Responses 61 Appendix 65 References 75 Notes 79 vi Contents Boxes Chapter 2 1. Defi nition of EU-SILC Variables Used in the Analysis 30 2. Stress Testing Household Indebtedness 32 3. EU-SILC and HBS Data on Household Debt: Comparisons with Other Sources 36 Chapter 3 4. Public Works Programs in ECA 62 Figures Executive Summary 1. Macroeconomic Shocks and Household Welfare: Stylized Transmission Channels xiv 2. Household Debt in ECA, 2008 xv 3. The Share of Vulnerable Households Before and After an Unemployment Shock xvi 4. Kyrgyz Republic: Poverty and the Food Price Crisis xvii 5. The Poor and Vulnerable Population, 2007­10 xviii Chapter 1 1.1 Macroeconomic Shocks and Household Welfare: Stylized Transmission Channels 2 1.2 Global Growth and Trade Slowdown 3 1.3 Export and Import Growth 3 1.4 Gross Capital Flows to Emerging and Developing Economies 4 1.5 Contraction in BIS Creditor Bank Foreign Claims 5 1.6 Commodity Price Developments 6 1.7 Sharp Contractions in Industrial Production 12 1.8 Rising Unemployment Rates 12 1.9 Sharp Deceleration in Formal Remittance Inflows 13 1.10 Growth in Remittance Inflows 13 1.11 Growth around Recent Crisis Periods: 1997­98 and 2008­09 14 1.12 "Credit-less Growth" in Emerging Markets and in the U.S. Great Depression 15 1.13 Local Currency Depreciations 16 1.14 International Food and Energy Price Movements and Local Currency Equivalent Indices 17 1.15 CPI Food Price Sub-indices: Selected Countries 18 1.16 Local Currency Equity Market Declines across ECA 18 1.17 Housing Prices in Selected ECA Countries 19 Chapter 2 2.1 Household Debt: Selected ECA Countries, 2008 25 2.2 Growth in Mortgage Debt: Selected ECA Countries, 2007 25 2.3 The Composition of Household Debt 26 2.4 Latvia Household Loan Delinquency Rates and Unemployment Rate 26 2.5 Foreign-Currency-Denominated Loans, 2008 27 2.6 Foreign-Currency-Denominated Loans in Ukraine, 2008 28 2.7 Mortgage Loans with Adjustable Interest Rates, 2006 29 2.8 Household Debt by Income Quintile 31 2.9 Household Income Used for Debt Repayments 31 2.10 Stress Testing Household Indebtedness: Selected EU-SILC Data 35 2.11 Stress Testing Household Indebtedness: Selected HBS Data 35 Contents vii 2.12 Energy Intensity and Food Consumption, 1970s­2000s 38 2.13 Food and Fuel Share of the CPI Basket 39 2.14 Food and Fuel Imports, 2006 39 2.15 Food Shares of Consumption 41 2.16 Utility/Energy Shares of Consumption 42 2.17 Kyrgyz Republic: Net Food Consumers and Net Food Producers 44 2.18 Kyrgyz Republic: Estimated Poverty Impact of the Food Crisis 44 2.19 Albania: Welfare Impact by Livelihood 45 2.20 Tajikistan: Welfare Impact by Livelihood 45 2.21 The Impact of the Crisis on Poverty and Vulnerability in the ECA Region 48 2.22 The Impact of the Crisis on Poverty and Vulnerability in the ECA Region: Sub-Regional Results 49 Chapter 3 3.1 General Government Balances in ECA, 2007 and 2009 56 Appendix A.1 The Impact of the Crisis on Poverty and Vulnerability 72 A.2 Financial Margins in Selected Countries 73 Tables Appendix A.1 Interest Rate Shock and Borrowers at Risk 65 A.2 Economic Shocks and Borrowers at Risk 66 A.3 Economic Shocks and Borrowers at Risk 66 A.4 Food Share of Consumption 67 A.5 Utility/Energy Share of Consumption 68 A.6 The Welfare Impact of a 10 Percent Food Price Increase 69 A.7 The Welfare Impact of a 10 Percent Fuel Price Increase 70 A.8 Summary Data: GDP Growth and Poverty Simulations 71 Acknowledgments This report was prepared by a core team led conducted an initial analysis of the Kazakhstan by Erwin R. Tiongson and including (in HBS data in July and August 2008. alphabetical order) Anna I. Gueorguieva, The team received valuable comments Victoria Levin, Kalanidhi Subbarao, Naotaka and suggestions at the concept note, decision Sugawara, Victor Sulla, and Ashley Taylor. draft, and other stages of the preparation pro- The report received generous fi nancial support cess from peer reviewers and numerous col- from the ECA Office of the Chief Economist. leagues. These include (in alphabetical order) This report was undertaken under the guid- Mohamed Ihsan Ajwad, Emanuele Baldacci ance of Indermit Gill (Chief Economist), Luca (IMF), Lawrence Bouton, R. Sudharshan Barbone (Sector Director), Asad Alam (former Canagarajah, Sanjeev Gupta (IMF), Ardo Sector Manager, currently ECCU3 Country Hansson, Jesko S. Hentschel, Valerie Herzberg, Director), and Benu Bidani (Sector Manager). Christos Kostopoulos, Kathy Lindert, Pradeep The report draws heavily from several back- Mitra, Fernando Montes-Negret, Pierella Paci, ground notes and papers including those pre- Stefano Paternostro, Bryce Quillin, Sophie pared by Victoria Levin (lessons from previous Sirtaine, and Marijn Verhoeven. Paloma Anos ECA crises) and Kalanidhi Subbarao (public Casero provided many useful suggestions dur- works programs). It also draws from an ongoing ing the early stages of this study. Salman Zaidi research project conducted by members of the offered generous advice at various stages of the team together with Thorsten Beck and Katie preparation process. M. Willem van Eeghen Kibuuka on household indebtedness in Europe provided valuable guidance in preparing the and the CIS. The research project is financed final version of this report. by the DECRG Research Support Budget A number of individuals generously shared (RF-P115252-RESE-BBRSB). Kechen Chen the results of their ongoing empirical analyses ix x Acknowledgments and/or their data, including (in alphabetical Micro Responses (held in Washington, DC, order) Carlo Azzarri (DECRG), Dániel Holló on February 15, 2008). The team is grateful to (Magyar Nemzeti Bank), Kotaro Ishi (IMF), the workshop participants, including presenters Alejandro Izquierdo (IADB), Sarosh Sattar from the European Bank for Reconstruction (ECSPE), Emil Tesliuc (HDNSP), Marijn and Development, the European Central Bank, Verhoeven (PRMPS), Olga Vybornaia (ECSPE), the International Monetary Fund, the Center and Alberto Zezza (FAO). This report also for Strategic Research (Moscow), the National draws from many of the ideas first discussed at Bank of Poland, and the Economic and Financial the World Bank Workshop on Macro Risks and Risk Unit of the World Economic Forum. Abbreviations xi Abbreviations ALMP Active Labor Market Programs BIS Bank for International Settlements CCT Conditional Cash Transfers CEMBI Corporate Emerging Markets Bond Index CIS Commonwealth of Independent States CPI Consumer Price Index DECPG World Bank Development Prospects Group DECRG World Bank Development Research Group EAP East Asia and the Pacific EBRD European Bank for Reconstruction and Development ECA Europe and Central Asia ECB European Central Bank ECCU3 World Bank South Caucasus Country Unit EMBIG Emerging Markets Bond Index Global EU European Union EU5 Czech Republic, Hungary, Poland, Slovak Republic, and Slovenia EU8 Member States of the EU (the Czech Republic, Estonia, Hungary, Latvia, Lithuania, Poland, Romania, Slovak Republic) EU10 New Member States of the EU (Bulgaria, the Czech Republic, Estonia, Hungary, Latvia, Lithuania, Poland, Romania, the Slovak Republic, and Slovenia) EU-SILC EU Survey of Income and Living Conditions FAO Food and Agriculture Organization FDI Foreign direct investment GDP Gross domestic product HBS Household Budget Survey HDNSP Social Protection Team IADB Inter-American Development Bank IMF International Monetary Fund LAC Latin America and the Caribbean MNB Magyar Nemzeti Bank (National Bank of Hungary) MSCI Morgan Stanley Capital International NGO Nongovernmental organization NPL National poverty line OECD Organisation for Economic Co-operation and Development PAYG Pay as you go PPP Purchasing power parity PRMPS World Bank Public Sector Governance Unit PUJ Publicly useful jobs SPJ Socially purposeful jobs WB World Bank WEO World Economic Outlook Executive Summary The Europe and Central Asia (ECA) region has economic and social impacts. Over the recovery been hit by a crisis on multiple fronts. Countries period following the 1998 Russian crisis through in ECA are facing major, interrelated, external 2006, more than 50 million people moved out macro-financial shocks. The first is the global of poverty in the region. Poverty fell through- growth slowdown leading to falling export mar- out all the sub-regions of ECA, with the mid- ket demand. In addition, the prospects for inflows dle-income countries of the Commonwealth of remittances to low-income countries have been of Independent States (CIS) experiencing the downgraded as economic activity in migrant host largest declines in poverty. Poverty reduction countries has declined. The second is the financial in ECA has been driven largely by growth in deleveraging by major banks and other financial mean incomes and rising real wages among the institutions in developed economies, which has working poor. However, the rapidly deteriorat- markedly reduced the availability, and increased ing global economic environment is eroding the cost, of external finance across public, corpo- the region's substantial recent gains, and is rate, and financial sectors. The third is the recent threatening the welfare of about 160 million commodity price changes, which have involved a people--close to 40 million people who are reversal of much of the commodity price boom poor and about 120 million people who are just of 2007 and 2008. As a result, countries whose above the poverty line. exports are focused on commodities have suf- fered adverse terms of trade pressures, in addition to the quantity shock to export demand. Across A. Objectives of the Study countries in the region, unemployment levels have risen while economic activities have collapsed. The main objective of the study is to understand The crisis risks reversing the region's recent the impact of these macroeconomic shocks on gains and exposes ECA to significant adverse household well-being. In particular, it seeks xiii xiv Executive Summary to understand the key macroeconomic shocks Neither does the diagram take into account the confronted by the region and the impact of such role of wealth effects as a transmission channel shocks on household welfare, including the of the crisis to households, such as via changes in effect on household income flows, consumption the prices of property, the value of equity hold- levels, and liabilities. It will also assess possible ings (directly or in pension funds), or indeed strategies to cope with the crisis and manage expectations of future labor income. Changes in the adverse social impact. wealth may directly lead to adjustments in the The report examines household vulner- consumption behavior of households or indi- abilities along three main transmission chan- rectly through the role of certain assets, such as nels. Figure 1 represents a stylized diagram property, as collateral that affects a household's for understanding the impact of macroeco- ability to access credit. The heterogeneous asset nomic shocks to date on household welfare. It positions of households mean that such wealth reflects a summary of the emerging concep- changes will lead to redistributions within the tual and empirical understanding of the social household sector, such as between those long or effects of macroeconomic crises experienced short in a particular asset. Unfortunately, lack of in various parts of the world over the past data on household wealth levels and composition three decades. In brief, the diagram focuses precludes detailed stress testing of such wealth on three main channels through which major effects. However, the build-up of mortgage macroeconomic shocks--such as the regional indebtedness detailed in the report provides growth slowdown or the credit crunch--are some indirect insight into the growing exposure transmitted to household welfare. These are of households' asset positions to property hold- the income and employment of members of the ings. The above diagram also does not address household; the relative prices of goods and ser- the role of government policy (including fiscal vices they purchase; and their access to fi nance and monetary) and social assistance (though (including the cost of credit and the burden of social assistance may be thought of as a source of servicing debt). income). Government policy can, in fact, either The diagram is stylized and abstracts from mitigate shocks or exacerbate them, depend- several important elements. It is simplified ing on how it is formulated and implemented. and ignores second-round effects and the con- A further transmission channel of the real and sequences of multiple shocks, and does not financial impacts of the crisis through to house- indicate how the social effects are distributed. hold welfare, though not explicitly addressed, is FIGURE 1 Macroeconomic Shocks and Household Welfare: Stylized Transmission Channels Income and Employment (Labor Market) Relative Prices Macro Shock Household Welfare (Product Market) Access to Credit (Financial Market) Executive Summary xv pension provision. The nature and magnitude of The microeconomic simulation in the report the effects of the crisis via this channel depend draws on a large, cross-country database of crucially on the structure of the pension system, household surveys. The report brings together in particular the mix between pay-as-you-go, for the first time comparable cross-country data funded, and voluntary pension systems. on household indebtedness for a large group The report analyzes household vulnerabili- of ECA countries using the European Union ties by examining credit markets, external prices Survey of Income and Living Conditions (food and fuel), and income shocks to date and and Household Budget Surveys. The report by assessing their impact on household welfare. also highlights newly updated information Because actual household survey data over the on household consumption from the ECA crisis period will typically not be available for Household Data Archives. Comparisons with some time to come, we use the most recent Western Europe and other advanced econo- pre-crisis household data along with aggre- mies are also used to inform the analysis when gate macroeconomic outturns to simulate the relevant data are available. impact on households of key economic shocks already taking place. The impact on household well-being is quantified as the change in the B. Main Findings household debt service burden, the fall in real income, or movements into poverty, as appro- The results of the analysis suggest that the priate. The report presents regional overviews adverse effects of the crisis on households--via along with cross-country comparisons and credit market shocks, food/fuel price shocks, contrasts. It also presents selected country and income shocks--are widespread. Both examples, depending on data availability and poor households and nonpoor households are relevant economic developments, to illustrate vulnerable depending on the economic shock, the incidence and distribution of specific vul- the specific transmission channel, and selected nerabilities within countries. household characteristics. FIGURE 2 Household Debt in ECA, 2008 In percent of GDP 50 40 30 Percent 20 10 0 Bulgaria Croatia Czech Rep. Estonia Hungary Latvia Lithuania Poland Romania Slovak Rep. Slovenia Belarus Kazakhstan Russian Fed. Ukraine Armenia Turkey New EU member states (incl. Croatia) Middle-income CIS Other xvi Executive Summary Credit Market Shocks will be unable to meet debt service obligations. The rapid rise in household indebtedness--in Interest rate shocks in Estonia, Lithuania, and the new EU member states as well as in some Hungary, for example, increase the share of Western Balkan countries, such as Albania and vulnerable households or borrowers at risk (in Serbia, and CIS countries such as Ukraine-- percent of all indebted households) by up to 20 has exposed households to a number of credit percentage points, depending on the magni- market shocks. There is no doubt that house- tude and severity of the shock. Unemployment hold debt holding has improved the lives of and exchange rate shocks also expand the share many, allowing them to smooth consump- of vulnerable households (out of all indebted tion and share risks, purchase durables, and households) by several percentage points. invest in housing stock. At the same time, the Many of those household borrowers at risk nature of household debt in ECA is such that of unsustainable debt burdens are from the households are now facing exchange rate and richer income quintiles. Although the shares of interest rate shocks, with few opportunities indebted households and households at risk in for hedging and with little prior knowledge of the ECA region still lag behind those of richer their vulnerability. countries, the aggregate effects of rising debt Within countries, household indebtedness service burdens are already being seen in ris- is more common than previously understood. ing household loan delinquency rates, as unem- Mortgage loans, in particular, have grown ployment has increased. rapidly in recent years among poorer and mid- dle-income households in a number of EU10 External Price Shocks countries. Both poor households and nonpoor The food and fuel crisis may not be over. Food households are exposed to the risks of unsus- and fuel prices have abated worldwide because tainable debt service burdens. of the worsening global financial crisis--the The results of stress tests on household economic recession or slowdown in many coun- indebtedness in selected countries suggest tries across regions--and, as a result, global that ongoing macroeconomic shocks may sig- demand for commodities has fallen. In addi- nificantly expand the pool of households that tion, increased agriculture production activity FIGURE 3 The Share of Vulnerable Households Before and After an Unemployment Shock In percent of indebted households 30 25 20 Percent 15 10 5 0 Estonia Hungary Lithuania Before the shock After the shock Executive Summary xvii FIGURE 4 Kyrgyz Republic: Poverty and the Food Price Crisis 20 15 Poverty headcount 10 5 0 Net food consumers Net food producers Before the crisis After the crisis led to a bountiful 2008 harvest and eased global suggest that, at least in principle, the poor commodity shortages. However, international are not necessarily the hardest hit. However, commodity price levels have not returned to the food share of total household consump- pre-2007 levels. Specialists have also pointed to tion typically falls with income; in some of the longer-term challenges in global food produc- low-income countries in the region, the food tion that are yet to be addressed. In addition, share of consumption among the poor is 70 to falling currencies in some EU10 countries are 80 percent. Moreover, in reality, the poor are resulting in a new round of price increases, the worst hit, as many of the poor in Albania, depending on the share of imported food and Kyrgyz Republic, and Tajikistan, for example, fuel in local consumption and the degree of pass- are also observed to be net consumers, with lim- through of exchange rate changes in domestic ited access to agricultural assets and inputs. prices. Finally, in a number of countries such as Belarus, Moldova, and Ukraine, the utility Income Shocks reform program remains largely incomplete. As a result, for reasons of economic efficiency Poverty will rise. Simulations suggest that by or fiscal consolidation, a number of countries 2010, there will be 11 million more people in will have to adjust their tariffs to cost-recovery poverty, and more than 23 million more people levels in the coming years. will find themselves just above ECA's interna- There is significant heterogeneity within tional poverty line, relative to baseline pre-cri- countries in the welfare impact of commod- sis projections. The growth in poverty would ity price shocks. The net effect of a food price represent a fifth of the ECA population who shock depends on whether households are net moved out of poverty between 1998 and 2006. producers or net consumers of food, it depends This is not surprising given that poverty in on their intensity of food consumption and ECA is shallow, characterized by large numbers the availability of cheaper substitutes, and it of individuals susceptible to falling into pov- depends on their livelihood strategies, access to erty even with modest falls in average income. agriculture assets and inputs, and their ability Alternatively, one could think of them as the to take advantage of profitable opportunities recent-poor, with tenuous links to the labor in agriculture. These multiple considerations market, with little precautionary savings, and xviii Executive Summary FIGURE 5 The Poor and Vulnerable Population, 2007­10 Using the $5-a-day measure of poverty 175 160.0 158.7 Population (in millions) 153.6 148.8 150 145.5 125 130.7 119.3 100 2007 2008 2009 2010 Pre-crisis growth projection Latest growth projection who are likely to have benefited from recent effects. On the other hand, general equilib- credit and construction booms. rium effects will either dampen or worsen The magnitude of the poverty impact var- some of these effects. ies by sub-regions. The middle-income CIS The second-order effects on human capital countries, on average, have seen the largest and accumulation and social capital will be signifi- most significant downward revisions to their cant. Lessons from the region's own experi- gross domestic product growth projections. ences suggest that transitory shocks' long-term As a result, and by construction, they are also toll on human capital has been substantial seeing the largest percentage point increases in because families curbed their education and the projected poverty headcount. They are fol- health investments in response to a bank- lowed closely by the low-income CIS. ing or exchange rate crisis. Crises may lead to The aggregate results mask the heterogene- increased social unrest, criminal activity, and ity of impact within countries, including the human trafficking; disrupt communal and eth- concentration of the poverty impact in selected nic relations; or bring down fragile govern- economic sectors. Country studies recently ments and fledgling democracies. completed suggest that for economic shocks transmitted primarily through the labor mar- ket, poverty will rise especially among house- Coping with the Crisis holds that have been dependent on remittance Compared to previous crises, the scope for inflows and those previously employed in households to engage in their traditional cop- booming construction sectors where economic ing strategies may be more limited. During activity is now projected to decline sharply. previous crises, households found secondary The results of the analysis are indicative employment, relied on transfers from friends of how vulnerabilities are distributed across and families, or left for work abroad to aug- countries and, within countries, across broad ment family income. Because of the global types of households. In some ways, the esti- nature of the crisis, and because macroeco- mated effects may be understated because nomic shocks are hitting households on mul- they capture only some of the fi rst-round tiple fronts, many of these coping strategies Executive Summary xix are no longer viable. For the poorest house- impact may be even larger when considering the holds, subsistence farming may still be fea- impact of such policies on production incentives sible, though evidence from the recent food and the likely spillover impacts of restrictive price shock suggests that many of the poorest trade policies on neighboring countries, thus households do not have access to agricultural exacerbating regional welfare consequences. assets and inputs. For some, transitions into The region's social assistance systems vary informal sector employment may be possible, in size and targeting performance, and not though for many households, earnings from every program can and should be scaled up. In informal sector activity will be insufficient to addition, some of these programs will have to offset the poverty impact of the crisis. be cut as revenues fall. The response to the cri- sis will vary across countries and may include, Policy Responses among other things, expanding some well- Fiscal policy responses in the short term performing programs and reforming relatively are constrained by rapidly falling revenues. less effective interventions. Some of those who Substantial government deficits are currently will fall into poverty because of the crisis-- projected for the region. It would be essential the "new poor"--may not be easily reached to determine the overall fiscal adjustment war- by existing social protection programs. For ranted for macroeconomic stability and debt example, returning migrants do not qualify for sustainability, taking into account initial con- unemployment insurance. ditions and the likely impact of the crisis on The ECA region should consider new public fi nances. Economies that experienced instruments of social protection. Social safety strong initial fiscal and external positions are nets in ECA need to be strengthened to likely to have more room for expansionary fiscal handle the challenges of global and domes- policy and can afford a fiscal stimulus package, tic risks. The experiences of other countries while those with weaker initial positions may suggest that programs such as workfare and require substantial fiscal adjustment. Where public works programs can be appropriate there are no new official or alternative sources instruments for protecting the vulnerable of fi nancing and little scope exists to mobilize from immediate as well as longer-term (sec- revenues, some countries will likely resort to ond-round) consequences of transitory shocks across-the-board cuts in spending. Although on nonincome dimensions of welfare, includ- social safety nets will be among those items ing human capital accumulation. There are a likely to be cut as revenues fall, protecting few, albeit limited, country experiences with these programs--and possibly expanding some workfare in ECA--in Bulgaria, Poland, the of them, where some reallocation of resources Slovak Republic, and Slovenia. Some insights is possible--will be an important element in from these country experiences can inform the response to the crisis. the broader application of workfare in ECA Inappropriate policy responses to economic so that they can be efficient instruments for shocks may have welfare consequences far larger social risk mitigation while minimizing dis- than the welfare losses resulting directly from placement effects. the shocks themselves. In 2007­08, some coun- The prioritization of labor-intensive public tries imposed trade restrictions and price con- investments could be an important response trols in response to rising food prices. Such poli- to the crisis while creating the conditions for cies redistribute income away from rural food medium-term growth. Such a strategy can cre- producers (who tend to be poorer) to urban con- ate employment opportunities, as in workfare sumers (who tend to be richer). The net social and public works programs, while creating the xx Executive Summary infrastructure that supports economic recovery some countries in ECA, recent growth per- and economic growth in the medium term. formance has been underpinned by economic activity concentrated in a few sectors, such as Longer-Term Policy Responses the housing or construction sector, or income flows from some dominant source, such as Over the longer term, there are various mea- migrant labor. sures for limiting the risks borne by households as financial markets deepen. On the demand Monitoring systems are important. side, promoting financial literacy may help Guaranteeing that statistical monitoring sys- households to understand the risks they expose tems are in place and that relevant household themselves to because of their consumption, data are collected regularly and made available employment, or borrowing choices. In addi- for analysis are important measures for ensur- tion, a whole host of macro-fi nancial policies, ing that household vulnerabilities to a range such as prudential norms as adopted by some of potential shocks are understood in a timely EU10 countries to limit foreign currency manner and that those households at risk can be exposures of households, can be used in combi- reached by a country's social protection system. nation to help mitigate the potential risks asso- The resilience of households to macroeco- ciated with increased exposures of households nomic shocks ultimately depends upon the to credit and fi nancial markets. economy's institutional readiness, the flexibility It is important that policy responses do of the economic policy regime, and the ability not conflict with the key longer-term reform of the population to adjust. Policy and institu- agenda. For example, authorities should guard tional preparedness is essential so that coun- against reversal of efforts to lower quasi-fiscal tries can manage the adverse social impacts of deficits in the energy sector, which is an ongo- macroeconomic shocks. This requires ex ante ing challenge in many ECA countries, and they analysis of risks, a good understanding of their should maintain an open and transparent trad- possible transmission channels if triggered, and ing regime. Some countries in ECA adopted their possible impacts on households; develop- restrictive trade and price controls in response ing approaches that ensure that the state does to the food price increases in 2007 but many of not intervene excessively in terms of detrimen- them have now been reversed. tal longer-term distortions to incentives or fis- Diversified sources of economic growth will cal sustainability; and having a comprehensive be critical in helping dampen ECA countries' social safety net system that provides for coun- vulnerability to macroeconomic shocks. In tercyclical and scalable interventions. CHAPTER 1 Macroeconomic Shocks A. Introduction B fi rst introduces a simple stylized framework of the transmission channels, which provides The Europe and Central Asia (ECA) region has the structure for the subsequent discussion. been hit by a crisis on multiple fronts. The first Section C examines the external shocks and is the global growth slowdown leading to falling their transmission channels to countries export market demand. In addition, the pros- within the region. The nature and extent of pects for inflows of remittances to low-income the transmission of these shocks through to countries have been downgraded as economic households depend crucially on an economy's activity in migrant host countries has declined. macroeconomic strengths and vulnerabilities, The second is the financial deleveraging by such as the degree of international integration, major banks and other financial institutions the strength of sectoral balance sheets, and in developed economies, which has markedly domestic policy stance, which are discussed reduced the availability, and increased the cost, in Section D. Finally, Section E outlines the of external finance across public, corporate, and main resulting macro shocks to household financial sectors. The third is the recent com- welfare, the micro implications of which are modity price changes, which have involved a analyzed in subsequent chapters. reversal of much of the commodity price boom of 2007 and 2008. This chapter examines how the exter- B. Macro Shocks and Household nal shocks arising from the global economic Welfare: Framework crisis can be transmitted through to macro shocks affecting the welfare of households, Major macroeconomic shocks are transmit- via their income, access to credit, wealth, ted through to household welfare via vari- and relative prices of food and fuel. Section ous mechanisms; this report focuses on three 1 2 The Crisis Hits Home--Stress-Testing Households in Europe and Central Asia FIGURE 1.1 Macroeconomic Shocks and Household Welfare: Stylized Transmission Channels Income and Employment (Labor Market) Relative Prices Macro Shock Household Welfare (Product Market) Access to Credit (Financial Market) main channels. Figure 1.1 represents a highly below. Neither does the diagram take into stylized diagram for understanding the impact account the role of wealth effects as a trans- of macroeconomic shocks on household wel- mission channel of the crisis to households, fare. It reflects a summary of the emerging such as via changes in the prices of property, conceptual and empirical understanding of the value of equity holdings (directly or in the social effects of macroeconomic crises pension funds), or indeed expectations of experienced in various parts of the world over future labor income. Changes in wealth may the past three decades. The main channels directly lead to adjustment in the consump- considered are the income and employment tion behavior of individual households or may of members of the household; the relative do so indirectly via the role of certain assets, prices of goods and services they purchase; such as property, as collateral that affects and their access to fi nancial market (includ- their ability to access credit. The heteroge- ing the cost of credit and the burden of servic- neous asset positions of households mean ing debt). As discussed below, in the current that such wealth changes are likely to lead to context the shocks to household welfare via redistributions within the household sector, these channels have arisen primarily because such as between those long or short in a par- of the impact of external shocks, such as to ticular asset. Unfortunately, lack of data on global income, credit conditions, and com- household wealth levels and composition pre- modity prices, whose effects depend crucially cludes detailed stress testing of such wealth on the domestic economy's macro strengths effects. However, the build-up of mortgage and vulnerabilities. indebtedness detailed in the report provides The diagram is stylized and abstracts from some indirect insight into the growing expo- a few important elements. It ignores second- sure of households' asset positions to property round effects (such as on human capital accu- holdings. The above diagram also does not mulation, access to social services, and disrup- address the role of government policy (includ- tions to communal ties) and the consequences ing fiscal and monetary) and social assistance of jointly occurring crises and, as drafted, (though social assistance may be thought of does not indicate how the social effects are as a source of income). Government policies distributed (along geographic, occupational, can, in fact, either mitigate shocks or exacer- sectoral, gender, or income lines), though all bate them, depending on how they are formu- these will be considered in varying degrees lated and implemented. Chapter 1--Macroeconomic Shocks 3 C. External Shocks and began in the summer of 2007, unfolded. The Transmission Channels1 IMF's April 2009 World Economic Outlook projected a contraction in world growth of 1.3 percent in 2009 with growth recovering to Global Income 1.9 percent in 2010 (figure 1.2). These shocks to Since 2006, the growth of the major developed global income may be transmitted to countries economy countries, and world export demand, within the region via trade flows and remit- has weakened as the global credit crisis, which tances. Demand for exports from the region FIGURE 1.2 Global Growth and Trade Slowdown Annual real growth 9 6 3 0 Percent -3 -6 -9 -12 -15 2007 2008 2009 2010 2007 2008 2009 2010 2007 2008 2009 2010 2007 2008 2009 2010 World GDP US GDP EU GDP World trade volume April 2009 WEO projections October 2008 WEO projections Source: IMF WEO October 2008 and April 2009. Note: 2009 and 2010 are projections. Trade defined as volume of goods and services. FIGURE 1.3 Export and Import Growth Growth year-on-year 50 25 Percent 0 -25 -50 Jan-03 Jan-05 Jan-07 Jan-09 ECA exports ECA imports World exports Source: WB DECPG and staff calculations. Note: Change in US dollar seasonally adjusted values. 4 The Crisis Hits Home--Stress-Testing Households in Europe and Central Asia has fallen sharply, with the GDP of advanced Global Credit Conditions economies projected to contract by 3.8 percent As the financial crisis deepened from September in 2009. At the same time, the credit crunch 2008, mounting concerns over liquidity risk, is also impeding the ability to fi nance export asset quality, and counterparty credit risk and trade credits. As a result, the World Bank's enhanced risk aversion have resulted in signifi- 2009 Global Economic Prospects predicts cant deleveraging and attempts to reduce port- that world trade will decline in 2009 for the folio risk by financial institutions. This has not first time since 1982 (with the IMF April 2009 only affected the ability of fi nancial institu- WEO projecting an 11 percent contraction in tions and corporates in developed economies the volume of world trade in goods and services to obtain financing but also led to a retrench- in 2009). The impact of these trends on trade ment of the international exposures of banks for ECA countries has resulted in a precipitous and, for many countries, a sharp reduction in their ability to access international fi nance. drop in export and import values (figure 1.3). Gross capital flows to emerging and develop- As discussed in more detail later, remittances ing economies have fallen significantly (figure have already fallen sharply in some countries 1.4). The contraction, which has been across and prospects for 2009 inflows of remittances asset classes, has been particularly marked to developing countries have been downgraded for ECA (whose share of such flows fell from as economic activity in migrant host countries around 40 percent in 2007 and 2008 to just has declined. For many countries in ECA, par- over 20 percent in 2009). In terms of interna- ticularly those in the former Soviet Union, this tional banking exposures, after peaking at end- reflects the impact of the slowdown in Russia March 2008, the consolidated foreign claims and the valuation effects of the nominal depre- of Bank for International Settlements (BIS) ciation of the ruble, that is, a second-round creditor banks, in real terms, on Organization regional shock, rather than being directly due for Economic Co-operation and Development to the slow-down in developed markets. (OECD) economies fell by around 20 percent FIGURE 1.4 Gross Capital Flows to Emerging and Developing Economies 200 US dollars (in billions) 150 100 50 0 Q1 Q2 Q3 Q4 Q1 Q2 Q3 Q4 Q1 Q2 2007 2008 2009 ECA LAC EAP Other Source: DECPG. Note: US dollar values deflated by US CPI. 2009 Q1 data are preliminary estimates. Chapter 1--Macroeconomic Shocks 5 FIGURE 1.5 Contraction in BIS Creditor Bank Foreign Claims In Q1 2009 prices 120 Index (end-March 2008=100) 100 80 60 Dec-06 Jun-07 Dec-07 Jun-08 Dec-08 ECA High-income OECD excluding ECA EAP LAC Sources: BIS, IMF International Financial Statistics and staff calculations. to end-March 2009 (figure 1.5). Real BIS credi- credit spreads also increased substantially. The tor banks' foreign claims on emerging econo- JP Morgan Emerging Markets Bond Index mies generally peaked at end-June 2008 and by Global (EMBIG) Emerging Europe sovereign end-March 2009 had fallen by 17 percent for spread increased from 275 basis points on 1 ECA compared with around 12.5 percent for September 2008 to over 900 basis points in late Latin America and the Caribbean and East October. Having declined to 740 basis points Asia and Pacific. by end-2008, the spread continued to nar- Regional equity and exchange rates row through 2009, with the average monthly came under pressure during late 2008 as spread reaching around 400 basis points in foreign investors drew back funds from July 2009. Similarly, the cost of credit default the region, and concerns over the domes- protection via credit default swaps and external tic impact of the fi nancial crisis mounted. corporate funding rates increased sharply. For Despite somewhat of a reversal in recent example, the JP Morgan Corporate Emerging months, the downturn in fi nancial markets in Markets Bond Index (CEMBI) external cor- ECA since early 2008, and since September porate spread for emerging Europe (covering 2008 in particular, has been widespread and Kazakhstan, Russia, and Ukraine) increased deep across asset classes and countries. These from 530 basis points to 1,560 basis points at declines unwound a large part of the gains end-2008 before declining to an average of made in equity and sovereign bond valuations around 730 basis points in July 2009. In addi- during the asset price run-up of preceding tion to the falls in bond and equity markets, years (in which the average dollar value of the interbank markets exhibited rising spreads in October and November reflecting interna- Morgan Stanley Capital International [MSCI] tional funding pressures and domestic liquid- Emerging Europe equity index increased ity concerns. by over seven times from October 2001 to December 2007). The reduced demand for emerging market assets also contributed to Global Commodity Prices marked local currency nominal depreciations The global growth slowdown has contrib- as discussed in detail below. Secondary market uted to a sharp easing in global food, fuel, 6 The Crisis Hits Home--Stress-Testing Households in Europe and Central Asia FIGURE 1.6 Commodity Price Developments 200 150 100 50 0 Jun-07 Dec-07 Jun-08 Dec-08 Jun-09 Oil price spot, US$ per barrel Energy price index (June 2007=100) Food price index (June 2007=100) Steel products price index (June 2007=100) Source: WB DECPG and staff calculations. Note: Oil price is a simple average of Brent, West Texas Intermediate, and Dubai crude oil prices. and other commodity prices since mid- fall in fertilizer prices having similar impact 2008. Although making projections is particu- in economies such as Belarus. larly difficult given the current macro-fi nancial uncertainty, the June 2009 World Bank Global Development Finance forecasts a 43 percent D. Context: Macroeconomic Strengths fall in the US dollar oil price in 2009 relative and Vulnerabilities to 2008, with a rise of 13 percent in 2010. Non- oil commodity prices are projected to fall by The impact on a country's economic outlook 30 percent in 2009 with a further 2 percent of the above external shocks depends upon decline in 2010. In recent months, there has its potential exposure, which can be mapped been some recovery in prices of some com- through different stages of the transmission modities, such as oil (figure 1.6). Clearly, mechanism. The first is the extent of interna- the overall impact of these price movements tional integration of the country via trade and on a country's external payment position is financial channels (including remittances). The dependent upon its net consumption mix and second is the structure and health of sectoral respective export and import price elastici- balance sheets, such as in terms of external ties. Countries whose exports are focused on financing requirements, currency, and matu- commodities have thus suffered adverse terms rity mismatches. The third stage is the ability of trade pressures, in addition to the quan- of policymakers to mitigate the impact of the tity shock to export demand. For example, shock through the policy stance in terms of while the weakening oil price has had material monetary policy, exchange rate flexibility, and, implications on external and fiscal positions in linked to the strength of public sector balance Russia and Kazakhstan, the downturn in steel sheets, the current fiscal stance and the abil- prices adversely affected the external outlook ity to use fiscal policy measures to absorb the for countries such as Ukraine with the rapid impact of the global shocks. Chapter 1--Macroeconomic Shocks 7 International Integration in the former sub-region and Southeastern The increasing international integration of Europe are also vulnerable to the commod- countries in ECA over the past decade via ity price and activity downturns (accounting trade, income, and capital flows has enabled for just under 10 percent of exports in both countries to benefit from the growth and cases). Trade patterns in the EU8 are strongly fi nancing of partners but also provides intra-industry and particularly focused on increasing channels through which global and consumer durables including products such regional shocks are transmitted to domestic as road vehicles (around 15 percent of exports economies. Over the past decade, the ratio and 10 percent of imports) and electrical and of the total US dollar value of merchan- telecom equipment which again are subject to dise trade (exports plus imports) of coun- strong demand shocks as consumption falls in tries in ECA to their gross domestic product partner regions. (GDP) has increased from around 45 per- Net capital flows to the region doubled cent in 1998 to 57 percent in 2007. This fol- as a proportion of GDP from 2002 to 2007, lows the general growth in global trade over financing consumption and investment while this period. However, there are marked varia- increasing the impact of any sudden stops to tions in the level of trade openness across sub- such flows. From a level of around 4 percent regions in ECA. For example, over this period of GDP in 2002­2004, net capital flows to the merchandise trade for the new European ECA increased to around 8 percent of GDP Union (EU) member states in Central Europe in 2007 buoyed by factors including the global and the Baltics (the EU8) rose from 70 percent search-for-yield, the EU accession process, to 107 percent while for middle-income CIS and foreign direct investment (FDI) related to the ratio was broadly flat over this period. commodity investments. This broad pattern of In addition to the variation in the level of growth in relative flows was present across sub- trade openness within the region, there are regions. The mean level of net capital flows to considerable differences in the patterns of GDP across countries reached over 15 percent trade partners and major trading products in Southeastern Europe and over 10 percent in across sub-regions, as would be expected from the EU8. Although the mean level fell in low- gravity-type models of trade flows. The impor- income CIS countries, this reflected net capital tance of the EU as an export market, and hence outflows from Azerbaijan. the exposure of countries to the contraction Increased cross-border lending and foreign in demand from the EU as a result of the cri- bank ownership have both contributed to the sis, is particularly marked in the EU members sharp rise in international banking claims on of ECA, Southeastern Europe, and Turkey. ECA. The level of gross international assets The extent of intra-regional trade linkages of BIS creditor banks in ECA, on a locational within ECA is particularly high between the basis, rose from around 13 percent of GDP at middle- and lower-income Commonwealth end-2004 to 21 percent of GDP at end-2007. of Independent States (CIS) countries. In There are considerable variations across sub- terms of major product categories, the exports regions with the CIS countries having net of these two groups of countries are highly claims on BIS creditor banks for most of the concentrated in petroleum and petroleum period from 2003. In contrast, the net assets products (around 44 and 53 percent of total of BIS creditor banks on the Baltics reached exports in 2007, respectively), whose inter- over 70 percent of GDP in early 2008 (up from national prices have dropped from their mid- 30 percent in mid-2005) with levels of around 2008 highs. Iron and steel export earnings 40 percent of GDP on the EU5 and Southeastern 8 The Crisis Hits Home--Stress-Testing Households in Europe and Central Asia Europe. These figures compare to net assets Remittance inflows have grown rapidly of BIS creditor banks on the five Asian cri- over the past five years and for lower-income sis countries (Indonesia, Korea, Malaysia, economies in ECA outweigh the intra- and Philippines, and Thailand) of around 20 per- extra-regional financial inter-linkages via pri- cent of GDP in mid-2007. vate capital flows. The dollar value of remit- The rise in consolidated foreign claims of tance inflows to ECA grew at an annualized BIS creditor banks (i.e., netting out intra-group growth rate of around 37 percent between 2003 exposures) in recent years has been particu- and 2007 versus around 19 percent per annum larly marked in ECA's EU member states and for developing countries as a whole. Of those Southeastern Europe, increasing the potential countries with high inflows of remittances for two-way spillovers between local and Western relative to GDP, annualized growth rates of European banking systems. Foreign claims on remittance inflows of 32 percent were seen for ECA, which include international claims (i.e., Moldova, 66 percent for Azerbaijan, 74 per- cross-border claims plus local claims of foreign cent for the Kyrgyz Republic, and 84 percent affiliates in foreign currency) and local claims for Tajikistan. In such countries, the level of of foreign affiliates in local currency, peaked at external financial inflows from remittances far around 45 percent of GDP at end-June 2008, up exceeds that of capital inflows, exposing coun- from around 30 percent in the two years from tries to reductions in employment and wages in June 2008, before falling to around 35 percent migrant host countries more than direct expo- of GDP by end-2008. In 2008, foreign claims sures to developments in international finan- on the Baltic countries peaked at almost 140 cial markets. For example, in Tajikistan and percent of GDP, Southeastern Europe at over Moldova in 2007 it is estimated that remittance 100 percent, and the EU5 at over 80 percent inflows to GDP were around 46 percent and of GDP. Much of this increase has been due to 34 percent respectively compared with net cap- the growth in local claims of western European ital inflows of around 10 percent of GDP. foreign affiliates. The level of foreign claims to For many countries, particularly in the lower- GDP on the EU8 and Southeastern European income CIS, the exposures to recent external countries at end-2008 remains considerably shocks come particularly via their second- higher than the averages for OECD countries round regional impact, for example in terms of and indeed that observed for the Asian-crisis remittances and trade flows with Russia. For countries at end-1997. International claims ECA's EU member states and Southeastern tend to be concentrated on the non-bank pri- Europe, fi nancial, particularly banking, and vate sector and banking sectors rather than trade integration developments highlighted the public sector with the share of short-term the increasing interdependences with eco- international claims relatively high in Turkey nomic developments in Western Europe and and the middle-income CIS countries. The global fi nancial conditions. However, the pat- geographic patterns of banking sector inter- terns of trade and importance of remittances linkages via foreign claims have evolved in a for lower-income CIS focus attention on the different manner across sub-regions. While potential for intra-regional spillovers arising all regions have seen a relative reduction in the from developments in Russia. For example, share of German banks in their foreign claims, exports to Russia accounted for around 35 per- Swedish banks are the primary foreign claim cent of Belarus' exports in 2007 (or roughly creditor in the Baltics and the share of Austrian 20 percent of GDP) and around a quarter of claims has risen in the Central European new the exports of Kyrgyz Republic, Moldova, member states of the EU, the middle-income Ukraine, and Uzbekistan. Similarly, in 2007 CIS, and Southeastern Europe. flows from Russia accounted for almost all the Chapter 1--Macroeconomic Shocks 9 remittance inflows for the highly remittance- conditions, and the increasing availability of dependent economies of Kyrgyz Republic and a broad range of mortgage instruments have Tajikistan and over half the remittances to underpinned it. For the new EU member coun- Moldova. As Russia's economy grew strongly tries, it has also been suggested that the conver- between 2003 and 2007 (with annual GDP gence in living standards toward the EU aver- growth in the range of 6.4 to 8 percent), these age has helped accelerate credit growth.3 Over inter-linkages led to strong positive spill- this same period, household financial assets overs to partner countries. The fl ipside is the also grew rapidly, though not at the same pace exposure of these countries to any downturn as household indebtedness. As a result, the net in Russia (with the World Bank's June 2009 financial assets relative to GDP of the house- Russian Economic Report forecasting a real hold sector have fallen in many countries in the GDP contraction of 7.9 percent in 2009 with past few years, although on a per capita basis growth of 2.5 percent in 2010). net financial assets have generally risen since 2000 with some decline in recent years because of the pace of accumulation of liabilities. To the Balance Sheet Strengths and Weaknesses extent that the rises in household liabilities are The composition and strengths of domestic associated with mortgages, they are likely to be sectoral balance sheets are crucial determi- matched by greater property assets, although nants of the impact on the economy of external unfortunately data on the overall balance sheet shocks transmitted via the various international positions of the household sector, including inter-linkages discussed above. Of particu- both financial and non-financial assets such as lar interest in ECA have been the interrelated property, are not available. developments in household and financial sector The rise in the gross fi nancial positions of balance sheets over the past five years relating the household sector, and their changing com- to increasing household indebtedness and their position, has both brought benefits and intro- exposures to currency and interest rate shocks. duced new sectoral vulnerabilities. As house- Household indebtedness has grown rapidly hold financial positions have grown, there has in many ECA countries. Between 2002 and been a shift toward housing loans or mortgages 2007, for example, household debt relative to on the liability side of the balance sheet and GDP grew at an annual average rate of 37 per- an increasing share of equities and pension cent in the newer member countries of the EU, and mutual funds on the asset side. On the while rising only by 7 percent in the older EU one hand, rising indebtedness reflects the ben- member countries. In the new EU members, efits of financial sector development, allowing household debt now represents a little over a households to smooth their consumption over quarter of GDP, although it remains below time and acquire home ownership without sig- the 65 percent level in older EU members. nificant savings. Changes in the asset side of Household indebtedness also has been grow- the balance sheet brought increasing diversi- ing rapidly in a number of countries in the CIS fication and exposure to higher yielding asset countries and the Western Balkans. The ris- classes than the traditional deposits and cur- ing trends in household indebtedness, and the rency. On the other hand, these developments associated risks and benefits, are analyzed in bring the potential for greater exposures of more detail in chapter 2. households' net financial positions to currency, The growth in household indebtedness fol- asset pricing, and interest rate risks. (This will lows the rapid expansion in credit to the private be discussed in greater depth in the next chap- sector more generally.2 Buoyant housing mar- ter.) If the respective risks are not hedged and kets, favorable macroeconomic and fi nancial they subsequently materialize, they may lead 10 The Crisis Hits Home--Stress-Testing Households in Europe and Central Asia to deteriorations in households' ability to ser- with net loans to consumer and short-term vice their debt obligations. This in turn can funding as of end-2007 in the range of 100 to adversely affect the health of financial sector 150 percent). The high level of intra-regional balance sheets with second-round implications linkages through the set of major parent banks for households in terms of the availability and raised concerns over the impact of group liquid- cost of credit. ity problems on local banking systems and the Banking sector balance sheets in ECA have related regulatory coordination issues. expanded rapidly in the past five years, par- Although the banking system-level picture ticularly in the Baltics and middle-income going into late 2008 from standard asset qual- CIS countries, funded increasingly by external ity and capital adequacy indicators revealed parent groups and wholesale markets. Credit only a limited number of weaker outliers, such growth has covered both the household sec- lagged aggregate indicators may not provide an tor and non-bank private corporate sector. As a accurate picture of the current health of bank- result, the mean private credit to GDP of coun- ing sectors. For example, national poverty line tries in ECA roughly doubled from 2003 to (NPL) ratios may be relatively low because of 40 percent in 2007 (with the median ratio ris- the recent rapid expansion of credit more than ing from 19 to 33 percent). In the Baltics, the offsetting reclassification of loans as problem- mean level of private credit to GDP tripled over atic. In addition, rising interest rates on their this period to 71 percent with the means for the own can cause a deterioration of capital ade- middle- and low-income CIS groupings dou- quacy ratios in a mark-to-market environment. bling to around 40 and 18 percent respectively. Furthermore, a more general weakening of These levels compare to means of 120 percent, asset quality may increase solvency concerns, 48 percent, and 33 percent for OECD coun- rather than the proximate liquidity concerns of tries, middle-income, and low-income develop- late 2008, in less robust banking sectors. ing country sub-samples respectively. As is well known, many of the economies Average bank credit-to-deposit ratios had in ECA, particularly in the Baltics and central reached around 120 percent in ECA by 2007, Europe, entered 2008 with substantial current leading to concern over funding and liquidity account deficits. Much of the funding for these risks. The growth in credit to GDP in general has come from the international bank flows that through 2002 to 2007 has been associated with are projected to decline markedly in 2009. Even rising credit-to-deposit ratios, that is, increased those countries more reliant on FDI financ- reliance on non-deposit funding sources. ing are likely to face increasing difficulties in However, both trends have been particularly funding given the general downturn in growth marked in countries within ECA. Some of these prospects in the region and corporate sector non-deposit funds reflect increased access to difficulties in developed economies. In some parental funding sources, as the trend toward of the cases of previous notable reversals cur- increased foreign ownership of banking sectors rent account deficits, such as in the Asian crisis, adjustment via the trade balance was possible in ECA has continued in recent years or, in the because of a relatively supportive external envi- case of many of ECA's EU member states, sta- ronment which unfortunately is not the case for bilized at high levels. Foreign operations have the current deficits in the ECA region. been attracted by relatively high returns within ECA in comparison to developed banking mar- kets, at least up to 2007. Some of the major par- Policy Stance ent banks in the region themselves also appeared The ability of policy measures to either support vulnerable to liquidity and funding risks (e.g., domestic balance sheets in the face of external Chapter 1--Macroeconomic Shocks 11 shocks or mitigate the transmission of these to household consumption and welfare. The shocks to the household sector is constrained next chapters examine the potential impact of by initial conditions in terms of fiscal space, shocks to household income via credit market exchange rate arrangements, and inflationary shocks, external prices (food and fuel), and outlook. Indeed the nature of the balance sheet income shocks. The macro context of each of strengths and weaknesses also guides the poten- these shocks is outlined below, in addition to tial policy responses to the external shocks. For a brief discussion of the potential for wealth example, a relatively weak banking sector with effects, which, as mentioned above, are not high levels of foreign liabilities may limit the included in the subsequent microanalysis for overall impact of exchange rate depreciation on data availability reasons. economic activity given the scope for adverse balance sheet effects to offset, or outweigh, any Income Shocks positive benefits in external trade positions. Shocks to household income arising from the Such considerations may caution against sig- economic crisis may arise through a variety of nificant exchange rate loosening. channels and, depending on the household, may On the fiscal side, the ability of economies to be viewed as temporary or permanent. First, use government spending to address the adverse labor income may fall as a result of the loss of income shock is generally constrained (or non- employment or falling real wages for those who existent) in the presence of large and harder to remain in work. Second, declining remittance finance current account deficits. In some cases, inflows are another channel through which it may also be limited through debt sustain- household income may fall. Finally, changes in ability concerns arising from structural deficits social protection policies as a result of the crisis and possible contingent liabilities arising from may also affect household incomes. banking sector recapitalizations. Although Though the exposures to external shocks oil exporters, such as Russia and Kazakhstan, and strengths of initial balance sheets and entered this period with stronger external and policy stances vary across countries, there has fiscal reserves than did other ECA countries, been a broad-based reduction in real activ- their scope to employ these funds has become ity in ECA as a result of the global economic more limited in the face of falling oil prices. and financial crisis. After turning negative in While reduced demand and lower commod- July 2008, the 3-month on 3-month growth ity prices have reduced overheating pressures in industrial production in ECA contracted at in certain countries, inflationary concerns an increasingly rapid rate from October 2008, remain in a number of countries during the reaching around ­40 percent on an annual- pass-through of exchange rate depreciation. In ized rate in January 2009. The severity of this other countries, the presence of fi xed exchange downturn has exceeded even the contraction rate regimes limits the usage of monetary pol- in the OECD countries. The scale of the real icy to respond to external shocks. Political con- slowdown is also evident in the magnitude of siderations, of course, add a further complexity the GDP contractions in 2009. For example, to these policy trade-offs. Ukraine suffered a 20 percent year-on-year real GDP contraction in Q1 2009 with con- tractions in the Baltics in the range of 12 to E. Shocks to Household Welfare 19 percent in Q1 and 17 to 23 percent in Q2 Depending on the strengths and weaknesses in (according to preliminary estimates. macro vulnerabilities, and policy responses, the As real activity has fallen, there is also global shocks may result in a range of shocks increasing evidence of the transmission of 12 The Crisis Hits Home--Stress-Testing Households in Europe and Central Asia FIGURE 1.7 Sharp Contractions in Industrial Production Percent change 3 month on 3 month, seasonally adjusted annualized rate 25 0 Percent -25 -50 Jan-07 Jul-07 Jan-08 Jul-08 Jan-09 ECA OECD EAP LAC Source: WB DECPG. FIGURE 1.8 Rising Unemployment Rates 20 15 Percent 10 5 0 Jan-07 Jul-07 Jan-08 Jul-08 Jan-09 Kazakhstan Ukraine EU5 (mean) Russian Federation Baltics (mean) Turkey Sources: UN Economic Commission for Europe, Turkish Statistical Institute, the Agency of Statistics of the Republic of Kazakhstan. Notes: Unemployment rates based on labor force surveys. Break in Turkey series methodology at January 2008. the crisis through to rising unemployment turned downward in a number of countries. numbers and declining real wage growth. For example, real wages in Ukraine fell by Unemployment rates have been gradually 12 percent year-on-year in Q1 2009 compared increasing since mid-2008. The Baltic coun- with growth of around 13 percent in Q1 2008. tries, where Latvia and Estonia experienced Remittance inflows have also taken a sharp GDP contractions in 2008, have shown the downturn, tracking developments in the major earliest and particularly steep rises in unem- sources of funds, in particular Russia and the ployment but rates have also started to show EU. Slower global and regional growth lowers an upward trend in many other countries demand for migrant workers from ECA. For across the region. Real wage growth has also example, growth in the Russian construction Chapter 1--Macroeconomic Shocks 13 sector, an important source of employment for of 29 percent year-on-year in Q1 and 10 per- regional migrants, has decelerated sharply. As cent in Q4. Formal remittance inflows, in credit conditions tighten further, construc- US dollars, for Tajikistan fell 36 percent tion sector activity is likely to continue to year-on-year in the fi rst five months of 2009 decline. Indeed, formal remittance outflows with inflows to Georgia and Moldova also from Russia to CIS countries contracted by down 21 percent and 32 percent year-on-year 31 percent year-on-year in US dollar terms in respectively in the fi rst half of 2009, track- Q1 2009 compared with growth of 12 percent ing declines in Russian construction activ- in Q4 2008. This compares to depreciation of ity. After annual growth of around 37 percent the average ruble to US dollar exchange rate per year from 2004 through 2007, the growth FIGURE 1.9 Sharp Deceleration in Formal Remittance Inflows Growth in 3 month moving average, percent year-on-year 100 50 Percent 0 -50 Jan-06 Jul-06 Jan-07 Jul-07 Jan-08 Jul-08 Jan-09 Jul-09 Georgia Tajikistan Moldova Russian construction (US$) Russian construction (roubles) Sources: National authorities, IMF International Financial Statistics, Datastream, and staff calculations. Notes: Remittances are from money transfer data. Russian construction is the value of works performed in current prices and is converted from Russian rubles into US dollars at the average exchange rate of period. FIGURE 1.10 Growth in Remittance Inflows Nominal US dollar 50 Annual growth, percent Forecasts for 2009 and 2010: 25 Base--solid line Low--dashed line 0 -25 2005 2006 2007 2008e 2009f 2010f ECA Developing economies Sources: WB DECPG Migration and Development Brief, 13 July 2009. 14 The Crisis Hits Home--Stress-Testing Households in Europe and Central Asia in the nominal dollar value of inward remit- April 2009 World Economic Outlook (WEO) tances to the region is estimated by the World forecasts are for a contraction of over 10 percent Bank to have declined to around 12 percent in the Baltics in 2009, around 6 percent in the in 2008 with a contraction of 15 to 17 percent middle-income CIS countries, and 5 percent in forecast for 2009. This is roughly double the Turkey. Combined with contractions of around projected baseline contraction for total remit- 2 to 3 percent forecast for Southeastern Europe tances to developing countries and compares and the Central European new member states to a contraction of 16 percent in 1999 during and some positive growth in the low-income the Russian crisis period. These effects will CIS, the overall contraction in GDP in ECA in be particularly felt in selected countries in 2009 is forecast to be just over 4 percent with a Southeastern Europe and the low-income CIS recovery to 1 percent growth in 2010. Of par- where remittances are the largest source of ticular interest, in terms of the potential driv- external fi nance and constitute large portions ers of economic recovery, is the global nature of GDP. of the downturn with the contraction in devel- Looking forward, the likely magnitude oped economies limiting the scope for export- and timing of these adverse income shocks is led recoveries, as pursued, for example, in the dependent upon the growth outlook for the Asian crisis countries. region. This outlook has been subject to contin- The quality of economic recovery also mat- ual, and significant, downgrades from October ters. In addition to substantial uncertainty 2008. As forecasts have been downgraded, regarding the duration and severity of the cri- the uncertainty around them (as reflected in sis, it is unclear whether economic growth--if the variation in the forecasts of contributors) and when the recovery begins--will necessarily has increased. Official growth projections of translate fully into growth in household con- the International Monetary Fund (IMF) and sumption. In part, the poverty impact of eco- World Bank have been continually revised since nomic recovery depends on whether renewed October in response to the changing national growth is accompanied by, for example, com- and global economic conditions. The IMF's mensurate increases in wage, employment FIGURE 1.11 Growth around Recent Crisis Periods: 1997­98 and 2008­09 Real GDP growth 12 12 6 6 Percent Percent 0 0 -6 -6 -12 -12 1997 1998 1999 2000 2001 2008 2009 2010 2011 2012 Advanced economies Asian crisis 5 Middle-income CIS Turkey EU5 Baltics Source: IMF WEO (April 2009) database and staff calculations. Note: Regional averages are weighted using country shares in global PPP GDP. Chapter 1--Macroeconomic Shocks 15 FIGURE 1.12 "Credit-less Growth" in Emerging Markets and in the U.S. Great Depression Emerging markets U.S. Great Depression 112 120 135 165 110 130 155 115 125 145 108 110 120 135 106 115 125 104 105 110 115 102 105 105 100 100 100 95 98 95 95 85 t-2 t-1 t t+1 t+2 1929 1930 1931 1932 1933 1934 1935 1936 GDP Credit Source: Calvo, Izquierdo, and Talvi 2006a. expansion, and renewed availability of credit the international fi nance lexicon as "credit- for households and enterprises. Some recent less growth," indicates that economic output research, the implications of which are yet to may recover without any measurable recovery be fully explored, suggests substantial hetero- in domestic or external credit.5 These studies geneity in countries' experience of the resump- have also documented comparable develop- tion of economic growth following a systemic ments in the United States following the Great financial crisis. Depression. Researchers speculate that this In the recovery period following the 1998 phenomenon may be driven in part by enterprises Russia crisis, some ECA countries experienced postponing their investment projects or, where what has come to be known as "jobless growth." investments have been observed to increase, by In part, this may have been due to rapid wage financing new investment projects out of earn- increases that outstripped productivity gains, ings or funds from informal credit sources. thus constraining job creation and squeezing Because these types of economic recovery have profits in the process. Not surprisingly, poverty just been recently documented, the household has been least responsive in countries where welfare implications of such experiences have employment creation has been limited and not yet been explored. One possibility would be where jobless men and women account for a sig- that private consumption may grow more slowly nificant share of the poor.4 Along with modest (compared with consumption growth following economic growth and stalled poverty reduction other types of recession), as households restore (in some cases, rising poverty), these countries their overleveraged balance sheets and increase have also experienced rising inequality. their precautionary savings.6 The experience of systemic financial crises It is not clear whether the ECA region is in emerging markets--including a number likely to experience such "credit-less growth" of ECA countries--also suggests the pos- when the region recovers. It has been suggested sibility of economic recovery without credit. that credit-less growth was made possible in The first set of results from recent pioneering the past in countries such as Argentina mostly research, in what has been incorporated into because of rapid export growth. As mentioned, 16 The Crisis Hits Home--Stress-Testing Households in Europe and Central Asia a major difference with the emerging market have dropped off, as policy rate and quantita- crises of the late 1980s is the global nature of tive easing have been adopted, and also likely the crisis, which effectively rules out an export- demand has dropped. However, of course, led recovery heavily reliant on growth in devel- these euro rates do not reflect the burden of oped markets.7 mortgage repayments in local currency, which increased markedly with the depreciations Credit Market Shocks during the fall of 2008. Ukraine's average US dollar exchange rate in July 2009 remained Funding pressures and rising credit risks depreciated by 40 percent compared with its within domestic banking sectors have resulted level at the beginning of September 2008, in a general tightening of domestic credit con- with Poland and Hungary depreciating by 22 ditions for households and the private sector percent and 13 percent, respectively, against more generally. On the quantity side, many the euro over the same period. Some coun- countries in ECA have seen a rapid decel- tries, such as Belarus and Kazakhstan, have eration in the expansion of domestic credit, undertaken step devaluations of their curren- reflecting the drying up of external funding cies (although the fi xed exchange rate regimes and concerns over potential credit risks. In in countries such as the Baltics and Bulgaria some countries, such as Ukraine, the level have remained fi rm). of nominal private credit outstanding has remained flat or declined in the fi rst half of 2009, with real credit now declining. On the External Price Shocks cost of fi nancing, the wide variety of house- Food and fuel prices rose sharply in many hold interest rates, by currency and maturity, ECA countries in 2007 and through the fi rst makes it difficult to provide a comprehen- three quarters of 2008. Between 2006 and sive assessment of changes in lending rates. 2008, global food and fuel inflation doubled. However, there was a gradual rise in rates The rapid increase in food prices was under- on euro housing loans through 2008. Since pinned by significant droughts in various parts October/November 2008, the euro rates of the world including in some countries in FIGURE 1.13 Local Currency Depreciations In US dollars per local currency 150 Foreign currency per local currency 125 (1 Jan 2008=100) 100 75 50 25 Jan-08 May-08 Sep-08 Jan-09 May-09 Hungary (vs. euro) Poland (vs. euro) Kazakhstan (vs. US dollar) Ukraine (vs. US dollar) Sources: Datastream and staff calculations. Chapter 1--Macroeconomic Shocks 17 the ECA region, shifts toward bio-fuel pro- impact of the food and fuel crisis on households duction, declining inventories and tight com- may not be over. The rationale for this view can modity market conditions, and rising demand be split into near- and medium-term factors. In in emerging markets. Over this same period, the near term, the significant depreciations of energy prices reached record highs. A number local currencies as detailed above serve as offset- of countries in the region were particularly ting factors against the decline in the US dollar hard hit, including Kazakhstan, the Kyrgyz international prices (see figure 1.14). In addition, Republic, and Tajikistan. In many of the CIS some have expressed concern that inflation risks countries, inflation rose close to 20 percent may persist with the continuing pass-through in 2008. of recent food and fuel price increases, due to As discussed above, as the global fi nancial lags in shipping and distribution.8 There is also crisis has worsened, food and fuel prices have some indication that falling global commod- abated worldwide. In large part, this has been ity prices have not translated into lower retail driven by the worsening global fi nancial cri- food prices locally in these economies, in part sis, the economic recession in many countries because hedge contracts may have previously across different regions and the economic locked in higher prices.9 In addition, while price slowdown, more generally, and, as a result, levels may have come down, they could still be falling global demand for commodities. In at levels substantially higher than their pre-2007 addition, the increased agriculture produc- or pre-2008 levels. The marked fall in interna- tion activity, following soaring food prices and tional food prices in US dollar terms from late higher returns to agricultural activity, led to 2008 has not been reflected in a number of food a bountiful 2008 harvest, easing global com- price Consumer Price Index (CPI) sub-indices modity shortages. (figure 1.15). Indeed, these sub-indices have For many countries in ECA, however, there are continued to rise over this period in a number reasons to believe that the impact of the adverse of countries. FIGURE 1.14 International Food and Energy Price Movements and Local Currency Equivalent Indices Food price index Energy price index 200 250 Index (Jan 2007 = 100) Index (Jan 2007 = 100) 200 150 150 100 100 50 50 Jan-07 Jul-07 Jan-08 Jul-08 Jan-09 Jan-07 Jul-07 Jan-08 Jul-08 Jan-09 Jul-09 Median local currency value of Food price index (US$) Median local currency value of Energy price index (US$) food price index (shaded areas energy price index (shaded areas are 10th­90th percentile range) are 10th­90th percentile range) Sources: Eurostat, State Statistics Committee of Ukraine, DECPG. Notes: US dollar indices converted into local currency value using monthly average exchange rates. 18 The Crisis Hits Home--Stress-Testing Households in Europe and Central Asia FIGURE 1.15 CPI Food Price Sub-indices: Selected Countries 200 Index (Jan 2007 = 100) 150 100 50 Jan-07 Jul-07 Jan-08 Jul-08 Jan-09 Jul-09 International food price index Baltics (average) Hungary Poland Turkey Sources: WB DECPG and staff calculations. FIGURE 1.16 Local Currency Equity Market Declines across ECA 100 Index (1 Jan 2008 = 100) 75 50 25 0 Jan-08 Apr-08 Jul-08 Oct-08 Jan-09 Apr-09 Jul-09 ECA median Developed markets (MSCI World Index) ECA interquartile range ECA 10th­90th percentile range Sources: Bloomberg, Datastream, and staff calculations. In some CIS countries, there are also pros- energy tariffs will have to rise to ensure these pects for further rounds of external energy price networks' financial viability.10 In the Western shocks in the near term as Russia moves toward Balkans, increasing tariffs to cost-recovery lev- full market pricing for its energy exports. The els will be an important component of electric- price of imported natural gas faced by con- ity sector reform. sumers in Moldova, for example, is still below As the global recession worsens, it may European market prices and is expected to con- undercut both public and private investments verge to European levels in the near future. In in the agriculture sector, thus curbing agricul- Ukraine, some have argued that energy net- tural production. The results of some simu- works have been historically underfunded and lations conducted at the International Food Chapter 1--Macroeconomic Shocks 19 FIGURE 1.17 Housing Prices in Selected ECA Countries Growth year-on-year 80 40 Percent 0 -40 Bulgaria Croatia Hungary Slovak Rep. Estonia Latvia Lithuania Russian Fed. 2005 Q2 2006 Q2 2007 Q2 2008 Q2 2008 Q4 Source: Knight Frank Global House Price Index, various issues. Policy Research Institute suggest that a global has, or will, lead to a significant wealth effect economic recession that depresses agricultural on consumption. There have also been major investment can be associated with cereal prices changes in the path of house prices in recent that are 30 percent higher over the longer term years in many ECA countries, as in developed than in the absence of a recession.11 markets. For example, in Estonia and Latvia house prices fell in Q4 2008 by around 16 per- cent and 36 percent year-on-year respectively Wealth Shocks compared with growth rates reaching roughly Households who were long in equities and 20 percent and 60 percent in Q1 2007. Such property or short in foreign currency were par- price changes lead to redistributions of wealth ticularly exposed to the shocks to hit their net between those long or short in housing stocks. worth position over the past six months. For These can then affect the distribution of con- example, the median fall in local equity indices sumption via direct wealth effects and via the from 1 January 2008 through to their July 2009 impacts of changing collateral values on credit average levels was 60 percent, with a 40 percent constraints. Indeed, household-level analysis median fall alone over the two months from 1 in the United Kingdom has found the largest September 2008. These compare with falls of elasticity of consumption with respect to hous- 38 percent and 25 percent respectively for the ing prices in older homeowners with an insig- MSCI developed equities index in local curren- nificant elasticity for younger renters.12 cies. As discussed above, households' fi nancial A further transmission channel of the real assets in the form of mutual fund and pension and financial impacts of the crisis through to fund holdings have increased in recent years household welfare is via pension provision. The in many countries in ECA, along with their nature and magnitude of the effects of the crisis holdings of residential property. However, it is via this channel depend crucially on the struc- unclear to what extent the reduction in asset ture of the pension system, in particular the values through the downturn in asset prices mix between pay-as-you-go (PAYG), funded, (or increase in the local currency value of and voluntary pension systems.13 For example, liabilities denominated in foreign currencies) the main transmission channel for public PAYG 20 The Crisis Hits Home--Stress-Testing Households in Europe and Central Asia systems is via the impact of the crisis on con- integral part of their mandatory pension tributions, through rising unemployment and schemes, appear most directly vulnerable to potentially reduced wage growth. Funded pen- the crisis. This group includes thirteen coun- sion systems are exposed to declining asset values tries in ECA, mainly EU members but also (particularly affecting those individuals reach- Croatia, Macedonia, Kosovo, Kazakhstan, ing retirement age during the crisis). Voluntary and Russia. However, in these countries the pensions may also suffer strong adverse effects near-term implications of the fall in asset via this channel, particularly via the wider prices for households in aggregate may be equity exposures of defined contribution funds, limited for a number of reasons. First, with or through the impact of declining corporate the exception of Kosovo and Kazakhstan, health on the defined benefit schemes. where 100 percent of contributions are in the Almost all countries in ECA, except for funded pillar, these countries place a heavy Kosovo and Kazakhstan, have some form of weight on public provision (with 6.7 percent PAYG system, helping mitigate the direct pen- to 35 percent of contributions in the funded sion impact of the crisis. However, the ability pillar). Second, in the near term relatively few of countries to absorb rising pension deficits workers are retiring with direct exposure to as a result of falling contributions is dependent such second-tier benefits (although clearly upon their fiscal space. Indeed, in some coun- there may be a greater future impact if asset tries that have been particularly affected by the prices remain depressed in the medium term). crisis, revisions to state pension provision may Looking forward, from a political economy form part of the fiscal adjustment. In Latvia, perspective, the fall in the value of funded for example, significant pension cuts have been pillar pensions because of the crisis may also recently proposed. have implications for the appetite for future Those countries, which adopted fully reforms or calls for changes in the current funded defi ned-contribution schemes as an structure of pension provision. CHAPTER 2 Household Vulnerabilities A. Introduction Interest rate shocks in Estonia, Lithuania, and Hungary, for example, can increase This chapter examines household vulnerabili- the share of vulnerable households or bor- ties by analyzing how macro shocks discussed rowers at risk (in percentage of all indebted in Chapter 1, namely (i) credit market shocks, households) by up to 20 percentage points, (ii) external price (food and fuel) shocks, and depending on the magnitude and sever- (iii) income shocks, impact their well-being. It ity of the shock. Household indebtedness treats each of these shocks separately and quan- has risen rapidly in ECA countries among tifies how the crisis is likely affecting the wel- both poor and nonpoor households. The fare of households, on average, and, whenever nature of household debt in ECA is such possible and drawing from country-specific that many households have likely exposed examples and illustrations, how such welfare themselves to various types of risks, effects may be distributed across households. including exchange rate and interest rate The results reported in this chapter sug- risks, with few opportunities for hedging. gest that household vulnerability in ECA is · The food and fuel crisis may not be over widespread. There are adverse effects on both and a new round of price increases will poor households and nonpoor households have substantial effects on household depending on the macroeconomic shock, the welfare. International commodity price specific transmission channel, and selected levels have not returned to pre-2007 lev- household characteristics. In brief, this chap- els and falling currencies in some ECA ter fi nds the following: countries are resulting in a new round of · The ongoing macroeconomic shocks will price increases, depending on the share of significantly expand the pool of house- imported food and fuel in local consump- holds that are unable to service their debt. tion and the degree of pass-through of 21 22 The Crisis Hits Home--Stress-Testing Households in Europe and Central Asia exchange rate changes in domestic prices. Section C discusses the empirical strategy, the The net effect of a food price shock depends scope and objectives, and the key limitations on whether households are net producers of the analysis. Section D presents the results or net consumers of food, it depends on of the analysis of household indebtedness and their intensity of food consumption and the likely impact of credit market shocks on the availability of cheaper substitutes, and the household debt service burden in selected it depends on their livelihood strategies, countries. Section E analyzes recent trends in access to agriculture assets and inputs, food and fuel prices, patterns of food and fuel and ability to take advantage of profitable consumption among households in the region, opportunities in agriculture. These mul- and the likely impact on households of a new tiple considerations suggest that, at least round of food and fuel prices increases. Section in principle, the poor are not necessarily F assesses the impact of the regional recession the hardest hit. However, the food share of and, in particular, the effect on household wel- total household consumption typically falls fare of falling incomes in the region. with income and, in reality, the poor are also likely to be the worst hit, as many of the poor in Albania, Kyrgyz Republic, and B. Context: Poverty and Vulnerability Tajikistan, for example, are also observed in the Pre-Crisis Period to be net consumers, with limited access to Over the recovery period following the 1998 agriculture assets and inputs. Russian crisis through 2006, some 50 million · Poverty will rise. The region now risks people moved out of poverty in the region.14 a major reversal of its gains in the years Poverty fell throughout all the sub-regions of of economic recovery following the 1998 ECA led by the middle-income countries of the Russian crisis. By 2010, about 11 million CIS, which experienced the largest declines in more people could fall into poverty and poverty, particularly in Russia where the share an additional 24 million people could of the poor and vulnerable has declined sharply find themselves vulnerable, or just above in percentage of the population. ECA's international poverty line, over the Poverty reduction in ECA has been driven next two years. The regional simulations largely by growth in average income. In par- mask the heterogeneity of impact within ticular, growth in mean income is calculated countries, including the likely concentra- to have contributed close to 90 percent of the tion of the poverty impact in selected eco- overall reduction in poverty experienced by the nomic sectors. Country studies recently region. A modest improvement in the distribu- completed suggest that for economic tion of income has also helped reduce poverty. shocks transmitted primarily through the The labor market has provided an important labor market, poverty will rise especially channel for poverty reduction in the region, among households that have been depen- largely through rising real wages among the dent on remittance inflows and those pre- working poor. In contrast, job creation has gen- viously employed in booming construc- erally not been an important factor for reducing tion sectors where economic activity is poverty. For a number of countries, the index of now projected to decline sharply. employment has been flat in recent years. The rest of the chapter is organized as fol- Notwithstanding the rapid decreases in lows: Section B sets out the context for the poverty, millions of people still remain poor or analysis in terms of the trends in poverty are just above the poverty line. More than two- reduction and vulnerability prior to the crisis. thirds of the poor live in the middle-income Chapter 2--Household Vulnerabilities 23 countries of the region, including Kazakhstan, The novel microeconomic analysis in the Poland, Romania, Russia, and Turkey. Not sur- report draws on a large, cross-country database prisingly, low-income countries in ECA have of household surveys. We use the most recent higher rates of poverty and vulnerability com- pre-crisis household data along with aggregate pared to other countries in the region, but as a macroeconomic outturns to simulate the impact group, they account for less than a quarter of on households of key economic shocks already the region's poor population. taking place. The report brings together for The growth slowdown will have significant the first time comparable cross-country data adverse consequences for the region, given its on household indebtedness for a large group poverty and vulnerability profile. Most of the of ECA countries using the EU Survey of poor in the ECA region are working adults and Income and Living Conditions (EU-SILC) children. Together, they represent about two- and Household Budget Surveys (HBS). The thirds of the poor population across countries report also highlights newly updated informa- in the region, with the working poor account- tion on household consumption from the ECA ing for anywhere from about a quarter (e.g., Household Data Archives. Comparisons with Turkey) to close to half (middle-income CIS) of Western Europe and other advanced econo- the poor population. The working poor repre- mies are also used to inform the analysis when sent a group that is directly exposed to the fall relevant data are available. in income and declining employment prospects projected throughout the region. In addition, Important Caveats because many of those currently employed have The analysis is not exhaustive. First, the choice just moved out of poverty, they are just above of emphasis has been guided in large part by the poverty line and highly susceptible to mod- the policy issues of the day when the project est falls in mean income and economic activity. was first designed. Thus, for example, it is an analysis of food and fuel price shocks, instead of rising prices of other household consump- C. Shocks to Household Welfare: tion items, reflecting the widespread concern Empirical Strategy over the food and fuel crisis. It is an analysis of household indebtedness and the welfare con- The rest of this chapter examines household sequences of rising debt burdens--rather than, vulnerabilities using micro data by exam- for example, the analysis of the welfare effects ining the potential impact of credit market of the continuing lack of access to credit among shocks, external price (food and fuel) shocks, certain households--given the policy interest and income shocks on household welfare. The in the new vulnerabilities created by the credit impact on household welfare is defined as the boom through 2007­2008 in many countries change in the household debt service burden, in ECA. Second, the country examples and the fall in real income, or movements into analyses have also been selective, depending on poverty, as appropriate. The report presents data availability and the relative country expo- regional overviews and simulations along with sure and risk. This is a primary reason why cross-country comparisons and contrasts. the potential impact on household welfare of It also presents selected country examples, wealth changes associated with the crisis, such depending on data availability and relevant as those related to property and equity price economic developments, to illustrate the inci- falls, is not analyzed. dence and distribution of specific vulnerabili- The analysis is not predictive. The GDP ties within countries. growth projections alone, on which many 24 The Crisis Hits Home--Stress-Testing Households in Europe and Central Asia of the simulations here are predicated, were The aim, then, is to provide a broad over- updated at least three times during the course of view of household vulnerabilities confronted the preparation of this report. The frequency by the region. This report seeks to assess how of updates reflects the substantial uncertainty such vulnerabilities are distributed across coun- regarding the severity and duration of the cri- tries and, within countries, across broad types sis and, by extension, its impact on households. of households. Some of the estimated effects In addition, the actual poverty impact can be may be understated, as they capture only some mitigated by private responses to the crisis and of the first-round effects. On the other hand, by various household coping strategies. This general equilibrium effects will either dampen study is unable to account for the interactions or worsen some of these effects. between channels, the behavioral changes This report is also cognizant of many ongo- (including coping strategies to mitigate the cri- ing, decentralized efforts throughout the region sis, discussed in the fi nal chapter), and the net to simulate the household welfare consequences higher-order effects. Policy responses, mean- of the crisis. Such efforts are able to model while, can either mitigate or exacerbate some these effects and interactions more fully using of these consequences. richer, more comprehensive country-specific The results are not additive. The analyses data. This report does not substitute for those of changes in household welfare in response efforts--some of whose preliminary results are to three types of shocks are treated as parallel also reported and referenced below--but instead exercises. First, the shocks have not been ana- provides a complementary regional overview. lyzed within a common general equilibrium framework. Second, the choice of country examples and illustrations vary across shocks, D. Households and Credit Market Shocks depending on data availability and the rel- evant country risk. More generally, the data Background are drawn from two different household data Household debt now represents a little over sources. Finally, there are methodological dif- a quarter of GDP in the EU10. Though non- ferences in the way vulnerability was evalu- trivial, this level is lower than the debt level in ated, consistent with the respective existing most of the older EU member countries, which literature. With household indebtedness, for on average is about 65 percent of GDP. Within example, the impact of the shocks is assessed the EU10, however, there is significant variation relative to indicative debt burden "thresholds" in aggregate household debt holdings, in both (payments in percentage of income). With their level and composition. Estonia is at the respect to the fall in income, the predicted higher end of the distribution by magnitude, household consumption levels are compared with household debt representing close to half to an international poverty line. However, the of GDP. Household debt in percent of GDP in real impacts in each country may very well be the CIS and other ECA countries, in turn, lags multi-channeled and cumulative. Countries behind the EU10, on average. Ukraine is an in Eastern Europe, for example, have to deal exception, with household indebtedness com- with fi nancial sector shocks and labor mar- parable to that of the EU10 average. ket shocks, as well as product market shocks As household financial positions have transmitted through currency declines. For grown, there has been a shift toward hous- a number of reasons, however, we treat these ing loans or mortgages on the liability side impacts independently using partial equilib- of the balance sheet and an increasing share rium analysis. of equities and pension and mutual funds Chapter 2--Household Vulnerabilities 25 FIGURE 2.1 Household Debt: Selected ECA Countries, 2008 In percent of GDP, end-period 50 40 30 Percent 20 10 0 Bulgaria Croatia Czech Rep. Estonia Hungary Latvia Lithuania Poland Romania Slovak Rep. Slovenia Belarus Kazakhstan Russian Fed. Ukraine Armenia Turkey EU10 (incl. Croatia) Middle-income CIS Other Sources: European Central Bank, National Central Banks, IMF, and UniCredit. FIGURE 2.2 Growth in Mortgage Debt: Selected ECA Countries, 2007 1.5 1.0 Percent 0.5 0 Bulgaria Czech Rep. Estonia Latvia Lithuania Hungary Poland Romania Slovenia Slovak Rep. Albania Serbia Armenia Ukraine Western EU10 Balkans CIS Source: European Mortgage Federation. on the asset side. For many of the countries in form of household loans. Elsewhere, such as in ECA, rising mortgage and housing loans have the CIS, housing loans are a much smaller share accounted for much of the growth in house- of all household loans compared to the EU10. hold liabilities and now, for example, account for Mortgage debt has increased multiple times over around 40 percent of total household liabilities the period 2002­2007 both in per capita terms in Bulgaria, Croatia, and Hungary and up to 60 and relative to disposable income. Some have percent in the Czech Republic. However, for suggested that government initiatives, such as some notable exceptions, such as Romania and construction or mortgage-related subsidy and Bulgaria, consumer credit remains the primary tax schemes, have contributed to this growth. 26 The Crisis Hits Home--Stress-Testing Households in Europe and Central Asia The growth in mortgages can also be viewed in accounted for close to 60 percent of all house- terms of longer-term convergence toward the hold loans. Mortgage debt in Serbia, for exam- financial norms of Western Europe. Indeed, as ple, grew at an astonishing rate of 96 percent in would be expected, the most rapid annualized 2007 while growing by 86 percent in Albania. In growth over the period 2002 to 2007 was seen Russia, the Central Bank reports that mortgage in those economies with relative low initial levels lending grew by a factor of 2.6 in 2007.15 CIS of mortgages per capita. At 2007, housing loans mortgage lending also grew very sharply. FIGURE 2.3 The Composition of Household Debt In percent of total EU10 2008 (end-period) Selected countries 2008 (end-period) 100 100 75 75 Percent Percent 50 50 25 25 0 0 Bulgaria Czech Rep. Estonia Hungary Latvia Lithuania Poland Romania Slovak Rep. Slovenia Russian Fed. Turkey Ukraine Housing loans Other Source: National Central Banks. Consumer credit Housing loans Other loans Source: European Central Bank. FIGURE 2.4 Latvia Household Loan Delinquency Rates and Unemployment Rate 20 15 Percent 10 5 0 Q3 Q4 Q1 Q2 Q3 Q4 Q1 Q2 Q3 Q4 Q1 Q2 Q3 Q4 Q1 Q2 2005 2006 2007 2008 2009 Share of household loans overdue by 31­90 days Share of household loans overdue more than 180 days Share of household loans overdue by 91­180 days Unemployment rate (seasonally adjusted) Sources: Eurostat and the Financial and Capital Market Commission, Latvia. Chapter 2--Household Vulnerabilities 27 The welfare consequences of rising house- adversely affect the health of financial sector hold indebtedness in the ECA region can be balance sheets with second-round implications significant. Rising indebtedness reflects the for households in terms of the availability and benefits of financial sector development, allow- cost of credit. Already, rising household loan ing households to smooth their consumption delinquency ratios are being observed in some over time and acquire home ownership without countries, as unemployment rates have risen. significant savings. In fact, there exists a grow- First, a large share of household debt is ing literature on the welfare impact of credit denominated in foreign currencies or is indexed constraints among households, particularly in a to foreign currencies, which has exposed house- downturn when households are unable to meet holds to recent exchange rate depreciations to their consumption needs.16 On the other hand, the extent that the currency composition of rapidly growing household indebtedness and their assets, particularly their labor income the exposure of the financial sector to vulner- flows, leaves them unhedged. Where foreign able households (or borrowers at risk) may have currency loans became popular in recent years, important consequences for financial stability. borrowers were typically obtaining loans in At the same time, the welfare and distribu- Euros and Swiss francs, attracted to relatively tional implications for households themselves lower nominal interest rates compared to loans can be large, particularly in a worsening mac- denominated in local currency.17 On the banks' roeconomic environment. side, at the height of the expansion in household Some characteristics of household debt in credit, there appeared little interest in reducing ECA expose these households to a number of their exposure to foreign-currency-denomi- macroeconomic shocks. Given these charac- nated loans because default rates were low and teristics, household debt service burdens may because of the ease of access, at the time, to for- increase in a difficult macroeconomic envi- eign currency funding via wholesale markets or ronment, and this in turn may lead to higher via Western European parent banks. Among default and delinquency rates. These in turn households borrowing in foreign currency, FIGURE 2.5 Foreign-Currency-Denominated Loans, 2008 In percent of bank loans to households 100 75 Percent 50 25 0 Bulgaria Croatia Czech Rep. Estonia Hungary Latvia Lithuania Poland Romania Slovak Rep. Armenia Belarus Kazakhstan Russian Fed. Ukraine FYR Macedonia Turkey EU10 CIS Other Sources: MNB and other National Central Banks. 28 The Crisis Hits Home--Stress-Testing Households in Europe and Central Asia FIGURE 2.6 Foreign-Currency-Denominated Loans in Ukraine, 2008 In percent of total, by currency 100 75 Percent 50 25 0 Consumer credit Housing loans Other Hryvnia US dollar Euro Russian rouble Other Sources: National Bank of Ukraine and staff calculations. however, there also seemed little awareness of Second, in some EU10 countries, mort- their exposure to currency risks although in gages with variable (adjustable) interest rates some countries a high share of foreign currency account for the largest share of lending, expos- deposits provides some hedging of the currency ing households to interest rate shocks. In these risk. These developments in the EU10 mirror countries, such variable interest rate mortgage recent trends elsewhere, particularly in the mid- debt represented over three-quarters of all dle-income CIS countries, where households mortgage debt, at least until recently, using also obtained loans denominated in US dollars available data. Households are vulnerable in a and other foreign currencies. In Serbia, few of financial downturn, in the event that banks pass the household loans are explicitly foreign-cur- on a higher cost of credit to them. However, rency-denominated. However, up to 81 percent as shown in other countries' experiences, this appear to be foreign-currency-indexed.18 may be mitigated to the extent that interest rate There are again substantial variations in the adjustments may be capped, as is the case, for foreign currency exposures of household debt example, in Denmark. across countries. The Baltics and Ukraine are Increasing mortgage indebtedness has at the higher end of this distribution, with for- exposed a rising share of households to the eign-currency-denominated loans accounting recent changes in house price trends in many for over 80 percent of bank loans to households of the EU10 countries. For example, in Estonia in Estonia and Latvia; the Czech and Slovak and Latvia house prices fell in Q4 2008 by Republics are at the lower end, with little or no around 16 percent and 36 percent year-on-year foreign-currency-denominated household debt. respectively compared with growth rates of Some have suggested that a few national policies around 20 percent and 60 percent in Q1 2007. may explain some of these differences across Such price changes lead to redistributions of countries--such as more restricted eligibility wealth between those long or short in housing requirements in 2004 for housing subsidies in stocks. These can then affect the distribution Hungary (which then prompted households to of consumption via direct wealth effects and substitute toward less expensive foreign currency via the impacts of changing collateral values loans) or regulatory measures to limit borrow- on credit constraints. Indeed, household-level ing foreign currency in the Czech Republic.19 analysis in the United Kingdom has found the Chapter 2--Household Vulnerabilities 29 FIGURE 2.7 Mortgage Loans with Adjustable Interest Rates, 2006 In percent of all housing loans 100 75 Percent 50 25 0 Ireland Hungary Slovakia Slovenia Belgium Croatia Denmark France Italy Netherlands Portugal UK EU10 Other countries Sources: IMF, OECD, and National Central Banks. largest elasticity of consumption with respect System that was first initiated in 2003, with the to housing prices in older homeowners with an new EU member countries undertaking their insignificant elasticity for younger renters.20 first surveys in 2005. Data are typically made available to the general public two years after Household Vulnerability: the survey, so we currently have data through 2007 and data for both older and newer EU Regional Overview members for 2005­2007. Among other vari- Microeconomic data can be a critical source ables, the EU-SILC collects information on of information on household indebtedness. the incidence of mortgage debt holding, inter- Current assessments of the credit risks faced est payments, arrears on mortgage interest pay- by the banking sector have been largely based ments, and disposable income. The analysis of on macroeconomic data. In general, little is EU-SILC data is supplemented with the analy- known about household indebtedness based sis of a few other ECA countries with relevant on household-level data in the EU10. The debt variables in their HBS. For these countries, profile could vary across household income there is information on total household debt, groups and by type of loan, such as mortgage including mortgage debt and other household and non-mortgage. In principle, such micro- loans, and on total debt service, including inter- economic data and profiles allow for a closer est payment and principal payment, in contrast monitoring of risks associated with selected to the EU-SILC data, which has information household groups. Where household borrow- on interest payment alone. ing is limited, indicators based on average There are a few notable patterns in the house- household indebtedness for all households as a hold survey data, suggesting likely adverse wel- whole mask the likely concentration of borrow- fare consequences during an economic down- ing among selected households. turn. These include patterns of household debt This section draws information from the holdings, including among those that are more databases of the EU-SILC, an annual household vulnerable or less able to service their debt in a survey anchored in the European Statistical difficult economic environment. 30 The Crisis Hits Home--Stress-Testing Households in Europe and Central Asia BOX 1 Definition of EU-SILC Variables Used in the Analysis Total disposable household income is the sum for all household members of gross personal income components (gross employee cash or near cash income; gross non-cash employee income; gross cash benefits or losses from self-employment [including royalties]; unemploy- ment benefits; old-age benefits; survivor's benefits; sickness benefits; disability benefits; and education-related allowances) plus gross income components at household level (income from rental of a property or land; family/children-related allowances; social exclusion not else- where classified; housing allowances; regular inter-household cash transfers received; inter- ests, dividends, profit from capital investments in unincorporated business; income received by people aged under 16) minus regular taxes on wealth; regular inter-household cash trans- fer paid; and tax on income and social insurance contributions. Interest paid on mortgage refers to the total gross amount, before deducting any tax credit or tax allowance, of mortgage interest on the main residence of the household during the income reference period. It excludes any other mortgage payments, either interest or principal, made at the same time, such as mortgage protection insurance or home and contents insurance; payments on mortgages to obtain money for housing purposes (e.g., repairs, renovations, maintenance) or for non-housing purposes; and repayments of the principal or capital sum. Reproduced from Eurostat, Description of SILC User Database Variables. Version 2007.1 (01-03-09). Permission being requested. First, debt holdings rise with household interest mortgages). If the mortgage payments income level but are spread across income represent a considerable share of a house- quintiles, including the poorer households. hold's disposable income, a rising debt burden In the Czech Republic, for example, over a may curtail the household's ability to protect third of households in the poorest quintile its welfare. hold some debt, rising to about 55 percent of Second, in some countries debt service is a households in the richest quintile. In addi- significant share of income, particularly among tion, on average among EU10 countries, the the poor. In Hungary, for example, data from share of mortgage holders across age groups the EU-SILC suggest that mortgage inter- first increases and then decreases with age, a est payments among the poor represent over pattern that is broadly consistent with the life 10 percent of their income. In Latvia, the share cycle theory of consumer behavior. Taken increases to almost 15 percent. If anything, together, these suggest that when macroeco- these estimates of the debt service burden may nomic shocks increase the financial burden due be understated. A recent, independent survey to mortgage debt, it is the poorest households for UniCredit Group indicates that for about and the youngest households with weaker ties 30 percent of all households, total household to the labor market who are among those most debt repayment covers more than a fifth of likely to suffer adverse shocks, in the absence the household budget. Another 20 percent of of a savings buffer. The shocks can be chan- households allocate 10 to 20 percent of their neled through income shocks, exchange rate household budget to debt repayments. shocks (if the mortgage is in foreign currency), Third, in some countries, mortgage inter- or interest rate shocks (in case of variable- est payments are a significant share of income Chapter 2--Household Vulnerabilities 31 FIGURE 2.8 Household Debt by Income Quintile In percent of households Selected EU10 countries, 2007 Other ECA countries, 2006 or 2007 80 80 60 60 Percent Percent 40 40 20 20 0 0 Czech Rep. Estonia Hungary Lithuania Latvia Poland Slovak Rep. Slovenia Georgia Kazakhstan Turkey Ukraine Q1 Q2 Q3 Q4 Q5 Sources: EU-SILC and staff calculations. FIGURE 2.9 Household Income Used for Debt Repayments In percent of all households 100 75 Percent 50 25 0 Bulgaria Czech Rep. Hungary Poland Romania Slovak Rep. Slovenia 1­5% of household budget 6­10% 11­15% 16­20% 21­30% >30% Don't know Source: UniCredit Group. Note: The graph should be read as follows: The bars represent shares of the household population. The colors represent the percentage of household budget spent on debt service. In Slovenia, for example, 10 percent of households spend 1 to 5 percent of their budgets on debt service. Another 15 percent of households allocate 6 to 10 percent of their budgets for debt service. among the youngest and oldest workers. In income to pay mortgage interest. In Lithuania, Hungary, the youngest workers (age 25 and interest payments as a share of disposable younger) allocate over a tenth of their dis- income are the smallest in the youngest work- posable household income to mortgage inter- ers and then rise with age, reaching close to est payments. Slovak workers in the youngest 6 percent of income among those aged 55 or group also spend almost 10 percent of their older. In the meanwhile, in Latvia, mortgage 32 The Crisis Hits Home--Stress-Testing Households in Europe and Central Asia interest payments occupy a large share of We ran similar analyses of mortgage debt household income across all age groups from on a few selected countries, using EU-SILC the youngest to oldest, which is, in most cases, data, and of total household debt, using HBS over 10 percent of disposable income. Large data. The choice of countries has been guided debt service ratios are also observed among by data availability and the degree of exposure those employed in economic sectors that have to interest rate, exchange rate, and unem- experienced some of the sharpest downturns in ployment rate shocks. The magnitude of the recent months (such as in construction). hypothetical shocks has been driven by actual changes in recent months--such as the dou- Stress Testing Household Indebtedness: bling of unemployment rates in some EU10 Country Illustrations countries--and is thus greater in magnitude than what has been previously assumed by the Introduction few stress tests that exist. We also simulated Stress tests of household debt using microeco- uniform shocks across countries--for exam- nomic data are rare, but the results of existing ple, 25 percent exchange rate depreciations-- tests suggest considerable welfare consequences. as explained more fully in Box 2. A recent stress test in Hungary suggests that a For two of the countries analyzed, the fi xed simultaneous fall in employment and an inter- exchange rate regime, the authorities' com- est rate shock would increase "risky loans" by mitment to the peg, and the credibility of this 8 to 12 percentage points. Though the banking regime protect households from the adverse sector is found to be resilient to these shocks, effects of exchange rate adjustments. Some the default risk is concentrated among the poor institutions have been examining the mer- households. A stress test in Poland suggests that its of alternative exchange rate regimes and unemployment shocks (compared to interest rate they argue that a regime change would lead to or exchange rate shocks) have the highest impact adverse social consequences.22 The results of the on probability of default. Modest increases in simulations here should be seen in a similar light, unemployment can increase the share of loans in as evidence of the likely welfare costs of abandoning default by over 5 percentage points.21 the peg. BOX 2 Stress Testing Household Indebtedness This section follows some recent attempts to stress test household debt holding and assess the share (or change in share) of vulnerable households. 23 The steps taken are outlined as follows: Calculating baseline figures. First, we use data from a sample of households from selected household surveys to calculate some baseline figures for the debt service burden (which may or may not include principal repayments, depending on the data source). Identifying vulnerable households. Second, the debt service burden is compared to an indicative threshold, against which a household may be classified as vulnerable. There are two principal ways of defining indicative thresholds for determining the increase in the number of vulnerable households (sometimes referred to as "borrowers at risk"): Chapter 2--Household Vulnerabilities 33 BOX 2 (continued) Stress Testing Household Indebtedness (i) One is based on the notion of a "financial margin," or the disposable income left after deducting debt payments and basic living costs.24 Households with negative margins are considered to be in financial distress and are likely to default on their debts. We use this benchmark in analyzing the HBS data on the total debt service burden. (ii) The other measure is an indicative threshold debt service burden that is more arbitrary, and based on previous findings regarding the level at which a household is more likely to become delinquent or be in arrears. In particular, we compare EU-SILC data on interest payments as a share of disposable income to an indicative threshold of 20 percent, follow- ing previous studies of mortgage delinquencies.25 We also use a 30 percent threshold in our sensitivity tests. Subjecting households to shocks. Third, the households are subjected to various mac- roeconomic shocks, such as an interest rate shock, an exchange rate shock, and an unem- ployment shock. The magnitude of the shock--such as some percentage point increase in the aggregate unemployment rate--depends on a country's own historical increase, over the most recent two-year period. 26 In addition, we use two hypothetical scenarios, as listed below. The selection of specific households subjected to a shock proceeds as follows: In the case of an unemployment shock, the selection of a particular household that is sub- jected to the shock--that is, whether a given household member becomes jobless--can be either random or based on a probability model of unemployment. In the case of a random assignment, we employ 1,000 draws and then calculate the average outcome, including the standard error.27 In the case of a probability assignment, the likelihood of being unemployed is drawn from a probit model of unemployment, where unemployment is a function of an indi- vidual's socioeconomic background, demographic characteristics, and geographic location. We employ a parsimonious model and assume that there are four key drivers of the probability of unemployment: educational level, age, gender, and place of residence.28 In the case of an interest rate shock, we use aggregate figures on the share of variable interest loans out of all household loans (e.g., 40 percent of all loans have adjustable interest rates). We adopt one critical assumption: we assume that loans with variable interest rates are proportionally distributed across indebted households, such that where, for example, 40 per- cent of all loans are variable interest rate loans, then we assume that 40 percent of all indebted households have variable interest rate loans. This is admittedly a crude assumption, as house- hold debt with variable interest rates may be concentrated among selected groups of house- holds. However, in the absence of more detailed information in household surveys on interest rate variability of each household's debt holding, this is the most reasonable assumption we could make.29 The share of households with variable interest loans is randomly selected using a similar routine outlined above, and these households are then subjected to the relevant inter- est rate shock. As before, we employ 1,000 draws and calculate the average outcome. In the case of an exchange rate shock, the selection of households follows the same general procedure outlined above in the preceding paragraph. 34 The Crisis Hits Home--Stress-Testing Households in Europe and Central Asia BOX 2 (continued) Stress Testing Household Indebtedness The magnitude of the hypothesized shock. As previously stated we simulate the impact of shocks based on actual changes in recent years as well as on uniform, hypothesized mag- nitudes. In the case of unemployment rates, for example, some Baltic countries have expe- rienced as much as 10 percentage point increases in the unemployment rate. We use these actual changes as base cases as well as a uniform 10 and 15 percentage point increase across countries. In the case of interest rate shocks, countries in our sample have experienced 2 to 4 percentage point maximum increases in recent years. In addition, we also simulate 3, 5, and 6 percentage point increases. A 3 percentage point stress test is typical in the household debt literature, but other have simulated more severe shocks such as 5 and 6 percentage point increases as well.30 Finally, in the case of exchange rate shocks, we use maximum actual changes (which have been substantial in a few cases) as well as a uniform 25 and 35 percent depreciation. For selected countries with SILC data, we are able to simulate inter- est rate, exchange rate, and unemployment shocks. For countries with HBS, we do not have the necessary information on the share of loans with variable interest rates. Recalculating the share of vulnerable households. As households are subjected to a shock--for example, as a member of the household becomes "unemployed" and ceases to receive any income from work31--the shock then results in the decrease in total household income and the household's ability to repay debt. In the case of an interest rate shock or an exchange rate shock, the size of debt (or interest) payment grows accordingly. In all cases, the debt service burden is recalculated and then compared to the relevant threshold. The share of vulnerable households then rises accordingly. Results interest rate hike--3 percentage points--using a The results of the analysis of EU-SILC data 30 percent threshold yields much lower but still suggest that current macroeconomic shocks nontrivial adverse consequences for household can significantly expand the pool of households well-being. Borrowers at risk increase by 4 to that are unable to service their debt. A severe 7 percentage points of all indebted households. 5 percentage point interest rate shock in Estonia, The results are in appendix table 1, appendix Lithuania, and Hungary, for example, can table 2, and appendix table 3. increase the share of vulnerable households or Unemployment shocks also expand the share borrowers at risk by up to 20 percentage points, of vulnerable households by several percentage depending on the magnitude and severity of points. The results hold, regardless of whether the shock (figure 2.10). A less restrictive thresh- the unemployment shock follows a probability old--that is, interest payment representing assignment or a random assignment, although 30 percent of disposable income--yields smaller a random assignment generally leads to higher welfare effects from interest rate shocks com- welfare costs. pared to this first set of estimates, but they are In the countries analyzed, interest rate still large. In particular, the share of borrowers shocks have the largest impact on household at risk can expand by 7 to 12 percentage points vulnerability. In part, this is due to the assumed in our sample of countries. A more modest magnitude of the shock. It is also driven by the Chapter 2--Household Vulnerabilities 35 FIGURE 2.10 Stress Testing Household Indebtedness: Selected EU-SILC Data Vulnerable households or borrowers at risk as a share of indebted households Unemployment shock Interest rate shock Exchange rate shock 30 30 30 20 20 20 10 10 10 0 0 0 Estonia Hungary Lithuania Estonia Hungary Lithuania Estonia Hungary Lithuania Before After Sources: EU-SILC data and staff calculations. Note: The simulated shocks are a 10 percentage point increase in unemployment rates, a 5 percentage point increase in interest rate shocks, and 25 percent depreciation in exchange rates. This refers to mortgage debt only. Vulnerable households are identified using a 20 percent interest payment threshold. See main text and Box 2. degree of initial exposure. In particular, while The analysis of HBS data assuming compa- only some indebted households will be hit by an rable unemployment and exchange rate shocks unemployment shock, all indebted households suggests a far more limited impact on house- with variable interest rates will see increasing hold welfare.32 In Belarus, Kazakhstan, Serbia, debt burdens from an interest rate hike. and Ukraine, exchange rate shocks increase, FIGURE 2.11 Stress Testing Household Indebtedness: Selected HBS Data Vulnerable households or borrowers at risk as a share of indebted households Unemployment shock Exchange rate shock 15 15 10 10 5 5 0 0 Belarus Kazakhstan Serbia Ukraine Belarus Kazakhstan Serbia Ukraine Before After Sources: HBS data and staff calculations. Note: The simulated shocks are a 10 percentage point increase in unemployment rates and a 25 percent depreciation in exchange rates. Vulnerable households are identified based on the "financial margin" measure. See main text and Box 2. 36 The Crisis Hits Home--Stress-Testing Households in Europe and Central Asia on average, the share of borrowers at risk by These two sets of estimates--one for SILC less than 1 percentage point. Unemployment and one for HBS--are not comparable. As rate shocks, on the other hand, were found to previously discussed, one is based on mort- increase vulnerable households by up to 5 to gage debt information and the other is on 6 percentage points, depending on the magni- total household debt. The two methodolo- tude of the shock and depending on whether gies employed for identifying borrowers at the shocks are distributed randomly or accord- risk--the use of a threshold level of interest ing to a probability assignment. payment burden in the case of SILC data BOX 3 EU-SILC and HBS Data on Household Debt: Comparisons with Other Sources This chapter uses household debt data drawn from EU-SILC and HBS data for those ECA countries for which the relevant data are available. Some caution is warranted in the interpre- tation of these household survey data, as they are not primarily designed to collect informa- tion on household liabilities. The focus of the SILC survey is not access to financial services or financing constraints, but rather social exclusion and income poverty. Because these are survey data, the volume of household loans and mortgages may not necessarily correspond fully to aggregate data from the banking sector. Nonetheless, in countries where SILC data have been analyzed, the data have provided a useful statistical portrait of the distribution of household debt.33 More important, the calculations used in this chapter are broadly consistent with other sources of macroeconomic and microeconomic information. First, the trends in the share of households with negative financial margins are consistent with aggregate macroeconomic developments. In Belarus, for example, between 2001 and 2008, the economy grew rapidly while poverty rates sharply. According to the National Bank's Bulletin of Banking Statistics, household and aggregate NPL ratios also declined. In our calculations of financial margins, the share of households with nega- tive margins was also small and fell steadily between 2001 and 2008 (appendix figure 2). In Kazakhstan, more recent household survey data obtained by the team indicates that the share of household with negative margins began rising as the crisis hit the country in late 2007. Second, the socio-economic and demographic patterns of household debt holding documented here are consistent with other existing analyses of household debt data. In Lithuania, for example, summary data from Bank of Lithuania indicate that the share of households with housing loans rises with income, consistent with EU-SILC pat- terns. More generally, the patterns of debt holding by income and by age in ECA countries for which data are available are fully consistent with those of Western Europe and other advanced economies. 34 Third, for the Western European countries for which we have SILC data, the relative rank- ing of countries in the share of indebted households is consistent with those of other sources, such as the European Central Bank and OECD. Greece and Italy are at the lower end of the distribution while Netherlands is at the higher end. Chapter 2--Household Vulnerabilities 37 BOX 3 (continued) EU-SILC and HBS Data on Household Debt: Comparisons with Other Sources The Share of Indebted Households: EU-SILC and ECB Data 80 60 40 20 0 Greece Italy Ireland Portugal Spain France Netherlands SILC ECB Nonetheless, in some countries there are significant differences in the estimated levels of household indebtedness as can be seen in the figure. There are also differences in the estimated debt service burden, where EU-SILC data for ECA countries can be compared with those of other studies. For example, the debt service numbers calculated by the Bank of Slovakia, particularly for the poorest households, are several factors higher than those reported in this chapter; it is not clear what is driving the differences. Meanwhile, the debt service burden estimated by UniCredit (figure 2.9) is much higher relative to estimates from the EU-SILC for the same group of countries. In part, it could reflect the deteriorating household financial conditions captured by UniCredit's more recent data collection efforts. They may also be explained by differences in survey design, though information on UniCredit's survey design is not readily available. The results of the stress tests in this chapter could then be possibly a lower bound. Debt Service: SILC and Unicredit Data 100 75 Percent 50 25 0 SILC Unicredit SILC Unicredit SILC Unicredit SILC Unicredit SILC Unicredit Czech Rep. Hungary Poland Slovak Rep. Slovenia > = 5% of household budget 6­10% 11­15% 16­20% 21­30% <30% 38 The Crisis Hits Home--Stress-Testing Households in Europe and Central Asia and the use of fi nancial margins in the case emerging and developing countries are now of HBS data, which is a less restrictive mea- substantially more energy-intensive in relative sure of vulnerability--preclude a meaning- terms.35 Similarly, food consumption is signifi- ful comparison of the two sets of outcomes. cantly higher in both emerging and developing Nonetheless, some of the differences in the countries compared to advanced economies. relative magnitudes may be driven in part Their food consumption levels (in percent of by the differences in average debt burdens household consumption) are almost three times in these countries. In many of the countries those of advanced economies. with HBS data, the share of borrowers at risk Where the price shock is transmitted is small, the average loan size is small, and through falling currencies, the economic con- the debt burdens are still sufficiently far from sequence depends on whether a country is a net critical thresholds. commodity importer. With respect to net food importing, some recent work based on an indic- ative threshold for vulnerability suggests that a E. Households and External Price Shocks few countries in ECA may be vulnerable. Soaring domestic food and fuel prices-- Economic and Welfare Impact: whether due to increasing global commod- Main Transmission Channels ity food prices or falling currencies--affects The economic and welfare impact of rising the national headline inflation based on these commodity prices depends on the intensity of commodities' share in a country's CPI. The use. One possible index is "energy intensity," relative importance of these commodities var- which is measured as energy consumption per ies across countries, though some comparisons unit of real GDP. Country averages over the are hampered by the differences in the defini- past four decades indicate that energy inten- tion of what constitutes "food" (e.g., including sity in advanced economies has fallen and both or excluding beverage, tobacco, and others) and FIGURE 2.12 Energy Intensity and Food Consumption, 1970s­2000s Energy intensity Food consumption In metric tons of oil equivalent per million of GDP In percent of household consumption in 2007 purchasing-power-parity dollars 320 50 40 240 Metric tons 30 Percent 160 20 80 10 0 0 Advanced Emerging Developing Advanced Emerging Developing economies economies economies economies economies economies 1970s 1980s 1990s 2000s Source: IMF 2008a. Chapter 2--Household Vulnerabilities 39 "fuel" (e.g., including gasoline, household utili- of the consumption basket in the world's poor- ties, and others). est countries. Fuel accounts for a much smaller Notwithstanding these measurement issues, share, although this does not capture the sec- these commodities typically account for a large ondary dimensions, that is, the use of fuel as share of the consumption basket, particularly input into the production of other items in the in poorer countries. Food represents about consumption basket. Evidence in some coun- 10 percent of the consumption basket in richer tries suggests that taken all together, both countries, while representing up to 80 percent the primary and secondary use of fuel could FIGURE 2.13 Food and Fuel Share of the CPI Basket 60 40 Percent 20 0 Armenia Bulgaria Croatia Czech Georgia Hungary Kyrgyz Lithuania Poland Romania Ukraine Middle- Rep. Rep. income CIS Food Fuel Sources: Kotaro Ishi and RES. FIGURE 2.14 Food and Fuel Imports, 2006 In percent of GDP 20 15 Percent 10 5 0 Bulgaria Croatia Czech Rep. Estonia Hungary Latvia Lithuania Poland Romania Slovak Rep. Armenia Azerbaijan Georgia Kyrgyz Rep. Moldova Belarus Kazakhstan Russian Fed. Ukraine Albania BiH FYRM Turkey EU10 (incl. Croatia) Low-income CIS Middle-income CIS Western Balkans Food Fuel Sources: Habermeier et al. 2009 and IMF 2008b. 40 The Crisis Hits Home--Stress-Testing Households in Europe and Central Asia account for about double its typical share of the Governments may prevent a less-than-full pass consumption basket.36 through of higher prices to consumers, through These aggregate economic effects hide lower fuel taxes or higher subsidies. Some coun- some of the likely distributional consequences tries use trade policy to address external price of commodity price shocks within countries. shocks. In 2008, for example, many countries First, within a given country, the food share enacted more restrictive food trade policy-- in the CPI consumption basket may under- such as through quantitative export restric- state the relative weight of food consumption tions and taxes on selected commodities--with among poorer households. Second, the poverty the stated objective of protecting food secu- and social consequences may vary depending rity and curbing price increases. Though this on the geographic location of the household may dampen overall price increases, the policy and depending on whether households are net redistributes income away from net food sell- food buyers or net food sellers. It may also ers to net food buyers. Where net food sellers matter whether they rely exclusively on food are mostly poorer, agricultural households, the purchases for food consumption and, if they policy impact can be regressive. do, whether they have access to cheaper sub- stitutes. The rural poor are thought to be more Regional Overview self-sufficient, able to produce food for own- The rise in food and fuel prices through 2008 consumption, compared to the urban poor. has had adverse economic effects on countries With respect to energy consumption, the poor in the region. For oil-importing countries, are also thought to have access to less expensive increases in energy prices led to widening trade sources of energy (though they are probably imbalances and escalating inflationary pres- dirtier sources and pose both environmental sures.38 Higher energy prices also led to higher and health risks). unit costs, which are perceived to undermine The indirect effects can also be large. Fuel competitiveness unless accompanied by pro- price increases can have a direct poverty effect, ductivity enhancements. The poverty impact through the household consumption of energy. of further increases in food and fuel prices The indirect effects on other products con- depends, as previously stated, on their rela- sumed by the households--using fuel as an tive importance in the consumption basket of intermediate input--can also be substantial. households. Such shares of total consumption Studies of non-ECA countries suggest that fuel are observed to vary across countries and geo- price increases can have net effects that are graphic locations and by household income, progressive (and thus have much larger welfare among other dimensions. consequences for urban and richer households This section draws from the ECA Household compared to poorer, rural households) mainly Data Archives to assess household welfare through their indirect effects.37 The indirect consequences in case of a new round of price effects through income can also be substan- shocks. The figures below, for example, report tial. For households with members who are the share of food and energy in total consump- wage-employed in the agriculture sector, the tion by household quintiles in selected ECA increase in earnings may partially offset the countries (figure 2.15 and figure 2.16). There welfare consequences of rising food prices. are a few notable observations: Policy responses, in turn, may mute the First, as expected, richer countries have effects of global commodity price shocks on lower food shares. For example, food represents domestic prices, though they may incur large about a third of total consumption among the fiscal costs or redistribute income regressively. new EU member countries, on average. Among Chapter 2--Household Vulnerabilities 41 FIGURE 2.15 Food Shares of Consumption In percent of total consumption, by quintiles of household consumption EU10 (incl. Croatia) Low-income CIS 100 100 75 75 Percent Percent 50 50 25 25 0 0 Bulgaria Croatia Estonia Hungary Latvia Lithuania Poland Armenia Azerbaijan Georgia Kyrgyz Rep. Moldova Tajikistan Uzbekistan Middle-income CIS Western Balkans 100 100 75 75 Percent Percent 50 50 25 25 0 0 Belarus Kazakhstan Ukraine Albania Bosnia and Macedonia, Montenegro Serbia Herzegovina FYR Q1 Q2 Q3 Q4 Q5 Source: Staff calculations. low-income CIS countries, food accounts for 5). In Moldova, for example, the food share close to two-thirds of total consumption. In of rural household consumption is close to fact, across sub-regional groups, the average 50 percent; among urban households, it is about share of food consumption tracks the level of a third. In Poland, food accounts for about development fairly well. There are of course 50 percent of consumption among households some exceptions. Ukraine, for example, which in the lowest quintile; among households in ranks as a middle-income country, has food the lowest quintile, the food share of consump- shares that are more or less comparable with tion is a little more than a fifth. some of the low-income CIS countries.39 Third, the energy shares of consumption do Second, within countries, there are sub- not reflect clear patterns along country, income, stantial variations across geographic space and or geographic lines. In part, this may be due to by household income (appendix tables 4 and the use of an imperfect proxy for energy share 42 The Crisis Hits Home--Stress-Testing Households in Europe and Central Asia FIGURE 2.16 Utility/Energy Shares of Consumption In percent of total consumption, by quintiles of household consumption EU10 (incl. Croatia) Low-income CIS 30 30 20 20 Percent Percent 10 10 0 0 Bulgaria Croatia Estonia Hungary Latvia Lithuania Poland Armenia Azerbaijan Georgia Kyrgyz Rep. Moldova Tajikistan Uzbekistan Middle-income CIS Western Balkans 30 30 20 20 Percent Percent 10 10 0 0 Belarus Kazakhstan Ukraine Albania Bosnia and Macedonia, Montenegro Serbia Herzegovina FYR Q1 Q2 Q3 Q4 Q5 Source: Staff calculations. of consumption. If the comparable consump- households are somewhat higher compared to tion aggregates from the ECA Household Data those of rural households. Furthermore, the Archives are used, the consumption component expenditure shares could reflect non-payment that comes closest to energy is the "utility" or payment arrears. They could also possibly share of household consumption, consisting reflect access to less expensive sources of fuel of expenditures on electricity, heat, gas, water, and energy. and sewerage. In addition, the state or quality Aggregate energy shares of consumption also of infrastructure matters as well. Where there likely mask important variations across energy is insufficient utility infrastructure, many sources. Connection rates, intensity of con- households may not be connected to central sumption, and payment behavior, among others, sources of energy. In fact, in many countries, are likely to vary across various energy sources, the utility shares of consumption among urban such as central heat, electricity, natural gas, and Chapter 2--Household Vulnerabilities 43 other sources of fuel. In Moldova, for example, needed to ensure that the households are as households connected to central heating units well off as they were prior to the price change are mostly urban households. In contrast, there or that household utility is kept constant. In is almost universal electricity connection.40 the case of the Laspeyres index, this is con- Some patterns in the utility shares of con- sistent with the items and quantities of con- sumption may also reflect country-specific pol- sumption remaining unchanged, with under- icies on utility tariffs, which can be lower than lying preferences characterized by a Leontief or equal to the relevant cost-recovery levels utility function (i.e., no substitution). In the depending on whether utility reform programs case of the geometric index of relative price have been completed. In fact, in countries change, it reflects Cobb-Douglas preferences known for energy tariffs that have not been and allows for substitution away from rela- completely adjusted to full cost-recovery, the tively more expensive goods but keeps house- utility shares of consumption are relatively low. hold utility constant. The utility shares of consumption in Belarus The welfare effects are of course propor- and Ukraine, for example, are about half or tional to the budget share of food consump- even a third of utility shares among some of tion, by construction. The results are in the the new EU member states. appendix (appendix tables 6 and 7). Across all An analysis conducted by the World Bank in countries in the region, the poor bear a greater the middle of the crisis suggested that the wel- burden of the welfare impact of a food price fare impacts of last year's food and fuel price increase. The differences along geographic increases were possibly very large. The study lines are also clear: rural households are hit found that for some ECA countries, a 5 percent harder compared to urban households. With relative increase in food prices could worsen respect to a 10 percent fuel price increase, poverty ratesby up to 3 percentage points.41 the distributional consequences are less clear. One simple numerical exercise for assess- In many countries, urban households and the ing the welfare impact of illustrative food (or more affluent households bear a greater burden energy price) increases would be to calculate of the welfare impact; this, however, is not true the fall in real income associated with these everywhere. Where there is some possibility of increases. This follows previous studies of fuel substitution, the welfare impact is smaller; this and food price increases, including selected is true for both food and fuel price increases. ECA countries.42 The calculations are made using two plausible scenarios. One scenario Country Illustrations assumes no substitution while another scenario Within countries, there is likely to be sub- allows for a limited degree of substitution. The stantial heterogeneity in the welfare impact of distribution of such welfare effects across all a price shock. The regional simulations above households can then be analyzed. ignore a number of household specificities that Previous studies thus essentially calculate determine the net poverty impact of food or two indexes of price change.43 One index, energy price increases. As previously stated, the Laspeyres price index, assumes that con- the welfare impact of rising food price depends, sumption quantities in the baseline period are in part, on whether households are net buyers fi xed. It does not allow for substitution toward or net sellers of food. The unfavorable conse- cheaper alternatives. The other price index, the quences of rising prices can also be offset by geometric price index, allows for some degree rising real transfers or real wage increases. of substitution. An alternative way to interpret This section illustrates some of these dimen- these calculations would be as compensations sions drawing from countries where the food 44 The Crisis Hits Home--Stress-Testing Households in Europe and Central Asia or fuel price increases that have taken place are income. However, net consumers still represent among the sharpest in the region. the majority of households in the poorest quin- In the Kyrgyz Republic, for example, the tile. More generally, for the country as a whole, food price increases are estimated to have had 53 percent of the population lives in house- substantial welfare effects on net consumers.44 holds characterized as net food consumers. Of Although the share of food in total consump- these net consumers, 35 percent are poor. As a tion falls with income (and is thus highest result, about 19 percent of the population are among the poor, as we saw from the regional net food consumers as well as living in poverty, overview in the previous section), the share and were estimated to have been hurt the most of net food consumers generally rises with by the food price increases in 2007. FIGURE 2.17 Kyrgyz Republic: Net Food Consumers and Net Food Producers In percent of households 100 75 Percent 50 25 0 Q1 Q2 Q3 Q4 Q5 Net consumers Net producers Source: Staff calculations. FIGURE 2.18 Kyrgyz Republic: Estimated Poverty Impact of the Food Crisis In percent of the populations 25 20 15 Percent 10 5 0 Wage +10% Wage +20% Wage +30% 2006 2007 Net consumers Net producers All Source: Staff calculations. Chapter 2--Household Vulnerabilities 45 At the same time, the welfare consequences equivalent to about 400,000 people falling of rising food prices were partially offset by into extreme poverty because of rising food rising wages in the Kyrgyz Republic. Not sur- prices. Urban poverty was also estimated to prisingly, the estimated net impact on abso- increase. This is not surprising considering lute poverty (with the poverty headcount at that about 30 percent of all net consumers live 40 percent in 2006) was inconclusive. Extreme in the capital. poverty, however, was estimated to increase In Tajikistan and Albania, the net pov- by up to 8 percentage points (from 9 percent erty impact is mediated by access to agricul- in 2006), depending on the degree of wage ture inputs, assets, and livelihood strategies. increase. This percentage point increase is In a recent study of 11 countries, including FIGURE 2.19 Albania: Welfare Impact by Livelihood Bubbles represent the size of the population affected 3 2 1 Percent 0 -1 -2 -3 1 2 3 4 5 Expenditure quintiles Other Agriculture Source: Zezza et al. 2008. FIGURE 2.20 Tajikistan: Welfare Impact by Livelihood Bubbles represent the size of the population affected 2 1 0 Percent -1 -2 -3 -4 1 2 3 4 5 Expenditure quintiles Other Agriculture Source: Zezza et al. 2008. 46 The Crisis Hits Home--Stress-Testing Households in Europe and Central Asia Tajikistan and Albania, the authors found that and private consumption projections and (ii) the poorest households were likely to be hurt poverty projections calculated from more the most by food price shocks (figure 2.19 and recent GDP and private consumption projec- figure 2.20).45 They did find some evidence that tions (released in January 2009 and April 2009). some of the poor households manage to ben- The difference between (i) and (ii) can be con- efit from rising prices of basic commodities, sidered a measure of the poverty impact of the depending on whether these households have global crisis. sufficient access to agriculture inputs, or assets This section uses household survey data and such as land. macroeconomic growth projections. Household Despite the moderating effects of livelihood survey data are drawn from the ECA Household and assets, the poor were nonetheless found Data Archives. We have survey data for 25 of to be the most likely to have been hit hard- the 29 countries in the ECA region, represent- est by the price shock. This was true for all ing 95 percent of the region's total population. countries, irrespective of location within each Economic growth projections are drawn from country. This is not completely surprising, as the IMF's WEO database.47 Growth projec- the poorest households also have the weakest tions at the country level are available for the access to agriculture inputs and assets. They pre-crisis period (April 2008) and for April also tend to be relatively less educated, with 2009. The January 2009 WEO growth projec- relatively lower productivity and limited capa- tions are available at the country level for some bility to take advantage of profitable activities ECA countries (such as Russia); however, for in agriculture. most ECA countries, only sub-regional aver- ages were made publicly available. These sub- regional averages were used to impute growth F. Households and Income Shocks projections at the country level. Finally, we use historical data (covering the three years pre- Data and Methodology ceding the crisis) for each country to estimate This section follows the methodology under- the ratio between private consumption growth pinning recent simulations of the poverty and GDP growth. On average, it is equal to one implications of economic growth projections.46 but there is some slight variation from country In brief, this section uses country-level pro- to country. (This assumption is discussed more jections for GDP and private consumption fully below.) The underlying country data are through 2010. Following existing exercises, we reported in appendix table 8. first assume that changes in GDP or income This numerical exercise uses a poverty line will not be passed on fully to private consump- of $PPP 2.50 per person per day and a poverty/ tion. The impact will depend on the relation- vulnerability line of $PPP 5.00 per person per ship between private consumption growth and day. This follows the most recent round of the GDP growth in each country. Next, the pro- International Comparison Program on pur- jected growth in private consumption is then chasing power parities (PPPs). Although the used to predict per capita household consump- World Bank recommends a new international tion in each country, using the household sur- poverty line of $PPP 1.25 per person, the $PPP vey data from the latest available year. Finally, 2.50 and $PPP 5.00 poverty/vulnerability lines the predicted household consumption is com- have been found to be more relevant for ECA, to pared to the relevant poverty line. take into account the higher cost of living asso- The exercise yields two sets of results: (i) ciated with the region's colder climate and the poverty projections using the pre-crisis GDP conditions in many middle-income countries Chapter 2--Household Vulnerabilities 47 throughout the region.48 In addition, this fol- losses. Similar results were found in an analysis lows the 2005 regional study that adjusted the of Russian households over this same period: region's international poverty line(s) to reflect The more children are present in a house- similar considerations.49 hold, the greater this household's poverty risk and the lower its ability to smooth consump- Important Caveats50 tion. In addition, households with higher ini- tial incomes and more assets were found to be These regional simulations are based on several better able to protect their welfare during and strong assumptions, which are explained further below. Although these assumptions are defensi- after a macroeconomic shock, in large part due ble, there is a large, inconclusive literature on, to their broader menu options of coping with among other things, whether these assumptions the shock, such as by drawing from their sav- hold on average, whether they hold linearly or ings or asset liquidation. non-linearly (depending on, for example, the The exercise does not assume a unitary pass- level financial sector development), and whether through from GDP growth to private consump- these assumptions hold depending on whether it tion growth as the economy contracts.52 These is a "normal" period or a crisis period, and, dur- simulations are based on the estimated ratio of ing a crisis period, whether it is a financial crisis private consumption growth to GDP growth or a different kind of economic crisis. using non-crisis data at the country level, which These simulations ignore the distributional may be higher or lower than one, depending on consequences of the crisis. The exercise assumes actual country outcomes. However, there is no that the growth in per capita household con- reason to think that this recent relationship will sumption is distribution neutral; that is, house- continue to hold over the crisis period. holds in every part of the income distribution are A large literature has emerged on the rela- all affected uniformly by the average growth or tionship between consumption growth and decline in consumption. On one hand, previous GDP growth. In particular, the literature has research suggests that there has been no change explored what is typically referred to as "excess in inequality, on average, during economic con- private consumption volatility" (relative to tractions, thus lending some support to this GDP or income growth volatility) and its pos- assumption. On the other hand, there may be sible drivers, including fi nancial intermedia- disproportional effects on the poor depending tion.53 The literature postulates that financial on the relative exposure of households to the deepening curbs consumption volatility, as economic shock. In fact, despite the zero change financial markets promote risk sharing and in inequality on average, substantial variations allow households to smooth consumption. exist from country to country. Some empirical studies have provided some A review of ECA's experience with previ- supporting evidence, but they have also found ous crises suggests that, in fact, certain types this relationship to be nonlinear, requiring of households tend to be more vulnerable dur- some sufficient level of financial deepening ing crisis periods.51 For example, an analysis before yielding any measurable relationship of vulnerability in Moldova reveals that the between consumption smoothing and fi nan- consumption of larger households experienced cial intermediation.54 In addition, some recent larger drops after the Russian fi nancial cri- studies of the fall in consumption during fi nan- sis; moreover, controlling for household size cial crises suggest that financial intermediation (among other individual and household charac- may exacerbate the decline in consumption, as teristics), the number of children in the house- consumption becomes more sensitive to the hold was associated with larger consumption availability of bank credit.55 48 The Crisis Hits Home--Stress-Testing Households in Europe and Central Asia The regional simulations based on GDP capital formation, the intergenerational trans- shocks abstract from the potential household mission of poverty, and the sustainability of income shocks arising from reduced remittances long-term economic growth. due to the impact of the crisis in migrant host The results presented below for the region countries. As highlighted in the previous chapter, as a whole should thus be interpreted with cau- many countries in ECA are heavily dependent tion. They should be treated as illustrative, tak- upon remittances as a source of foreign finance ing into account many of the limitations of the at the macro level. At the household level, such calculations behind them. In some cases, the inflows can be important sources of funds for poverty impact may be understated, such as in consumption expenditures, health and educa- the event that consumption substantially lags tion, and investment. For example, in Tajikistan behind GDP growth in 2010. In other cases, the in 2007 around 60 percent of the yearly con- poverty impact may be overstated, where there sumption of the median household was financed are opportunities for consumption-smoothing by remittances (World Bank 2009c). As outlined and offsetting income shocks. in the country studies detailed below, it is thus important to tailor the specific income shocks to country characteristics. Regional Overview: Main Results These simulations also ignore the many The results suggest that poverty will rise. By non-monetary and nonincome dimensions of 2010, the poverty headcount (using the $PPP poverty likely resulting from the crisis. As has 2.5 poverty line) for the region as a whole is been seen in previous crises, the social conse- expected to be about 2.4 percentage points quences of crises can be significant, as house- higher than it would have been, relative to holds cut back on their health care spending, baseline projections of income or GDP growth pull children out of school, and curb many (figure 2.21). The share of the poor or vulner- other essential expenditures.56 Such social con- able population also rises by 7.2 percentage sequences may have immediate consequences, points. In 2009, these shares rise by 1.6 and as well as longer-term implications for human 5.9 percentage points, respectively. In absolute FIGURE 2.21 The Impact of the Crisis on Poverty and Vulnerability in the ECA Region Poverty and vulnerability projections 2007­2010, in percent of the population $PPP 2.5 a day $PPP 5.00 a day 10 40 8.0 33.5 33.3 8 7.8 7.6 32.2 31.1 Percent Percent 7.1 7.1 30 30.8 6.9 30.5 6.8 29.4 6 6.2 27.4 25.0 5.2 4 20 2007 2008 2009 2010 2007 2008 2009 2010 Pre-crisis projections January projections Latest projections Source: Staff calculations. Chapter 2--Household Vulnerabilities 49 terms, an additional 11 million people will fall differences across sub-regions reflect the sig- into poverty by 2010. An additional 23 million nificant revisions to GDP growth projections people will find themselves vulnerable because (figure 2.22). Many of the middle-income CIS of the crisis.57 In 2009, 7 million more people countries, for example, have experienced sub- are in poverty and 28 million more are either stantial downward revisions in their economic poor or vulnerable. growth prospects between April 2008 and April Turkey and the middle-income CIS coun- 2009, with growth prospects switching from tries are driving the percentage point increases expansion to recession. Russia's 2009 growth in poverty and vulnerability, followed closely projections, for example, fell from +6.8 to ­6.5. by the EU10 plus Croatia. Relative to baseline Ukraine's growth prospects in 2009 declined projections, the share of the poor and vulnera- from +4.9 to ­7.3. In contrast, a number of low- ble population rises by over 8 percentage points, income CIS countries still expect their econo- on average, in these countries. The relative mies to expand by modest amounts in 2009. FIGURE 2.22 The Impact of the Crisis on Poverty and Vulnerability in the ECA Region: Sub-Regional Results Poverty and vulnerability projections 2007­10, in percent of the population EU Western Balkans $PPP 5.00 a day $PPP 5.00 a day 25 35 23.6 32.7 22.3 21.3 22.3 32.3 30.5 Percent Percent 30.4 30.2 20.9 20.9 20 30 19.8 29.6 29.5 28.1 19.0 28.1 17.0 25.8 15 25 2007 2008 2009 2010 2007 2008 2009 2010 Low-income CIS Middle-income CIS $PPP 5.00 a day $PPP 5.00 a day 80 25 74.3 21.7 70.2 22.3 70 68.7 20 19.1 21.2 66.9 Percent Percent 68.8 18.9 18.5 64.6 17.0 63.2 60 15 59.4 15.4 13.1 50 10 2007 2008 2009 2010 2007 2008 2009 2010 Baseline projections January 2009 projections Latest projections Source: Staff calculations. 50 The Crisis Hits Home--Stress-Testing Households in Europe and Central Asia These include Azerbaijan (+1.7), Georgia (+1.0), for foreign labor in Russia and Kazakhstan. As Tajikistan (+1.4), and Uzbekistan (+4.9), though mentioned in Chapter 1, the growth of remit- all their GDP growth numbers have been tance outflows from Russia to CIS countries revised downward as well. At the lowest end contracted in Q4 2008, tracking the declines in of these adjustments is Uzbekistan, with 2009 Russian construction activity. The latest figures GDP growth numbers essentially unchanged indicate that the dollar value of recorded remit- from the pre-crisis period to April 2009. tances to Tajikistan (via money transfers) fell 36 percent year-on-year in the first five months Some Sensitivity Tests of 2009, with similar contractions seen in the heavily remittance-dependent economies of The core simulations are based on the histori- Georgia and Moldova. In Tajikistan the results cal relationship between GDP growth and pri- of simulating the impact of up to a 50 percent vate consumption. We re-ran the simulations fall in remittances suggest that the poverty above to allow for a unitary pass-through from headcount can rise by up to 7 percentage points, GDP growth to private consumption (and thus assuming households do not adjust (e.g., return- to household consumption). The key results are ing migrants do not find jobs in local labor mar- essentially unchanged--both for the region as kets), or by 3.5 percentage points, allowing for a whole and for the sub-regions. Selected fig- some adjustment on the part of households. ures are presented in appendix figure 1. In Armenia,60 the poverty headcount is esti- The simulations were also re-calculated using mated to increase by over 5 percentage points the elasticity of poverty to growth in consump- between 2008 and 2010, primarily driven by tion per capita calculated for ECA countries shocks transmitted through labor market chan- over the period 1998­2003.58 Because the period nels. Many will find themselves out of work refers to the economic recovery period follow- or earning substantially less than in previous ing the Russian crisis, the poverty elasticity years, particularly in sectors such as the con- estimates may not be appropriate. Nonetheless, struction and mining sectors. In addition, the the results are essentially unchanged. impact on remittance-dependent households could be large. The results of recent analysis Country Illustrations suggest that the poverty headcount may rise The regional overview masks the likely het- from 18 percent to 27 percent in households erogeneity of impact within countries. This that receive remittances from sources other regional analysis should not be a substitute for than immediate family members (a peculiar country-specific poverty analysis. As previously feature of household well-being in Armenia is discussed, the results of the preceding numeri- the large share of remittances from non-imme- cal exercises ignore many country specifici- diate family members). ties and the likely concentration of vulnerable In Bulgaria,61 the economic slowdown and households, depending on the nature of the the fall in remittances is estimated to lead to a economic downturn. In addition, while some 1.2 percentage point increase in poverty. economies are still projected to grow by mod- Although more modest than the expected impact est amounts, the welfare risks are not shared in other countries, the labor market is again an uniformly across all households. important transmission channel, particularly Tajikistan is a case in point.59 While the among workers in the construction sector. country as a whole is still projected to grow Declining remittances underpin about a quar- by a very modest amount through 2009, many ter of the overall poverty increase, and a much households are vulnerable to falling demand larger share of the rise in extreme poverty. Chapter 2--Household Vulnerabilities 51 In Russia,62 the results of recent simula- to rural households, households with children tions suggest that the poverty headcount in and pensioners are at the highest risk of falling rural areas will likely rise by over 5 percentage into poverty. The growth in unemployment points. The simulations suggest that in addition levels will also drive increases in poverty. CHAPTER 3 Coping with the Crisis A. Introduction in the policy response to the crisis, drawing from some recent analyses of fiscal space and The resilience of households to macroeconomic the availability of fiscal resources and a recent shocks ultimately depends upon the economy's assessment of the performance of social protec- institutional readiness, the flexibility of the tion programs in ECA countries. Finally, it will economic policy regime, and the ability of provide examples of possible policy responses the population to adjust. Policy and institu- to mitigate the impact of the crisis, by type tional preparedness is essential so that coun- of shock to households. The options for pol- tries can manage the adverse social impacts of icy responses covered here are by no means macroeconomic shocks. This requires ex ante exhaustive. They are discussed below mainly analysis of risks, a good understanding of their for illustrative purposes. possible transmission channels if triggered, and their possible impacts on households; approaches that ensure that the state does not B. Household Responses: Lessons intervene excessively in terms of detrimental from the ECA Experience longer-term distortions to incentives or fis- cal sustainability; and a comprehensive social Over the transition period, a growing literature safety net system that provides for countercy- has documented patterns of self-insurance, clical and scalable interventions. informal insurance, and informal risk pooling This chapter looks at how the impact of the in ECA. They have also chronicled household various shocks arising from the crisis on house- strategies for coping with economic shocks, hold welfare may be offset by households' own including borrowing, migration, substitution coping strategies and by social safety net sys- of consumption toward less expensive goods, tems. It will also assess some key constraints and engaging in risky or illegal activities. Such 53 54 The Crisis Hits Home--Stress-Testing Households in Europe and Central Asia strategies and risk mitigation mechanisms may Migration. As local employment opportu- be disproportionately concentrated among cer- nities decline, one option would be to relocate tain groups of households, depending on their to a region with better labor market conditions. region of residence, income, and social capital. Labor migration can be internal or interna- Households in ECA have employed a vari- tional, and it can involve either the whole house- ety of coping strategies to smooth consump- hold or only some of its working-age members. tion during previous crisis periods. Some of The necessary conditions for the effectiveness these strategies correspond to what is referred of this coping strategy are labor market flexibil- to in the literature as "risk management."63 ity and the ability and willingness of workers to These include household members holding move to locations with jobs plus a host region jobs with uncorrelated risks, either domesti- likely to provide sufficient labor income gains to cally or through migration of family members offset migration costs. In the years following the to foreign countries. Other strategies, such Russian crisis, for example, many Moldovans as dissaving, including asset liquidation, and migrated abroad in search of employment borrowing where possible, smooth consump- opportunities and better living conditions. tion over time and still other strategies share Studies of household behavior during the early risks across households (inter-household pri- transition period in Kazakhstan also found that vate transfers). Generally, the empirical litera- migration responded almost immediately to ture rejects the existence of full consumption movements in relative exchange rates and to sys- insurance, whereby temporary income fluctua- temic crises, such as the 1998 Russian financial tions have no effect on consumption, but does crisis. However, wage differentials and differ- provide some support to partial consumption ences in construction activity have been found smoothing, whereby consumption changes are to take a longer time in influencing migration smaller than income changes. This implies patterns. A recent study of Turkey found that that, in general, households in the region were rural-to-urban migration has been the most successful, albeit only partially, in protecting important informal coping strategy, leading to their welfare through crisis times by relying increased urbanization in all regions of Turkey. on a variety of coping strategies as described in The broad-based nature of the ongoing down- more detail below.64 turn, affecting both traditional host and source Labor supply adjustments. In response to countries, is clearly an important determinant falling income due to the effects of a macro- of the effectiveness of this coping strategy in the economic shock, a household can increase its current economic environment. involvement in the labor market. Those already Subsistence Farming. Households that employed can increase their hours worked, or have access to private land plots can use them to find secondary employment. Other members supplement their income by selling the home- of the household can transition out of inactiv- produced goods or to augment their own con- ity in order to supplement household income. sumption with such home production. Studies Existing studies of Bulgaria, Russia, and Turkey of household behavior in Bulgaria, Russia, and indicate that, in general, labor supply adjust- Turkey during previous crises observe that ments have not allowed households in the ECA subsistence farming has often been employed region to preserve their pre-crisis welfare lev- by households to supplement their food con- els. The proportion of households that found sumption. However, it has also been found to secondary employment is small and the earn- be an ineffective coping strategy to address ings were generally insufficient to compensate vulnerabilities in non-food consumption, or in for the loss of income from primary jobs. lifting households out of poverty. Chapter 3--Coping with the Crisis 55 Dissaving/Borrowing and Asset Liquid- household will be able to use this strategy effec- ation. By saving a portion of its income flow in tively, since a widespread shock would, by defi- good times and dissaving (or borrowing) in bad nition, affect most people in that household's times, a household can minimize consumption social network. Moreover, some marginalized variability in the presence of income fluctua- groups, such as ethnic minorities, are often tions. The same principle applies to the accu- excluded from the informal support networks mulation of assets (such as housing, durables, available to other households. production equipment) in good times and the Compared to previous crises, the scope for sale of such assets in bad times. Of course, the households to engage in their traditional cop- usefulness of this strategy depends on house- ing strategies may be more limited. During holds' access to credit markets and their abil- previous crises, households found secondary ity to dispose of their assets after a macroeco- employment, relied on transfers from friends nomic shock. For example, many households and families, or left for work abroad to augment in Turkey turned to selling their assets during family income. Because of the global nature of the 2001 crisis. About 20 percent of households the crisis, and because macroeconomic shocks reported selling assets and valuables. However, are hitting households on multiple fronts, these it has also been suggested that proceeds from coping strategies may no longer be feasible. such asset sales were meager due to insuffi- For the poorest households, subsistence cient demand for assets held by the poor. This farming may be a viable strategy, though evi- is similar to the phenomenon of asset "fire- dence from the recent food price shock sug- sales" in financial markets potentially yielding gests that many of the poorest households prices below fundamental values (and a wealth do not have access to agricultural assets and redistribution from net sellers to buyers). In inputs. For some, transitions into informal sec- addition, as highlighted in the discussion of tor employment may be possible though, for household indebtedness, household gross asset many households, earnings from informal sec- positions may be associated with financial lia- tor activity will likely be insufficient to offset bilities such as mortgages, thus limiting the the poverty impact of the crisis. potential net gains from such sales (or even leading to potential net losses as in the case of negative mortgage equity). C. Context: Policy Response, Government Private transfers. Besides turning to the Resources, and Constraints government for assistance, households suffer- ing from the impacts of an economic shock can Overall Fiscal Envelope turn for help to friends and family. If the shock The ability of many governments in ECA to does not affect households equally--that is, respond to the crisis--such as by increasing more technically, if the idiosyncratic compo- social transfers--is generally constrained by nent of the economic shock is large compared to rising government deficits. Between 2007 and the covariate component--then households can 2009, on average, deficits in percent of GDP are reduce consumption volatility through inter- projected to rise by about 3 percentage points. household transfers, flowing from less affected There are marked variations across countries, households to more affected households. In with Estonia, Montenegro, and Russia at the ECA, private transfers have been found to be higher end of this distribution, while Belarus, an important buffer against household income Hungary, and Georgia are at the lower end of the and consumption volatility. However, macro- distribution. Nonetheless, for many countries, economic crises decrease the likelihood that a fiscal policy responses to the crisis will likely be 56 The Crisis Hits Home--Stress-Testing Households in Europe and Central Asia muted by rising deficits that have become much regard to relative fiscal constraints, ECA is in more difficult to finance. It would be essential a better position. However, there is still poten- to first determine the overall fiscal adjustment tially a large shortfall in education and health warranted for macroeconomic stability and spending among countries with little fiscal debt sustainability, taking into account initial capacity and it is not likely to be financed by conditions and the likely impact of the crisis on donor resources. public finances.65 Economies that experienced Where there are no new official or alterna- strong initial fiscal and external positions are tive sources of financing, or where there is lit- likely to have more room for expansionary fiscal tle scope to mobilize revenues, some countries policy and can afford a fiscal stimulus package, may resort to across-the-board cuts in public while those with weaker initial positions may spending.67 Although social safety nets will be require substantial fiscal adjustment. Other among those items likely to be cut as revenues features of a country's macroeconomic policies fall, protecting these programs--and possibly are also important. For example, countries with expanding some of them, where some realloca- fi xed exchange rates will have to depend more tion of resources is possible--will be an impor- on fiscal policy, rather than monetary policy, tant element in the response to the crisis. for adjustment. The prioritization of labor-using invest- A recent analysis in fact suggests that there ment expenditures--either from countries' are likely large shortfalls in education and own budgets or from resources provided by health spending worldwide due to the growth donors--could be one option for addressing the slowdown.66 ECA countries, as a group, will labor market consequences of the crisis while require the largest outlays, compared with accounting for constrained fiscal resources. other regions, to protect their planned expen- Such investments could include rural roads ditures in education and health services. With projects or irrigation systems rehabilitation FIGURE 3.1 General Government Balances in ECA, 2007 and 2009 In percent of GDP 8 4 Percent 0 -4 -8 Montenegro Bulgaria Croatia Czech Rep. Estonia Hungary Latvia Lithuania Poland Romania Slovak Rep. Slovenia Armenia Azerbaijan Georgia Kyrgyz Rep. Moldova Tajikistan Turkmenistan Uzbekistan Belarus Russian Fed. Kazakhstan Ukraine Albania Bosnia FYR Macedonia Serbia Turkey Middle- EU10 (incl. Croatia) Low-income CIS income CIS Western Balkans 2007 2009 Source: IMF WEO database (April 2009). Chapter 3--Coping with the Crisis 57 projects that can create short-term employ- introducing new programs as appropriate. ment opportunities while creating the condi- The experiences of other countries suggest tions for longer-term growth. Improving the that programs such as conditional cash trans- efficiency of public spending may also create fers (CCTs),69 workfare schemes, and public some additional fiscal space. works programs can be effective instruments for protecting the vulnerable from immediate Existing Social Protection Systems68 as well as longer-term (second-round) conse- quences of transitory shocks on nonincome Countries in the region operate a combina- dimensions of welfare, including human capi- tion of safety net programs. The programs are tal accumulation. typically in the form of cash transfers with an emphasis on family allowances (such as child allowances), social pensions, heating and hous- ing allowances, and targeted anti-poverty pro- D. Immediate Policy Responses: grams. Some countries in ECA have yet to Some Illustrations reform a range of categorical benefits and sub- This section provides illustrations of potential sidies left over from the pre-transition period. policy responses and instruments for mitigat- Across countries, multiple programs exist, ing the poverty and social impact of the crisis. leading to the fragmentation and duplication The examples are organized by type of shock of benefits. to households. The treatment of the topic is not The region's social protection systems cur- comprehensive and the options listed are by no rently vary in size and targeting performance means exhaustive. The options are instead dis- across countries. The results of comparing the cussed below for illustrative purposes. targeting performance of selected social assis- Across various policy instruments and tance benefits across countries suggest that social protection programs, there are a num- Lithuania, Ukraine, and Turkey are among ber of important considerations. The primary the countries with the bulk of social benefits consideration would of course be the appro- reaching the poorest households. At the lower priate role for the government--whether the end are countries where only 40 percent of government responds to the crisis through the social benefits reach the poorest quintile. expenditure policy, tax policy, or regula- However, most countries in the region have at tory policy. Another important consider- least one targeted safety net program that can ation would be the fiscal cost of a program, possibly be scaled-up in response to the crisis. the administrative ease with which it can be Expanding such programs can take place either implemented, and the incentives it creates. It by increasing the value of benefits they provide will also be important to consider whether a or by expanding their coverage to reach those program should have universal coverage, or households still currently outside the system. whether it should be narrowly targeted or self- However, in some countries, including Belarus, targeted. As explained below, country-specific Bosnia, Hungary, Kazakhstan, Moldova, and resources and institutions will drive the rela- Russia, the targeting performance of existing programs remains weak. tive merits of these instruments. Depending on a country's initial conditions, the response to the crisis in terms of social Credit Market Shocks assistance may involve expanding some well- In countries where households are experiencing performing programs, reforming relatively rising debt service burdens, governments may less effective interventions, or, alternatively, have to consider facilitating the restructuring 58 The Crisis Hits Home--Stress-Testing Households in Europe and Central Asia of household debt in default.70 In many ECA an interest rate shock are 0.44, 0.19, and 0.17 countries where banks currently have limited percent of GDP, respectively, equivalent to capital buffers, bank responses to rising non- about 29 percent, 6 percent, and 12 percent of performing loans have focused on extending the social assistance budget, respectively. In grace periods. However, without the certainty Hungary and Ukraine, the compensation for of a rapid economic recovery, these restructur- an exchange rate shock is 0.04 and 0.22 per- ing strategies effectively postpone problems cent of GDP, respectively, equivalent to about into the near future. This creates substantial 2 and 15 percent of the social assistance bud- risks of under-provisioning and inadequate rec- get, respectively.72 In general, these are rela- ognition of losses and thus of over-estimating tively small sums, except in Estonia. The share bank solvency. of indebted households in these countries that There is a role for governments to pro- is still not as large as those in more advanced vide incentives for proper debt restructuring. economies likely drives this. Defining the right framework is challenging On the other hand, the fiscal cost of assisting as it requires balancing competing pressures severely indebted households may be an under- on banks, households, and the government in estimate for a number of reasons. First, the a way that is fiscally affordable, creates mini- SILC-based calculations (Estonia, Hungary, mal market disruption, is socially acceptable, and Lithuania) include mortgage interest pay- and allows banks to remain solvent and able to ments only, because SILC data allows us only resume lending in the medium term. to assess the rising cost of mortgage debt ser- A template for government-assisted house- vice. Taking all of household debt service hold debt restructuring has been proposed costs together may then yield a much bigger recently.71 The authors advocate a restructur- sum. Second, the risk exposure varies by type ing program that reflects some essential fea- of household debt. In Hungary, in particu- tures including simplicity and limited scope, lar, we know from central bank data that the as well as participation on a voluntary basis, exchange rate exposure of consumer loans is among other features. They consider two gen- much larger than the exchange rate exposure of eral approaches, one involving the creation of housing loans (60 percent versus 84 percent at a legal and institutional framework that can the end of 2008). The SILC and Ukraine data underpin case-by-case debt restructuring. The are for 2007 and our calculations are based on other approach is based on some form of finan- households reporting themselves as indebted. cial assistance by the government. Between 2007 and 2008, the pool of indebted households may have expanded further. The estimated cost of providing fi nancial assistance to indebted households at risk of default is relatively modest, on average, though External Price Shocks with variations across countries. Simple calcu- Increases in the domestic prices of energy and lations can be made for the cost of compensat- food pose both short- and long-term challenges ing indebted households who are subjected to for policymakers.73 In the current environ- an interest rate shock in Estonia, Hungary, ment, the short-run inflationary pass-through and Lithuania and an exchange rate shock in of higher imported food and energy prices, Hungary and Ukraine. This is based on the including those due to exchange rate effects, assumption that all households are fully com- may be offset by falling domestic demand as pensated for the increase in the debt service economic activity weakens. However, managing burden resulting from the shock. In Estonia, inflation using appropriate policy instruments Hungary, and Lithuania, the implied costs of remains important. The full pass-through of Chapter 3--Coping with the Crisis 59 price increases to consumers avoids introduc- improving energy efficiency, and strengthen- ing distortions into productive incentives, ing social safety nets and other risk mitigation with support for vulnerable households to be systems. Finally, communication to the public provided by appropriate, well-targeted social of the policy choices adopted to address the assistance within the constraints imposed by a impact of higher energy and food prices, and country's fiscal space. A country may also opt the various trade-offs which are involved, may for a gradual phase-in of energy tariff adjust- help build broader constituencies in support of ments, where the price increases required by the adopted policies. adjustment toward full cost recovery may be too steep. Income and Employment Shocks The ECA region's previous experience with The unemployment insurance system, the main energy tariff adjustments suggests that direct tool for addressing rising unemployment, may transfers or tariff-based subsidies can play an not be sufficient to mitigate the impact of the important role in protecting poor households.74 crisis. In fact, the unemployment insurance sys- There is ongoing debate on the comparative tem itself may suffer from several limitations, merits of direct transfers and tariff-based subsi- including weak incentives for reducing welfare dies, such as a lifeline tariff or charging a lower dependency or for job search. Benefit dura- tariff for an initial minimum level of energy tions are long and can be expensive. Moreover, consumption. Opponents of lifeline tariffs sug- none of the existing programs are available for gest that they are expensive and, for at least the returning migrant workers, for example due to initial block of, for example, electricity con- a lack of work history necessary for unemploy- sumption, such lifeline tariffs subsidize both ment insurance or a lack of permanent residence poor and nonpoor consumers alike. However, required for social assistance. This suggests supporters of lifeline tariffs argue that where that it may be worthwhile to consider alterna- poverty is prevalent, where there is close to tive approaches to social protection. universal access to network energy, and where More generally, the global crisis will prob- social transfers are not well targeted, there is a ably create "the new poor," or households that case to be made for tariff-based subsidies. may be among those in higher income quin- It is also important that policy responses do tiles in the pre-crisis period but made poor by not conflict with the key longer-term reform shocks to their income flows, liabilities, or con- agenda. For example, authorities should guard sumption. These households will likely not be against reversal of efforts to lower quasi-fiscal reached immediately by existing social protec- deficits in the energy sector, which is an ongoing tion programs. challenge in many ECA countries. It will also Public works programs were an important be critical to maintain an open and transpar- component of the overall safety net package in ent trading regime. Although some countries Argentina, Mexico, Korea, and Thailand dur- within ECA adopted restrictive trade and price ing the so-called Tequila Crisis of 1995 and the controls in response to the food price increases 1997 East Asian crisis. Public workfare pro- in 2007, many of them have now been reversed. grams generally played an important role in For example, Ukraine eliminated its export mitigating the negative effects of the macroeco- quotas in June 2008 and Kazakhstan lifted its nomic crises in these countries, though its rela- export ban in September 2008. At the same time, tive role varied depending on the effects of the work needs to continue toward medium- to lon- crisis on the labor market. With the exception ger-term goals for improving the policy envi- of Mexico, income gains from public workfare ronment for agricultural productivity growth, programs were significant in these cases. For 60 The Crisis Hits Home--Stress-Testing Households in Europe and Central Asia example, in Korea the main reason for launch- when communities are fully informed about ing a workfare program was the very sharp program goals. Sixth, the term "public works" increase in unemployment, in both formal and is often mistakenly associated with "forced informal sectors, because of the crisis. The labor." Indeed, in some countries the two rate of unemployment increased from about terms have been used synonymously for years. 2 percent prior to the crisis to 9.3 percent barely Not surprisingly, workfare programs are not six months into the crisis. Not surprisingly, viewed favorably in the region. Finally, at the public works in Korea played a dominant role implementation level, it is common practice in providing immediate short-term employ- to use contractors. This creates its own prob- ment opportunities at low wages. In contrast, lems, and requires careful attention to the in Mexico the labor market impact of the crisis tendering process and even greater attention was a dramatic fall in the real incomes of work- to monitoring. ers, with relatively more limited job losses. A Given these constraints, what role can pub- cash income support program (Progresa) played lic works program play in ECA? With respect a major role in social protection, with public to the low-income countries in the region, works playing only a subsidiary role. public works program may have considerable The ECA region's experience with work- merit. These countries have significant sea- fare has been relatively limited. To date, sim- sonal shortfalls in employment (during agricul- ple, non-randomized evaluations are available tural slack seasons). In addition, given very low for only four countries: Bulgaria, the Slovak yields of main agricultural crops, productivity- Republic, Slovenia, and Poland.75 These coun- enhancing works, including the construction tries' experiences with workfare are summa- of rural infrastructure, have a role to play to rized in Box 4. promote agriculture productivity. Moreover, A broader use of workfare in the region the design features of public works programs faces a number of constraints. First, workers can be modified and adjusted relatively easily in most countries in the region have access to to suit this class of countries. unemployment insurance. Second, unemploy- In contrast, the introduction of public ment rates are quite high in some countries works programs in the lower-middle- and even during non-crisis periods. Third, most upper-middle-income countries of the region unemployed are long-term unemployed, with requires more careful design and adjustment. some looking for jobs for over two years, in For example, the central design feature of pro- contrast to many developing countries. Fourth, grams--the wage rate--needs to be set care- a uniform application of any policy across the fully, so that the level is higher than the unem- region is not feasible given tremendous diver- ployment benefit but lower than the prevailing sity within the region in terms of income, market wage rate. In order for the program to degree of urbanization, dependence on agri- be attractive to skilled and semi-skilled labor, culture or manufacturing, and so on. Fifth, in projects need to be carefully selected to accom- many countries over a long period, there has modate labor of varying skills, as was done, for been substantial emphasis on workers' rights, example, in Korea. trade unions have been strong and influential, The use of contractors to execute public and the attitude toward a downward adjust- works programs may pose a challenge, but there ment of wages has been generally hostile. are suitable alternatives. In general, countries By contrast, in more decentralized countries have followed two approaches: One option is to such as Argentina (and South Africa) there avoid using contractors at all (such as in Korea was less resistance to low wages, particularly and Argentina) and let local governments and Chapter 3--Coping with the Crisis 61 communities implement the program instead. a high unemployment rate, contractors have to Alternatively, a program can use contractors use available local labor, and only if such local under a regulatory framework, which has taken labor is unavailable are contractors legally several forms. For example, governments could allowed to bring in labor from outside a region. fi x the share of labor in each specific contract This is not easy and tendering bids may be (say 30 percent wage cost in a road repair proj- slower than otherwise. Clearly, if contractors ect) in the tendering process, and enforce it. are used, "self-selection" as a method of target- Another option would be to provide a list of ing may not be feasible and so other approaches laborers in each locality that contractors could to targeting need to be explored. hire for any specific activity. Contractors pro- In sum, the following sequential decisions vide a weekly report to the government on the need to be taken: (i) Decide on whether a pub- number and names of persons working on a lic works program is the right option. If so, specific project, and the government will then what is the balance between public works and transfer wages direct into the bank accounts of other programs already in place? (ii) When laborers. A third approach is to provide appro- public works is chosen, decide on the specific priate incentives in the tendering process: ask design aspects critical for success for the cho- contractors to specify the share of labor cost sen country. (iii) Decide on the implementa- in each project that they plan to execute, with tion modalities. (iv) Finally, institute a credible the proviso that this would be one of the selec- monitoring and evaluation system. Regardless tion criteria. The chosen bid can then be the of which program is chosen, post-crisis settings one that promises to meet quality standards require that full attention be paid to selecting and promises to use the highest labor share in the right targeting instrument to reach the total cost. This needs to be enforced and moni- intended groups, exploring possibilities for tored on a weekly or monthly basis. In all of quickly scaling up the program throughout the the above approaches, the government fi xes the country without compromising on the quality, level of the wage and contractors have to follow and considering exit strategies upfront before accordingly. Clearly, in any of the above sce- launching any program. narios, contractors do not have the freedom to offer a wage that is different from the one fi xed by the government. E. Longer-Term Policy Responses Another difficulty could be market wages Over the longer term, there are various mea- that are below the statutory minimum wage. sures for limiting the risks borne by households Because governments cannot offer a wage less as financial markets deepen. These include than the statutory minimum wage, this becomes measures affecting both the demand- and a challenge. Argentina stated in its program supply-side of household financial products. In document that what is provided is not a wage, terms of the demand-side, promoting financial technically, but compensation. However, con- literacy may help households to understand the tractors may refuse to offer a wage less than risks they expose themselves to because of their the minimum wage, even if the minimum wage consumption, employment, borrowing, and were higher than the prevailing market wage. asset portfolio choices. Addressing the supply- Finally, contractors may bring in their own side dimension potentially incorporates a whole labor gangs to work on a project, instead of host of macro-fi nancial policy measures. using local labor. To avoid this scenario, the Detailed examination of such measures is appropriate requirements need to be specified not within the scope of this report, but in its clearly in the tender. In a depressed region with discussion of the generic policy implications of 62 The Crisis Hits Home--Stress-Testing Households in Europe and Central Asia BOX 4 Public Works Programs in ECA Bulgaria. Bulgaria launched its public works program in 2002 in response to high long-term unemployment and a perceived heavy dependence of government beneficiaries on social assistance. The program aimed to promote a more "active" approach to social assistance. The benefit comprised a minimum monthly salary plus social insurance plus health insur- ance, but not unemployment insurance. Wages were set around 125 Leva (around $60) when the minimum wage was 110 Leva, but those employed had to give up social assistance benefits (40 Leva, on average, in 2002). The net wage cost of the program to the govern- ment was about 80 Leva per participant. Projects executed under the public works program included community infrastructure (construction and renovation), afforestation, and other environmentally useful tasks. Tripartite labor councils at the local and regional level selected project proposals based on bidding procedures. A training component was included, but few workers chose to participate. The long-term unemployed were given priority, and refusal to participate in workfare meant being dropped out of the social assistance beneficiary list. Several ministries are involved in its implementation. In 2003, 117,800 temporary jobs were created, and 79,400 persons were absorbed. The program was recently evaluated76 and several findings emerged. First, the program reduced unemployment spells among the long-term unemployed, thus preventing further deterioration of skills and work ethics. Second, by far the largest positive gain observed was with respect to self-confidence and job-search motivation. Third, completed projects and services were found to be of use to communities. One reason for the attractiveness of the program was that (eligible) participants received pensions, and all participants were covered by health insurance. At the same time, the program also suffered from many weaknesses. For example, only 8 percent of participants managed to get a job quickly in the private sector, mainly because of a lack of attention to training. Employers preferred those who underwent on-the-job training (only about 8 percent underwent such training). In addition, private firms and nongovernmental organizations (NGOs) that implemented public works were found to have trained their workers more compared to municipalities. The Slovak Republic introduced public works in 1992, known as "socially purposeful jobs" (SPJ) and "publicly useful jobs" (PUJ). The two programs continued with some modifica- tions through 1997­98. In the beginning, there was no emphasis on training but in 1997, both programs were reformed. Retraining was emphasized, and more attention was paid to targeting specific groups such as the disabled, long-term unemployed, young workers, and school leavers. SPJ and PUJ were then formally combined into a single program. Jobs were provided for workers of all skill levels. An assessment77 of this program found that the vast majority of workers (90 percent of the total) had a higher transition rate to a regular job after participation in the program, although a few workers (10 percent of the total) had a lower transition rate. In general, female, lower educated, and older unemployed workers have had much greater difficulty in accessing regular jobs. The exit rate from PUJ jobs was quite high, mainly due to the positive effect of the retraining provided. Chapter 3--Coping with the Crisis 63 BOX 4 (continued) Public Works Programs in ECA Slovenia also introduced public works during 1992­96, in response to rising unemployment in the early 1990s. There was heavy reliance on public works, partly because of the limited success of other programs. The program consisted of creating jobs for the unemployed under the auspices of a public or a non-profit (NGO) organization. Projects were required to provide useful services or build infrastructure of use to communities. Jobs were for a period of one year. The objectives were to help workers maintain their workforce attachment and prevent the erosion of their human capital. The contractors (selected via a tendering process) orga- nized and carried out public works as well as provided mentoring and training as needed. In comparison with programs in Hungary and Poland, Slovenian public works attracted signifi - cantly more educated participants and younger workers, and the program was more innova- tive. An assessment78 concluded that Slovenian public works participants immediately found a job upon completing the program. Poland implemented active labor market programs (ALMP) consisting of training and retrain- ing and programs to enhance human capital, along with public works. A systematic evalua- tion of Polish ALMP showed training and retraining performed well, with the post-treatment employment rates of both female and male participants observed to be higher than those who did not participate in the program. By contrast, the public works program suffered from major distortions, mainly because of "benefit churning" and lack of attention to proper calibra- tion of various benefits (unemployment compensation) and wage levels. In addition, program officials felt male heads of households deserved sustained income support, which in turn adversely affected the functioning of the public works program.79 Some important lessons emerge from the Polish experience. First, greater emphasis should have been given to the training of the long-term unemployed while they were employed in public works. Second, the careful monitoring of employer behavior should have been incor- porated into the program, to curb abuses of hiring cheap labor under the program. Third, distortions in the functioning of programs should have been minimized by carefully calibrating program benefits (wage levels) with other benefits such as unemployment compensation. household credit growth in emerging market as in sharing credit information and effective countries, the IMF highlighted five key policy enforcement of collateral. The fourth area areas that can help to limit related fi nancial relates to enhanced data availability on the risks stability risks.80 These include prudent macro- in household credit portfolios both in aggre- economic management to minimize the poten- gate and individually. Finally, as highlighted tial future likelihood of interest rate, exchange in the various trade-offs discussed throughout rate, and income shocks to households. The this report, authorities should recognize the second set of measures relates to the usage of potential impact on households of traditional macro-prudential norms, such as on loan-to- economic management policy measures, such value or debt-service-to-income ratios. The as exchange rate and interest rate adjustments. third area concerns improvements to the over- Many of these generic policy messages all legal environment and infrastructure, such have clear resonance for countries within 64 The Crisis Hits Home--Stress-Testing Households in Europe and Central Asia ECA. For example, while a number of coun- Diversified sources of economic growth will tries within the EU10 have taken measures to also be critical in helping dampen ECA coun- limit the extent of foreign currency borrow- tries' vulnerability to macroeconomic shocks. ing by households, a recent analysis found that In some ECA countries, recent growth per- they might have had only limited effects.81 In formance has been underpinned by economic particular, the analysis finds that while such activity concentrated in a few sectors, such as measures appeared to influence the extent of the housing sector, or income flows from some foreign currency credit channeled through the dominant source, such as migrant labor. domestic fi nancial system, they may have had Monitoring systems are important. the effect of shifting borrowing toward non- Guaranteeing that statistical monitoring sys- resident financial institutions. This highlights tems are in place and that relevant household the need to adopt a combination of policies, at data are collected regularly and made avail- both the macro-financial and household levels, able for analyses are important measures for to help mitigate the potential risks associated ensuring that household vulnerabilities are with increased household credit growth, for understood in a timely manner. Such monitor- example relating to foreign currency or vari- ing systems can also help identify households able interest rate loans. The exact combination at risk and ensure that they can be reached of policy measures that is appropriate to allow by a country's social protection system. The for the benefits of household credit growth monitoring of vulnerabilities could include while limiting the potential associated risks risks in household credit portfolios, as dis- will depend on country circumstances. cussed previously. Appendix TABLE A.1 Interest Rate Shock and Borrowers at Risk Percentage change in share of borrowers at risk, in percent of indebted households Interest Rate Shock 20% Threshold 30% Threshold Historical 3 pp 5 pp 6 pp Historical 3 pp 5 pp 6 pp Estonia 16.80 11.06 17.95 22.35 12.02 6.76 12.45 13.61 (0.03) (0.03) (0.03) (0.04) (0.03) (0.02) (0.03) (0.03) Hungary 5.70 7.62 11.82 12.75 3.25 4.93 8.71 10.12 (0.04) (0.05) (0.06) (0.06) (0.03) (0.04) (0.05) (0.06) Lithuania 8.39 5.50 11.28 13.34 4.58 3.73 6.95 7.52 (0.06) (0.04) (0.06) (0.07) (0.04) (0.04) (0.04) (0.05) Source: Staff calculations. Note: Absolute errors in parentheses. "pp" stands for percentage points. 65 66 The Crisis Hits Home--Stress-Testing Households in Europe and Central Asia TABLE A.2 Economic Shocks and Borrowers at Risk Percentage change in share of borrowers at risk in percent of indebted households Exchange Rate Shock Unemployment Rate Shock Probability Assignment Random Assignment Historical 25% 35% Historical 10 pp 15 pp Historical 10 pp 15 pp Estonia ... 5.51 9.02 9.05 11.64 16.40 10.86 14.31 20.98 ... (0.04) (0.05) (0.08) (0.09) (0.11) (0.10) (0.11) (0.13) Hungary 5.40 3.72 6.10 3.31 12.04 17.42 3.30 12.37 18.03 (0.04) (0.03) (0.05) (0.04) (0.07) (0.08) (0.04) (0.07) (0.08) Lithuania ... 1.89 4.31 5.76 4.29 6.70 8.68 6.64 10.16 ... (0.05) (0.10) (0.14) (0.12) (0.15) (0.17) (0.15) (0.18) Source: Staff calculations. Note: Absolute errors in parenthesis. 20% threshold is used. "pp" stands for percentage points. TABLE A.3 Economic Shocks and Borrowers at Risk Percentage change in share of borrowers at risk in percent of indebted households Exchange Rate Shock Unemployment Rate Shock Probability Assignment Random Assignment Historical 25% 35% Historical 10 pp 15 pp Historical 10 pp 15 pp Belarus 0.12 0.11 0.19 0.21 2.68 3.81 0.26 3.47 5.39 (2008) (0.00) (0.00) (0.00) (0.00) (0.01) (0.01) (0.00) (0.02) (0.02) Kazakhstan 0.13 0.18 0.36 0.46 2.99 4.56 0.69 4.52 6.97 (2005) (0.00) (0.00) (0.00) (0.00) (0.01) (0.01) (0.00) (0.01) (0.01) Serbia 0.05 0.05 0.05 0.65 2.36 3.70 0.77 2.82 4.43 (2008) (0.00) (0.00) (0.00) (0.01) (0.01) (0.02) (0.01) (0.01) (0.02) Ukraine 0.52 0.20 0.34 1.12 3.86 5.66 1.19 4.09 6.26 (2007) (0.00) (0.00) (0.00) (0.01) (0.01) (0.01) (0.01) (0.01) (0.02) Source: Staff calculations. Note: Absolute errors in parenthesis. "pp" stands for percentage points. Appendix 67 TABLE A.4 Food Share of Consumption In percent of total consumption Share of Food Location Consumption Quintiles All Rural Urban Q1 Q2 Q3 Q4 Q5 EU10 (including Croatia) Bulgaria 47 54 44 51 50 47 44 41 Croatia 25 28 22 32 27 25 22 18 Estonia 35 38 34 48 40 35 31 22 Hungary 28 31 26 36 31 28 25 18 Latvia 32 36 29 45 36 32 25 19 Lithuania 41 51 37 57 47 41 36 26 Poland 36 41 32 49 40 35 31 23 Low-income CIS Armenia 64 70 61 77 72 66 60 45 Azerbaijan 56 57 55 65 60 56 53 45 Georgia 70 80 61 78 74 73 68 57 Kyrgyz Republic 55 56 54 66 60 56 50 43 Moldova 42 48 33 56 49 42 37 25 Tajikistan 76 78 71 72 79 79 79 74 Uzbekistan 75 79 69 85 84 79 72 56 Middle-income CIS Belarus 32 34 30 35 33 32 30 27 Kazakhstan 46 52 42 55 50 47 43 35 Russian Federation 42 51 38 56 49 43 36 24 Ukraine 57 65 53 66 61 57 54 47 Turkey Turkey 36 43 32 51 40 35 30 24 Western Balkans Albania 52 55 48 59 55 54 49 43 Bosnia and Herzegovina 39 37 41 48 43 40 36 30 Macedonia, FYR 56 59 54 77 64 56 46 36 Montenegro 66 65 69 80 72 67 62 51 Serbia 40 43 37 50 43 39 36 31 Source: Staff calculations. 68 The Crisis Hits Home--Stress-Testing Households in Europe and Central Asia TABLE A.5 Utility/Energy Share of Consumption In percent of total consumption Share of Fuel Location Consumption Quintiles All Rural Urban Q1 Q2 Q3 Q4 Q5 EU10 (including Croatia) Bulgaria 19 18 20 26 20 18 17 16 Croatia 9 10 9 10 10 9 9 8 Estonia 17 13 20 18 20 19 16 15 Hungary 18 18 18 20 20 19 18 15 Latvia 12 9 14 17 14 12 10 8 Lithuania 13 10 15 13 15 15 13 11 Poland 20 17 22 18 20 20 21 21 Low-income CIS Armenia 7 6 8 7 7 7 8 7 Azerbaijan 6 7 6 6 6 6 6 7 Georgia 8 4 12 8 9 8 8 8 Kyrgyz Republic 11 11 11 9 10 11 13 10 Moldova 14 13 16 12 13 14 14 17 Tajikistan 2 2 1 3 2 2 2 1 Uzbekistan 1 1 2 2 1 1 1 2 Middle-income CIS Belarus 8 8 8 8 8 8 8 7 Kazakhstan 10 10 10 11 11 10 10 9 Russian Federation Ukraine 12 8 14 12 12 13 12 12 Turkey Turkey 12 11 12 10 12 12 13 12 Western Balkans Albania 12 11 14 12 12 12 12 14 Bosnia and Herzegovina 10 11 10 13 11 11 10 8 Macedonia, FYR 8 6 9 2 6 8 10 13 Montenegro 6 7 5 3 6 7 8 9 Serbia 13 11 14 12 13 13 13 12 Source: Staff calculations. Appendix 69 TABLE A.6 The Welfare Impact of a 10 Percent Food Price Increase Laspeyres Price Index Geometric Price Index Location Consumption Quintiles Location Consumption Quintiles Rural Urban Q1 Q5 Rural Urban Q1 Q5 EU10 (including Croatia) Bulgaria 5.38 4.37 5.06 4.08 5.26 4.26 4.94 3.97 Croatia 2.80 2.22 3.25 1.81 2.70 2.14 3.14 1.74 Estonia 3.76 3.40 4.83 2.20 3.65 3.29 4.71 2.12 Hungary 3.10 2.58 3.61 1.81 3.00 2.49 3.50 1.74 Latvia 3.65 2.93 4.50 1.91 3.54 2.83 4.38 1.84 Lithuania 5.06 3.68 5.67 2.57 4.94 3.57 5.56 2.48 Poland 4.10 3.22 4.91 2.28 3.99 3.11 4.79 2.20 Low-income CIS Armenia 6.98 6.11 7.65 4.49 6.88 6.00 7.57 4.37 Azerbaijan 5.69 5.47 6.48 4.52 5.57 5.35 6.37 4.40 Georgia 7.99 6.13 7.79 5.67 7.92 6.01 7.71 5.55 Kyrgyz Republic 5.59 5.37 6.61 4.32 5.47 5.25 6.51 4.21 Moldova 4.84 3.34 5.64 2.53 4.72 3.23 5.52 2.44 Tajikistan 7.83 7.14 7.19 7.44 7.74 7.04 7.09 7.35 Uzbekistan 7.87 6.92 8.53 5.56 7.79 6.81 8.47 5.44 Middle-income CIS Belarus 3.40 3.04 3.54 2.70 3.29 2.94 3.44 2.61 Kazakhstan 5.17 4.18 5.46 3.53 5.05 4.07 5.35 3.42 Russian Federation 5.07 3.82 5.61 2.37 4.95 3.71 5.49 2.29 Ukraine 6.54 5.33 6.62 4.73 6.43 5.21 6.51 4.61 Turkey Turkey 4.30 3.18 5.11 2.38 4.18 3.08 4.99 2.29 Western Balkans Albania 5.53 4.79 5.89 4.29 5.41 4.67 5.78 4.17 Bosnia and Herzegovina 3.67 4.14 4.76 3.05 3.56 4.02 4.65 2.95 Macedonia, FYR 5.94 5.38 7.69 3.57 5.82 5.26 7.60 3.46 Montenegro 6.47 6.88 7.99 5.06 6.36 6.77 7.91 4.95 Serbia 4.33 3.74 5.03 3.12 4.21 3.63 4.91 3.01 Source: Staff calculations. 70 The Crisis Hits Home--Stress-Testing Households in Europe and Central Asia TABLE A.7 The Welfare Impact of a 10 Percent Fuel Price Increase Laspeyres Price Index Geometric Price Index Location Consumption Quintiles Location Consumption Quintiles Rural Urban Q1 Q5 Rural Urban Q1 Q5 EU10 (including Croatia) Bulgaria 1.8 2.0 2.6 1.6 1.73 1.91 2.47 1.57 Croatia 1.0 0.9 1.0 0.8 0.93 0.85 0.99 0.75 Estonia 1.3 2.0 1.8 1.5 1.28 1.90 1.69 1.42 Hungary 1.8 1.8 2.0 1.5 1.75 1.78 1.90 1.46 Latvia 0.9 1.4 1.7 0.8 0.87 1.33 1.66 0.74 Lithuania 1.0 1.5 1.3 1.1 0.97 1.45 1.22 1.06 Poland 1.7 2.2 1.8 2.1 1.60 2.10 1.73 1.98 Low-income CIS Armenia 0.6 0.8 0.7 0.7 0.55 0.77 0.68 0.64 Azerbaijan 0.7 0.6 0.6 0.7 0.63 0.59 0.57 0.65 Georgia 0.4 1.2 0.8 0.8 0.38 1.14 0.81 0.73 Kyrgyz Republic 1.1 1.1 0.9 1.0 1.01 1.06 0.86 0.96 Moldova 1.3 1.6 1.2 1.7 1.22 1.58 1.19 1.62 Tajikistan 0.2 0.1 0.3 0.1 0.21 0.13 0.29 0.14 Uzbekistan 0.1 0.2 0.2 0.2 0.10 0.21 0.15 0.19 Middle-income CIS Belarus 0.8 0.8 0.8 0.7 0.72 0.74 0.74 0.71 Kazakhstan 1.0 1.0 1.1 0.9 0.97 0.97 1.05 0.85 Russian Federation 0.0 0.0 0.0 0.0 Ukraine 0.8 1.4 1.2 1.2 0.80 1.37 1.16 1.18 Turkey Turkey 1.1 1.2 1.0 1.2 1.08 1.20 0.99 1.16 Western Balkans Albania 1.1 1.4 1.2 1.4 1.07 1.33 1.11 1.31 Bosnia and Herzegovina 1.1 1.0 1.3 0.8 1.07 0.95 1.25 0.75 Macedonia, FYR 0.6 0.9 0.2 1.3 0.59 0.82 0.18 1.22 Montenegro 0.7 0.5 0.3 0.9 0.66 0.51 0.28 0.85 Serbia 1.1 1.4 1.2 1.2 1.08 1.32 1.13 1.19 Source: Staff calculations. Appendix 71 TABLE A.8 Summary Data: GDP Growth and Poverty Simulations Estimated Growth Baseline Poverty GDP Growth in Per Capita Private Headcount1 Projection for 2009 Consumption for 2009 Country $PPP 2.5 $PPP 5.00 Baseline Apr 2009 Baseline Apr 2009 Albania 15.0 60.0 5.6 -0.1 6.0 -0.1 Armenia 30.0 84.0 5.6 -7.1 4.9 -6.2 Azerbaijan 1.0 71.0 14.7 1.7 10.6 1.2 Belarus 1.0 13.0 7.4 -3.4 6.9 -3.1 Bosnia and Herzegovina 1.3 8.0 4.6 -3.3 4.7 -3.3 Bulgaria 3.1 20.0 5.6 -1.3 5.9 -1.4 Croatia 0.0 2.0 4.0 -3.5 3.9 -3.5 Estonia 2.0 18.0 3.9 -10.1 4.0 -10.3 Georgia 39.0 76.0 9.8 1.0 10.8 1.1 Hungary 0.0 7.0 2.7 -3.2 2.8 -3.3 Kazakhstan 7.0 54.0 6.9 -2.1 7.3 -2.2 Kyrgyz Republic 52.1 88.1 6.9 -2.1 7.0 -2.1 Kosovo 38.1 82.0 5.7 -2.0 5.8 -2.0 Latvia 1.0 12.0 0.8 -11.7 0.8 -11.9 Macedonia 10.0 37.1 4.7 -2.3 4.9 -2.4 Moldova 30.0 77.0 8.0 -3.4 10.1 -4.3 Montenegro 10.2 49.2 5.7 -2.0 5.8 -2.0 Poland 2.0 20.0 4.7 -0.7 4.5 -0.7 Romania 7.0 45.0 5.1 -3.8 4.8 -3.6 Russian Federation 3.0 20.0 6.8 -5.6 7.7 -6.3 Serbia 2.0 17.1 5.7 -2.0 5.4 -1.9 Tajikistan 56.0 89.0 6.4 1.4 6.8 1.5 Turkey 15.0 49.0 3.1 -6.1 3.0 -6.0 Ukraine 1.0 18.0 4.9 -7.3 6.0 -8.9 Uzbekistan 19.0 67.0 5.4 4.9 5.5 5.0 Sources: ECA Regional Household Data Archives, IMF World Economic Outlook database, and staff calculations. 1 Latest year for which data are available. 72 The Crisis Hits Home--Stress-Testing Households in Europe and Central Asia FIGURE A.1 The Impact of the Crisis on Poverty and Vulnerability Allowing for full pass-through from GDP growth to private consumption growth, in percent of the population $2.5-a-day poverty rates projections 2007­2010 $5-a-day poverty and vulnerability projections 2007­2010 10 40 8.0 33.5 33.0 8 7.9 7.6 32.3 Percent Percent 7.2 30 31.0 6.9 30.4 6.8 29.5 6 6.3 27.7 25.3 5.3 4 20 2007 2008 2009 2010 2007 2008 2009 2010 Pre-crisis projections January projections Latest projections Source: Staff calculations. Appendix 73 FIGURE A.2 Financial Margins in Selected Countries Evidence from HBS Data, in local currency and in percent of all households, by years Household vulnerability, Belarus Household vulnerability, Kazakhstan Cumulative distribution of household financial margins1 Cumulative distribution of household economic margins 2 100 100 75 75 50 50 25 25 0 0 -10 -5 0 5 10 15 20 25 30 -500 -250 0 250 500 750 1000 1250 1500 Millions of rouble Thousands of tenge 2001 2002 2003 2004 2001 2002 2003 2005 20063 2005 2006 2007 2008 Household vulnerability, Serbia Household vulnerability, Ukraine Cumulative distribution of household economic margins 4 Cumulative distribution of household economic margins 5 100 100 75 75 50 50 25 25 0 0 -1000 0 1000 2000 3000 -30 -15 0 15 30 15 60 75 Thousands of dinar Thousands of hryvnia 2006 2007 2002 2003 2004 2005 2006 Source: Staff calculations. Note: Financial margin is calculated as M i = DIi ­ BLCi ­ DSEi where I = household; M i = margin of the i 'th household; DIi = disposable income of the i 'th household; BLC i = basic living costs (defined by country) of the i 'th household; and DSE i = debt service expenditure of the i 'th household, as explained in the main text. BLC (basic living costs) is based on the subsistence-level budget/expenditure in Belarus, Kazakhstan, and Ukraine, and on the national poverty line in Serbia. 1. Based on subsistence-level budget. 2. Based on subsistence level. 3. Quarterly data are analyzed. 4. Based on poverty line of $5.00/person/day in 2005 price. 5. Based on subsistence level of average expenditure. References Alam, A., S. Kathuria, and O. Vybornaia, 2008a, Brownbridge, Martin, and Sudharshan Canagarajah, "Rising Food Grains and Energy Prices in ECA: 2009, "How Should Fiscal Policy Respond Some Economic and Poverty Implications, and to the Economic Crisis in the Low Income Policy Responses," mimeo, World Bank. Commonwealth of Independent States? Alam, A., S. Kathuria, and O. Vybornaia, 2008b, Some Pointers from Tajikistan," World Bank "Food Grains and Energy Prices in ECA: An Policy Research Working Paper No. 4970 Update," mimeo, World Bank. (Washington, DC: The World Bank). Alam, Asad, and Victor Sulla, 2009, "An Update on Calvo, Guillermo, Alejandro Izquierdo, and Ernesto Income Poverty and Inequality in Eastern Talvi, 2006a, "Phoenix Miracles in Emerging Europe and the Former Soviet Union," mimeo Markets: Recovering without Credit from (Washington, DC: The World Bank). Systemic Financial Crises," NBER Working Paper No. 12101 (Cambridge, Great Britain, Alderman, Harold, and Christina H. Paxson, National Bureau of Economic Research). 1992, "Do the Poor Insure? A Synthesis of the Literature on Risk and Consumption _____, 2006b, "Sudden Stops and Phoenix Miracles in in Developing Countries," World Bank Emerging Markets," American Economic Review, Policy Research Working Paper No. 1008 Vol. 96 (2): pp. 405­410. (Washington, DC: The World Bank). Campbell, John Y., and Joao F. Cocco, 2007, "How Beer, Christian, and Martin Schurz, 2007, Do House Prices Affect Consumption? "Characteristics of Household Debt in Austria: Evidence from Micro Data," Journal of Monetary Does Household Debt Pose a Threat to Economics, Vol. 54 (3): pp. 591­621. Financial Stability?" Monetary Policy and the Central Bank of the Russian Federation, 2008, Economy, 2nd Quarter: pp. 58­79. Banking Supervision Report 2007 (Moscow: Bekaert, Geert, Campbell R. Harvey, and Christian Central Bank of the Russian Federation). Lundblad, 2006, "Growth Volatility Chen, Kechen, and Mali Chivakul, 2008, "What and Financial Liberalization," Journal of Drives Household Borrowing and Credit International Money and Finance, Vol. 25 (3): pp. Constraints? Evidence from Bosnia and 370­403. Herzegovina," IMF Working Paper No. 75 76 The Crisis Hits Home--Stress-Testing Households in Europe and Central Asia 08/202 (Washington, DC: International Holló, Dániel, 2007, "Household Indebtedness and Monetary Fund). Financial Stability: Reasons to Be Afraid?" Coady, David, and others, 2006, "The Magnitude and MNB Bulletin, November (Budapest: National Distribution of Fuel Subsidies: Evidence from Bank of Hungary). Bolivia, Ghana, Jordan, Mali, and Sri Lanka," International Monetary Fund (IMF), 2009a, World IMF Working Paper No. 06/247 (Washington, Economic Outlook, April (Washington, DC: DC: International Monetary Fund). International Monetary Fund). de Koning, Jaap, Mariana Kotzeva, and Stoyan _____, 2009c, Europe Regional Economic Outlook, April Tzvetkov, 2007, "Mid-Term Evaluation of (Washington, DC: International Monetary the Bulgarian Programme `From Social Fund). Assistance to Employment,'" in Jaap de _____, 2009d, Global Financial Stability Report, April Koning, ed., Employment and Training Policies (Washington, DC: International Monetary in Central and Eastern Europe (Amsterdam: Fund). Dutch University Press). _____, 2009e, "Republic of Latvia: Request for Dudwick, Nora, Elizabeth Gomart, and Alexandre Stand-By Arrangement--Staff Report; Marc, with Kathleen Kuehnast, 2003. When Staff Supplement; Press Release on the Things Fall Apart: Qualitative Studies of Poverty Executive Board Discussion; and Statement in the Former Soviet Union (Washington, DC: by the Executive Director for the Republic World Bank). of Latvia" (Washington, DC: International European Central Bank (ECB), 2009, Financial Monetary Fund). Stability Review, June (Frankfurt: European _____, 2008a, World Economic Outlook, October Central Bank). (Washington, DC: International Monetary _____, 2007, EU Banking Sector Stability, November Fund). (Frankfurt: European Central Bank). _____, 2008b, World Economic Outlook, April Fankhauser, Samuel, Yulia Rodionova, and Elisabetta (Washington, DC: International Monetary Falcetti, 2008, "Utility Payments in Ukraine: Fund). Affordability, Subsidies and Arrears," Energy _____, 2008c, "Food and Fuel Prices--Recent Policy, Vol. 36 (11): pp. 4168­4177. Developments, Macroeconomic Impact, and Policy Responses," June 19 (Washington, DC: Freund, Caroline L., and Christine Wallich, 1997, International Monetary Fund). "Public Sector Price Reforms in Transition Economies Who Gains? Who Loses? The _____, 2006b, "Household Credit Growth in Case of Household Energy Prices in Poland," Emerging Market Countries," Global Financial Economic Development and Cultural Change, Vol. Stability Report, September, Chapter II. 46 (1): pp. 35­59. Johansson, Martin W., and Mattias Persson, 2006, Ghosh, Atish R., Marcos Chamon, Christopher "Swedish Households' Indebtedness and Ability Crowe, Jun I. Kim, and Jonathan D. Ostry, to Pay: A Household Level Study," Sveriges 2009, "Coping with the Crisis: Policy Options Riksbank Economic Review, No. 3: pp. 24­40. for Emerging Market Countries," IMF Staff Kang, S. J., and Y. Sawada, 2008, "Credit Crunch Position Note No. SPN/09/08 (Washington, and Household Welfare: The Korean Financial DC: International Monetary Fund). Crisis," Japanese Economic Review, Vol. 59 (4): Gupta, Sanjeev, Marijn Verhoeven, Robert pp. 438­458. Gillingham, and others, 2000, Equity and Karasulu, Meral, 2008, "Stress Testing Household Effi ciency in the Reform of Price Subsidies: A Debt in Korea," IMF Working Paper No. Guide for Policymakers (Washington, DC: 08/255 (Washington, DC: International International Monetary Fund). Monetary Fund). Habermeier, Karl, nci Ötker-Robe, Luis Jacome, Kluve, Jochen, Hartmut Lehmann, and Christoph M. Alessandro Giustiniani, Kotaro Ishi, David Schmidt, 1999, "Active Labor Market Policies Vávra, Turgut Kiinbay, and Francisco Vazquez, in Poland: Human Capital Enhancement, 2009, "Inflation Pressures and Monetary Policy Stigmatization, or Benefit Churning?" Journal Options in Emerging and Developing Countries: of Comparative Economics, Vol. 27 (1) (March): A Cross Regional Perspective," IMF Working pp. 61­89. Paper No. 09/1 (Washington, DC: International Kose, M. Ayhan, Eswar Prasad, and Marco E. Monetary Fund). Terrones, 2003, "Financial Integration and References 77 Macroeconomic Volatility," IMF Staff Papers, World Bank, 2009a, "The Expected Impact of Vol. 50, Special Issue: pp. 119­142. the Global Financial Crisis on the World's Laeven, Luc, and Thomas Laryea, 2009, "Principles Poorest," mimeo (Washington, DC: The World of Household Debt Restructuring," IMF Staff Bank). A note prepared by the World Bank's Position Note No. SPN/09/15 (Washington, Development Economics Vice Presidency in DC: International Monetary Fund). February 2009. Lampietti, Julian, and others, 2007, People and _____, 2009c, "Tajikistan: Poverty Assessment" Power: Electricity Sector Reforms and the Poor (Washington, DC: The World Bank). in Europe and Central Asia (Washington, DC: _____, 2009d, "Armenia: Implications of the Global The World Bank). Financial Crisis for Poverty," Report No. _____, ed., 2004, "Power's Promise: Electricity 47770-AM (Washington, DC: The World Reforms in Eastern Europe and Central Asia," Bank). Working Paper No. 40 (Washington, DC: The _____, 2009e, "Bulgaria: Poverty Implications of World Bank). the Global Financial Crisis for Poverty" (Washington, DC: The World Bank). Lubyova, Martina, and Jan C. van Ours, 1999, "Effects of Active Labor Market Programs on _____, 2009f, "Protecting Core Public Spending the Transition Rate from Unemployment into during the Global Economic Crisis," Briefi ng Regular Jobs in the Slovak Republic," Journal of Note prepared by the World Bank's PREM Comparative Economics, Vol. 27 (1): pp. 90­112. Vice Presidency. May, Orla, and Merxe Tudela, 2005, "When Is _____, 2009g, Russian Economic Report No. 18 Mortgage Indebtedness a Financial Burden (Washington, DC: The World Bank). to British Households? A Dynamic Probit _____, 2009h, "Social Protection and Economic Approach," Bank of England Working Paper Shocks in ECA: The `Social Side' of the Global No. 277 (London: Bank of England). Crisis," Social Protection Team (Washington, National Bank of Serbia (NBS), 2008, "2007 DC: The World Bank). Financial Stability Report" (Belgrade: _____, 2009j, "In Focus: Domestic Credit National Bank of Serbia). Developments," EU10 Regular Economic Rosenberg, Christoph B., and Marcel Tirpák, 2008, Report, February (Washington, DC: The "Determinants of Foreign Currency Borrowing World Bank). in the New Member States of the EU," IMF _____, 2008a, "Rising Food Prices: Policy Options Working Paper No. 08/173 (Washington, DC: and World Bank Response," Background. International Monetary Fund). _____, 2008b, "PREM Guidance Note on the Sawada, Y., K. Nawata, M. Ii, and J. Lee, 2007, "Did Financial Crisis" (Washington, DC: The the Credit Crunch in Japan Affect Household World Bank). Welfare? An Augmented Euler Equation _____, 2005, Growth, Poverty, and Inequality: Approach Using Type 5 Tobit Model," CIRJE Eastern Europe and the Former Soviet Union Discussion Papers F-Series No. 498 (Tokyo: (Washington, DC: The World Bank). University of Tokyo). _____, 2004, "Growth, Employment, and Living U.S. Bureau of Labor Statistics, 1997, "The Standards in Pre-Accession Poland," Report Experimental CPI Using Geometric Means No. 30078 (Washington, DC: The World (CPI-U-XG)" (Washington, DC: U.S. Bureau Bank). of Labor Statistics). Zezza A., B. Davis, C. Azzarri, K. Covarrubias, L. Vatne, Bjřrn Helge, 2006, "How Large Are the Tasciotti, and G. Anriquez. 2008, "The Impact Financial Margins of Norwegian Households? of Rising Food Prices on the Poor," ESA An Analysis of Micro Data for the Period 1987­ Working Paper, 08-07 (Rome: FAO). 2004," Norges Bank Economic Bulletin, December ochowski, D., and S. Zajczkowski, 2008, "Stress 2006, Vol. LXXVII (4): pp. 173­180. Testing Household Indebtedness: Impact of Vodopivec, Milan, 1999, "Does the Slovenian Public fi nancial vs. Labor market shocks," World Bank Work Program Increase Participants' Chances Workshop, Macro Risks and Micro Responses. to Find a Job?" Journal of Comparative Economics, _____, 2007, "The Distribution and Dispersion Vol. 27 (1) (March): pp. 113­130. of Debt Burden Ratios among Households von Braun, Joachim, 2008, "The Food Crisis Isn't in Poland and Its Implications for Financial Over," Nature, 456 (11): p. 701. Stability," IFC Bulletin No. 26 (Basel: BIS). Notes 1. For a discussion of the origins of these shocks 13. This and the next paragraphs draw heavily from and their global implications, see, for example, World Bank (2009h). IMF WEO (October 2008a, 2009a), IMF Global 14. This section draws heavily from the work of Alam Financial Stability Report (October 2008, April and Sulla (2009). Their analysis is based on the 2009) and, for a particular focus on commodity same internationally comparable consumption price movements, see the World Bank Global aggregates and poverty and vulnerability lines Economic Prospects (2009). used in this chapter. 2. See also World Bank (2009j). 15. Central Bank of the Russian Federation (2008). 3. See OECD (2006). 16. This problem has been modeled as a problem of 4. See also World Bank (2004 and 2005). More for- inter-temporal consumption smoothing under a mally, the (total) elasticity of the poverty head- stochastic income process. Using this framework, count with respect to growth in average per capita some papers (Kang and Sawada, 2008; Sawada, consumption has been about ­1.3 in these coun- Nawata, Ii, and Lee, 2007) estimate the welfare tries, compared with ­3.4 in the middle-income costs of a credit crunch. For example, over the CIS countries. period of the Korean "twin" crises of 1997­1998, 5. See Calvo, Izquierdo, and Talvi (2006a and 2006b). the marginal utility loss due to the credit crunch Among the ECA countries in their sample are the was found to be higher among lower-income Russian Federation, Slovenia, and Ukraine. households, ranging from 29.3 percent for the 6. See also IMF (2009a). bottom quartile to 14.2 percent for the richest 7. See Ghosh et al. (2009). quartile. In Japan in 1998, the marginal utility 8. IMF (2008b, 2008c). loss ranged from 10.3 to 2.4 percent for the bot- 9. "Rising Food Prices Hit Eastern Europe," The tom and top income quartiles, respectively. Wall Street Journal, March 12, 2009. 17. Rosenberg and Tirpák (2008) suggest that EU 10. Fankhauser et al. (2008). Under the ongoing IMF- membership promotes borrowing in foreign cur- supported program, energy tariff increases in rency indirectly, such as through capital account Ukraine are expected to be phased-in (IMF, 2009c). liberalization that then facilitates access to for- 11. Von Braun (2008). eign funds. In addition, they also observe that EU 12. Campbell and Cocco (2007). membership seems associated with greater private 79 80 The Crisis Hits Home--Stress-Testing Households in Europe and Central Asia sector confidence in the stability of the exchange the unemployment shock. For example, if the rate and the eventual adoption of the euro. rural unemployment rate is double that of urban 18. National Bank of Serbia (2008). unemployment, the ratio is kept constant after 19. See Rosenberg and Tirpák (2008) for a brief sur- the new "unemployed" are selected. vey of some of these policies. But the results of 29. Neither can we estimate a probability model for their analysis suggest that the observed cross- taking on loans with variable interest rate, as the country differences are in large part explained by relevant information does not exist in the house- interest rate differentials. hold survey. 20. Campbell and Cocco (2007). 30. Holló (2007), for example, simulates a 5 percent- 21. ochowski and Zajczkowski (2007 and 2008). age point interest rate shock in Hungary. 22. See, for example, the discussions on Latvia sum- 31. It should be pointed out that only those who were marized in IMF (2009e). previously employed are subjected to the shock. 23. See, for example, Holló (2007), ochowski and Unemployed individuals as well as unpaid family Zajczkowski (2008), Johansson and Persson workers and agricultural workers and farmers pro- (2006), and others. ducing for their own consumption are excluded. 24. The fi nancial margin is calculated as Mi = DIi ­ 32. At the time these stress tests were conduct, the BLCi ­ DSEi where i = household; Mi = margin of study team was unaware of publicly available the i'th household; DIi = disposable income of the aggregate data on the share of household loans i'th household; BLCi = basic living costs (defi ned with variable interest rates. by country) of the i'th household; and DSEi = debt 33. See, for example, Beer and Schurz (2007). service expenditure of the i'th household. See also 34. See, for example, OECD (2006) and ECB (2009). Johansson and Persson (2006) and Vatne (2006). 35. IMF (2008a). Nationally defi ned minimum consumption 36. Coady et al. (2006). expenditures, minimum consumption baskets, or 37. Coady et al. (2006). poverty lines have been typically used to proxy 38. There are of course offsetting developments basic living expenses. that are difficult to quantify in net terms. Rising 25. See May and Tudela (2005) on how this threshold oil prices through late 2008 bolstered the GDP interest payment level is determined. It may be growth of oil-exporting countries such as Russia argued that a 20 percent threshold seems overly to about 7.3 percent on average in 2007. In turn, restrictive because other stress testing exercises the growing demand from the region's oil- have used a 30 percent threshold instead (e.g., Beer exporting countries has increased exports among and Schurz, 2007; ECB, 2007; Karasulu, 2008). smaller oil-importing countries. However, those exercises were based on compre- 39. This was previously noted in World Bank (2005) as hensive data on household debt (including mort- well. See the report's data and methodology appendix. gage debt, consumer loans, and other household 40. Baclajanschi et al. (2007). debt) and household debt service (covering both 41. Alam et al. (2008a and 2008b). interest payments and principal repayments). In the 42. See, e.g., Baclajanschi et al. (2007); Freund and case of SILC data, we have information on mort- Wallich (1997); and Coady et al. (2006). gage interest payments, thus requiring the relevant 43. The Laspeyres index is also known as an arith- interest payment threshold. Nonetheless, we also metic mean index, calculated as the sum of use a 30 percent threshold as a sensitivity test. weighted prices in time t+1, retaining base period 26. This refers to the analysis of EU-SILC data, consumption shares. The second is a geometric which are for 2007, and thus the two-year period index of relative price change (RG) using the base for these EU countries captures the recorded period consumption shares as weights: RG = increase in unemployment rates between 2007 and (Pt+1,i / Pt,i) Wt,i where Pt+1,i and Pt,i are the prices 2009. In the case of HBS data, which are for CIS of the i'th consumption items in the new and base and other countries and refer to older years (e.g., time periods, respectively, and Wt,i is the share 2006), we use a longer time horizon (e.g., 5-year of the i'th consumption item in total consump- period) to capture historical increases (rather than tion. See also Pollak (1989), Gupta et al. (2000), decreases) in unemployment rates. U.S. Bureau of Labor Statistics (1997), and the 27. The procedure is carried out in Stata using the r Technical Appendix in Baclajanschi et al. (2006). binomial routine. 44. This is based on Sarosh Sattar's unpublished work 28. It is assumed that the ratio of unemployment rates and data, which she generously shared. Any errors within each subgroup is constant before and after of interpretation are the study team's own. Notes 81 45. Zezza et al. (2008). The data were generously pro- 64. The full set of references is available on request. vided by the authors. 65. See World Bank (2008b) for a more comprehen- 46. See, for example, World Bank (2009a and 2005). sive discussion. 47. Ideally, private consumption growth projections 66. World Bank (2009f). The calculations are based from the Global Economic Prospects database on the April 2009 WEO and are consistent with would have been preferable. However, country- other WEO-based calculations elsewhere in this level projections were not readily available. The report. calculations in World Bank (2009a) are based on 67. See, for example, Brownbridge and Canagarajah the Global Economic Prospects database. (2009). 48. See also Alam and Sulla (2009). 68. This and subsequent paragraphs are drawn heav- 49. World Bank (2005). ily from World Bank (2009h). The report pro- 50. This section draws in part from the caveats out- vides an excellent, comprehensive review of the lined in World Bank (2009a). region's social protection system. 51. This and other references throughout the report 69. According to a recent report (World Bank 2008b), to the welfare costs of previous crisis in ECA are Turkey initiated a CCT program during the 2001 drawn from the comprehensive review of the lit- crisis and the program has been found to be effec- erature conducted by Victoria Levin. tive. FYR Macedonia is now in the process of 52. In fact, there are two underlying assumptions: (i) developing a CCT, with a special focus on youth the first concerns the relationship between GDP and infants. growth and average per capita consumption and (ii) 70. We thank Sophie Sirtaine for raising these the second concerns the relationship between the points. growth in average per capita private consumption 71. See Laeven and Laryea (2009). as measured in national accounts and the growth 72. If fi nancial assistance is provided only to the in per capita household consumption as measured households in the lowest quintile, the cost of in household surveys. The first (i) is estimated providing assistance is of course substantially as explained in the data section and the second smaller. (ii) assumes a full pass-through. As discussed in 73. For a detailed discussion of these issues, see Alam World Bank (2009a), the poverty impact may be et al. (2008a and 2008b). overstated if the actual fall in household consump- 74. See Lampietti et al. (2007) and Lampietti (2004) tion is less than the fall in national accounts-based for a more comprehensive discussion. private consumption. 75. The region's experience with workfare may be 53. See, for example, Bekaert, Harvey, and Lundblad broader than these four evaluations may sug- (2006). gest. For example, during the 1998­2000 period, 54. See, for example, Kose, Prasad, and Terrones participation rates (as percentage of registered (2003). In fact, up to a certain threshold, rela- unemployed) in public works were about 15 per- tive consumption volatility may rise with greater cent in Bulgaria, 11 percent in Estonia, 27 percent fi nancial intermediation. in Hungary, 19 percent in Russia, 3.2 percent in 55. IMF (2009a). Slovakia, 2.2 percent in Poland, and 11.6 percent 56. See also Dudwick et al. (2003). in Ukraine. Unfortunately, this rich experience 57. This is based on the estimated total ECA population has not been evaluated. of about 477 million, using UN population data. 76. De Koning, Kotzeva, and Tzvetkov (2005). 58. World Bank (2005). See Table 2.1. 77. Lubyova and van Ours (1999). 59. World Bank (2009c). 78. M. Vodopivec (1999). 60. World Bank (2009d). 79. J. Kluve, H. Lehmann, and C. M. Schmidt 61. World Bank (2009e). (1999). 62. World Bank (2009g). 80. IMF (2006b). 63. Alderman and Paxson (1992). 81. Rosenberg and Tirpák (2008). The global financial crisis, which has led to a sharp slowdown in economic activity everywhere, including the emerging markets in Europe and Central Asia, now risks reversing the substan- tial gains and improvements in living standards achieved by the region over the past few years. It threatens the well-being of millions of people in the region who are poor or who are living just above the poverty line and at risk of easily falling into pov- erty as economies contract. This book seeks to understand the key macroeconomic shocks confronted by the region and the impact of such shocks on household welfare, including the effect on household income flows, consumption levels, and liabilities. It draws on a large, cross-country database of recent house- hold surveys and presents regional overviews as well as relevant country examples to illustrate the incidence and distribution of specific vulnerabilities. ISBN 978-0-8213-8222-6 SKU 18222