85557 Hardship and Vulnerability in the Pacific Island Countries Hardship and Vulnerability in the Pacific Island Countries Contents Acknowledgements. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . iii executive summary. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . v Many People across the Pacific Are Living in Hardship Many Others Are Vulnerable to Hardship in the Future. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . v People in the Pacific Are Uniquely Vulnerable to Aggregate Shocks. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . vi Traditional Systems and Government Responses Help but Are Not Enough. . . . . . . . . . . . . . . . . . . . . . . . vii The Way Forward: Govern Prudently, Invest in Data, and Enable Households and Communities. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . viii Chapter 1: Concepts and Context. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1 Basic Concepts. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1 A Unique Set of Features in the Pacific Shape Hardship and Vulnerability.. . . . . . . . . . . . . . . . . . . . . . . . . . 3 Reducing Hardship and Vulnerability through Risk Management. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5 Pacific Approaches to Reducing Hardship and Vulnerability. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7 References. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8 Chapter 2: A Snapshot of Hardship in the Pacific. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9 Background. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9 Poverty Measurement in the Pacific. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10 The Incidence of Hardship Based on National Poverty Lines. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11 Hardship Differentials across Households. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13 Inequality in the Pacific Island Countries. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 22 Key Messages. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 25 References. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 25 Technical Annex: Important Cross-Country Differences in Poverty Analysis. . . . . . . . . . . . . . . . . . . . . . . . 27 Chapter 3: Vulnerability and the Impacts of Shocks on Pacific Islanders. . . . . . . . . . . . 29 Pacific Islanders Are Vulnerable to Many Aggregate Shocks. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 29 Aggregate Economic Shocks Are Particularly Important.. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 31 Impacts of Aggregate Economic Shocks on Households: Results from Microsimulations. . . . . . . . . . . . . 38 Health Shocks Are Also an Important Source of Vulnerability. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 44 Other Types of Shocks Also Affect Pacific Islanders but Are More Difficult to Measure. . . . . . . . . . . . . . 45 Key Messages. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 46 References. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 47 Technical Annex: Details of Microsimulation Modeling. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 49 Chapter 4: Current Approaches to Reducing Hardship and Vulnerability. . . . . . . . . . . . 51 The Role of Traditional Networks in Reducing Hardship and Vulnerability.. . . . . . . . . . . . . . . . . . . . . . . . . 52 Government Social Insurance Programs. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 59 Government Provision of Basic Services. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 60 Government Policies to Manage Aggregate Economic and Natural Shocks. . . . . . . . . . . . . . . . . . . . . . . . 62 Key Messages. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 65 References. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 66 Chapter 5: Implications for Policy: The Way Forward. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 69 Govern Prudently and Proactively Manage Aggregate Shocks. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 69 Invest in Data. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 74 Enable Households and Communities.. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 75 References. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 77 Hardship and Vulnerability in the Pacific Island Countries Acknowledgments iii Acknowledgments T his report was prepared by a team from the Drees-Gross, Sudhir Shetty, Xiaoqing Yu, and Michael World Bank and the Australian Department of Carnahan (DFAT). Foreign Affairs and Trade (DFAT). The team was led by Melissa Adelman, together with Oleksiy For their valuable comments and inputs, the team Ivaschenko, under the guidance of Truman Packard would like to thank David Abbott, M. Ihsan Ajwad, and Vivek Suri. The members of the core team were Jehan Arulpragasam, Timothy Bulman, Cesar Thomas Bowen, Caesar Cororaton, Imogen Halstead Calderon, Shubham Chaudhuri, Nathan Dal Bon (DFAT), Jae Kyun Kim, David Knight, and Tatiana (DFAT), Tobias Haque, Virginia Horscroft, Philip Sviridova. Proofreading and production assistance O’Keefe, Anna Maria Oviedo, Lucy Pan, Alban Pinz were provided by Sarah Harrison. The report was (DFAT), Clinton Pobke (DFAT), and Manohar Sharma. prepared under the overall guidance of Franz iv | HARDSHIP AND VULNERABILITY IN THE PACIFIC ISLAND COUNTRIES Acronyms CBN Cost of Basic Needs CIF Consolidated Investment Fund GDP Gross Domestic Product GER Gross Enrollment Rate HIES Household Income and Expenditure Survey NCD Noncommunicable Disease OMTV Overseas Medical Treatment Visits PIC Pacific Island Country PPP Purchasing Power Parity REP Rapid Employment Project TTF Tuvalu Trust Fund TYPSS Ten Year Pacific Statistics Strategy UNDP United Nations Development Programme Executive Summary v Executive Summary I n many Pacific Island Countries (PICs), meeting Family and community networks are central to life nonfood basic needs is a growing challenge and in most PICs, providing critical support to members further complicated by substantial economic in need and acting as safety nets when individuals and environmental risks. Of the 11 PICs that are or households experience losses from shocks. members of the World Bank, six are lower middle However, in many PICs, people are moving away from income countries, with gross national income per isolated areas and islands and coalescing in urban capita of less than $4,085 in 2012. The incidence centers. These changes may put strain on traditional of extreme deprivation (hunger) is traditionally low networks and reduce access to land for subsistence. In throughout the Pacific region, but urbanization and addition, when shocks are “covariate,” affecting entire monetization of the economy are creating new forms communities or countries, traditional networks may be of hardship, particularly for meeting the costs of less able to mitigate their impacts. nonfood needs. Small populations and remoteness limit economic diversification and increase the impacts The primary objective of this report is to present solid of economic shocks, while location and topography empirical evidence of hardship, vulnerability to shocks, expose Pacific Islanders to a range of natural shocks, and risk management in the Pacific region. The report which will likely be exacerbated by climate change. is primarily a stocktaking exercise that brings together In addition, many PICs are heavily reliant on external existing evidence and new analysis of available data inflows of money and goods, and this further exposes using a consistent framework. The report takes a them to the volatility in neighboring large economies “micro-” perspective—that of the individual and and global markets. household—but accounts for the important role of communities, the state, and international partners. As Hardship and vulnerability are increasingly prominent such, it is a complement to the studies that analyze concerns in PICs, but the knowledge base to guide the macroeconomic context of the Pacific or that policy making is limited.1 Household Income and analyze rich subnational data. The purpose of building Expenditure Surveys have been conducted in many this solid evidentiary base is to catalyze needed PICs, but internationally comparable statistics on future work, including new data collection, rather than hardship and vulnerability are available for only a to provide the final answer to any of the important few. Although the aggregate or macroeconomic questions addressed in the report. impacts of negative shocks have been relatively well studied, much less is known about the impacts on household well-being, in large part because of data Many People across the Pacific limitations. In addition, the effectiveness of prevailing risk management mechanisms is not well understood. Are Living in Hardship, Many Others Are Vulnerable to Hardship in the Future 1. In several PICs, the label of poverty is considered culturally Across the Pacific, many people are living in hardship, inappropriate because it is viewed as implying a failure of traditional, community-based safety nets. As such, it is not meaning they are unable to meet their basic needs. discussed on political platforms, although it is viewed as a The evidence presented in the first two sections of concern by relevant government departments and agencies. chapter 2 shows that more than 20 percent of people Therefore, in this study, the term “hardship” will be used to in most PICs live in hardship, meaning they are unable refer to the welfare concept commonly termed “poverty.” to meet their basic food and nonfood needs. The vi | HARDSHIP AND VULNERABILITY IN THE PACIFIC ISLAND COUNTRIES incidence of hardship is highest in Papua New Guinea, leaves policy makers and development partners where 40 percent of the population lives in hardship. without timely information about people’s well-being These results accord with many other measures of that could help them make the best use of limited well-being that have been made for the Pacific. PICs public resources. Even when data exist, significantly have had mixed success with making progress toward different methodological choices in the poverty the Millenium Development Goals, and the United measurement and analysis as well as limited sharing Nations Development Programme categorizes all the of data makes it difficult to clearly interpret and included PICs as having either “medium” or “low” communicate the results. In a context in which the human development. Taken together, all of these notion of “poverty” is already controversial, these results show that hardship is a real challenge that challenges mean that measures of hardship too often merits the attention of policy makers in the Pacific. end up being ignored in policy making. Within countries, many factors have a bearing on hardship, including location, educational attainment, People in the Pacific Are work status, gender, and age. In Fiji and Papua New Guinea, people living in urban areas are much less Uniquely Vulnerable to likely to live in hardship. In general, households Aggregate Shocks headed by individuals who have limited education or People in the Pacific are uniquely vulnerable to who do not work are more likely to live in hardship. aggregate economic and natural shocks because Households headed by the elderly are more likely of their countries’ combination of small size, to live in hardship than those headed by younger isolation, and other geographic features. The people, and households with more children are also location and topography of PICs exposes them to more likely to live in hardship. The relationships a disproportionate number of natural shocks, and between these characteristics and hardship vary in several are among the most vulnerable countries in strength and are driven by many underlying factors. the world in terms of relative natural disaster losses. However, identifying these relationships is an Fuel and food imports, tourism, remittances, and important first step in understanding the challenge of international aid all contribute to the well-being of hardship across the Pacific. Pacific islanders and help countries overcome the limitations on development caused by geography. Many people not currently in hardship remain However, PIC economies are still small and vulnerable to it, and levels of inequality in the Pacific undiversified, so negative shocks to these external are comparable to those in East Asian countries. flows can have very large impacts. Across PICs, much of the population that lives above their respective national poverty lines does not Commonly occurring price shocks to commodity consume much more than those in hardship, meaning imports and exports increase hardship substantially. they are vulnerable to falling into hardship in the People in the small countries of the Pacific are highly future. In all countries, the most well-off people (the exposed to high and volatile global commodity top 20 percent) consume many times more than the prices. Microsimulation analysis for Kiribati, Papua least well-off. As measured by the Gini coefficient, New Guinea, and Tonga finds that shocks to the prices inequality is highest in the Solomon Islands, Papua of imported food and fuel, agricultural commodity New Guinea, and Fiji. Moreover, within most exports, and remittances push many people into countries, inequality in rural areas is equal to or higher hardship and deepen the severity of hardship for than inequality in urban areas (Fiji is the prominent many others. The impacts of import price shocks exception). When considered in a global context, are particularly severe in the small atoll nations that Pacific levels of inequality are not extremely high, but rely heavily on imports for staple foods and fuel. For to the extent that inequality affects economic growth example, in Kiribati, simultaneous spikes in the prices and social cohesion, they may still be a matter of of rice, wheat, and oil are estimated to push 6 percent concern for policy makers. of the country’s entire population into hardship. This impact of a commonly occurring set of shocks (20 to Infrequent surveys, significant methodological 30 percent price increases, with a likelihood of about variation, and cultural rejection mean that hardship 33 percent in any given year) is close to the estimated currently plays a limited role in policy making. impacts of the 2007 global food and fuel crisis on 20 Unfortunately, across the Pacific, representative of the worst affected countries in the world. household surveys are infrequent. This lack of data Executive Summary | vii The growing epidemic of noncommunicable diseases PICs. Evidence from household surveys shows that a (NCDs) is an aggregate health shock with significant minority of households hold savings accounts, loans, consequences for the well-being of people in the or insurance policies. Without access to these risk Pacific. NCDs reduce productivity and quality of life management tools, households are likely relying too and are very expensive to treat. Increase in NCDs much on coping—for example, by drawing down has already eroded life expectancy in Tonga. Most on productive assets such as livestock or reducing PICs are facing this epidemic while also dealing their investments in human capital. At the same time, with continued threats from communicable diseases growth in financial access without effective regulation and maternal and child mortality. With limited fiscal and consumer education can lead households into resources, trying to manage this “double burden” of excessive debt, which is a concern in some countries disease is a major challenge for Pacific governments. including Fiji. In addition to aggregate shocks, people in the Pacific Governments provide little social insurance, but some face many idiosyncratic and local shocks, but little programs show promise within country constraints. data are available to identify their frequency and Across the Pacific, Fiji is the only country with a impacts. Crop failure, job loss, violence, and many national hardship-targeted cash transfer program. other idiosyncratic or localized shocks are likely to However, many other countries provide transfers or occur in the Pacific, as they do around the world. subsidies to small groups of people identified as Some striking evidence on the prevalence of domestic being in need. Broader measures to support those violence and unwanted pregnancy shows that these in hardship or experiencing shocks are generally not personal shocks are much more common in the Pacific in place because of fiscal and capacity constraints, as than in neighboring East Asian countries. However, well as lack of information, particularly in the smaller existing national household surveys are not designed islands. Two programs that have been tried by some to capture information on the full range of shocks that governments and development partners in the Pacific occur or their impacts. Much more could be learned if are funds for the elderly and cash for work schemes. future surveys asked specific questions about shocks These show promise in part because of their lower and followed households over time to measure the information requirements for targeting participants. impacts of shocks. Government funding for basic services is under fiscal pressure because of rapidly rising costs of coping with Traditional Systems and NCDs. Health care expenditures in the Pacific largely go to coping with health shocks: curative, palliative, Government Responses Help and rehabilitation care absorbs 80 to 90 percent of but Are Not Enough national health expenditures. This focus on coping is fiscally unsustainable because of NCDs, which are Traditional systems do not eliminate hardship and can spreading quickly and are costly to treat. Greater provide only partial insurance. Although traditional emphasis is needed on knowledge and protection systems of resource sharing and self-subsistence are measures to slow their increase, but changing important to the well-being of many Pacific islanders, people’s behavior is difficult. In addition, funding both hardship and vulnerability are still major challenges. ex-ante knowledge and protection measures for the Traditional systems do not reach everyone, and future, while dealing with the present costs of coping, evidence from household surveys suggest that those is a major fiscal challenge. in deepest hardship may be the least likely to be part of gift-giving networks. In addition, cultural and social pressures seem to require greater generosity than Managing aggregate economic shocks through many households feel they can truly afford. At the coping actions has limited effectiveness, and more same time, traditional systems cannot insure against protective measures hold promise. Few ex-post the many aggregate shocks that are common in the responses to economic shocks in the Pacific have Pacific. Governments therefore have a role to play proven to be effective in reducing the negative in complementing traditional systems with hardship impacts on households while also being fiscally reduction and risk management efforts. sustainable. Some ex-ante measures that provide protection or insurance against shocks are being explored, but shocks will continue to be part of Households have limited access to market instruments the Pacific landscape. Therefore, some of the most that can help them manage risks. In particular, access important actions governments can take are to pursue to formal financial instruments is limited in most viii | HARDSHIP AND VULNERABILITY IN THE PACIFIC ISLAND COUNTRIES prudent macroeconomic policy, including building up more household surveys in the past decade, which is savings in good times and actively mobilizing revenue a promising start toward regular data collection. For to have resources to deploy during shocks. the countries with more sporadic surveys and where reliance on external financing is greater, funding regular surveys, including the technical capacity to The Way Forward: Govern implement and analyze them, should be a priority for development partners. Additional consideration of Prudently, Invest in Data, survey methods and the content of surveys could help and Enable Households increase data quality and lower costs, which would increase the financial feasibility of regular surveys. and Communities For example, the use of local market price surveys, shorter consumption diary data collection periods, Govern Prudently and technology-assisted survey methods could all be Governments and development partners have considered. important roles to play in managing risk through sound policy, particularly in managing risks from Conducting surveys is not enough: Public access aggregate shocks. Broadly speaking, government to data and analysis is equally critical to getting it policy should aim to avoid being a source of instability used by policy makers and partners. When data itself and should actively recognize and account for are made accessible, researchers and others use it risks in all areas. For the aggregate economic and to provide policy-relevant insights and also often natural risks described in this report, households and help improve the quality of the data by pointing communities cannot fully manage on their own and out problems that might not have been recognized. need systematic support. Several Pacific countries and their regional partners have restricted access to the collected data and to the Good government policy in all areas should factor analysis conducted with the data, including poverty in risks. For example, urbanization is changing risk measurement. In general, only summary statistics are profiles and presenting new challenges, as well as available to the public, and no established processes opportunities. Active planning can help to address are in place for obtaining access. Consequently, little the challenges, and maximize the opportunities, use is made of Pacific data outside of the few groups by identifying and managing the associated risks. that have access to it, and Pacific countries are often Strengthening the ties and coordination between excluded from external databases and analyses. government bodies responsible for infrastructure Increased data accessibility is one of the objectives provision and building codes and those responsible of the Ten Year Pacific Statistics Strategy, and this for disaster risk management and climate change accessibility should go beyond summary statistics. adaptation is one important step in this area. In addition, given the importance of land in the social, In addition, active communication about the cultural, and economic life of people in the Pacific, meaning and implications of poverty measurement governments should aim to support good land is needed to get hardship on the policy agenda. management, including enabling communities to use Poverty measurement has been carried out using it as a resource to manage the risks they face. household survey data in many Pacific countries. However, government and the general public are Development partner activities should also factor in often unfamiliar with the analysis and its implications, risks and seek to reduce rather than add to volatility. so the information receives limited consideration in Flexible funding arrangements that are responsive to policy making. Active communication led by national changed expenditure priorities in the light of major statistics offices and supported by development shocks can help. PIC governments with stretched partners is needed to help clarify what poverty implementation capacity would also benefit from a measurement actually measures and to increase focus on the practicalities of implementation support awareness of its implications about hardship. following natural disasters. Enable Households and Communities Invest in Data Identify the important roles of governments and Governments and their development partners should development partners in supporting the efforts of invest in fielding regular household surveys for all households and communities to reduce hardship countries. Some countries have conducted two or and vulnerability: In particular, selective investments Executive Summary | ix in social programs are needed to provide safety nets to do their work more productively, whatever their and relief from hardship. Governments in the Pacific work may be, and therefore earn more. Given the are also well placed to effectively build human capital challenges to private sector growth in the Pacific, by improving the quality of education that often is education that is “portable” is likely to be the most already publically provided and by creating productive valuable. For example, governments can work with economic opportunities. In addition, governments development partners (who often receive migrants have an important role to play in increasing people’s as well) to adopt education qualifications that will access to a broader set of risk management tools. be recognized in major migrant-receiving countries. Simultaneously, regulatory barriers that restrict Strongly consider expanding the role of government opportunities for migration should be removed to and development partners in social protection provide greater numbers of people in the Pacific the programs: Households and communities in the opportunity to access overseas labor markets. Pacific are not able to fully manage the risks they face or to eliminate hardship. In particular, traditional Foster more opportunities for productive work at networks do not reach many households experiencing home, but only where it makes economic sense: the deepest hardship and appear to provide only Jobs and increased income not only raise living partial insurance to households suffering shocks. In standards in good times but also enable households addition, traditional networks cannot manage local or to better manage the risks they face. Given the unique aggregate shocks that affect most of their members. challenges in the Pacific, realistic expectations about These findings make it clear that Pacific governments, the potential for job growth led by the private sector and their development partners, need to consider are needed. While remaining supportive of viable an expanded role in social protection that takes into economic growth sectors, expenditure should be account traditional networks and is mindful of fiscal carefully focused on investments with high expected and capacity constraints. economic returns and relatively low risks, given the limited means of most Pacific governments. Increase the quality, accessibility, and portability Development partners can play a role in financing of education and create more opportunities for these investments and in sharing experiences on what migration: High-quality education equips people works from other parts of the world. Concepts and Context 1 Chapter 1 Concepts and Context Basic Concepts to obtain basic services. The characteristics of people experiencing hardship will be explored in chapter 2. Hardship and vulnerability are related, but distinct, concepts: Hardship is about having low current People in the Pacific are vulnerable to many shocks, well-being, and vulnerability is about expectations meaning they face high risks to well-being. Shocks of reductions in future well-being. People can be are occurrences, often difficult to predict, that have said to experience hardship based on many different positive or negative impacts on well-being. Risks are measures, such as lacking access to services or the expected negative impacts from shocks, which living in low-quality housing. A common measure of depend on both the likelihood of shocks occurring hardship around the world, which this report applies and the impacts they have if they do occur. On the to the Pacific, is the inability to meet the basic needs other hand, opportunities are the expected positive of life as measured by consumption. On the other impacts that some shocks can have (World Bank hand, vulnerability is based on expectations about the 2013a). This report focuses on risks but accounts for future. Specifically, vulnerable people face high risk, the role of opportunities when possible. Experiencing or a high probability, of a reduction in their well-being hardship in the present may increase vulnerability by in the future, possibly to the point of experiencing limiting the ability to avoid negative shocks or guard hardship or deepening existing hardship. against their impacts. Table 1.1 provides explicit definitions for these concepts, which will be utilized Many people in the Pacific experience hardship, for repeatedly throughout this report. Because it deals many different reasons. Global or national factors, with the uncertain future, vulnerability cannot be such as commodity prices or natural resource easily measured without data that follow the same endowments, can affect hardship and are discussed individuals or households over time. This type of in the next section. In addition, local, household, data, also representative at the national level, does and individual characteristics are often important not exist in the Pacific. In chapter 3, proxies for determinants of who faces difficulties in meeting the vulnerability are used instead, including the frequency basic needs of life. For example, people with limited of different types of shocks among households and education may be unable to earn a living, whereas the magnitudes of the impacts of shocks. people who live in remote, rural areas may be unable Table 1.1: Definitions of Key Concepts Term Definition Hardship Low level of current well-being Vulnerability Facing high risks to future well-being Risk Expected negative impact of shock = probability of shock occurring × impact given occurrence Opportunity Expected positive impact of shock Shock Events or occurrences that are difficult to predict Uncertainty Lack of knowledge about future outcomes (including shocks, risks, and opportunities) Source: Adapted from World Bank 2013a. 2 | HARDSHIP AND VULNERABILITY IN THE PACIFIC ISLAND COUNTRIES Table 1.2: Shocks Defined by Source and Level of Covariance Covariance Idiosyncratic Local Aggregate Source (Individual or Household) (Community or Region) (National or International) Harvest failures, floods, Natural Death of livestock Cyclones, earthquakes, droughts landslides Commodity price shock, foreign aid Economic Job loss Sector collapse reduction Health Injury, illness Epidemic Pandemic Sociopolitical or Theft, abandonment Community disputes Ethnic conflict, political violence, coups personal Source: Adapted from World Bank 2012a. Shocks are defined by four main sources (what source and level of covariance, these characteristics causes them): natural, economic, health, and determine both the risks associated with shocks sociopolitical. Natural shocks include disaster and the best approaches to managing those risks. events (such as volcanic eruptions and tsunamis), Even among shocks from the same source and at ongoing environmental damage (such as erosion and the same level of covariance, large differences can salination), and other natural events that are difficult be found in the rate of onset, intensity of impacts, to predict (such as pestilence and excessive rainfall). or expected probability of occurring. For example, Economic shocks include unpredictable changes a transient illness striking a household member in employment, income streams, prices, and other may be a health shock that occurs fairly frequently factors. Health shocks are equally wide-ranging and in a household but with relatively small impacts, include injury, illness, and death. Finally, sociopolitical while a permanently disabling injury to a household shocks include theft, violence, and ethnic conflict. member is a health shock that is more rare but also more severe. All these characteristics determine the Shocks are also defined by their level of covariance risks to well-being that shocks present, but in many (how widespread their impacts are): from idiosyncratic, cases, substantial uncertainty surrounds shocks and to local, to aggregate.1 Idiosyncratic shocks strike their associated risks. Such uncertainty can limit the particular individuals or households, but not entire attempts individuals (or communities or countries) communities or countries at the same time. The make to address risks, as well as the effectiveness death of a household member, the loss of a job, and of any attempts that are made, thereby increasing the failure of a household’s crop are all examples of vulnerability. A systematic approach to managing risks idiosyncratic shocks. Local shocks, such as flooding effectively that begins with increasing knowledge is or community disputes, strike entire communities at discussed later in this chapter. the same time. Aggregate shocks have the broadest reach, striking many communities, islands, or countries Hardship and vulnerability are closely linked concepts, simultaneously. Some idiosyncratic and local shocks and experiencing hardship can increase vulnerability. can expand to become national, for example, an Households experiencing hardship often do not illness that first strikes only a few households, but have the resources to effectively prepare for shocks then spreads, or a community dispute that grows into or may be forced to make decisions that increase broader civil unrest. Table 1.2 presents a typology of their vulnerability. For example, some households shocks, by both source and level of covariance, and in hardship may consume everything that they earn provides additional examples of each type. Because of or produce, leaving little or nothing for savings that data limitations, the evidence presented in chapter 3 could be used in response to future negative shocks. focuses on idiosyncratic and aggregate shocks. Others may decide to live on unsafe land because it is the most affordable option: In Fiji, for example, the Shocks also differ in their rate of onset, intensity of least well-off migrants to urban areas tend to settle on impacts, and probability of occurring. Along with some of the most undesirable lands—waste dumping sites, flood-prone areas, and unstable hillsides (Mohanty 2006). 1. More specifically, the covariance of a shock can be defined as how closely related the probability of a shock impacting a single individual or household is to the probability that Vulnerability can also increase hardship. Vulnerable the same shock impacts other individuals, households, or households face substantial negative impacts when communities. shocks occur, such as the loss of assets or access Concepts and Context | 3 to income production. These impacts can push centers of economic activity and from each other (see households deep into hardship that can be difficult to figure 1.2). For example, the 100,000 or so inhabitants recover from. In addition, actions taken by vulnerable of Kiribati are scattered across islands that are spread households to manage risk can indirectly increase over an area larger than India. In addition to isolation, hardship. For example, many households around the location and topography of many Pacific islands the world without access to insurance try to reduce create both risks, in the form of natural disasters, and the risks to their incomes, by choosing safer but less opportunities, in the form of natural resources. productive crops, or by diversifying into many income- generating activities, which can lower not only risk but Smallness and isolation contribute to hardship and also opportunity (Morduch 2004). After shocks occur, vulnerability by limiting the size and diversity of the the coping responses of households with few options private sector. Private enterprises face significant can also increase hardship, for example, if households challenges in achieving economies of scale, have to draw down on their productive assets like specialization, and innovation needed to grow, livestock or reduce their investments in human capital. because domestic supply chains, labor markets, and consumer demand are limited (World Bank 2013b). Geographic isolation exacerbates these challenges: The movement of goods and people across long A Unique Set of Features in distances is costly and thus makes most production the Pacific Shape Hardship in Pacific island countries uncompetitive in world markets (Winters and Martins 2004). Therefore, jobs in and Vulnerability the private sector, and the opportunities they provide Geographic features common to most Pacific island to reduce hardship, are limited (see figure 1.3). In countries (PICs), including small size, large distances, addition, these constraints on private sector growth and topography, contribute to the hardship and can lead to one or two sectors dominating the vulnerability of people in several ways (see figure 1.1). economy, making households vulnerable to downturns With the exception of Papua New Guinea, PICs in these sectors. For example, in Palau, where tourism all have populations under 1 million, with many accounts for almost half of gross domestic product populations numbering in the tens of thousands. (GDP), the 2008 bankruptcy of one Taiwanese airline These countries are scattered across the equivalent caused a 17 percent contraction in tourism between of over 15 percent of the earth’s surface, far from 2008 and 2010 (Colmer and Wood 2012). Figure 1.1 Distribution of PICs 4 | HARDSHIP AND VULNERABILITY IN THE PACIFIC ISLAND COUNTRIES Figure 1.2 Size and Distance from Markets of PICs versus Other Countries Average distance from markets (km) 16,000 14,000 12,000 10,000 All countries Pacific 8,000 6,000 4,000 9 11 13 15 17 19 21 23 Log population Source: Cororaton and Knight 2013. Figure 1.3 Working-Age Population by Sector of Employment in Kiribati, Solomon Islands, and Vanuatu Kiribati Public sector Solomon Islands Other formal employment Vanuatu Rest of working age population 0% 20% 40% 60% 80% 100% Source: World Bank 2013b. Economic openness, while needed to mitigate the and indirectly through the transport costs that make effects of smallness and geography, contributes to up a large part of import and export costs. vulnerability by exposing Pacific islanders to the vicissitudes of global markets. This vulnerability is Smallness and distance also affect hardship and most pronounced in the coral atoll islands, including vulnerability through their impacts on the public Kiribati, the Marshall Islands, and Tuvalu, where sector. In small countries, the functions of government limited land mass and underdevelopment constrain (including administration, service delivery, and agricultural productivity. Households are consequently infrastructure) are provided to small populations and dependent on imports of basic foods and vulnerable paid for by a small number of taxpayers, making to rising and volatile global prices. In these countries, it difficult to realize economies of scale (World the ratio of food imports to GDP is three to five Bank 2011). The dispersion of populations across times the global average for developing countries large swathes of ocean (as in Kiribati) or across difficult (Cororaton and Knight 2013). Households in most PICs terrain (as in Papua New Guinea) increases expenses are also highly vulnerable to global energy prices, even further. In addition, the specialized capacities both directly through their reliance on fuel imports required for administration are relatively scarce Concepts and Context | 5 because of small population sizes and may be difficult a shock. At the same time, the existence of such to obtain in part because of geography. Weakness programs can reduce the hardship-perpetuating in public provision of services and infrastructure impacts of vulnerability, by reducing the need for contribute to hardship by limiting the ability of people to cope in destructive ways. Examples of household members to build their human capital such programs, established well ahead of crises and (through schooling and health care) and by limiting effectively serving both roles, are starting to emerge households’ access to economically productive around the world. infrastructure like roads and telecommunications. Public sector efforts to reduce risks and manage A systematic approach to risk management is opportunities are often weak, increasing the needed to effectively reduce the vulnerability of vulnerability of households. Pacific islanders. As detailed in the preceding sections, Pacific peoples are vulnerable to many Location and topography make many Pacific islanders risks. Every risk cannot be removed, so how can vulnerable to natural shocks. Several PICs are located households, governments, and international partners along the Pacific Ring of Fire, the area along the edge decide which risks to address and how? The 2014 of the Pacific Ocean where tectonic plates collide, or in World Development Report, Risk and Opportunity: areas of high typhoon activity. Eight of the 20 countries Managing Risk for Development, provides a in the world with the highest annual average losses (as framework for risk management consisting of four a share of GDP) from disasters are PICs (World Bank components: knowledge, insurance, protection, 2012b). Many of the coral atoll islands are also facing and coping (figure 1.4). The characteristics of the existential threats from the long-run shocks of rising sea shocks experienced and the relative costs and levels and ocean acidification. benefits of these components, as well as how they interact, all help determine the best approach to risk On the other hand, geography (natural resources) management. Each component of the framework is combined with traditional practices helps to described in detail below. limit extreme hardship, while offering significant opportunities in some PICs. Most Pacific islanders Knowledge, of both the shocks that may occur and have access to and practice subsistence agriculture ways to reduce the associated risks, is the critical first or aquaculture, obtaining at least some of their step toward reducing vulnerability, but it is not enough. households’ food from cultivated gardens, wild The first barrier to managing risks is uncertainty: vegetation, and the ocean. Coupled with traditions of Without knowledge about the types of shocks that are communal support, these resources help to limit the possible and the nature of the risks associated with prevalence of hunger in the Pacific, although nutrition these shocks, little effective action can be taken. In and obesity are significant problems (WHO 2010). In addition, knowledge of measures to reduce risks and addition to food resources for household use, some the capacity to assess and implement these measures PICs are endowed with large amounts of resources. is needed. For example, across the Pacific, only about For example, Papua New Guinea is rich in natural gas half of households have access to improved (protected and minerals, and the waters of Kiribati’s exclusive and sanitized) drinking water, putting populations at economic zone provide valuable fishing license risk for a range of diseases and even death, particularly revenues. However, governments still face challenges in for young children (WHO 2008). Providing households effectively realizing the benefits of natural endowments with information about the dangers of contaminated while minimizing damage to local livelihoods. water and the ways in which they can manage the risks is important, but research from around the world has found that even with this knowledge, households often do not take action. Costs and behavioral factors, Reducing Hardship and such as the effort associated with remembering to Vulnerability through Risk take extra steps to purify water, can have large effects on whether or not households utilize their knowledge Management (Ahuja, Kremer, and Zwane 2010). Because hardship and vulnerability are closely linked, reducing one can reduce the other. For example, Based on the right knowledge, protection measures targeted programs that provide support or services reduce risks by lowering the probability that negative for those experiencing hardship also can reduce impacts from shocks occur. Protection can take many vulnerability by serving as safety nets, expanding to forms, depending on the types of shocks that are serve individuals or entire communities experiencing faced. The probability of some shocks occurring, 6 | HARDSHIP AND VULNERABILITY IN THE PACIFIC ISLAND COUNTRIES Figure 1.4 Key Components of Risk Management KNOWLEDGE of shocks, vulnerability, EX ANTE potential impacts INSURANCE PROTECTION to reduce the size of to reduce probability negative impacts from that negative impacts shocks that do occur from shocks occur COPING to recover from EX POST negative impacts of shocks Source: Adapted from World Bank 2013a. such as cyclones or global commodity price spikes, different probabilities of a shock occurring reduces cannot be directly changed by protection measures. the overall risks and makes insurance less costly. For However, the probability that these shocks will cause example, small island nations may not be able to negative impacts can be reduced: Wind-resistant afford disaster insurance individually, but pooling construction and early warning systems can lessen the initiatives in which nations join together to purchase probability of damage and loss of life from cyclones, regional insurance have shown early success in the and purchasing agreements and domestic agricultural Caribbean and promise in the Pacific.2 production can limit the impacts of commodity price spikes. For other shocks, the probability of occurrence After a shock occurs, coping encompasses the actions can be directly reduced through protection. For that are taken in response. Coping can include many example, appropriate chlorination of drinking water, different actions, such as taking out loans or increasing which has low financial costs, drastically reduces the amount of time spent working. Such actions are bacterial contamination and illnesses. common and can be relatively benign. However, in many cases detrimental coping actions must be taken Insurance, also implemented prior to shocks because households were unable to manage risks occurring, reduces the size of negative impacts ex ante. These actions, such as selling productive from shocks by transferring resources across time or assets or taking out high-interest loans, can increase people. Two forms of insurance exist, self-insurance hardship in the long run. and pooling, and the optimal use of each depends on many factors, including the types of shocks faced. The optimal mix of components depends on the Self-insurance means transferring one’s own resources characteristics of the shocks experienced. The over time (through savings or borrowing) to the probability of shocks occurring, the intensity of impacts time after a shock occurs. Households self-insure when they do occur, and the level of covariance are by accumulating cash savings, storing food, having important for determining the best risk management lines of credit, and a range of other actions that will approach (Ehrlich and Becker 1972; Gill and Ilahi help them reduce the negative impacts of shocks. On the other hand, pooling means sharing risks with others, and it can occur at multiple levels, from 2. The Caribbean Catastrophe Risk Insurance Facility informal pooling within a family network to formal made its first payouts in 2007 and currently has 16 member insurance purchased in the private market. Pooling countries. The Pacific Disaster Risk Financing Program pilot across households, communities, or nations that face was launched in January 2013. Concepts and Context | 7 2000). For example, coping may be the best response remittances do help to reduce inequality in urban to rare shocks with minimal impacts, because the Papua New Guinea and Tonga, but not in rural Papua costs of additional risk management would outweigh New Guinea. Gibson also finds some evidence that the benefit of reducing the small risks. On the other transfers act as insurance: In Papua New Guinea, net hand, likely shocks with large impacts may be too transfers increase when households experience a costly to insure through market pooling insurance. loss of cash income or a birth. Interwoven with these Many other factors are also critical to determining traditional networks, churches are at the heart of the best approach, including the availability, costs, many communities across the Pacific: Many churches and characteristics of each component and how they provide various forms of support to their members interact with each other. For example, protection but also require substantial commitments of time and is complementary to insurance purchased in the money (Barker 1996). private market when the price of insurance responds to protection measures. A common instance of this However, traditional networks may be less effective in is private automobile insurance providers offering responding to local or aggregate shocks. When many reduced prices for those with alarm systems or other households within the same network are affected by safety features in their cars. the same shock, such as a community experiencing disputes, the ability of households to assist one Shock characteristics also help determine who another can be significantly reduced. When Tuvalu should take the lead in risk management. Although suffered a drought in 2011, for example, international households actively manage the idiosyncratic risks partners had to airlift in desalination equipment and they face, large, covarying shocks may require water. International migration can increase the ability government or international assistance to manage, of these networks to respond to covariate shocks by because households and communities often can diversifying the shocks to which network members do little prevention and cannot afford to insure are exposed. For example, after the 2009 tsunami in themselves. Even when shocks do not covary, a role Samoa, remittances surged by more than $10 million is often still found for the broader community and above their historical levels (Gibson 2010). government to play because of many challenges to risk management, including those discussed above.3 In addition, traditional networks and churches may also place burdens on people that perpetuate hardship, and shifts toward urbanization and Pacific Approaches to Reducing monetization may leave some households without support. The traditional systems in the Pacific require Hardship and Vulnerability exchange and contribution for many reasons other In the Pacific, extended kinship networks play the than helping the less well-off, including for ceremonial largest role in supporting people experiencing events and support of communal activities. In hardship and are believed to both help equalize qualitative participatory assessments conducted welfare and act as insurance. Traditional expectations across dozens of Pacific communities in the early of reciprocity and generosity in the exchange of food 2000s, participants in Samoa, Tonga, and Tuvalu and other items, culturally specified and formalized cited the burdens of meeting community and church in different ways across the Pacific, help to minimize obligations as one of the top causes of hardship hunger and severe deprivation (Ratuva 2010). Few (Abbott and Pollard 2004). In addition to these quantitative studies of the effectiveness of traditional pressures, migration out of rural areas into cities and systems have been conducted because of lack of overseas may be weakening traditional ties, while data. In one of the few, Gibson (2006) utilizes data increasing use of cash for purchasing necessities from Papua New Guinea and Tonga in the 1980s and reduces the role of food and material exchange in 1990s and finds that interhousehold transfers and reducing hardship. Most Pacific governments are not involved in direct transfer programs to the less well-off but focus on 3. This report focuses on the roles of households, communities, governments, and international partners in providing free or low-cost social services to the managing risks. The 2014 World Development Report (World population. Governments are often reluctant to Bank 2013a) reviews the roles of all these actors, as well as become involved in direct transfer programs to the enterprise sector and the financial system, and provides those experiencing hardship, for fear of undermining additional guidance on who is best placed to manage traditional systems. In addition, the egalitarian ethos different types of risks. of most Pacific societies makes means-targeted 8 | HARDSHIP AND VULNERABILITY IN THE PACIFIC ISLAND COUNTRIES programs unattractive to policy makers and viewed Pacific Island Economies.” Working Paper 1, Pasifika as potentially unfair. As a consequence, most Pacific Interactions Project, Wellington, New Zealand. governments focus on providing free or low-cost ———. 2010. “Recent Shocks and Long-Term Change health, education, and other services to the broader in the Samoan Economy.” Pacific Economic Bulletin population. The quality and accessibility of these 25 (3): 7–23. public services vary across the Pacific, and private Gill, Indermit, and Nadeem Ilahi. 2000. “Economic alternatives may be of better or worse quality, if Insecurity, Individual Behavior, and Social Policy.” they exist. For example, in Kiribati, the government World Bank, Washington, DC. provides free education through the junior secondary level, but the eight church-run schools offering junior Mohanty, Manoranjan. 2006. “Squatters, Vulnerability, secondary education are generally considered of and Adaptability of Urban Poor in a Small Island higher quality than the 24 government-run junior Developing State: The Context of Fiji Islands.” secondary schools. Families therefore often make University of the South Pacific, Suva, Fiji. concerted efforts to place their children in the church Morduch, Jonathan. 2004. “Poverty and Vulnerability.” schools, such as finding the cash needed to pay American Economic Review 84 (2): 221–25. fees and sending children to live with relatives on Ratuva, Steven. 2010. “Back to Basics: Towards Tarawa, where most church-run schools are located Integrated Social Protection for Vulnerable Groups (Republic of Kiribati MOE 2011). Chapter 4 will in Vanuatu.” Pacific Economic Bulletin 25 (3): 40–63. describe in further detail both current approaches to Republic of Kiribati Ministry of Education. 2011. Digest hardship reduction as well as what is known about risk of Education Statistics 2011. South Tarawa. management in the Pacific. WHO (World Health Organization). 2008. Sanitation, Hygiene, and Drinking Water in the Pacific Island Countries: Converting Commitment into Action. References Geneva: WHO. Abbott, David, and Steve Pollard. 2004. Hardship and ———. 2010. “Pacific Islands Pay Heavy Price for Poverty in the Pacific. Manila: Asian Development Abandoning Traditional Diet.” Bulletin of the World Bank. Health Organization 88: 484–85. Ahuja, Amrita, Michael Kremer, and Alix Peterson Winters, L. Allen, and Pedro Martins. 2004. “When Zwane. 2010. “Providing Safe Water: Evidence Comparative Advantage Is Not Enough: Business from Randomized Evaluations.” Annual Review of Costs in Small Remote Economies.” World Trade Resource Economics 2: 237–56. Review 3 (3): 347–83. Barker, John. 1996. “Winds of Change (Review).” World Bank. 2011. Discussion Note: Pacific Futures. Contemporary Pacific 8 (1): 234–36. Sydney: World Bank. Colmer, Patrick, and Richard Wood. 2012. “Major ———. 2012a. Addressing Vulnerability in East Asia. Economic Shocks and Pacific Island Countries.” Washington, DC: World Bank. Paper presented at the High-Level Conference on ———. 2012b. Acting Today for Tomorrow: A Pacific Island Countries, Apia, Samoa. Policy and Practice Note for Climate and Disaster Cororaton, Caesar, and David Knight. 2013. Resilient Development in the Pacific Islands Region. “Economic Shocks and Their Impacts on Washington, DC: World Bank. Households: Micro Simulation Results from Three ———. 2013a. World Development Report 2014: Risk Pacific Island Countries.” Background paper for this and Opportunity—Managing Risk for Development. report. Washington, DC: World Bank. Ehrlich, Isaac, and Gary Becker. 1972. “Market ———. 2013b. At Work in East Asia Pacific. Insurance, Self-Insurance, and Self-Protection.” Washington, DC: World Bank. Journal of Political Economy 80 (4): 623–48. Gibson, John. 2006. “Are There Holes in the Safety Net? Remittances and Inter-household Transfers in A Snapshot of Hardship in the Pacific  9 Chapter 2 A Snapshot of Hardship in the Pacific Background is that extreme deprivation is not widespread in the Pacific, and the data presented in this chapter support Across the Pacific, many people are unable to meet this. In addition, the existence of such deprivation their basic needs. The purpose of this chapter is to would imply the failure of the traditional networks shed light on the well-being of people in the Pacific of support that exist across the Pacific, making it and the incidence of and characteristics related to a culturally sensitive topic. In addition, networks being unable to meet basic needs. Over the past are considered to still be strong and active in most decade, most PICs have conducted household surveys countries. and used the data to analyze the well-being of their populations. However, little attempt has been made Finally, the notion of measuring well-being in to consider this analysis and its implications in a monetary terms through consumption or income is comparable manner across countries. still relatively foreign because a significant share of the population still relies on self-produced or procured “Hardship” is the preferred concept in the resources from land (often communal) and sea. In Pacific, because “poverty” is considered culturally addition, interhousehold transfers of goods and inappropriate in several countries. Hardship refers services are widespread. However, the share of cash to the same condition that is called poverty in other transactions has been rising, driven by urbanization regions: having a low level of current well-being or, and the need to pay for things such as school fees and more specifically, being unable to meet one’s basic telephones. This renders the population increasingly needs. However, the term “poverty” is not well accepting of the notion of “monetary” poverty.1 accepted by many governments and societies in the Pacific, for three main reasons detailed below. This chapter is structured as follows. The second section discusses the results of the poverty First, understanding is often limited on the part measurement carried out in several PICs, focusing on of policy makers and the public of the meaning of the use of consumption as the measure of welfare and poverty measures. Across PICs, the analysis carried the setting of poverty lines.2 The next section presents out using household survey data has established the hardship rates resulting from these analyses, poverty lines and poverty rates, but these results have and the fourth section considers how hardship rates had limited acceptance or impact on policy making. are related to various factors within countries. The Poverty measurement is technically demanding, fifth section analyzes the level of inequality within and the details of the analysis not easily accessible to nonexperts. National statistical offices that produce the measures (often with the assistance of development partners) have limited resources to 1. Although hardship is a multidimensional concept, the invest in the communication that would be needed to emphasis for this report is on monetary measures of hardship familiarize both policy makers and the public with the because they are increasingly relevant and yet poorly measures and their meaning. understood and underutilized in the Pacific at present. 2. Published reports containing the results of this poverty Second, “poverty” in many PICs has been associated measurement are not all explicitly cited in this chapter but with hunger or extreme deprivation. Received wisdom are included in the reference list of this chapter. 10 | HARDSHIP AND VULNERABILITY IN THE PACIFIC ISLAND COUNTRIES Figure 2.1 Daily Food Poverty Lines 2.5 2.0 2.0 Food poverty line, 2005 PPP ($) 2.0 1.9 PPP $2 1.5 1.5 1.4 1.1 1.0 PPP $1.25 0.5 0.0 Solomon Fiji 2009 Papua New Kiribati Samoa Vanuatu Islands Guinea 2006 2008 2010 2006 2010 Sources: World Bank staff estimates based on Household Income and Expenditure Surveys (HIESs) and existing poverty analysis (SINSO and UNDP 2008; KNSO and UNDP 2010; SBS and UNDP 2010). Note: PPP = purchasing power parity. countries, and the final one concludes with a summary The daily values of food poverty lines vary substantially of the key findings. across countries but are mostly above the extreme poverty line of $1.25 used for international comparisons. To make international comparisons, Poverty Measurement national poverty lines in local currencies can be converted to purchasing power parity (PPP) dollar in the Pacific terms for each country for which a PPP conversion Poverty measurement in PICs has been based on the factor has been calculated. These conversion factors Cost of Basic Needs approach. The Cost of Basic are from 2005 and are regression-based estimates, Needs approach is a commonly used method that rather than price-based, in all PICs except Fiji (ICP attempts to define the minimum resources needed 2005).4 Across countries included in this analysis, the for long-term physical well-being, usually in terms of Solomon Islands has the lowest food poverty line, consumption (Haughton and Khandker 2009). Using slightly below the international line of $1.25 per day this approach, a poverty line is defined as the amount in PPP terms (see figure 2.1). Other PICs set their of spending required to obtain those resources. A daily food poverty lines above this level, with Samoa list of “basic needs” defines the minimum resources and Vanuatu having the highest lines at $2.20 and and consists of food and nonfood (clothing, shelter, $2.30, respectively. This pattern is consistent with the services) items.3 The poverty line reflecting food needs expectation that lower income countries would have is called a food or “extreme” poverty line. The poverty lower costs for similar amounts of food. line reflecting both food and nonfood needs is called a total poverty line. Daily values of total poverty lines also vary substantially. Figure 2.2 shows that the value of basic 3. It is this list of items and their treatment (i.e., how they are aggregated) in household surveys that often differs greatly 4. Some PICs, including Tuvalu, do not have PPP conversion across countries and makes consistent comparisons across factor estimates and are therefore not included in this countries challenging. section. A Snapshot of Hardship in the Pacific | 11 Figure 2.2 Daily Total Poverty Lines 4.0 Total poverty line, 2005 PPP ($) 3.5 3.3 3.4 3.0 3.1 3.0 2.5 2.0 2.0 PPP $2 1.5 1.5 1.0 PPP $1.25 0.5 0.0 Solomon Papua New Vanuatu Fiji 2009 Samoa Kiribati Islands Guinea 2010 2008 2006 2006 2010 Sources: World Bank staff estimates based on HIESs and existing poverty analysis (SINSO and UNDP 2008; KNSO and UNDP 2010; SBS and UNDP 2010). Note: PPP = purchasing power parity. needs poverty lines range from the Solomon Islands across countries. Additional detail on these differences at $1.50 up to Samoa and Kiribati at $3.30–3.40. These is provided in the technical annex to this chapter. lines were derived using a similar methodology across countries: The cost of essential nonfood needs is estimated using a reference group of households, and The Incidence of Hardship this cost is added to the value of the food poverty line. Across the countries analyzed, food costs constitute Based on National Poverty Lines about 60 to 70 percent of the total costs required to The incidence of food hardship ranges from 2 percent meet basic needs. Fiji, at 46 percent, is substantially of the population in Vanuatu to 27 percent in Papua lower than the other countries, possibly because it New Guinea. Using the national food poverty lines is more urbanized, and food costs as a share of total described in the previous section, 5 percent or less costs tend to shrink in urban areas. Overall wealth also of the population is living below these lines in Fiji, plays a role, with a lower share of food costs in total Samoa, and Vanuatu. These countries also have costs an outcome of Fiji’s higher average income. relatively high average consumption compared to other countries in the region. On the other hand, in What do the poverty lines measure? The Papua New Guinea and the Solomon Islands, where methodological approach to defining both food and average consumption is relatively low, the percentages basic needs poverty lines is similar across the Pacific, of population living below the food poverty lines but important differences are found. The definition of are substantially higher. Figure 2.3 illustrates this a sufficient amount of energy, the group of households relationship, with an additional dimension: Countries used to determine typical food and nonfood needs, with higher average consumption set their food and the prices used to determine the costs of food poverty lines at higher levels and also have relatively and nonfood needs all vary substantially. For example, low shares of their populations living below the food the definition of a sufficient amount of energy ranges poverty lines. from 2,100 calories per person in Samoa, the Solomon Islands, and Vanuatu, to 2,200 calories per adult in The incidence of total hardship, including food and Papua New Guinea. These types of differences exist other basic needs, is high in most PICs. The share of across all countries in the world but should be kept in the population unable to meet their total basic needs, mind when comparing the values of the poverty lines including food and nonfood, is over 20 percent in all 12 | HARDSHIP AND VULNERABILITY IN THE PACIFIC ISLAND COUNTRIES Figure 2.3 Food Hardship Rates in Selected PICs, Percentage of Population 350.0 Mean consumption (monthly, PAE), 2005 PPP ($) 300.0 Samoa 2008 1.1% Vanuatu 2010 2.2% 250.0 Fiji 2009 4.5% Kiribati 2006 4.9% 200.0 Solomon Islands 2006 150.0 10.6% Papua New Guinea 2010 26.5% 100.0 50.0 — 10.0 20.0 30.0 40.0 50.0 60.0 70.0 80.0 Food poverty line (monthly), 2005 PPP in ($) Sources: World Bank staff estimates based on HIESs and existing poverty analysis (SINSO and UNDP 2008; KNSO and UNDP 2010; SBS and UNDP 2010). Note: Bubble size = the share of the population below the food poverty line. PAE = per adult equivalent; PPP = purchasing power parity. the countries included in the analysis except Vanuatu. falling into hardship as a result of the idiosyncratic and Figure 2.4 highlights a similar relationship as seen with aggregate shocks that will be discussed in chapter 3. the food poverty lines: Countries with higher average consumption have total poverty lines set at higher In recent years, hardship has decreased in Fiji, but levels, but the shares of the population living below less is known about hardship over time in other the total poverty line are more similar to countries countries. Poverty measurement was carried out in with lower average consumption. This may reflect a similar fashion on the two most recent nationally in part the greater demands for nonfood expenses representative household surveys conducted in Fiji on households in more urbanized countries such as (2002/2003 and 2008/2009). This analysis shows that Fiji, where housing, school fees, and other costs are total hardship declined from nearly 40 percent of the important needs that require cash and may not easily population in 2003 to 35 percent in 2009 (World Bank be met through traditional methods. 2011). Before the 2010 survey in Papua New Guinea, a nationally representative household survey had not There may be a sizeable portion of the population been carried out since 1996. Little is known about how that is consuming only marginally more than those the well-being of the population changed during that in hardship, and they are especially vulnerable to 14-year interval, but the incidence of total hardship hardship in the future. The hardship rates illustrated in in Papua New Guinea in 2010 was 40 percent, figures 2.3 and 2.4 describe who is living in hardship almost equal to the incidence in 1996. In other PICs, at present, as defined by the national poverty lines. methodological questions make it less straightforward However, numerous people may be consuming to compare hardship rates over time. Detail on these enough to only hover above the poverty line. Such questions is provided in the technical annex to this people are often among the most vulnerable to chapter. A Snapshot of Hardship in the Pacific | 13 Figure 2.4 Total Hardship Rates in Selected PICs, Percentage of Population 350.0 Samoa 2008 Mean consumption (monthly, PAE), 2005 PPP in ($) 13.4% 300.0 250.0 Vanuatu 2010 Kiribati 2006 12.7% 21.8% 200.0 Solomon Islands 2006 Fiji 2009 22.7% 35.2% 150.0 100.0 Papua New Guinea 2010 39.9% 50.0 — 0.0 20.0 40.0 60.0 80.0 100.0 120.0 140.0 Food poverty line (monthly), 2005 PPP in ($) Sources: World Bank staff estimates based on HIESs and existing poverty analysis (SINSO and UNDP 2008; KNSO and UNDP 2010; SBS and UNDP 2010). Note: Bubble size = the share of the population below the total poverty line. PAE = per adult equivalent; PPP = purchasing power parity. Hardship Differentials across characteristics that cause hardship. Characteristics are also considered one at a time when, in reality, they are Households likely to be related to each other, as well as to hardship. Within countries, whether or not a household As such, caution must be used in the interpretation of experiences hardship is influenced by various factors. these bivariate relationships. The following analysis presents a “hardship differential”5 that compares the incidence of total hardship in a Hardship Differentials across Places country to the incidence of total hardship among Households in rural areas of Fiji and Papua New households with a specific characteristic. This analysis Guinea are more likely to live in hardship than simply identifies household characteristics that correlate households in urban areas; the opposite is true in with hardship and does not attempt to identify other PICs, but methodology affects interpretation of these results. Figure 2.5 shows the hardship differential between urban and rural areas by country. In Papua New Guinea, the incidence of hardship in 5. Hardship differential: the difference between the national urban areas is 35 percent lower than the national incidence of total hardship (presented in the previous average, and in Fiji’s urban areas it is 26 percent lower section) and the incidence of total hardship among than the national average. A very different picture of households with a specific characteristic, expressed as a percentage difference. A percentage difference is used rural-urban differences emerges for other countries: to move the analysis away from differences in hardship In Kiribati, Vanuatu, and the Solomon Islands, the incidence between countries and focus instead on incidence of hardship in urban areas is substantially differences between households within countries, in a way higher than in rural areas. For example, in Vanuatu, that allows for comparison between countries. the incidence of hardship in urban areas is 62 percent 14 | HARDSHIP AND VULNERABILITY IN THE PACIFIC ISLAND COUNTRIES Figure 2.5 Urban-Rural Differences in Hardship Rates in Selected PICs 80 between urban/rural hardship rates and the Hardship differential (% change difference 60 62.1 national hardship rate) 40 34.3 Urban 25.0 Rural 20 15.0 13.1 5.0 4.8 0 –3.9 –5.5 –6.5 –10.9 –20 –21.9 –25.6 –31.5 –40 Fiji Papua Samoa Tuvalu Solomon Kiribati Vanuatu New Islands Guinea Sources: World Bank staff estimates based on HIESs and existing poverty analysis (SINSO and UNDP 2008; KNSO and UNDP 2010; SBS and UNDP 2010). Note: Hardship differential is the percentage change difference between urban/rural hardship rates and the national hardship rate. Refer to box 2.1 for a discussion of the methodological issues in data collection that may have influenced the urban/rural variation in hardship. Box 2.1 Measurement Matters: Reference Households and Spatial Differences in Hardship Another important consideration in poverty measurement is how to define a reference group of households to estimate the cost of basic nonfood needs. This estimate of nonfood needs is added to the cost of basic food needs to arrive at a total poverty line. Countries around the world define this reference group in different ways, but a key principle helps to guide this choice. When measuring poverty in a country, the poverty lines used should represent the same level of well- being across all the places being compared within that country (Ravallion 1992; Haughton and Khandker 2009).1 In many countries in the Pacific, including Kiribati, Samoa, the Solomon Islands, Tonga, Tuvalu, and Vanuatu, this principle does not hold. In these countries, the reference group of households is defined as all the households in the bottom 30 percent of the consumption distribution in every region of the country. This means that in less well-off regions, the consumption of the reference group of households represents a lower level of well-being than the consumption of the reference group of households in wealthier regions. In other words, the bottom 30 percent of households in a region with low consumption are less well-off than the bottom 30 percent of households in a region with high consumption. Note that this is different from the practice of setting different poverty lines across regions This methodological approach means that the incidence of hardship in rural areas, where consumption is usually lower, is likely estimated to be lower than it would be if the same level of well-being across regions were being measured. This should be kept in mind when considering the urban-rural differences and spatial variation in hardship presented in this chapter. In addition, using a relative measure affects the interpretation of hardship incidence over time, because as populations become better or worse off, the value of the poverty line will also automatically change. 1. This does not apply to relative measures of poverty, which focus on identifying households that are less well-off relative to other households. higher than the national average. Methodological household surveys conducted in recent years in differences across countries affect how these measures Fiji allow for comparisons to be made over time. should be interpreted, as detailed in box 2.1. Figure 2.6 shows that in 2003, the incidence of hardship among rural households was 12 percent In Fiji, the hardship differential between rural and higher than the national average. By 2009 the urban households is growing over time. The two differential had grown to 25 percent, because the A Snapshot of Hardship in the Pacific | 15 Figure 2.6 Change of Urban-Rural Difference over Time in Fiji 30 25.0 between urban/rural hardship rates and the Hardship Differential (% change difference 20 12.3 national hardship rate) 10 0 Urban Rural –10 –13.3 –20 –25.6 –30 2002–03 2008–09 Sources: World Bank staff estimates based on HIESs and existing poverty analysis (SINSO and UNDP 2008; KNSO and UNDP 2010; SBS and UNDP 2010). Note: Hardship differential is the percentage change difference between urban/rural hardship rates and the national hardship rate. Refer to box 2.1 for a discussion of the methodological issues in data collection that may have influenced the urban/rural variation in hardship. incidence of hardship stayed the same in rural areas hardship differentials described above. For example, but declined in urban areas (World Bank 2011).6 in Suva, the capital city of Fiji, the incidence of hardship is among the lowest in the country (map 2.1).7 In countries with rapidly growing urban centers, Conversely, in Vanuatu, where the incidence of the incidence of hardship is higher among urban hardship is higher in urban areas than in rural areas, households. In general, urbanization accompanies the capital city of Port Vila has one of the highest economic development and allows for scale incidences of hardship in the country (map 2.2).8 This economies and thicker markets, and increased depiction of the spatial dispersion of hardship in these specialization. However, urbanization that occurs as a PICs highlights the difficulties faced in delivering result of poor services in rural areas, food insecurity, services to the areas most in need. However, better or land shortages offers few economic benefits understanding this variation within a country is a and places people at increased risk of hardship in crucial first step in helping policy makers to better congested urban slums (World Bank 2013). Figure 2.7 target their resources. suggests that this has been the case for some PICs, Hardship and Characteristics including the Solomon Islands and Vanuatu. These of the Household Head countries are still primarily rural and offer limited economic opportunities in urban areas for those that Across countries, households with more educated migrate. household heads are less likely to live in hardship. Education increases human capital and individual Socioeconomic maps for Vanuatu and Fiji confirm the presence of a significant amount of spatial variation in the incidence of hardship within each country. The spatial dispersion of hardship pictured in these maps 7. However, the sample frame is likely not wholly provides detail in support of the urban and rural representative and undersamples those living in Suva. People living in squatter settlements around Suva, for example, may not have been adequately represented, leading to a degree of underreporting of hardship there. 6. This analysis does not reveal the fact that urban settlement 8. Although the comparison between Fiji and Vanuatu areas have higher hardship incidence than other urban areas. is illustrative, they are not strictly comparable in that the In fact, in those areas the incidence of hardship is often hardship levels are reported as outcomes of different national higher than in rural areas. poverty lines. 16 | HARDSHIP AND VULNERABILITY IN THE PACIFIC ISLAND COUNTRIES Figure 2.7 Urban Hardship versus Rural Hardship in PICs with Rapidly Growing Urban Centers 80 Vanuatu 60 Hardship Differential (% change difference between the urban hardship rate and the Solomon Islands 40 national hardship rate) Samoa 20 Kiribati Tuvalu 0 –20 Papua New Guinea Fiji –40 –60 –15 –10 –5 0 5 10 15 20 25 30 35 Urbanization rate of change (% change, 2002–12) Sources: World Bank staff estimates based on HIESs and existing poverty analysis (SINSO and UNDP 2008; KNSO and UNDP 2010; SBS and UNDP 2010). Note: Hardship differential is the percentage change difference between urban hardship rate and the national hardship rate. Bubble size = urban share of total population, 2012 (Fiji, the largest at 52.6 percent; Papua New Guinea, the smallest at 12.6). Map 2.1 Hardship Rates in Fiji: Hardship Head-Count Ratio at the Tikina Level, 2007 A Snapshot of Hardship in the Pacific | 17 Map 2.2 Areas with High Rates of Hardship in Vanuatu: Hardship Head-Count Ratio at the Area Council Level, 2010 Sources: Population and Housing Census 2009 and Household Survey 2010. Note: The median values for each category are 1.5, 4.3, 5.8, 10.2, 12.0, 16.0, and 24.9 percent, respectively. 18 | HARDSHIP AND VULNERABILITY IN THE PACIFIC ISLAND COUNTRIES Figure 2.8 Hardship by Level of Education 60 44.8 between the hardship rates by level of education 40 25.8 25.6 27.7 Hardship Differential (% change difference completed and the national hardship rate) 20 14.0 8.8 6.3 0.2 0 –4.2 –8.0 –10.9 –10.8 None/Primary –20 –18.7 –24.6 Secondary –30.2 –31.7 –40 –35.4 Tertiary –39.3 –60 –54.8 –70.8 –80 –74.8 –100 Kiribati Samoa Tuvalu Papua Solomon Vanuatu Fiji New Islands Guinea Sources: World Bank staff estimates based on HIESs and existing poverty analysis (SINSO and UNDP 2008; KNSO and UNDP 2010; SBS and UNDP 2010). Note: Hardship differential is the percentage change difference between hardship rates by level of education completed and the national hardship rate. productivity and in the Pacific is often required for New Guinea at 60 percent.9 Yet relatively few students access to formal sector jobs that pay well (World Bank continue to secondary and tertiary education. This 2012a). Figure 2.8 shows that each level of education is in part because of lack of access to secondary that the household head completes is associated with schools in rural areas and high costs of attendance. In decreasing hardship differentials, to varying degrees, addition, the low quality of primary education in most across countries. In particular, households with heads countries leads to high failure rates on secondary who have completed primary school or less are much school entrance exams and reduces the human more likely to live in hardship. For example, in Tuvalu, capital gained from attending school. For example, the incidence of hardship among these households is nearly 26 percent higher than the national incidence, whereas households with a tertiary-educated head are 9. Gross enrollment rate (GER) retrieved from the World Development Indicators. GER is the number of pupils 39 percent less likely to live in hardship. enrolled in a given level of education regardless of age expressed as a percentage of the population in the However, few household heads have completed theoretical age group for that level of education. The GER secondary or tertiary schooling. Across most may be greater than 100 percent when students younger or countries, primary school gross enrollment rates older than the official age for a given level of education are exceed 100 percent, with the exception of Papua enrolled in that level. A Snapshot of Hardship in the Pacific | 19 Figure 2.9 Hardship Level by Head of Household Employment Status 60 55.5 between hardship rates by labor status and 50 Hardship Differential (% change difference 45.2 40 the national hardsip rate) 30 20 18.6 17.0 10 Employed 7.2 Not Employed 0.9 0 –2.9 –4.1 –4.0 –4.9 –10 –13.3 –20 –21.4 –30 Vanuatu Tuvalu Fiji Papua New Kiribati Solomon Guinea Islands Sources: World Bank staff estimates based on HIESs and existing poverty analysis (SINSO and UNDP 2008; KNSO and UNDP 2010; SBS and UNDP 2010). Note: Hardship differential is the percentage change difference between hardship rates by labor status and the national hardship rate. in the Solomon Islands, a recent analysis found that households with a head who did not work in the only 22 percent of students in primary school are 30 days preceding the survey, the incidence of functionally literate (World Bank 2012c). hardship is 56 percent higher than the national incidence. Among household heads who do work, Households with a working household head are less hardship incidence varies by the type of work done. likely to live in hardship. Although the definitions In Fiji, households headed by informal sector workers of work differ, in all countries except the Solomon (unpaid family workers or self-employed workers) Islands the incidence of hardship is much higher have a hardship incidence 15 to 40 percent higher among households with a head who is not working than the national incidence. (figure 2.9). 10 For example, in Vanuatu, among Demographic Characteristics: 10. Definitions of “employed” are taken from the Household Gender, Age, and Exposure Survey questionnaires, with questions asking the following: to Hardship “Did you do any work in the last 7 days? yes or no” (Kiribati, Households headed by women are more likely to live Papua New Guinea, Tuvalu)/“Did you do any work in the last 30 days? yes or no (Vanuatu)/“Is anyone in this household in hardship in Kiribati and the Solomon Islands but currently working for pay in a job, business or profession? yes not in other countries. Across the Pacific, the share of or no, OR Did any member of the household receive regular households headed by women is typically very low: income from any of the following commercial activities from the Solomon Islands at 5.4 percent to Tuvalu at during the last 12 months?” (Samoa and Solomon Islands). 18.7 percent. With this in mind, figure 2.10 shows that 20 | HARDSHIP AND VULNERABILITY IN THE PACIFIC ISLAND COUNTRIES Figure 2.10 Hardness Level by Head of Household Gender 40 household head and the national hardship rate) between hardship rates by the gender of the 31.9 Hardship Differential (% change difference 30 22.0 20 10 4.3 1.7 1.6 0.9 1.5 0 Male –0.2 –1.8 Female –4.8 –5.5 –6.2 –10 –11.1 –20 –17.8 –30 Solomon Kiribati Vanuatu Samoa Fiji Papua Tuvalu Islands New Guinea Sources: World Bank staff estimates based on HIESs and existing poverty analysis (SINSO and UNDP 2008; KNSO and UNDP 2010; SBS and UNDP 2010). Note: Hardship differential is the percentage change difference between hardship rates by gender of head of household and the national hardship rate. female-headed households have a lower likelihood of elderly-headed households is about 77 percent and hardship in four countries and a higher likelihood in 66 percent, respectively, higher than the national two. Over time, the hardship differential for female- averages. The hardship differential is about half as headed households has fallen substantially in Fiji, large in Vanuatu, at 33 percent. These differences from 18.6 percent higher than the national average across countries are likely to be related to social in 2003 to 6.2 percent lower in 2009. These differing insurance systems for the elderly. Most PICs have relationships between gender of the household contribution-based pension schemes that cover only head and hardship across countries merit further the small number of people working in government or study, because they could be related to many factors. the formal private sector. This topic will be revisited in These factors include the existence of economic chapter 4. opportunities for women and rates of migration and remittance sending. The Effect of Household Size on the Hardship Differential Households headed by the elderly are more likely It is worrisome that households with more children to live in hardship. The share of households headed are more likely to live in hardship. As is observed in by a person age 65 and over varies substantially most countries around the world, larger households across countries, from 6 percent in Papua New in the Pacific are more likely to live in hardship. Of Guinea to 29 percent in Samoa. However, in all great concern is the fact across PICs, the incidence of countries, the incidence of hardship is substantially hardship is higher for households with more children. higher for elderly-headed households compared Figure 2.12 traces how the hardship differential with the national average. Figure 2.11 illustrates changes with each additional child for four of the the relationship between age of household head countries in the sample.12 In each of the countries, and hardship differential for three countries.11 In the incidence of hardship among households with no Kiribati and Tuvalu, the incidence of hardship among children is 17 to 50 percent lower than the national 11. Similar results are found for the other countries in the 12. Similar results are found for the other countries in the sample and have been removed from the figure for ease of sample and have been removed from the figure for ease of presentation. presentation. A Snapshot of Hardship in the Pacific | 21 Figure 2.11 Hardness Level by Head of Household Age 80 76.8 66.1 60 between hardship rates by household head Hardship Differential (% change difference age and the national hardship rate) 40 33.4 29.5 Vanuatu 20 Tuvalu Kiribati Fiji 0 –11.7 –14.1 –20 –16.9 –29.7 –40 25–34 35–44 45–54 55–64 65+ Household Head's age Sources: World Bank staff estimates based on HIESs and existing poverty analysis (SINSO and UNDP 2008; KNSO and UNDP 2010; SBS and UNDP 2010). Note: Hardship differential is the percentage change difference between hardship rates by age of head of household and the national hardship rate. Figure 2.12 Hardness Level by Number of Children in Household 80 between hardship rates by household head Hardship Differential (% change difference 61.3 age and the national hardship rate) 60 44.8 40 27.2 26.8 20 Fiji Vanuatu 0 Tuvalu Kiribati –16.9 –20 –21.3 –40 –42.5 –50.3 –60 no children 1 2 3 or more children Number of Children Sources: World Bank staff estimates based on HIESs and existing poverty analysis (SINSO and UNDP 2008; KNSO and UNDP 2010; SBS and UNDP 2010). Note: Hardship differential is the percentage change difference between hardship rates by number of children and the national hardship rate. 22 | HARDSHIP AND VULNERABILITY IN THE PACIFIC ISLAND COUNTRIES average. For households with three or more children, consumption shares of the second and even third the incidence of hardship is substantially higher than quintiles across countries suggests that people in the national average in all countries. these quintiles are vulnerable, in that negative shocks could push them into hardship in the future. Evidence High fertility rates and young populations mean that from other countries supports this possibility. For many households have large numbers of children. example, in Indonesia, the poverty rate in 2011 was Across the Pacific, 33 to 41 percent of country 12.5 percent, yet between 2008 and 2010, 25 percent populations are age 14 or younger. Fiji has the lowest of the population was below the poverty line in at fertility rate among the countries considered here, least one year (World Bank 2012b). at 2.6 children per woman (World Development Indicators, data from 2011). In other countries, fertility Gini coefficients for some PICs are relatively high rates are much higher, ranging from 3.5 in Kiribati compared to East Asian neighbors. 13 Figure 2.14 to 4.2 in the Solomon Islands (ibid.). If the observed displays the Gini coefficients of several countries, relationships continue to hold, large numbers of graphed against GDP per capita. East Asian children will continue to grow up in hardship. countries, other island nations, and countries known for particularly high or low inequality are also graphed for comparison. Among the PICs, the Solomon Inequality in the Pacific Island Islands, Papua New Guinea, and Fiji have the three highest Gini coeffcients. Comparing these three PICs Countries to East Asian countries with similar income per capita Prosperity is not widely shared in most PICs. This levels, the Gini coefficients for the Solomon Islands section examines how consumption is distributed and PNG are notably higher than in Cambodia, the across the populations of several countries, utilizing Lao People’s Democratic Republic, and Vietnam. several measures of inequality. These measures all The Gini coefficient in Fiji is also slightly higher suggest that inequality is substantial, particularly in than, for example, that of Indonesia. Although the Fiji, Papua New Guinea, and the Solomon Islands. Gini coeffcients indicate notable inequality in these The extent to which prosperity is shared is important, three PICs, they are lower than the Gini coefficients because high levels of inequality may hamper for other small island developing states such as São economic growth and efforts to reduce hardship. Tomé and Principe and Cape Verde, pictured in In addition, high levels of inequality can threaten figure 2.14. social cohesion, a particular concern for ethnically diverse countries that have histories of conflict (World Within countries, inequality tends to be higher in Bank 2013). rural areas. Figure 2.15 displays the Gini coefficient separately for rural and urban areas. In many countries Across countries, the top 20 percent of the population around the world, urban inequality is higher, reflecting consumes 6 to 12 times as much as the bottom a widening of outcomes for low- and high-skilled 20 percent. Figure 2.13 shows the share of total labor in the labor market, as opposed to relatively consumption that goes to each quintile of the homogeneous returns to agricultural labor in rural population by country. The patterns are strikingly areas. However, it appears that this is not the case similar across countries, with the bottom 20 percent for Kiribati, Solomon Islands, and Tuvalu, where accounting for very little of total consumption (from inequality is higher in rural areas than it is in urban less than 5 percent in Papua New Guinea to about areas. In Papua New Guinea and Vanuatu, urban 7 percent in Vanuatu). Interestingly, in most countries, and rural inequality are similar. Fiji stands out as the each of the quintiles consumes about 1.5 times as only country analyzed here where urban inequality is much as the quintile below it, except for the top substantially higher than rural inequality. As with the quintile, which consumes more than two times more hardship rates, this difference in Fiji may be related to than the next quintile down. The largest differences its more developed economy. between the bottom and top quintiles are in Papua New Guinea and the Solomon Islands. 13. The Gini coefficient is a widely used measure of inequality, comparing the distribution of a variable (in Many people may be vulnerable to falling into this case, consumption) with a uniform distribution that hardship. The data to estimate the rates at which represents equality. Zero represents perfect equality in the people fall into and escape from hardship over time distribution, and 1 complete inequality (here values have do not exist in the Pacific. However, the relatively low been multiplied by 100). A Snapshot of Hardship in the Pacific | 23 Figure 2.13 Share of Total Consumption for Each Quintile within PICs Papua New Guinea Fiji bottom 20%: 4.6% bottom 20%: 6.6% 20–40%: 9.9% 20–40%: 10.6% 40–60%: top 20%: 15.1% top 20%: 40–60%: 47.8% 48.3% 14.4% 60–80%: 60–80%: 22.7% 20.2% Solomon Islands Vanuatu bottom 20%: 5.2% bottom 20%: 7.2% 20–40%: 9.6% 20–40%: 11.8% 40–60%: top 20%: top 20%: 14.0% 40–60%: 42.6% 50.5% 16.2% 60–80%: 60–80%: 20.7% 22.2% Tuvalu Samoa bottom 20%: 6.4% bottom 20%: 6.7% 20–40%: 10.3% 20–40%: 10.1% top 20%: 40–60%: 40–60%: 14.6% top 20%: 14.2% 47.1% 48.5% 60–80%: 60–80%: 21.5% 20.5% Kiribati bottom 20%: 6.1% 20–40%: 11.1% top 20%: 40–60%: 45.2% 15.5% 60–80%: 22.1% Sources: World Bank staff estimates based on HIESs and existing poverty analysis (SINSO and UNDP 2008; KNSO and UNDP 2010; SBS and UNDP 2010). Note: These measures of consumption include gifts of food or other items exchanged between households (given away and received) for all countries except PNG. See the Technical Annex to this chapter and the analysis of gifts in chapter 4. 24 | HARDSHIP AND VULNERABILITY IN THE PACIFIC ISLAND COUNTRIES Figure 2.14 Gini Coefficients for PICs versus Other Countries 64 South Africa 59 Haiti Brazil 54 Säo Tomé and Príncipe Cape Verde Pacific Gini Index, latest year 49 Dominican Republic Mexico East Asia Papua Solomon Islands Jamaica Malaysia 44 New Guinea Other "SIDS" Philippines China Fiji Other comparators 39 Thailand Kiribati Indoneisa Cambodia Lao PDR Mongolia 34 Vietnam Vanuatu 29 Finland Ukraine 24 1000 2000 4000 8000 16000 32000 GDP per capita, PPP (constant 2005 $), 2012 Sources: World Bank staff estimates based on HIESs and existing poverty analysis (SINSO and UNDP 2008; KNSO and UNDP 2010; SBS and UNDP 2010). Note: PPP = purchasing parity power; SIDS = small island developing states. Figure 2.15 Gini Index for Rural versus Urban Areas in PICs 46 45.09 44 43.90 42.31 42 41.65 41.12 41.17 40.46 40.30 40.12 40 Gini Index 38.53 38.50 38.64 Total 38 37.57 Urban Rural 36 34.96 35.13 35.04 34 33.87 32.04 32 30 Solomon Papua Fiji Tuvalu Kiribati Vanuatu Islands New 2008–09 Guinea Sources: World Bank staff estimates based on HIESs and existing poverty analysis (SINSO and UNDP 2008; KNSO and UNDP 2010; SBS and UNDP 2010). A Snapshot of Hardship in the Pacific | 25 Key Messages plays little role in policy making. Unfortunately, across the Pacific, many years usually elapse between Across the Pacific, many people are living in hardship, nationally representative household surveys. This meaning they are unable to meet their basic needs. lack of data leaves policy makers and development The evidence presented in the first two sections of partners without timely information about people’s this chapter showed that over 20 percent of people well-being that could help them make the best use in most PICs analyzed live in hardship, meaning they of limited public resources. Using the existing data, are unable to meet their basic food and nonfood significantly different methodological choices in the needs. The incidence of hardship is highest in Papua poverty measurement and analysis that has been New Guinea, where 40 percent of the population lives carried out, as well as limited sharing of data, it has in hardship. These results accord with many other been difficult to clearly interpret and communicate the measures of well-being that have been made for the results. In a context in which the notion of “poverty” Pacific. PICs have had mixed success with making is already controversial, these challenges mean that progress toward the Millenium Development Goals, measures of hardship have largely been ignored in and the United Nations Development Programme policy making. (UNDP) categorizes all the included PICs as having either “medium” or “low” human development (PIFS 2013; UNDP 2013). Taken together, all of these results show that hardship is a real challenge that merits the References attention of policy makers in the Pacific. Deaton, Angus, and Salman Zaidi. 2002. “Guidelines for Constructing Consumption Aggregates for Within countries, many factors are related to the Welfare Analysis.” Living Standards Measurement incidence of hardship, including where people live, Survey Working Paper 135. World Bank, their educational attainment, work status, gender, and Washington, DC. age. In Fiji and Papua New Guinea, people living in Haughton, Jonathan, and Shahidur Khandker. 2009. urban areas are much less likely to live in hardship. Handbook on Poverty and Inequality. Washington, Households headed by individuals who have limited DC: World Bank. education or who do not work are more likely to live in ICP (International Comparison Program). 2005. Global hardship. Households headed by the elderly are more Purchasing Power Parities and Real Expenditures. likely to live in hardship than those headed by younger Washington, DC: World Bank. people, and households with more children are also more likely to live in hardship. The relationships KNSO (Kiribati National Statistics Office) and UNDP between these characteristics and hardship vary in (United Nations Development Program). 2010. strength and are driven by many underlying factors. Kiribati: Analysis of the 2006 Household Income and However, identifying these relationships is an Expenditure Survey. Suva, Fiji: KNSO and UNDP important first step in understanding the challenge of Pacific Center. hardship across the Pacific. PIFS (Pacific Islands Forum Secretariat). 2013. 2013 Pacific Regional MDGs Tracking Report. Suva, Levels of inequality in the Pacific are comparable to Fiji: PIFS. those in East Asian countries. Across PICs, the most Ravallion, Martin. 1992. “Poverty Comparisons: A well-off people (the top 20 percent) consume many Guide to Concepts and Methods.” Living Standards times more than the least well-off. As measured by Measurement Survey Working Paper 88. World Gini coefficients, inequality is highest in the Solomon Bank, Washington, DC. Islands, Papua New Guinea, and Fiji. Moreover, within SBS (Samoa Bureau of Statistics) and UNDP (United most countries, inequality in rural areas is equal to Nations Development Program). 2010. Samoa: A or higher than inequality in urban areas (Fiji is the Report on the Estimation of Basic Needs Poverty prominent exception). When considered in a global Lines, and the Incidence and Characteristics of context, these levels of inequality are not extremely Hardship and Poverty. Suva, Fiji: SBS and UNDP high but, to the extent that inequality affects Pacific Center. economic growth and social cohesion, may still be a SINSO (Solomon Islands National Statistics Office) matter of concern for policy makers. and UNDP (United Nations Development Program). 2008. Solomon Islands: Analysis of the 2005/06 Infrequent surveys, significant methodological Household Income and Expenditure Survey. Suva, variation, and cultural rejection mean that hardship Fiji: SINSO and UNDP Pacific Center. 26 | HARDSHIP AND VULNERABILITY IN THE PACIFIC ISLAND COUNTRIES UNDP (United Nations Development Program). 2013. ———. 2012c. Skills for Solomon Islands. Sydney: 2013 Human Development Report. New York: World Bank. UNDP. ———. 2013. At Work in East Asia Pacific. ———. 2011. “Poverty Trends, Profiles and Small Washington, DC: World Bank. Area Estimation (Poverty Maps) in Republic of Fiji ———. n.d. World Development Indicators (2003–2009).” World Bank, Washington, DC. (database). World Bank, Washington, DC. ———. 2012a. World Development Report 2013: http://data.worldbank.org/data-catalog/ Jobs. Washington, DC: World Bank. world-development-indicators. ———. 2012b. Targeting Poor and Vulnerable Households in Indonesia. Public Expenditure Review. Washington, DC: World Bank. A Snapshot of Hardship in the Pacific | 27 Technical Annex: Important gifts, whereas gifts received (except cash) are included in the consumption aggregate of the households that Cross-Country Differences receive the gifts. In Papua New Guinea, gifts given are in Poverty Analysis excluded from the consumption aggregate. All the analysis included in the chapter is based on The second expenditure category, rents, is also Household Income and Expenditure Surveys (HIESs) an important issue. In many countries around the carried out in each country at different points in time. world, rents are imputed for households that own These surveys are representative at the national level their dwellings, to avoid making renters (who spend and are similar in their structure and content across on rent) look better off than owners. The standard countries.14 However, the HIESs are not identical and method for imputation is to use a hedonic regression reflect the variation in country contexts. that estimates the rental value of owner-occupied dwellings based on the dwellings’ characteristics and This annex focuses on the major differences the rental price of similar dwellings. This method can in the poverty analysis that was carried out by be very difficult to implement in areas with thin or national statistical offices (often in conjunction with nonexistent rental markets, which is the case in much development partners) in the countries included in of the Pacific. However, rent imputations have been this chapter.15 In particular, the annex highlights major made in many PICs and included in the consumption differences in how consumption aggregates are aggregates. This is the case in Kiribati, Tonga, and constructed and how poverty lines are defined. Tuvalu. In Fiji, where rental markets are more active, imputed rent was calculated using the standard Although these methodological differences create method and included in the consumption aggregate. challenges in interpreting the results of the poverty In contrast, in Papua New Guinea, neither imputed analysis that has been conducted, particularly across rents nor actual rents paid were included in the countries, it is important to keep in mind that such consumption aggregate. differences exist across all countries in the world and are not unique to the Pacific. The prices used to estimate the costs of basic needs Consumption Aggregates also vary across countries. In countries where detailed price surveys exist, they are often used to assign Two major areas of cross-country differences in the prices to consumption items. However, in much of Pacific can be identified regarding the construction of the Pacific, such surveys are only carried out in urban consumption aggregates: (1) the set of expenditure areas, where prices can be very different from prices in categories to be included in the aggregate and (2) the rural areas. The poverty analysis in different countries prices used to cost out the consumption. has dealt with this challenge in different ways. In some countries, the “unit values” (the prices implied in the Two particularly important expenditure categories are household survey data) are used instead of actual treated differently across the Pacific: gifts and rents. prices.16 In others, prices from price surveys are used Gifts, or interhousehold transfers, are an important in urban areas, and unit values are used in rural areas. part of traditional networks in the Pacific, so their treatment in poverty measurement is important. In Poverty Lines general, experts have suggested excluding gifts Across the Pacific, poverty analysis has defined both a given from consumption aggregates, because gifts food poverty line and a total, or basic needs, poverty that are given are not consumed by the households line. The food poverty line is based on the cost of a that give them, and double-counting should be specific basket of foods that is estimated to provide a avoided (Deaton and Zaidi 2002). However, in most of sufficient number of calories per person. An estimated the PICs studied in this chapter (Fiji, Kiribati, Tonga, cost of nonfood basic needs is then added to the food Tuvalu, and Vanuatu), gifts given are included in the poverty line to arrive at the total poverty line. When consumption aggregate of the households that give conducting this analysis, two important considerations arise: what number of calories constitutes a sufficient 14. In many countries, the surveys are not strictly nationally representative, because extremely remote areas had to be 16. The household surveys record both the amount of excluded from the sample frames because of costs. each item that households report consuming as well as the 15. These are in addition to the differences highlighted in the amount of money they report spending on the item, if it was box within the chapter. purchased. 28 | HARDSHIP AND VULNERABILITY IN THE PACIFIC ISLAND COUNTRIES number per person, and which households will be the In addition to the choices highlighted here and in the reference group for defining the basket of food and chapter box, many other choices must be made when the cost of nonfood basic needs. conducting poverty analysis, ranging from the use of multiple food baskets to reflect differences within In Fiji, Tuvalu, and Vanuatu, 2,100 kcal is defined as countries to the treatment of other expenditures such the minimum sufficient amount of food intake per as durable goods and health. person, whereas Papua New Guinea uses 2,200 kcal per adult equivalent. For nonfood needs, Kiribati, Although no universal standards are in place for poverty Tonga, Tuvalu, and Vanuatu use the households in measurement, general guidelines have been produced the bottom 30 percent of total consumption as the by the World Bank for many years. A handbook that reference group, whereas Papua New Guinea uses the includes these guidelines is freely available online at households whose consumption on food equals the http://go.worldbank.org/4WJH9JQ350. level of the food poverty line. Vulnerability and the Impacts of Shocks on Pacific Islanders 29 Chapter 3 Vulnerability and the Impacts of Shocks on Pacific Islanders P eople in the Pacific are vulnerable to many Pacific islanders are extremely vulnerable to the different shocks, ranging in type, covariance, effects of natural disasters, which destroy assets and other attributes, and people already and cause physical harm, reducing well-being and experiencing hardship tend to be the most possibly increasing hardship. Between 1980 and 2009, vulnerable. Vulnerability varies across countries and 2.3 percent of the world’s reported natural disasters also depends on both personal and household occurred in the Pacific islands, which represent only characteristics, but households in the Pacific are about 0.1 percent of the world’s population (EM-DAT among the most vulnerable in the world to certain 2013). Figure 3.1 details the range of natural shocks shocks. This chapter provides an overview of the most that have affected PICs (or areas within countries) common shocks to which people in the Pacific are since 1980. The impacts of disasters include asset vulnerable as well as, where possible, estimates of damage or destruction, disruption of economic the impacts of these shocks. The first section reviews activity, displacement, injury, and death, any of which the different types of aggregate shocks common can significantly reduce the well-being of households in the Pacific. The second focuses on aggregate and may push some into, or further into, hardship. In economic shocks, and the third presents estimates fact, at the country level, Pacific islands are among of the household-level impacts of such shocks. The the most vulnerable to negative impacts from natural fourth section then presents evidence of the evolving disasters. Eight of the top 20 countries by annualized nature of health shocks across the Pacific, and the fifth relative losses from natural disasters are PICs, with summarizes the limited evidence available on other many experiencing economic losses of several types of shocks. The last section summarizes the key percentage points of GDP (figure 3.2). findings. People experiencing hardship may be more vulnerable to certain natural disasters. For example, Pacific Islanders Are Vulnerable informal or unauthorized settlements around urban centers in the Pacific are often the areas to Many Aggregate Shocks that experience flooding, such as the flooding that Perhaps the two most important sources of aggregate occurred after Cyclone Evan hit Samoa in 2012 shocks in the Pacific are natural and economic. (Samoa MNRE 2013). Disasters may affect only some Aggregate shocks are those that covary at the regions within countries, for example, a few islands in regional, national, or international level, affecting an archipelago. When disasters hit remote regions, large numbers of households at the same time. As the people living there may receive limited and less detailed in chapter 1, the unique combination of small timely assistance because of logistical and information populations and geography that characterizes many challenges. PICs helps shape the vulnerability of their people, by exposing countries to international economic Vulnerability to climate change is a longer-term shocks as well as to natural disasters in the short and concern, but negative impacts are already being felt. long term. In addition, health shocks are increasingly Rising seas are already encroaching on some islands’ affecting the Pacific at the aggregate level and will be freshwater lenses, and coral erosion from warming and discussed in section 4. acidifying seawater threatens both the food sources 30 | HARDSHIP AND VULNERABILITY IN THE PACIFIC ISLAND COUNTRIES Figure 3.1 Natural Disasters by Type in PICs, 1980–2009 110 100 90 80 70 60 50 40 30 20 10 0 Tropical Earthquake Flood Volcanic Landslide Drought Epidemic Wildfire cyclone and/or eruption tsunami Source: EM-DAT (the OFDA/CRED International Disaster Database). Figure 3.2 Annual Average Economic Losses from Cyclones, Earthquakes, and Tsunamis 7.0 6.0 Percentage of GDP 5.0 4.0 3.0 2.0 1.0 0 . a s oa u u e i tu e ga s ts ji s at ne Fi nd nd nd la al ag iu .S ua rib m n Pa v N ui la la la er To Tu Sa n ed G Ki Is Is Is Va av ,F ew k ll on ha N oo ia N m EA s es C lo ar a AS pu on So M Pa r ic M Source: World Bank 2012a. Note: ASEAN = Association of Southeast Asian Nations. Pacific islanders rely on and—for residents of coral Households in the Pacific are heavily dependent on atolls in particular—the very land they stand on (World external flows (of money and goods) for their well- Bank 2013a). Pacific islands are also forecasted to being. Large proportions of households in the Pacific experience unprecedented temperature extremes. rely on external financial flows, including tourism, Two recent World Bank reports, Acting Today for remittances, and international aid. These flows help Tomorrow (2012a) and Strong, Safe, and Resilient provide jobs, support household consumption, and (2013b), extensively review natural disasters in the contribute to government budgets. At the same Pacific and East Asia, and a third, Turn Down the Heat time, many Pacific households are deeply dependent (2013a), reviews the forecasts and expected impacts of on international commodity markets. Some are climate change globally. heavily reliant on commodity imports, including for Vulnerability and the Impacts of Shocks on Pacific Islanders | 31 basic foods, and others are exporters of agricultural impacts. The next section goes into further detail commodities and natural resources. These economic about these types of shocks and their effects on features are critical to the well-being of people in the people. Pacific, given the constraints on growth caused by the unique features of the region discussed in chapter 1. At the same time, the smallness and geography of Aggregate Economic Shocks the Pacific mean that people are exposed to external volatility via these routes. Are Particularly Important External financial flows, including tourism, The smallness and geography of the Pacific mean that remittances, and aid, are important sources of people are particularly vulnerable to shocks to these funding for most PICs. People in the Federated sources of well-being. External financial flows tend States of Micronesia and Polynesia benefit most to be concentrated both by type and source country, from remittances, because of bilateral migration in part because PICs are so small. This concentration agreements or histories of labor agreements with makes people in the Pacific more vulnerable to specific industries. Tourism is important in another shocks to these flows. For example, a shock to a set of countries: Figure 3.3 shows that in Fiji, Palau, single export commodity can have large effects on Samoa, and Vanuatu international tourism receipts government revenues and economic activity and accounts for 20 to nearly 50 percent of GDP. Finally, affect a wide section of society. In countries that are aid from international development partners is an reliant on commodity imports, domestic alternatives important source of funding for government budgets, are limited, and price shocks can have significant particularly in the smallest and most isolated PICs. effects on consumer price inflation, eroding people’s Each of these sources of external income can be buying power. In addition, governments tend to volatile and present risks for households. However, lack the financial capacity to reduce such negative these risks are often worth taking and need to be Figure 3.3 Remittance Inflows, tourism Receipts, and Aid, 2000–2010 Average 55% 50% 45% 40% Percentage of GDP 35% Remittances 30% 25% Tourism receipts 20% International 15% assistance 10% 5% 0% s. ji i u oa s a u u at a Fi St nd la ng al at ne rib m Pa v nu la To d. Tu ui Sa Ki Is Va Fe G on ew a, m si N lo ne a So pu ro ic Pa M Sources: World Development Indicators; World Bank, data for 2011. 32 | HARDSHIP AND VULNERABILITY IN THE PACIFIC ISLAND COUNTRIES managed as well as possible, rather than avoided. flow of remittances (figure 3.4a). For example, the This is discussed further in chapter 4. recent global economic downturn has reduced the employment and remittances sent from I-Kiribati and External financial flows in each Pacific country tend Tuvaluan seafarers, who mostly work in the global to come from a small number of sources, with shipping industry. Development assistance can also households in each country vulnerable to shocks be influenced by domestic conditions of donor transmitted along these lines. International migrants countries, and PICs tend to be dependent on a small from individual PICs tend to concentrate in a limited number of donors—the majority receive more than number of countries and sectors. For example, more 70 percent of aid from one or two bilateral donors. than 95 percent of Tongans overseas (a group that is Most tourism and international aid originates from the estimated to account for about half the total Tongan same few countries that receive migrants—Australia, population) live in one of three countries: Australia, New Zealand, and the United States—meaning that New Zealand, or the United States (Taufatofua 2011). households in some PICs are particularly vulnerable This type of exposure means that slowdowns or to the economic performance and policy choices in political changes in receiving countries and downturns these countries (figure 3.4b). in specific sectors can have large impacts on the Figure 3.4 a. Remittance Inflows 40% 35% Fiji 30% Kiribati Percentage of GDP 25% Papua New Guinea 20% Samoa 15% Solomon Islands 10% Tonga 5% Vanuatu 0% 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 2011 2012 (estimated) Figure 3.4 b. Majority Country Donors as Share of Total International Aid, 2011 100% 90% 80% 70% Other donors 60% United States 50% New Zealand 40% 30% Australia 20% 10% 0% a lu u ji i s. ds u a oa s at Fi nd ne ng at la va St rib an m Pa nu la ui To Tu Sa . Ki sl ed Is G Va lI ,F on ew al sh ia m N s lo ar a ne So pu M ro Pa ic M Sources: World Development Indicators; World Bank, data for 2011. Vulnerability and the Impacts of Shocks on Pacific Islanders | 33 Figure 3.5 Food and Mineral Fuel Imports in PICs 25% Percentage of GDP 20% 15% Food imports 10% Low- and middle-income average 5% World average 0% i St a, lu u s ji oa a te at Fi on at ng va d, si es rib m s. nu Fe one m To Tu -L Sa Ki lo Va or icr So m M Ti 16% 14% Percentage of GDP 12% 10% 8% 6% 4% 2% 0% Kiribati Solomons Tuvalu Fiji Tonga Samoa Papua Vanuatu (free on New board) Guinea Source: World Development Indicators. External dependency exacerbates vulnerability to labor nationals.1 The greater opportunities come from market shocks. Small, undiversified, and often niche, increased government revenues through royalties, export markets are vulnerable to instability, which can corporate taxes, and other revenues. If spent well, put large numbers of people out of work. For example, these extra resources can benefit a wide segment depressed demand for automotive parts following the of the population (World Bank 2013d). Managing onset of the global economic crisis in 2009 led to the opportunities will be revisited in chapter 4. loss of fifteen hundred formal jobs in one factory in Samoa, which represented 13 percent of total formal Global commodity prices are also an important source private sector employment in the country (World Bank of vulnerability in the relatively open economies 2013c). Most of these jobs have not yet been reinstated. of the Pacific. In recent years, food and fuel prices have been high and volatile, and most PICs lack But external exposure can also bring positive shocks. the resources and size to insulate their people from Cash-cropping households stand to benefit from these shocks. These vulnerabilities are particularly higher commodity prices, although the benefits they important for small atoll nations such as the Federated capture depend on middlemen passing on better States of Micronesia, Kiribati, and Tuvalu, where prices. Oil and gas developments in countries such as the ratio of food imports to GDP is between three Papua New Guinea have the potential to bring some and five times higher than the developing country local jobs, although in practice benefits tend to be average (figure 3.5). In addition to being relatively the greatest during the exploration and construction large importers, many PICs are exporters of primary phases. For example, the liquefied natural gas plant currently under construction in Papua New Guinea is estimated to be employing 14,300 workers, 1. International Monetary Fund, Papua New Guinea: 2012 of which 60 percent are Papua New Guinean Article IV Consultation. 34 | HARDSHIP AND VULNERABILITY IN THE PACIFIC ISLAND COUNTRIES products. Exports of cash crops are important for substantial shares of GDP in Papua New Guinea and many Pacific islanders—more so than GDP figures the Solomon Islands, but shocks to these exports have may suggest, because they represent one of the few relatively less immediate impact on households, few of sources of jobs and cash incomes for many people. which benefit directly from these industries. Figure 3.6 Conversely, mineral and fossil fuel exports account for describes international price trends of some of the Figure 3.6 International Major Commodity Volatility, 1995–2012 International rice price volatility 600 500 400 Rice: Actual 300 Rice: Trend Rice: 1 SD interval 200 100 0 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 2011 2012 International oil price volatility 140 120 100 Oil: Actual 80 Oil: Trend 60 Oil: 1 SD interval 40 20 0 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 2011 2012 International coffee price volatility 700 600 500 Coffee: Actual 400 Coffee: Trend 300 Coffee: 1 SD interval 200 100 0 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 2011 2012 International copra price volatility 1,400 1,200 1,000 Copra: Actual 800 Copra: Trend 600 Copra: 1 SD interval 400 200 0 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 2011 2012 Source: IMF Primary Commodity Prices. Vulnerability and the Impacts of Shocks on Pacific Islanders | 35 Figure 3.7 Annual Average CPI Change and Change in Current Account Balance as Percentage of GDP for PICs Micronesla, Fed. Sts. Papua New Guinea Solomon Islands Developing EAP Marshall Islands Annual average CPI differential 12 Peak inflation Change in current account 10 EAP Developing LMIC Vanuatu 8 Samoa balance (% GDP) Kiribati Tuvalu Tonga LMIC 6 Fiji 4 0 2 –5 0 –10 –15 Papua New Guinea Micronesia, Fed, Sts. Marshall Islands Solomon Islands Kiribati Tuvalu Samoa Tonga Timor-Leste Fiji Vanuatu –20 –25 –30 Note: CPI = consumer price index; EAP = East Asia and the Pacific; LMIC = low- and middle-income countries. most important commodities for the Pacific and their public expenditures and lower taxes, adding to fiscal volatility.2 sustainability pressures (figure 3.8). Problems of import dependence are magnified in Within countries, vulnerability to economic shocks also small, remote archipelagos. Small populations must varies, depending on where people live (urban or rural be served regularly by low-volume shipping routes areas), initial well-being, and other factors. In many for critical supplies, which are often operated or PICs, households in hardship tend to self-produce subsidized by the state. As well as the added cost of a larger share of their food than other households shipping, interruptions to supply can lead to shortages do (figure 3.9). At the same time, households in of food and other basic goods, which is yet another hardship also spend a bigger proportion of their total source of risk for Pacific islanders. budget on food, so that, despite own-production, these households may be as vulnerable to volatility in These aggregate economic shocks have a consumer food prices. Urban dwellers experiencing disproportionate effect at the macroeconomic level hardship or close to it may be the most vulnerable on Pacific islands. During the recent global food to consumer food price shocks, because they rely and fuel price crisis, the spike in consumer prices largely on purchased foods (figure 3.10), and without experienced in the Pacific tended to be well above access to land and sea resources, such shocks can the increases seen in East Asia and other developing push them into or deeper into hardship. Conversely, countries, and current account balances deteriorated rural households are more vulnerable to cash crop rapidly (figure 3.7). In addition to price increases, price shocks, and in some countries, rural households other adverse economic shocks over the period 2007 in hardship are more dependent on cash crop to 2009 included drops in remittances and tourism. income than those not in hardship (figure 3.11). These shocks, combined with the fiscal response in Shocks to external financial flows are also likely some countries to mitigate their effects, led to higher to be unevenly distributed across the population, depending on which households receive remittances, work in tourism, or benefit more from international aid (figure 3.12). The next section seeks to provide some insight into the question of how exogenous 2. The intervals are based on one standard deviation bounds economic shocks might affect households in the on annual variation over the last 20 years, which means that Pacific by simulating the effects of common shocks on a year-to-year variation of this magnitude has occurred over household welfare in Kiribati, Papua New Guinea, and the last 20 years with probability of about 30 percent. Tonga. 36 | HARDSHIP AND VULNERABILITY IN THE PACIFIC ISLAND COUNTRIES Figure 3.8 Average Change in Net Taxes on Products and Change in Government Expenditures for PICs Change in government expenditures, % GDP 2 Average change in net taxes on 1 20 0 15 Change in expenditure –1 EAP Developing 10 products (%) –2 –3 5 –4 0 –5 –6 –5 Papua New Guinea Solomon Islands Timor-Leste Tuvalu Samoa Vanuatu Tonga Micronesia Kiribati Fiji Marshall Islands –7 –8 2007 2008 2009 Source: IMF 2013. Notes: Change in net taxes on products averaged over countries with data available: Fiji, Kiribati, Micronesia, Fed. Sts., Palau, Papua New Guinea, Tonga, and in constant domestic prices. Change in government expenditure compares periods 2006–2008 and 2009–2011. EAP = East Asia and the Pacific. Figure 3.9 Food Expenditures for Households with and without Hardship 80% 70% Food expenditure as (% of total) 60% 50% 40% 30% 20% 10% 0% Nonhardship Hardship Nonhardship Hardship Nonhardship Hardship Nonhardship Hardship Nonhardship Hardship Nonhardship Hardship Fiji Papua Solomon Tonga Samoa Kiribati New Guinea Islands Purchased Self-produced Source: World Bank staff estimates from HIESs. Vulnerability and the Impacts of Shocks on Pacific Islanders | 37 Figure 3.10 Food Expenditures of Households in Hardship 100% % of food expenditure of hhlds in hardship 90% 80% 70% 60% 50% 40% 30% 20% 10% 0% Fiji Papua New Guinea Solomon Islands Tonga Samoa Kiribati Fiji Papua New Guinea Solomon Islands Tonga Samoa Kiribati Rural Urban Purchased Self-produced Source: World Bank staff estimates from HIESs. Figure 3.11 Rural households in Papua New Guinea Growing Cash Crops 90% Proportion of rural households growing cash crops 80% 70% 60% 50% 40% 30% 20% 10% 0% Hardship Nonhardship Hardship Nonhardship Hardship Nonhardship Hardship Nonhardship Hardship Nonhardship Papua New Kiribati Vanuatu Fiji Tonga Guinea Source: World Bank staff estimates from HIESs. 38 | HARDSHIP AND VULNERABILITY IN THE PACIFIC ISLAND COUNTRIES Figure 3.12 Households Receiving Remittances in Select PICs 100% 90% Households receiving remittances 80% 70% 60% 50% 40% 30% 20% 10% 0% Nonhardship Hardship Nonhardship Hardship Nonhardship Hardship Samoa Tonga Fiji Source: World Bank staff estimates from HIESs. Impacts of Aggregate Economic Across all three countries, shocks to major food commodities, oil, cash crops, or remittances are Shocks on Households: Results estimated to push between 1 and 6 percent of the from Microsimulations population into hardship and to deepen hardship substantially for many others. Shocks that are This section presents analytical estimates of the simulated are of moderate magnitudes that have impacts of economic shocks on household welfare occurred in the past and are likely to occur again on a and on rates of hardship for three PICs with different frequent basis. Even at these magnitudes, household vulnerabilities: Kiribati, Papua New Guinea, and well-being is substantially affected by increases in Tonga. The preceding section showed that PICs the cost of living or decreases in income (depending are vulnerable to a variety of aggregate economic on the type of shock). Urban residents are relatively shocks, especially reductions in external financial flows more affected by commodity price shocks, but since including remittances, price increases in imported all three countries are still primarily rural, the majority commodities, and price decreases in cash crops. of the people pushed into hardship by these shocks However, this aggregate perspective tells us little live in rural areas. Falls in cash crop prices affect rural about the impacts on households. To look in more residents and deepen hardship for those already depth at the likely impacts of these types of shocks on experiencing it, particularly in Papua New Guinea. households, microsimulation analysis utilizes nationally Negative shocks to remittances also increase hardship representative household data and simulates the in Tonga. The following subsections provide an effects of different shocks on household income and overview of the analytical approach, details of the consumption, across the distribution of households model, and country-by-country results. in each country. These three countries were chosen from the subset of Pacific countries for which recent, detailed household-level data are available, and to Overview of the Microsimulation represent a range of country contexts in the Pacific— Analysis in terms of size, cultural area, and average income. The analysis simulates shocks to prices and to Additional detail on the analytical model can be remittance income at magnitudes that can and found in the annex to this chapter, and a complete have occurred relatively frequently in the Pacific. description of the analysis and results can be found in Therefore, the results are not meant to describe Cororaton and Knight (2013). extremes that occur only very rarely, but scenarios Vulnerability and the Impacts of Shocks on Pacific Islanders | 39 that have a roughly one in three chance of occurring their cost of living and reduce their real expenditure. in any given year. The simulated price shocks are Similarly, a drop in income, be it from cash crop applied to specific food commodities that are earnings or remittances, will also tend to lead to a commonly consumed and cash crops that are sold fall in real expenditure. Direct effects on household rather than broad aggregates. Rice and wheat are consumption dominate in the rice and wheat sectors, important imported food commodities in parts of where there is negligible domestic production. Cash the Pacific, and the impacts of price shocks on these crops are primarily exported and so have limited two foods are simulated in all three countries. Oil, impact outside their direct effect on growers. Oil both as a commodity in itself and as a major input to price shocks are felt in a more diffuse way, which is transport and energy costs, is crucial to many Pacific reflected in the simulation of this shock as described countries, and an upwards price shock to oil prices in table 3.2. The simulations estimate changes in real will be simulated. A fall in coffee prices is simulated expenditure relative to country-specific basic needs in Papua New Guinea, and in Kiribati the price of poverty lines. copra is shocked. The impact of a drop in the inflow of remittances is simulated in Tonga. The sizes of the The model assesses impact effects only and does price shocks modeled, which are based on historical not incorporate any substitution by households away variations, are given in table 3.1. The primary data from higher priced food items or lower value crops. sources are the same household surveys described in In this respect, the impact estimates are an upper chapter 2. bound, because households will rationally substitute away from higher priced goods to reduce the welfare The model takes a microsimulation approach to impact of a price shock. However, for the shocks estimate the impacts of macroeconomic shocks on simulated, the ability of households to substitute in households. The model is based on the World Bank many cases is likely to be constrained by the very Poverty Reduction and Economic Management Food limited domestic agricultural supply in many Pacific and Fuel Simulation model that is publically available,3 islands, and consequently little potential to raise although it has been significantly customized for this production in the short term, so households have analysis. Similar approaches have been used in recent little choice but to rely on imported food. In other papers on the simulated impacts of commodity price countries, including Papua New Guinea, alternative shocks, for instance, Anderson, Ivanic, and Martin own-grown or collected foods may be a more viable (2013) and Wodon and Zaman (2008). alternative to imported foodstuffs. As authors such as Gibson and Kim (2013) have pointed out, in such The model is based on a partial equilibrium approach. countries households in or near hardship may have Therefore, it does not require data and assumptions a capacity to switch to cheaper and lower quality that establish the relationships between sectors and foods in response to a price shock. By considering the actors in an economy, which makes it more suited to presubstitution effects of shocks, this analysis seeks the limited data available for many PICs. Increases to identify households that are likely to be adversely in the cost of items that households consume raise affected and does not attempt to account for this kind of food switching. In countries such as Papua New Guinea, where cash crops are a major source of household income, mature stocks represent a significant fixed asset for households, and they are Table 3.1: Magnitudes of Simulated Shocks unlikely to remove plantations to grow other crops Commodity Price Shock Coffee 20% price decrease Copra 20% price decrease Table 3.2: Sectoral Price Changes in Response Crude oil 30% price increase to 30 Percent Oil Price Shock Rice 20% price increase Sector Papua New Guinea Kiribati Wheat 20% price increase Electricity 15% 30% Petroleum products 30 30 Public transport 30 30 Transport costs (all goods) 30 30 3. http://go.worldbank.org/3C2XG5B1G0. 40 | HARDSHIP AND VULNERABILITY IN THE PACIFIC ISLAND COUNTRIES in response to year-to-year price variations. Similarly, A shock to wheat prices is predicted to have a smaller coffee and cocoa plants take 15 to 20 years to mature, effect on hardship. Although significant consumption so additional supply would respond only very slowly to of bread and other wheat-based foods is seen in price incentives. These limited supply responses are urban areas, it is not as important a food as rice, and a borne out by recent research that suggested that the shock to its price alone leads to only a relatively small supply elasticity for various cash crops in Papua New number of people moving below the poverty line. Guinea is negligible in the short to medium term (Aba, Table 3.3 summarizes the results of these, as well as Aipi, and Irau 2012; Aipi, Irau, and Aba 2012). the oil and combined shocks. Cost-of-Living Simulations: Papua New An increase in world oil prices is felt across a range of Guinea expenditure items by households and is estimated to The microsimulation modeling reveals that shocks to push the greatest number of people into hardship. food commodity prices would push large numbers An oil price shock leads directly to higher petrol and of people into hardship, and deeper into hardship, kerosene prices. It also increases the cost of public in Papua New Guinea. An increase in imported rice transport and electricity and pushes up the cost of prices of 20 percent is likely to place about 35,000 most goods by increasing the cost of transport both more people into hardship, with a total welfare loss to and within the country. The large size of the shock to those in hardship of around $7.5 million (in 2012 and the diffused impacts via transportation costs of prices). In urban areas, where fewer alternatives to goods means that a price shock entails more severe rice are found, 0.8 percent of the population would impacts on those in hardship. The estimated impact be pushed into hardship. Rural populations, including of an increase in oil prices by 30 percent is to push those living in periurban areas, are also affected. The 116,000 people below the poverty line—1.6 percent of majority of the country’s less-well-off population live the total population. Again, urban populations would in rural areas, so 0.5 percent falling below the poverty be the hardest hit, because they spend larger fractions line there translates to 28,000 people, compared with of their budget on imported goods and transport to the 8,000 affected in urban areas. Households already get around urban areas to access services and work in hardship in both rural and urban areas would also opportunities. The amount that would be needed to be adversely affected and would have to find an compensate those in hardship for the price rise would additional 0.5 percent (urban) and 0.8 percent (rural) be $26.7 million. Consumer prices are estimated to of their budgets to finance the same consumption rise by 4.9 percent. basket. The increase of rice prices would be sufficient to push up national inflation by 1.4 percentage points, In recent years the prices of basic commodities with urban consumer price index likely to rise by more. including food and fuel have increased together, and Table 3.3: Summary Results for Papua New Guinea Cost-of-Living Simulations Increase in Rice Price Increase in Wheat Increase in Oil Price All Three (20%) Price (20%) (30%) Commodity Shocks People pushed into 35,000 (0.5%) 4,000 (0.1%) 116,000 (1.6%) 178,000 (2.5%) hardship (% of population)  Of which rural 28,000 (0.5%) 2,000 (0.1%) 93,000 (1.6%) 143,000 (2.4%)  Of which urban 8,000 (0.8%) 2,000 (0.2%) 22,000 (2.2%) 35,000 (3.4%) Annual welfare loss, total $7.5 million $2.2 million $26.7 million $38.7 million hardship Proportion of hardship 0.7% 0.2% 2.6% 3.6% rural household budget Proportion of hardship 0.4% 0.2% 1.3% 2.1% urban household budget Change in hardship 2.1% 0.8% 7.4% 10.5% severity Inflation 1.4 0.6 4.9 6.9 Note: Welfare loss is annualized, expressed in 2012 prices. Vulnerability and the Impacts of Shocks on Pacific Islanders | 41 a simulated simultaneous shock to oil, rice, and wheat would increase sharply, by 19 percent. Table 3.4 is estimated to push 2.5 percent of the population summarizes these results and for the following into hardship. This is equivalent to moving 178,000 simulations. people into hardship, at a welfare cost of $38.7 million. Many of those already below the line would A simultaneous shock to the three basic commodities fall deeper into hardship. A measure of the severity of (rice, wheat, and oil) in Kiribati is particularly harmful, hardship, which captures how far below the poverty pushing 6 percent of the population into hardship. line some households are, would increase by more The impact would be more severe than might be than 10 percent. Finding the right policy levers to suggested by the sum of individual shocks: Many compensate people is a major challenge, but even if more households that would just make ends meet a perfect mechanism did exist, the government would if the price of one commodity rose on its own are struggle with the fiscal cost of alleviating this increase pulled into hardship when faced with multiple price in hardship, because the cost to those in hardship increases. A total of 5,200 people would be expected of a simultaneous shock being equivalent to about to fall into hardship. Given the small population 3.4 percent of government discretionary spending. size, the total welfare effects of the shocks are much smaller, illustrating the relatively small sums of money Cost-of-Living Simulations: Kiribati that would be needed to ameliorate adverse impacts An oil price shock of 30 percent is estimated to push (assuming an efficient transfer mechanism could 3,000 people or 3.7 percent of the population into be found, which in chronically capacity-constrained hardship. An increase in the price of rice is estimated countries such as Kiribati is a major challenge). The to place around 600 people below the poverty line. equivalent cost for government would be 1 percent of A wheat price shock would have a more moderate discretionary spending. effect. Interisland shipping costs are subsidized in Kiribati, but higher transport costs are still be felt by Cost-of-Living Simulations: Tonga households from international shipping. An oil price In Tonga, 1 percent of the population would be shock would hit Kiribati particularly hard, given the pushed into hardship from an oil price shock, higher shipping costs of reaching its remote location. whereas a simultaneous shock to oil, wheat, and rice A total of 3,200 people, 3.7 percent of the population, prices would push about twice as many people into would fall into hardship as a result, with those in South hardship. The types of imported food consumed Tarawa being hardest hit. Similarly, the adverse effect in Tonga are markedly different from Kiribati and of urban households already in hardship is larger, with Papua New Guinea. Rice is not a major staple, additional costs equating to nearly 2 percent of their with meat and wheat products being the mainstay total expenditure, and hardship severity nationwide of the diet, along with local vegetable produce. Table 3.4: Summary Results for Kiribati Cost-of-Living Simulations Increase in Rice Price Increase in Wheat Increase in Oil Price All Three (20%) Price (20%) (30%) Commodity Shocks People pushed into 630 (0.7%) 170 (0.2%) 3,200 (3.7%) 5,200 (6.0%) hardship (% of population)  Of which rural 460 (1.0%) 170 (0.4%) 900 (1.9%) 1,200 (2.9%)  Of which urban 170 (0.4%) 0 2,300 (5.9%) 4,000 (10.2%) Annual welfare loss, total $92,000 $24,000 $323,000 $574,000 hardship Proportion of hardship 0.5% 0.1% 0.9% 1.7% rural household budget Proportion of hardship 0.5% 0.2% 1.8% 2.8% urban household budget Change in hardship 3.9% 1.9% 18.6% 25.1% severity Inflation 1.6 0.8 7.3 9.7 42 | HARDSHIP AND VULNERABILITY IN THE PACIFIC ISLAND COUNTRIES Table 3.5: Summary Results for Tonga Cost-of-Living Simulations Increase in Rice Price Increase in Wheat Increase in Oil Price All Three (20%) Price (20%) (30%) Commodity Shocks People pushed into — 300 (0.4%) 1,200 (1.4%) 1,600 (1.9%) hardship (% of population)  Of which rural — 250 (0.4%) 800 (1.0%) 1,200 (1.5%)  Of which urban — 50 (0.2%) 400 (1.4%) 400 (1.7%) Annual welfare loss, total $0.04 million $0.9 million $1.8 million $2.9 million hardship Proportion of hardship — 0.3% 0.5% 0.8% rural household budget Proportion of hardship — 0.2% 0.7% 1.0% urban household budget Change in hardship — 2.6% 6.3% 9.1% severity Inflation — 0.9 3.2 4.2 Note: Welfare loss is annualized, expressed in 2012 prices. — = not available. Therefore, a rice price shock has little impact, but an households tend to be less well-off than the average increase in wheat prices directly affects households, household, with the hardship head-count ratio for particularly in rural areas. Approximately 1,200 cash-crop households at 45 percent compared with people would be expected to fall below the poverty 40 percent for all households. A drop in coffee prices line following an increase in oil prices, and 1,600 by 20 percent would force 230,000 people below the people would be affected by a simultaneous price poverty line, 3.3 percent of the country’s population. shock to oil, wheat, and rice. These shocks would This is equivalent to one in every 20 people who also push those already in hardship further into it. are not already in hardship. The potential impacts Overall, the welfare cost to households in hardship on households already in hardship are particularly would be $2.9 million for the simultaneous shock, disastrous—with these households needing to equivalent to 1 percent of total consumption and find alternative income to finance 6.3 percent of expenditure of urban households and 0.8 percent expenditure, and hardship severity for cash-cropping for rural households, and 3.2 percent of government households increasing by more than 50 percent. The discretionary spending. These results are welfare cost borne by households in hardship is also summarized in table 3.5. It is worth noting that the large at $81.3 million and amounts to in excess of analysis does not consider the indirect effects of 7 percent of the Papua New Guinea government’s changes in grain prices on meat, but because grain annual discretionary budget. feed is an important input into livestock rearing, prices often move together. The estimates are As in Papua New Guinea, a large proportion of therefore likely to understate the overall effect of households outside the urban area rely on cash price shocks on household expenditure. crop income in Kiribati. A major source of income in the outer islands is copra, which benefits Income Simulations from a guaranteed purchase scheme operated It is not only via expenditure that households are by government that in effect heavily subsidizes exposed to external price shocks. In PICs, many households to produce copra. Across all outer income sources are also highly dependent on island households, agricultural cash incomes, which external factors. Declines in the price that cash are primarily copra, make up 37 percent of cash crops can be sold for can also be a major source income (excluding nonmonetized income like home- of impoverishment in PNG. A third of households grown food). In response to a fall in copra prices, in PNG grow coffee, which is often the only source it is expected that 800 people, 1.2 percent of the of cash income in a household. Cash-cropping population, would move below the poverty line. Vulnerability and the Impacts of Shocks on Pacific Islanders | 43 Table 3.6: Summary Results for Income Simulations across Countries Decrease in Coffee Price: Decrease in Copra Price: Decrease in Remittances: Papua New Guinea (20%) Kiribati (20%) Tonga (20%) People pushed into hardship 230,000 (3.3%) 800 (1.2%) 900 (1.1%) (% of population)  Of which rural 230,000 (3.7%) 800 (1.6%) 400 (1.0%)  Of which urban 0 0 500 (1.4%)  Of which affected households 230,000 (9.8%) 800 (3.0%) — Proportion of hardship hhd budgeta 6.3% 0.2% 1.3% Annual welfare loss, total hardship $81.3 million $56,000 $2.1 million Change in hardship severitya 50.7% 17.0% 2.3% Note: Welfare loss is annualized, expressed in 2012 prices. — = not available. a. For affected households only. In Tonga, remittances from overseas form an 2007 to 2011, remittances in Tonga fell by far more important source of household income. The results than the shock simulated here, in part because of of the microsimulations show that a 20 percent fall the global economic crisis. There was widespread in remittances to households would put 1.1 percent concern that the fall in remittances would substantially of the population into hardship, of which more than reduce household well-being, but data have not half are based in the urban center of Nuku’alofa at a been available to estimate its actual impacts. Box 3.1 total welfare cost to those in hardship of $2.1 million, provides a comparison of these results for Tonga as which is around 4 percent of government discretionary well as Papua New Guinea and Kiribati with global spending (see table 3.6). It is worth noting that from price shocks. Box 3.1 PICs Face the Equivalent of a Global Food and Fuel Crisis Every Few Years The microsimulation results illustrate that Pacific island countries might expect to see the kind of impacts that are roughly comparable to the global food and fuel crisis in other countries every few years. Several studies, including Ivanic and Martin (2008), Dessus, Herrera, and De Hoyos (2008), and De Hoyos and Medvedev (2011), have estimated the impact of the 2007–2008 spike in commodity prices on poverty in developing countries (see table B3.1.1). De Hoyos and Medvedev estimate that the increase in food prices would put 2.4 percent of households below the poverty line. Ivanic and Martin estimate this to be 3 percent. Dessus et al. indicate that for the 20 worst affected countries in the world, the equivalent increase in poverty head-count ratios is 5 percent. Table B3.1.1: Studies of Changes in Head-Count Ratio Change in Head-Count Study Ratio (Percentage Points) Assumptions De Hoyos and Medvedev 2011 2.4% Based on all food consumption, for 21 low- and middle- income countries Ivanic and Martin 2008 3.0 Based on major food commodities, for 9 low-income countries Dessus et al. 2008 5.0 Estimate for 20 worst affected countries from a dataset of 72 low- and middle-income countries Current report (including oil) 1.9–6.0% Four commodities including oil. Shock magnitude one-fifth the size of above studies. Prices of food and fuel increased by more than 100 percent over the period examined by these studies, so it is important to note that their estimates are based on food price shocks at least five times greater than the simulated price movements studied in this report. These Pacific estimates also look at only two food commodities, whereas the global estimates look at a broader basket of goods. On the other hand, the global estimates do not generally account for the impact of fuel price increases on transport and energy costs, as the Pacific estimates do. Keeping in mind these methodological differences, the results for commodity price shocks are of a comparable magnitude to the global averages. 44 | HARDSHIP AND VULNERABILITY IN THE PACIFIC ISLAND COUNTRIES Health Shocks Are Also shows that NCDs cause the majority of deaths in almost all PICs, and the most common NCDs are an Important Source cardiovascular diseases. In Tonga, the rise of NCDs of Vulnerability has already contributed to a reduction in average life expectancy (World Bank 2012b). This rise has been In addition to aggregate economic and natural caused in part by reliance on low-quality imported shocks, health shocks are an important source of foods, leading to high rates of obesity, and the vulnerability for Pacific islanders at the aggregate widespread use of tobacco and alcohol (WHO 2010). level. People in the Pacific face a “double burden” Figure 3.14 shows how widespread these conditions of disease: continued threats from communicable and behaviors are across countries: About two-thirds disease and poor maternal and child health, as well as of adults in Kiribati smoke daily, and more than half of high and growing rates of NCDs. In many countries, adults in Samoa and Tonga are obese. NCDs have reached epidemic proportions. Figure 3.13 Figure 3.13 Estimated Leading Cause of Death, 2008 100% 90% 80% 70% Injuries 60% Infectious diseases, 50% maternal, and 40% perinatal conditions 30% NCDs 20% 10% 0% ji i ds s. a oa s a lu at Fi nd ne ng St va rib an m la ui To Tu Sa d. Ki sl Is G Fe lI on ew al a, sh m N si lo ar u ne So pa M ro Pa ic M Source: WHO Global Health Observatory. Note: NCD = noncommunicable disease. Figure 3.14 Population Shares with Risk Factors for NCDs 80% 70% 60% 50% 40% 30% 20% Smoking daily 10% Obese 0% ji i ds s. oa s a u at Fi nd ng at St rib an m nu la To Sa . Ki sl ed Is Va lI ,F on al sh ia m s lo ar ne So M ro ic M Source: WHO Global Health Observatory. Note: NCD = noncommunicable disease. Vulnerability and the Impacts of Shocks on Pacific Islanders | 45 The rise in NCDs has large impacts at the country private sector activity and, through these channels, level, through costly treatment and lost productivity. people’s livelihoods. In most PICs, governments finance the majority of health care, with assistance from international People in the Pacific are also vulnerable to localized development partners. The rapid and widespread and idiosyncratic shocks from all sources, but sparse increase in NCDs threatens the financial sustainability data limit how much can be learned about the of this system, because the costs of treating NCDs are prevalence and impacts of these shocks. As in other often substantially higher than treating other types countries, everything from community disputes, of disease and injury (World Bank 2012b). In Samoa, to landslides, to crop failures, occur in the Pacific. for example, dialysis treatment is funded by the However, data on these shocks are very limited: government and costs over $38,000 per patient per Nationally representative household surveys are year, more than 10 times GDP per capita (World Bank collected several years apart, and households are not 2012b).4 In addition, NCDs affect many people during followed over time. This makes it difficult to identify their working ages, reducing the productivity of the shocks, which are inherently about changes over time. working-age population through illness, disability, and In addition, few surveys ask questions specifically premature mortality. about past shocks, making it difficult to even estimate the prevalence of shocks. Within these limitations, Health shocks at any level of covariance have a range some interesting results can still be identified about of negative impacts on individuals and families, idiosyncratic shocks, because existing available data including on household budgets. For example, in do not allow for the identification of locally covarying Fiji, about 27 percent of households in the 2009 HIES shocks. reported that at least one member is not working because he or she is disabled or ill. When a household In some PICs, violence is a commonly experienced member is unwell, other members may need to personal shock, and the rates of violence against forgo income-generating activities to care for the sick women are among the highest in the world. In Papua person. In addition to missing work, disease and other New Guinea, 18 percent of households reported health shocks can have large impacts on people in experiencing a theft, physical assault, or domestic many other ways. For example, even when treatment violence in the year preceding the survey. Interestingly, is free, households may have to migrate at their own these personal shocks are similarly common for expense to an urban center to receive it. households in and out of hardship. Across PICs, the rate of domestic violence experienced by women is very high. The share of women age 15 to 49 reporting ever experiencing either physical or sexual violence Other Types of Shocks Also perpetrated by an intimate partner (a husband or Affect Pacific Islanders but Are boyfriend) reaches well over 50 percent in Kiribati, the Solomon Islands, and Vanuatu (figure 3.15). More Difficult to Measure Sexual violence against children is also prevalent In addition to aggregate economic, natural, and in many PICs and may be related to the high rates health shocks, people in the Pacific are also vulnerable of unplanned pregnancies among young women. to aggregate sociopolitical shocks. Civil unrest, Violent shocks have direct physical, psychological, and violence, and political instability have occurred in emotional impacts that reduce well-being and are also several countries over the past decade, including likely to have effects on a broad range of economic Fiji, Papua New Guinea, the Solomon Islands, and outcomes. Research from around the world shows Tonga. These shocks affect people in many ways, that abused women earn less, costs to care for victims including physical violence, destruction of assets, and (when care is available) are high, and patterns of reductions in economic activity. For example, during violence are often passed from one generation to the civil unrest in 2006, 80 percent of the central business next (World Bank 2012c). district in Nuku’alofa was destroyed (World Bank 2008). Sociopolitical shocks often reduce confidence Unplanned pregnancy is another personal shock and create fear, which can affect tourist arrivals and that is much more common in the Pacific than in neighboring East Asia. Across the Pacific, fertility rates are high, and many pregnancies, particularly 4. The management of health risks and the fiscal challenges those occurring to young women, are unplanned presented by NCDs will be discussed further in chapter 4. (figure 3.16). These unplanned pregnancies can 46 | HARDSHIP AND VULNERABILITY IN THE PACIFIC ISLAND COUNTRIES Figure 3.15 Women Aged 15–49 Reporting Physical or Sexual Violence from an Intimate Partner, Select PICs and Asian Countries 80% 70% 60% 50% 40% 30% 20% 10% 0% Solomon Islands Thailand Timor-Leste Philippines Japan Cambodia Vanuatu Marshall Islands Kiribati Samoa Tuvalu East Asia Pacific Source: World Bank 2012c (compiled from Demographic and Health Surveys). Figure 3.16 Mistimed and Unwanted Pregnancies among Women under 20 Years of Age, Select PICs and Asian Countries 100% 90% 80% 70% Missing 60% % unwanted pregnancies 50% 40% % pregnancies wanted later 30% % pregnancies wanted 20% 10% 0% ds s lu ru oa s a a te m nd ne di si va es na au an m ne bo la pi Tu -L Sa et N sl Is ilip do am or Vi lI on In Ph al m C sh Ti m lo ar Sa M Source: World Bank 2012c (compiled from Demographic and Health Surveys). have adverse health and economic impacts, and the location and topography of PICs exposes them to a economic impacts may be particularly important for disproportionate number of natural shocks, and several young women who are still in school or just beginning are among those countries in the world with the most to work. That many women are unable to meet their relative disaster losses. Fuel and food imports, tourism, own preferences regarding pregnancy, as well as remittances, and international aid all contribute to the high rates of domestic violence and low rates of the well-being of Pacific islanders and help countries women’s representation in political leadership, are all overcome the limitations on development caused by indicative of the substantial gender inequality in the geography. However, their economies are still small region (World Bank 2012c). and undiversified, so negative shocks to these external flows can have very large impacts. Commonly occurring price shocks to commodity Key Messages imports and exports increase hardship substantially. People in the Pacific are uniquely vulnerable to People in the small countries of the Pacific are highly aggregate economic and natural shocks because exposed to high and volatile global commodity of their countries’ combination of small size, prices. Microsimulation analysis for Kiribati, Papua isolation, and other geographic features. The New Guinea, and Tonga finds that shocks to the prices Vulnerability and the Impacts of Shocks on Pacific Islanders | 47 of imported food and fuel, agricultural commodity Anderson, K., M. Ivanic, and W. Martin. 2013. “Food exports, and remittances push many people into Price Spikes, Price Insulation, and Poverty.” Policy hardship and deepen the severity of hardship for Research Working Paper 6535. World Bank, many others. The impacts of import price shocks Washington, DC. are particularly severe in the small atoll nations that Cororaton, Caesar, and David Knight. 2013. “Economic rely heavily on imports for staple foods and fuel. For Shocks and Their Impacts on Households: Micro example, in Kiribati, simultaneous spikes in the prices Simulation Results from Three Pacific Island of rice, wheat, and oil are estimated to push 6 percent Countries.” Background paper for this report. of the country’s entire population into hardship. This De Hoyos, Rafael, and Denis Medvedev. 2011. impact of a commonly occurring set of shocks (with “Poverty Effects of Higher Food Prices: A Global a likelihood of about 33 percent in any given year) Perspective.” Review of Development Economics is close to the estimated impacts of the 2007 global 15 (3): 387–402. food and fuel crisis on 20 of the most severely affected countries in the world. Dessus, Sebastien, Santiago Herrera, and Rafael De Hoyos. 2008. “The Impact of Food Inflation on Urban Poverty and Its Monetary Cost: Some The growing epidemic of noncommunicable Back-of-the-Envelope Calculations.” Agricultural diseases is an aggregate health shock with Economics 39 (S1): 417–29. significant consequences for the well-being of people in the Pacific. NCDs reduce productivity Gibson, John, and Bonggeum Kim. 2013. “Quality, and quality of life and are very expensive to treat. Quantity, and Nutritional Impacts of Rice Price Growth in NCDs has already eroded life expectancy Changes in Vietnam.” World Development 43: in Tonga. Most PICs are facing this epidemic 329–40. while also dealing with continued threats from IMF (International Monetary Fund). 2013. World communicable diseases and maternal and child Economic Outlook April 2013. Washington, DC: IMF. mortality. With limited fiscal resources, trying to ———. n.d. Primary Commodity Prices (database). manage this “double burden” of disease is a major Washington, DC: IMF. challenge for Pacific governments and will be Ivanic, Maros, and Will Martin. 2008. “Implications discussed in chapter 4. of Higher Global Food Prices for Poverty in Low- Income Countries.” Policy Research Working Paper In addition to aggregate shocks, people in the Pacific 4594. World Bank, Washington, DC. face many idiosyncratic and local shocks, but little Samoa Ministry of Natural Resources and Environment data are available to identify their frequency and (with GHD Ltd). 2013. “Vaisigano Flood of impacts. Crop failure, job loss, violence, and many December 2012: Report on Damage and Possible other idiosyncratic or localized shocks are likely to Causes.” Apia, Samoa. occur in the Pacific, as they do around the world. Some striking evidence of the prevalence of domestic Taufatofua, Pita. 2011. “Migration, Remittance, and violence and unwanted pregnancy shows that these Development in Tonga.” Food and Agriculture personal shocks are much more common in the Pacific Organization of the United Nations, Apia, Samoa. than in neighboring East Asian countries. However, Université Catholique de Louvain. EM-DAT: The existing household surveys are not designed to capture OFDA/CRED International Disaster Database. the full range of shocks that occur, particularly their Université Catholique de Louvain, Brussels. http:// impacts. Much more could be learned if future surveys www.emdat.be. asked specific questions about shocks and followed WHO (World Health Organization). n.d. Global Health households over time to measure the impacts of shocks. Observatory (database). WHO, Geneva, Switzerland. http://www.who.int/gho/. ———. 2010. “Pacific Islands Pay Heavy Price for References Abandoning Traditional Diet.” Bulletin of the World Health Organization 88: 484–85. Aba, L. A., B. Aipi, and T. Irau. 2012. “Supply Response Wodon, Q., and H. Zaman. 2008. “Rising Food Prices of Coffee in Papua New Guinea.” Bank of Papua in Sub-Saharan Africa: Poverty Impact and Policy New Guinea Working Paper. Responses.” Washington, DC: World Bank. Aipi, B., T. Irau, and L. A. Aba. 2012. “Supply Response World Bank. 2008. Tonga: Trade Brief. Washington, of Cocoa in Papua New Guinea.” Bank of Papua DC: World Bank. New Guinea Working Paper. 48 | HARDSHIP AND VULNERABILITY IN THE PACIFIC ISLAND COUNTRIES ———. 2012a. Acting Today for Tomorrow: A Policy ———. 2013b. Strong, Safe, and Resilient: A Strategic and Practice Note for Climate and Disaster Policy Guide for Disaster Risk Management in East Resilient Development in the Pacific Islands Region. Asia and the Pacific. Washington, DC: World Bank. Washington, DC: World Bank. ———. 2013c. Program Document: Samoa ———. 2012b. The Economic Costs of Non- Development Policy Operation. Washington, DC: Communicable Diseases in the Pacific Islands. World Bank. Washington, DC: World Bank. ———. 2013d. At Work in East Asia Pacific. ———. 2012c. Toward Gender Equality in East Asia Washington, DC: World Bank. and the Pacific. Washington, DC: World Bank. ———. n.d. World Development Indicators (database). ———. 2013a. Turn Down the Heat: Climate Extremes, World Bank, Washington, DC. http://data.worldbank Regional Impacts, and the Case for Resilience. .org/data-catalog/world-development-indicators. Washington, DC: World Bank. Vulnerability and the Impacts of Shocks on Pacific Islanders | 49 Technical Annex: Details of all sectors is linked to the price of the transport sector, which is in turn affected by the world price of oil. Microsimulation Modeling The microsimulation model used in this chapter Changes in the international price of commodities calculates a range of parameters and estimates for may not necessarily feed through to the same price three scenarios: the previous period, the baseline, and change for final consumption goods. To the extent the simulation scenario (see figure 3.A.1). The baseline that the consumer price is made up of other input and simulation scenarios are compared to consider costs and profit margins, these may dilute the impact. the effects of a shock. In addition to the standard Pricing behavior, industrial organization of the sector, sectoral breakdown of agriculture, industry, and and the sensitivity of suppliers and consumers to services, the model breaks down the economy into prices changes will all have an impact. However, it is further sectors, guided by the commodity groupings difficult to quantify these effects with precision, so for that are directly affected by the simulated shocks, simplicity, price shocks are directly applied to the final including coffee, cocoa, coconut and derivatives, palm market prices. In most cases, commodity price shocks oil, rice, wheat products, fruit and vegetables, fats and are directly applied to the commodity sector defined oil, fish, meat, other food and drink, fuel, energy, and in the model. An exception to this is an oil price electricity, transport, and other expenditure. shock, which feeds though to a variety of different sectors. An adjustment is made to the price of the Prices are set to a unitary value in the previous and electricity sector in Papua New Guinea to account baseline period. Although active baseline forecasts for the significant hydroelectric generation capacity can be incorporated, for purposes of this modeling in that country. Baskets of consumption weights Xi,t,k exercise it was not considered necessary. Prices for are established for i sectors and k groups, which are the relevant sectors are then altered in the simulation (1) national, (2) below the appropriate regional poverty scenario to reflect the price shock. Final consumer lines, and (3) for specific subgroups such as cash crop prices (PC) are comprised of producer prices, sales farmers below the poverty line. In line with assuming and excise tax rates (itx), import duties (tm), and no substitution effects, the weights are unchanged transport costs (trsp). The change in transport costs for between the baseline and simulation. With the Figure 3.A.1 Stylized Microsimulation Model Logic World prices, tariffs, indirect taxes, Baseline data transport costs Calibrate simulation For example, to price of specified commodity Baseline Price computation Simulation Income and poverty Rural/urban, region, subgroups line calculation Poverty head count Inflation Rural/urban, region Poverty depth Poverty severity 50 | HARDSHIP AND VULNERABILITY IN THE PACIFIC ISLAND COUNTRIES baseline poverty line decomposed into prices and poverty lines, and the results are presented as weightings, the poverty line is then reaggregated in hardship head-count, gap, and severity ratios. The the simulation scenario as follows: impact on consumer inflation is also estimated. Income shocks feed directly into the household’s real i 51 expenditure constraint, such that a change in total Et,k 5  PC X n i,t i,t,k income has an equivalent proportional effect on total where Xi 5 1 for n goods. expenditure. For example, if cash crop income drops by 20 percent, and made up half of total income, In the final stage, the model assesses household expenditure would fall by 10 percent. expenditure against both the baseline and simulation Current Approaches to Reducing Hardship and Vulnerability 51 Chapter 4 Current Approaches to Reducing Hardship and Vulnerability H ardship and vulnerability have many causes, People in the Pacific benefit from strong traditional ranging from challenges at the country level systems, but governments are increasingly recognizing to individual characteristics. As described the need to address hardship and vulnerability and in chapter 1, geographic features particular their underlying causes. This chapter provides an to the Pacific limit economic growth, reduce the overview of the strengths and limitations of some of effectiveness of government, and expose people the current approaches that households, communities, to many types of aggregate risks. These factors and governments take to reduce hardship and all contribute to the hardship and vulnerability vulnerability, utilizing the risk management framework experienced across PICs. At the same time, at the presented in chapter 1. Figure 4.1 groups these individual and the household level, many factors that approaches according to their function in the risk are common around the world contribute to hardship management framework. The first section reviews the and vulnerability, including limited human capital and main ways in which households and communities work lack of productive assets such as land (see chapter 2). to reduce hardship and manage risks, including the Figure 4.1 Current Approaches to Risk Management in the Pacific Basic services (Section 3) KNOWLEDGE of shocks, vulnerability, potential impacts Basic insurance INSURANCE PROTECTION (Section 3) to reduce the size of to reduce probability Basic services negative impacts from that negative impacts (Section 3) Traditional Networks shocks that do occur from shocks occur (Section 1) COPING to recover from negative impacts of shocks Traditional Networks Responses to Basic services (Section 1) economic shocks (Section 3) (Section 4) Primary actors Government and international partners Households and communities 52 | HARDSHIP AND VULNERABILITY IN THE PACIFIC ISLAND COUNTRIES limited quantitative evidence of their effectiveness. countries.1 Melanesian wantoks are defined in part Sections two through four then focus on the various by ethnic identity and are often led by “big men.” In efforts of governments and their development the ethnically diverse countries of Melanesia, tribe partners to reduce hardship and vulnerability. The and language are important determinants of who is second section provides an overview of the programs in the same wantok. The term wantok itself comes through which Pacific governments directly assist from the Tok Pisin word for “one talk” or sharing those in hardship and provide social insurance. The the same language. In some countries, including third section reviews the much more widespread Fiji, political alliances also play a role in defining provision of free services by governments, which networks. In Melanesian countries, networks tend can address individual causes of hardship, as well to operate around “big men” or male chiefs who as reduce vulnerability by increasing knowledge, lead their communities. Power can be based on protection, and facilitating coping. This section hereditary inheritance or on individual merit and focuses in particular on Pacific governments’ role accomplishments, and this varies from community to in providing health care. The fourth section then community. Across countries, “big man” systems are discusses Pacific governments’ current approaches to well integrated into formal political structures. managing risks from aggregate shocks, with a focus on economic shocks. The final section summarizes the In Micronesia and Polynesia, network structures range key findings. from egalitarian to chief-based leadership. In Kiribati, for example, decision making tends to be more communal, but this differs across islands. Southern The Role of Traditional Networks islands are traditionally led in a more egalitarian manner, with the elders of each community (unimane) in Reducing Hardship and ruling by consensus. In the central and northern Vulnerability islands, chiefly leadership is more common. Across Kiribati, the physical representation of this type of Traditional networks in the Pacific may reduce group leadership is the maneaba, a meeting house hardship and vulnerability through exchange and in the center of each community where people come subsistence resources. Across most PICs, relationships together for decision making, celebrations, and other between people based on blood relation, village, activities. In contrast, Samoa has a strictly hereditary language, and other factors have long been system of mainly male chiefs known as matai. The established by custom (Nanau 2011; Sviridova 2013). matai are intrinsic to Samoa’s modern government: These relationships are the basis of networks that Most local legal matters are handled by councils of take different forms across countries, such as wantok matai, and only matai are allowed to serve in Samoa’s in utu in Kiribati and Melanesia. Norms of behavior national legislature.2 and access to land are based on these networks, as is political power in many countries (Fukuyama 2008). Norms of behavior help define one’s obligations to A strong ethos of resource sharing is common to others, including how resources should be shared traditional networks across the Pacific. Asking for between individuals, households, and communities. goods and services from one’s network members in The majority of land in many PICs is owned by custom, times of need is common practice, with the general giving everyone access to some land in principle. Land understanding that this behavior will be reciprocated. provides livelihoods, helps define identity, and also Meeting these requests, often called gift giving, holds deep spiritual meaning for people. Across most is considered an important obligation, and failing of the Pacific, land and ocean resources continue to provide for large shares of household food, housing, and other needs. Although these traditional networks 1. This section draws on the sourcebooks produced by the have long histories, it is important to remember that International Labor Organization’s Sub-regional Initiative on they are dynamic and continue to adapt to changing Social Security in Pacific Island Countries, as well as Sviridova social and economic realities. (2013). 2. As an example of the dynamic nature of tradition, a greater Exchange in Traditional Networks number of matai have been created in recent years, thus Family forms the core of traditional networks, but expanding the share of the population that is eligible to network structures vary widely across and within serve in the legislature and other governing roles. Current Approaches to Reducing Hardship and Vulnerability | 53 to do so brings shame and can weaken network members. In Fiji, these large-scale collections are ties. In some countries, expectations for resource called solevu, and matrilineal relatives are expected sharing differ based on network relationships. For to contribute more. In addition to ceremonies, example, individuals may expect different levels of communities also come together to raise resources generosity from matrilineal and patrilineal relatives. for community-level needs such as utilities or public Few limitations are in place on what resources can be buildings. In Samoa, fa’alavelave encompasses the shared, but three specific methods of resource sharing resource sharing obligations related to major events, are particularly important: specialized exchange, as well as broader principles of showing respect for generalized reciprocity, and communal collection. others through gift giving. In many PICs, this type of resource sharing is now also widely utilized by Specialized exchange occurs when individuals or churches and religious organizations. households exchange goods or services of similar value with each other. In the isolated and small The effectiveness of traditional networks in reducing communities of the Pacific, this type of exchange hardship and vulnerability is not fully understood. helps diversify the consumption goods available to Norms of resource sharing are still strong in the each member of a community and helps allocate labor Pacific, and anthropological research and anecdotal to where it is needed. For example, in Fiji, solesolevaki reports suggest that individuals and households refers to the exchange of collective labor, which can facing economic, personal, or other difficulties benefit be used when households undertake construction from resources shared by better-off members of their of a home or engage in other activities. Households networks. Failures to fulfill gift-giving obligations in often exchange smaller amounts of labor or goods these networks could result in exclusion and inability on a frequent basis, such as when women look after in the future to receive assistance when needed. each other’s children or exchange food and household Traditional networks therefore both support those items. Specialized exchange also takes place across experiencing ongoing hardship, as well as act as networks, such as when inland and coastal tribes informal insurance. At the same time, three related exchange goods through barter in the Solomon considerations point to the limitations of traditional Islands. In general, specialized exchange occurs networks: Networks do not reach everyone, network bilaterally, but in communities with chiefs or big men, obligations may help perpetuate hardship in some these leaders can play a role in overseeing these cases, and urbanization and monetization may weaken activities and resolving disputes. or obviate traditional ties over time. As Morauta (1983: 8) states, “Transfers in the wantok system are Generalized reciprocity occurs when an individual or not transfers of charity or in a state welfare program. household provides resources to another individual They are part of a system of personal obligation, and or household in need. This principle of resource some people who badly need transfers have nobody sharing goes by different names across the Pacific, to help them.”3 such as kerekere, which means “to ask” in Fiji, and bubuti, which means “request” in Kiribati. As with all Analysis of gifts and remittance data sheds some forms of resource sharing, reciprocation is generally light on the role of traditional networks in reducing expected. However, in practice, this type of exchange hardship and insuring people against negative is thought to be redistributive. This is because better- impacts from shocks. Few quantitative studies of off individuals and households are expected to share the impact of traditional networks on hardship have resources with less well-off relatives and community been conducted because of lack of data. In one of members, who are often unable to reciprocate fully. the few, Gibson (2006) utilizes data from the 1980s and 1990s and finds that interhousehold transfers and Communal collection includes contributions made for remittances do help to reduce inequality in Tonga and ceremonial events, as well as resource collection for urban Papua New Guinea, but not in rural Papua New community-wide needs. Across the Pacific, households Guinea. The same study also finds that indicators of are expected to provide food or other resources for ceremonies and celebrations, including weddings and funerals. These resources are shared with all those 3. Morauta studied both urban squatter settlements and rural attending, and, depending on the event, remaining villages in Papua New Guinea and found limited evidence resources may be partitioned among community that wantok obligations reduced hardship (Gibson 2006). 54 | HARDSHIP AND VULNERABILITY IN THE PACIFIC ISLAND COUNTRIES Figure 4.2 Household Gift Giving. percentage of total households Fiji Kiribati 100% 100% 90% 90% 80% 80% 70% 70% 60% 60% 50% 50% 40% 40% 30% 30% 20% 20% 10% 10% 0% 0% 1 2 3 4 5 6 7 8 9 10 1 2 3 4 5 6 7 8 9 10 Gift Received Gift given Gift Received Gift given Papua New Guinea Samoa 100% 100% 90% 90% 80% 80% 70% 70% 60% 60% 50% 50% 40% 40% 30% 30% 20% 20% 10% 10% 0% 0% 1 2 3 4 5 6 7 8 9 10 1 2 3 4 5 6 7 8 9 10 Gift Received Gift given Gift Received Gift given Tonga Tuvalu 100% 100% 90% 90% 80% 80% 70% 70% 60% 60% 50% 50% 40% 40% 30% 30% 20% 20% 10% 10% 0% 0% 1 2 3 4 5 6 7 8 9 10 1 2 3 4 5 6 7 8 9 10 Gift Received Gift given Gift Received Gift given Source: World Bank staff estimates using HIESs. Note: Data come from 14-day consumption diaries in each country, meaning that they capture relatively high-frequency exchanges. Two exceptions: for Fiji, data include gifts received in the 12 months prior to survey, and for Kiribati, data include remittances received from abroad. recent shocks such as births and unemployment show Many households participate in gift exchange, but, in no clear relationship with the net receipt of gifts. To some countries, households experiencing the deepest contribute to the limited evidence base, analysis of hardship are the least likely to participate. Figure 4.2 gift and remittance data in the HIESs of Fiji, Papua shows the share of households reporting that they New Guinea, Samoa, and Tuvalu was conducted. In either gave or received any gift of goods or cash in most PICs, these surveys ask about the source of food the two-week period of consumption diary keeping, and other items that households report consuming, by expenditure decile.5 The prevalence of gift activity and they ask if gifts of cash or large items were varies across countries: Giving or receiving gifts is less received. Households are also asked whether they common in Fiji and most widespread in Papua New gave any gifts, in cash or in kind. Survey data are Guinea, where more than 80 percent of households highly imperfect but do provide some useful insights, report participating. In four of the six countries which are detailed below.4 studied, households in the lowest expenditure decile are the least likely to report receiving any gifts. This suggests that many households experiencing the 4. HIES are not generally designed with the specific objective most severe hardship are not receiving assistance of learning about traditional networks. Most importantly, survey questions related to gifts are asked using different time frames for gifts given versus gifts received. For example, in Fiji, households are asked to recall gifts received in the 5. To calculate expenditure deciles, gifts received and given are 12 months preceding the survey but are not asked about first subtracted from expenditures, and remaining expenditures gifts given in that time period. are normalized by adult equivalent household size. Current Approaches to Reducing Hardship and Vulnerability | 55 Figure 4.3 Relation to Expenditures on Gift Giving to Household Expenditures in Select PICs, percentage of total household expenditures Fiji Papua New Guinea 20% 40% 18% 35% 16% 30% 14% 12% 25% 10% 20% 8% 15% 6% 10% 4% 2% 5% 0% 0% 1 2 3 4 5 6 7 8 9 10 1 2 3 4 5 6 7 8 9 10 Gift Received Gift given Gift Received Gift given Samoa Tonga 10% 20% 9% 18% 8% 16% 7% 14% 6% 12% 5% 10% 4% 8% 3% 6% 2% 4% 1% 2% 0% 0% 1 2 3 4 5 6 7 8 9 10 1 2 3 4 5 6 7 8 9 10 Gift Received Gift given Gift Received Gift given Source: World Bank staff estimates using HIESs. Note: Data come from 14-day consumption diaries in each country, meaning that they capture relatively high-frequency exchanges. For Fiji, data also include gifts received in the 12 months prior to survey. from traditional networks. Several factors could Gift exchange may act as an informal insurance as be related to these households’ inability to obtain well. On average, households in Kiribati, Papua New assistance, including a lack of family members (which Guinea, and Tuvalu with one or more morbid (sick or could contribute to both hardship and disconnection injured) members are more likely to participate in gift from networks) or previous failure to meet network exchange than healthy households. Figure 4.4 shows obligations. that a slightly greater share of these households both give and receive gifts, compared with households with For households that do participate, gift exchange members who are all healthy. However, little evidence is appears to transfer net resources to those available to suggest that the net value of gifts received experiencing the deepest hardship. Figure 4.3 shows by households experiencing morbidity is any higher the average value of gifts given and received by decile than for other households. The positive correlation of household expenditure, for all the households between gift giving and morbidity might suggest reporting that they either gave or received any that households are more active in gift exchange gifts. Across the four countries with sufficient data, networks around the time at which a member falls households in the lowest deciles tend to receive ill. For example, households might seek to “repay” gifts of greater value than they give, whereas the gifts that were received at a time when someone was opposite is true for households in the highest deciles. sick soon after that person has recovered. The data These results suggest that for households with active might also suggest that households prone to ill-health traditional networks, gift exchange does help to self-select into gift-exchange support networks; this transfer resources to those experiencing hardship. phenomenon—termed “adverse selection”—is not However, in Samoa, gift exchange does not appear uncommon in insurance markets. In any case, the to favor households in the lowest deciles, who link between morbidity and gift exchange highlights report receiving about twice as much as they gave, the contribution social support networks can make in which is similar to households much higher up in the helping Pacific households to cope with idiosyncratic expenditure distribution. negative shocks. 56 | HARDSHIP AND VULNERABILITY IN THE PACIFIC ISLAND COUNTRIES Figure 4.4 Share of Households Participating in Gift Exchange, percentage of total 90% 80% 70% 60% 50% 40% 30% 20% 10% 0% Receiving Giving Receiving Giving Receiving Giving Papua New Guinea Kiribati Tuvalu Households with a sick or injured member Healthy households Source: Halstead 2013. Gift exchange is unlikely to be effective in insuring accord with anecdotal evidence that all members are against aggregate or repeated shocks. For example, expected to contribute, and those who cannot afford when shocks covary at the local or aggregate level, to contribute directly often put in a great deal of time network members are likely to all be impacted within to fundraise. These types of efforts are less likely to be a similar time frame, making it difficult to transfer captured in the survey data, which suggests that the resources to each other. Similarly, shocks that occur donation figures are underestimates. Churches and with some frequency can be too costly for networks other organizations utilize the support they receive to insure. More broadly, the limited global evidence for many purposes, but these are not captured in the that exists finds that informal insurance does not come available survey data. close to replacing lost resources (Udry 1994). Gifts and donations help determine people’s social In addition to gift exchange, many households also standing, and obligations may sometimes contribute give donations to their churches or to community to hardship. Traditional network obligations extend causes. Religion is an important part of life for far beyond caring for those in need and include gifts many people in the Pacific, and churches and other for important family events, such as weddings, and religious organizations often rely on the financial community efforts, such as building new structures. and in-kind support of their members. Across three Contributing generously can provide individuals countries with available data—Papua New Guinea, or households with social approval and esteem, Tonga, and Tuvalu—a large share of households whereas failing to do so can provoke disapproval report donating cash, goods, or services to religious and remonstrance from friends and neighbors. organizations. Although the households deepest in These social pressures sometimes push households hardship (in the bottom expenditure decile) are the to contribute more than they can truly afford. The least likely to report donating, the share of households evidence presented above that many of the least reporting donating does not increase substantially well-off households report giving gifts and donations with expenditures (figure 4.5). In Samoa, almost all provides suggestive support for this possibility. In households report giving to churches, community qualitative surveys carried out in the early 2000s, causes, or fa’alavelave, regardless of the household’s participants in Samoa, Tonga, and Tuvalu cited own hardship status. Across countries, donations are the burdens of meeting community and church sizeable, ranging from 2 to 10 percent on average of obligations as one of the top causes of hardship total annual household expenditures. These results (Abbott and Pollard 2004). A recent indication of these Current Approaches to Reducing Hardship and Vulnerability | 57 Figure 4.5 Share of Households Contributing to Religious Organizations in Past 12 Months in Tonga and Tuvalu 100% 90% 80% 70% 60% Papua New Guinea 50% Tonga 40% Tuvalu 30% 20% 10% 0% 1 2 3 4 5 6 7 8 9 10 Source: World Bank staff estimates using HIESs. challenges comes from an April 2013 Samoa Observer of which was seen in chapter 3). However, in some headline that reads “Put children first, fa’alavelave cases, such as when young people move to new second, says Samoa’s Minister of Education.”6 locations on their own, network ties weaken and obligations become less relevant, possibly leaving an Indirectly, these obligations may contribute to increasing number of households without traditional hardship by potentially reducing incentives for sources of assistance. economic success. Newspaper accounts and anecdotes assert that immigrants are often the most Traditional networks do not eliminate hardship or successful businesspeople in the Pacific because they vulnerability, but more data are needed to better are not obligated to share their wealth or grant favors, understand their role. The results presented above while native civil servants and others employed in the show that the available data suggest many households relatively small formal sector struggle under the heavy experiencing hardship do not receive gifts through obligations imposed by their traditional networks. traditional networks. In addition, the results from chapter 2 show that hardship can exist at high levels The relationship between traditional networks and in countries where gift exchange is widespread, such hardship may be changing as people migrate to urban as Papua New Guinea. However, these results provide areas or overseas. Movement of large numbers of only a partial picture of gift exchange, because people away from rural villages to capital cities or to comparable data on giving and receiving are collected international destinations (most often, Australia, New only over very short time periods. In addition, other Zealand, or the United States) has and will continue to important aspects of traditional networks, such as shape traditional networks in the Pacific. Sometimes, the exchange of services, are not easily captured by such as when many members of the same network standard household surveys. Given that traditional move to the same new location, practices from home networks are often cited as a reason for limiting villages or islands are continued. Many international the role of government in supporting households migrants maintain ties with their networks and help to experiencing hardship, better understanding of the reduce hardship through remittances (the importance current context is important. Land and Subsistence Many people in the Pacific rely on land and ocean 6. Byline April 3, 2013 http://www.samoanews.com/ resources to meet their basic needs, either on an ?q=node/73675. ongoing basis or as a coping mechanism. Although 58 | HARDSHIP AND VULNERABILITY IN THE PACIFIC ISLAND COUNTRIES geography varies tremendously across PICs, in most, complex issues, which can create challenges for land is an important social, cultural, and economic broader economic development. asset (Ratuva 2010). In part, the importance of land comes from the fact that it allows people to provide Urbanization and monetization are also affecting the for themselves and their families. With the small size sustainability of subsistence practices. Although most of domestic industries and high and variable costs of Pacific peoples still live in rural areas, many are moving imports, self-production is an important resource for to cities, where the concentration of people creates households to provide for many of their basic needs. challenges for self-production. Higher population The majority of households report self-producing densities and lack of customary access to land means (cultivating, growing, hunting, or fishing) some of that many urban residents have little to no room for the food they eat, even in many urban areas (see cultivation. In addition, inadequate infrastructure in chapter 3). In a survey of communities in the Solomon urban areas means that land and ocean resources Islands and Vanuatu, increasing the share of food are becoming increasingly polluted. One of the most sourced from home gardens and reefs were the extreme examples is in South Tarawa, Kiribati, where two most commonly reported coping responses by poisoning from fish caught in the lagoon, illness from households to negative shocks (Feeny et al. 2012). polluted groundwater, and other adverse outcomes However, several factors are affecting the role of are on the rise (McKay 2009). Pacific economies are self-production. also becoming more monetized, because imported food and goods are in demand and obtainable Arable land and coastal fisheries are limited and only with cash. Imported food, for example, is often straining under the pressures of population growth. preferred to local food, because of the ease in Figure 4.6 shows that the average amount of arable preparation and relatively higher fat, salt, and sugar land per capita in most PICs is well under the average content. In addition to their health implications, these for both low- and middle-income countries as well as shifts may reduce the role of subsistence resources small states. Land is becoming scarcer over time, as over time as methods of production begin to be relatively fast population growth continues in many forgotten. countries. For example, in Tonga, all men over age 16 have the constitutional right to an allotment of land, to Additional Mechanisms That be provided by landowning members of the nobility, Households Use to Reduce but currently fewer than 50 percent have received their Hardship and Vulnerability allotments because of a lack of available land (Tonga Many people migrate in search of better economic Ministry of Finance and National Planning 2012). opportunities. Given the limitations to private sector In most PICs, rights to land and its use are deeply growth in much of the Pacific, opportunities for Figure 4.6 Arable Land per Capita, hectares 0.25 0.20 0.15 Low- and middle- 0.10 income average 0.05 Small states average 0 ji a te u u a oa ls s i . at ts Fi on ng at la ne l es ha rb .S m Pa nu m To ui -L Sa Ki s ed G lo Va ar or So F M ew m a, Ti N si a ne pu ro Pa ic M Source: World Development Indicators. Current Approaches to Reducing Hardship and Vulnerability | 59 work outside of the public sector and subsistence eaten as common coping mechanisms in qualitative activities are few. Pacific islanders have a long history surveys, but their actual prevalence in the Pacific is not of migrating overseas to work, on either a temporary clearly understood (Abbott and Pollard 2004; Feeny or permanent basis. Many of these migrants remit et al. 2012). To the extent that they do occur, these substantial amounts of money, helping to reduce coping mechanisms can increase hardship in the long hardship. These shared earnings usually go directly term, as they erode health and human capital. to individuals and households, who utilize them for consumption, investment in assets, and other purposes (Gibson, McKenzie, and Stillman 2011; Government Social Insurance Gibson and McKenzie 2014).7 In addition, migrants can provide informal insurance when negative Programs shocks hit family and friends at home. Although Few government programs in the Pacific directly Pacific migrants tend to concentrate in industries target households experiencing hardship. Fiji is the or countries (as discussed in chapter 3), the large only Pacific country with a national cash support benefits possible from migration suggest that more program targeted to households experiencing opportunities should be sought, while accounting for hardship. This Family Assistance Program benefited the risks associated with concentration. For example, about 13 percent of the population in 2009 by governments could encourage households to save providing cash payments to households experiencing a share of remittances received, as self-insurance hardship and other difficulties such as disability (World against future drops. Bank 2011). In other PICs, smaller programs benefit select groups of people, such as education grants for In some countries, traditional stores of value can be children with disabilities (Sviridova 2013). In addition, used by households lacking cash to obtain needed some countries have recently begun programs that goods and services. For ni-Vanuatu, pig tusks, woven target people vulnerable to violence (see box 4.1). mats, and shells with specific characteristics hold substantial value. For rural households with few Public sector capacity, fiscal constraints, and equity opportunities to earn cash, producing these valuable concerns are three of the main barriers to creating items can allow them to obtain services they may not such programs. In many countries, administrative otherwise be able to access. For example, some rural data on households or individuals are very limited, schools and health centers accept these items as and national household surveys and censuses are forms of payment (Huffman 2005). carried out infrequently. Without accurate and timely data, identifying households most in need Few private sector options are available for helping of assistance is very difficult. In addition to lacking households insure against or cope with shocks. In data, many governments in the Pacific would also many PICs, formal financial markets are not well face challenges in identifying the personnel and developed, and relatively few households have savings financing needed to establish and administer new accounts, purchase insurance, or obtain bank loans. In national programs. Across the Pacific, governments Kiribati, only 2 percent of households in the 2006 HIES have very limited fiscal space with most already at reported paying premiums on any type of insurance high risk of debt distress. Half are expected to run policy, and only 3 percent of households in the Papua large budget deficits in the current fiscal year.8 As New Guinea 2010 HIES reported taking a loan from such, funding for any new programs would likely need a bank or moneylender. With limited options in the to come from international development partners. private market, many people cite taking children out Even if data and fiscal space were available, some of school and reducing the amount and quality of food policy makers express concern that hardship-targeted programs could undermine traditional networks or be inequitable. Little direct evidence exists regarding the possibility of crowding out, but in one study of several 7. The study by Gibson et al. (2011) of the impacts of countries, including Papua New Guinea, Gibson, permanent migration on sending households finds that Olivia, and Rozelle (2006) find that increased public consumption and other indicators of well-being may decrease for the household members left behind, because remittances do not fully compensate for the loss of migrants’ earnings at home. However, seasonal migration schemes 8. Information taken from joint IMF–World Bank debt seem to significantly increase the well-being of sending sustainability analysis, various countries; IMF World Economic households (Gibson and McKenzie 2014). Outlook April 2013. 60 | HARDSHIP AND VULNERABILITY IN THE PACIFIC ISLAND COUNTRIES Box 4.1 Reducing Vulnerability through a Rapid Employment Program in the Solomon Islands Government social insurance programs can be targeted in many different ways, and recent projects financed by the World Bank in Papua New Guinea and the Solomon Islands have focused on targeting groups vulnerable to violence. In the Solomon Islands, a Rapid Employment Project (REP) was created in 2010 to mitigate the impacts of planned government austerity measures and to reduce the probability of social unrest.1 The REP provides basic training and work (approximately 50 days per year) in basic public maintenance activities such as pothole repair and grass cutting. The REP targets urban youth, who have been key actors in previous violence, as well as women, who have fewer labor market opportunities and are also vulnerable to gender-based violence. In addition, REP activities are geographically targeted by being based in communities within Honiara and surrounding periurban areas where violence had occurred in the past. Periodic assessments of the REP have found that the project is succeeding in targeting vulnerable groups: More than 50 percent of participants are youth (16 to 34 years old), and more than 50 percent are women. It is important to note, however, that targeting people vulnerable to violence or other shocks does not equal targeting people in hardship. Estimates find that about 40 percent of REP participants come from the bottom 40 percent of the per capita consumption distribution of Honiara. In other words, participants are about equally likely as the average person to be in hardship. 1. From 1998 to 2003, Solomon Islanders lived through the “Tensions,” a period marked by conflict and social unrest. Violent riots broke out again in 2006, and an international peacekeeping force is still present in the country. Sources: World Bank 2010; World Bank staff estimates. transfers would not be likely to significantly reduce improve the schooling and nutritional outcomes of private transfers. Concerns about equity are difficult to children in the same household, so among the elderly address without robust data that can make clear which living with extended family, the benefits may be shared households are truly in need of assistance. Chapter 5 (Yoong, Rabinovich, and Diepeveen 2012). provides additional discussion of how some countries have overcome these types of barriers. Government Provision Many governments do administer contributory insurance programs for the elderly, but these of Basic Services programs reach only a small number of people. Many governments in the Pacific provide basic Almost all Pacific governments mandate that formal services to the whole population, which can help to sector employees and their employers, including civil reduce hardship and vulnerability. Across most PICs, servants, contribute to pension funds (Sviridova 2013). basic education and health services are intended to However, because formal sector employment is so be provided at low or no cost. Such provision, when small, these funds cover very few people. In Kiribati, effective, helps people maintain their health and fewer than 4 percent of households in the HIES report productivity and also build their human capital to either contributing to or receiving payments from the reduce hardship in the future. These services can also National Provident Fund, and fewer than 1 percent of build people’s knowledge of different types of shocks households do in Papua New Guinea. and lead them to take better protection and insurance measures to reduce their vulnerability. For example, in Some governments have created noncontributory, or a country on the other side of the Pacific, Colombia, social, insurance programs for the elderly. In Kiribati, young people who received effective sex education the Elderly Fund provides monthly payments to all at school were more likely to utilize vouchers for individuals age 67 and older, regardless of their work condoms (Chong et al. 2013). history. These types of funds are relatively simple to administer, because proof of age is the primary Accessibility and quality are still lacking in many requirement for eligibility. In addition, because countries. The physical isolation of many communities qualification ages are set relatively high, the pool of within the Pacific makes it costly and logistically eligible individuals remains small and costs limited. At difficult to deliver services to everyone. Even when the same time, such funds have the potential to reduce free services are accessible, their quality may be hardship, because households headed by elderly inferior to the for-fee services provided by religious people are more likely to be in hardship (see chapter 2). organizations or the private sector. In addition, in Transfers to elderly women have also been found to several countries, services are in reality not available Current Approaches to Reducing Hardship and Vulnerability | 61 for free. In some cases, service providers charge sick or injured household member reported having fees because they are not adequately funded or sought treatment (Halstead 2013). regulated by the government. In other cases, there are additional costs that people must incur to receive Health care expenditures are mostly curative and services. For example, families must usually pay for the often support costly programs that benefit a limited school supplies, uniforms, and transportation required number of people. Figure 4.8 shows the breakdown for children to attend school. of health expenditures in three PICs by type. The majority of expenditures are for curative and Health Care rehabilitative care, occurring after negative health Most basic health care services are provided at low shocks have occurred. Given the small size of their or no cost and on a universal basis in the Pacific. health sectors, several Pacific governments fund Health care can protect people from negative shocks overseas treatment for serious health shocks, but (such as vaccinations against common diseases) as these come at significant expense. For example, in well as help them cope after a shock has occurred Samoa, the spending on each patient in the Overseas (such as dialysis treatment for individuals with kidney Medical Treatment Visits (OMTV) scheme is about failure). Across all health care provided in most 4.5 times GDP per capita. The scheme funds care PICs, the majority is publicly funded, and individuals for 0.1 percent of the population at a cost of about bear about none of the cost to 20 percent directly 15 percent of total annual health expenditures (figure 4.7). These direct costs are primarily out-of- (World Bank 2012b). Data on patient characteristics pocket, because private market insurance does not are not available, but there is reason to believe that exist in many countries, and where it does exist, it these types of programs tend to benefit better-off is held by only a small number of people. However, households. In Samoa, the types of illnesses treated these national health expenditures are unlikely to overseas are much more commonly reported among capture all health spending, because many people in the well-off, and the costs borne by patients for the Pacific continue to utilize and pay for the services overseas treatment (including paying for part of their of traditional healers. In addition, free care is not airfare) may be large enough to dissuade those in specifically targeted to the least well-off households, hardship from seeking treatment (World Bank 2012b). and many people may be forgoing needed treatment. For example, in Papua New Guinea, local health The fiscal sustainability of publicly provided curative centers charge fees, in part because they do not care is a growing concern, because PICs face costly receive adequate funding from the government, and aggregate health shocks. In particular, the rapid rise slightly fewer than 60 percent of households with a of NCDs is raising difficult questions of how to fund Figure 4.7 Composition of National Health Expenditures, 2010–2011 100% 90% 80% Other private 70% 60% Out-of-pocket 50% 40% General government 30% Share of total spending 20% from external sources 10% 0% Fiji Kiribati Papua Samoa Tonga Tuvalu Vanuatu LMIC UMIC New Guinea Sources: World Bank 2012a; Halstead 2013. Note: LMIC = lower-middle-income country; UMIC = upper-middle-income country. Other private expenditures include private insurance and nonprofit institutions. Data for Samoa, Tonga, and Vanuatu are for 2010. Data for Fiji, Kiribati, PNG, and Tuvalu are for 2011, and share of spending for external sources for those countries is an average from 2007 to 2011. 62 | HARDSHIP AND VULNERABILITY IN THE PACIFIC ISLAND COUNTRIES Figure 4.8 National Health Expenditures by Type 100% 90% Prevention and public health 80% services 70% Ancillary services, administrative, 60% and other 50% Pharmaceuticals and medical 40% durables 30% Long-term nursing care 20% Curative and rehabilitative care 10% 0% Fiji Micronesia, Fed. Sts. Vanuatu Sources: FSM Department of Health and Social Affairs 2010; Hyoung-sun and Rannan-Eliya 2010. Figure 4.9 Households Accessing on Unprotected Water as Drinking Source 80% 70% 60% 50% 40% 30% 20% 10% 0% Papua New Guinea Kiribati Fiji Tuvalu Source: World Bank staff estimates using HIESs. available treatments and determine beneficiaries. is unsustainable, and expanding coverage of market Relative to countries at similar income levels, PIC insurance would be too costly. governments already spend substantially more on health care (Halstead 2013). These costs are At the same time, in some countries, basic increasing, as the incidence of NCDs continues to infrastructure to protect people’s health still does not rise. For example, the costs of the OMTV scheme exist. For example, in Papua New Guinea, the majority in Samoa grew 30 percent between 2008 and 2012 of households did not have access to protected (World Bank 2012b). In Vanuatu, the annual cost of drinking water, and in Kiribati that figure reached medicines alone for a person newly diagnosed with about 40 percent (figure 4.9). In addition, only about type 2 diabetes is five times more than per capita 20 percent of households in Papua New Guinea have government expenditure on health care (World access to an improved toilet facility (Halstead 2013). Bank 2012a). Between 2011 and 2030, the rate of diabetes among adults in Vanuatu is forecasted to rise by 20 percent. Already, PICs have three of the Government Policies to five highest rates of diabetes in the world: Kiribati, Manage Aggregate Economic the Marshall Islands, and Nauru (World Bank 2012a). In risk management terms, because NCDs are a and Natural Shocks high-impact, aggregate, and long-term shock, Governments and their development partners are coping through government funding of treatments increasingly pursuing active strategies to reduce the Current Approaches to Reducing Hardship and Vulnerability | 63 impact of aggregate economic and natural shocks Pacific are very vulnerable to these types of shocks. on households. This section reviews the common Commonly used approaches are first reviewed, approaches that governments take to managing followed by an overview of less common measures that the risks associated with commodity price shocks may hold promise for the region. The use of national in particular. As detailed in chapter 3, people in the funds to manage risks is highlighted in box 4.2. Box 4.2 Using National Funds to Manage Risks and Opportunities Half the countries of the Pacific have national trust funds or sovereign wealth funds, although they differ considerably in their structure and objectives. Kiribati, Palau, and Tuvalu have funds that aim to provide long-term dividends and fiscal smoothing, although the level of government control of disbursements varies. The Federated States of Micronesia and the Marshall Islands both have trust funds that are still being capitalized and so are not yet providing financing. Papua New Guinea is currently in the process of launching a sovereign wealth fund capitalized with natural resource receipts, and Timor-Leste holds significant assets from petroleum receipts in a trust fund. The extent to which these funds serve as buffers in the face of shocks varies and depends on the management structure and the level of discretion the government has to increase withdrawals. Tuvalu’s is one that is structured to balance the needs for long-term financing with responding to short-term shocks. The Tuvalu Trust Fund (TTF) was established in 1987. The TTF initially comprised a single capital account, which was capitalized by donors (Australia, New Zealand, Japan, the Republic of Korea, and the United Kingdom) and the government. The TTF’s capital has grown over the years through reinvestment of its own earnings and contributions by the government during surplus periods. The TTF is not fully sovereign, with development partners represented on its board. The TTF’s capital account aims to generate a real rate of return of 4.5 percent in excess of the Australian Consumer Price Index. Recognizing the volatility associated with the returns on the capital account and with key revenue streams (particularly fishing licenses), in 1991 the TTF Board of Directors endorsed the creation of a new account under the government’s full control, the Consolidated Investment Fund (CIF), to provide a fiscal buffer against budget deficits. The CIF aims to accumulate a minimum balance equivalent to 16 percent of the TTF’s real maintained value (or around 45 percent of GDP), estimated to be sufficient to see the government through a spell of up to four years of bad times, during which time the CIF could finance significant budget deficits. In years when the market value of the TTF exceeds its real maintained value, the surplus is transferred to the CIF, and during negative shocks, government makes larger withdrawals from the CIF. There are no set procedures for these activities, which depend on the current fiscal policy framework, but the Articles of Agreement dictate that assets of the capital account can be withdrawn under extraordinary circumstances. Figure B4.1.1 Evolution of TTF and CIF Balances 140 A$ millions 16 A$ millions 14 130 12 120 TTF Year- end Target 10 110 Value 8 CIF 6 Balance 100 4 CIF Balance TTF Year- 90 (without additional end Actual 2 donor contributions) 80 Value 0 2005 2006 2007 2008 2009 2010 2011 2012 2005 2006 2007 2008 2009 2010 2011 2012 Note: CIF = Consolidated Investment Fund; TTF = Tuvalu Trust Fund. As figure B4.1.1 shows, in “good times” the government maintained the value of the TTF above the target value and built up assets in the CIF. With the onset of the global financial crisis the value of the TTL dropped, and increased fiscal deficits necessitated drawdowns from the CIF. However, although the TTF value was progressively built back up beginning in 2008, funds in the CIF would have been depleted if not for additional donor contributions. This illustrates the important role of development partners when even well-designed risk management frameworks are unable to cope with the scale of a major shock. Source: Tuvalu Trust Fund Advisory Committee Annual Reports. 64 | HARDSHIP AND VULNERABILITY IN THE PACIFIC ISLAND COUNTRIES In recent years, policy responses in the Pacific to exceeded 100 percent in recent years. Reducing taxes commodity price shocks have tended to focus on can also exacerbate problems of fiscal sustainability, coping through macroeconomic and aggregate fiscal which are already a major challenge for many actions. In particular, exchange rate movements, tax PICs. Even when tax reductions are intended to be policies, subsidies, and price controls have all been temporary, it can be politically difficult to raise them implemented in different PICs in recent years in again at a later date, and such measures can easily response to price shocks. Although to some extent become permanent. effective in reducing the impact on households, these measures may be poorly targeted and expensive, Some countries have used producer subsidies to and they have adverse consequences on fiscal reduce the pass-through of fuel price increases sustainability and economic prospects. to consumer prices, but subsidies often benefit the better-off at high fiscal cost. Subsidies have For those countries with independent monetary most often been provided to electricity utilities, policies, exchange rate movements can offset adverse including in Fiji, Kiribati, and Samoa. To the extent terms of trade shifts but result in winners and losers. that these subsidies are targeted for specific import Indeed, during the food and fuel price crisis of usages—in this case electricity generation—and 2007–09, the exchange rates of many Pacific island are provided on a clearly defined basis, they can be countries appreciated notably (see figure 4.10). effective in lowering the impact of fuel price shocks However, exchange rate policy also has negative on households. However, as with tax reductions, impacts, with exporters losing competitiveness. they can be difficult to withdraw and can become Exchange rate adjustment is also unlikely to be very costly relative to the limited fiscal resources of effective in addressing shocks that disproportionately most PICs. Subsidies are also often poorly targeted, impact subgroups within a country, such as in that they benefit households that are better-off. households experiencing hardship. Specifically, in many PICs, households in hardship are less likely to use electricity than better-off Tax reduction has commonly been used in response households. Similar arguments can be made for to commodity price increases, but this is often a one- subsidies that protect primary cash crop producers time use policy that may not be fiscally sustainable. from market prices. They can easily become a major Between 2007 and 2013, the Marshall Islands, Samoa, burden on the budget (see box 4.3 on the Kiribati the Solomon Islands, Tonga, and Vanuatu have all copra subsidy scheme), and government resources reduced import duties on certain food and fuel might be better used to fund other policies that are commodities as prices have spiked. But these tax less distortionary and less costly. changes have only partially offset price rises, because existing taxes on basic imported goods were low to Several PICs regulate the prices of some commodities, begin with and quickly reach a lower bound of zero. but regulation is costly to implement effectively and In contrast, the price increases for some goods have potentially distortionary. Countries such as Fiji and Figure 4.10 Real Effective Exchange Rate for Select PICs, 2004–2010 110 Real effective exchange rate 105 100 95 Fiji 90 Papua New Guinea Samoa 85 Solomon Islands Tonga 80 2004 2006 2007 2008 2009 2010 2005 Source: World Bank World Development Indicators. Current Approaches to Reducing Hardship and Vulnerability | 65 Box 4.3 Kiribati’s Coveted Coconuts Copra (dried coconut meat) is the main cash crop produced in Kiribati. The Kiribati Copra Subsidy Scheme has been in operation for more than 30 years and helps secure livelihoods and income for many I-Kiribati living on the outer islands. Under the scheme, the government maintains a price floor, which is a price at which it guarantees to purchase all copra produced. This guarantee means that the government bears all the risks from volatility in the world market price for copra. The government also absorbs the costs of processing and transporting the copra from outer islands to South Tarawa for processing and export. In most years, the purchase price (before shipping and processing) has been held above the international market price. Although this subsidy scheme provides protection to copra-producing households against negative price shocks, it also comes with considerable drawbacks. The fiscal costs of the scheme are very high, accounting for between 5 and 7.5 percent of recurrent government revenue over recent years. A recent review of the sector found that the scheme leads to substantial economic inefficiency. The state-owned mill that processes the copra operates inefficiently and pays no dividends to the government. In addition, the scheme is not well targeted to households in hardship, many of whom live in South Tarawa and do not have access to coconuts. Finally, weak systems have led to substantial leakage, which could be increasing costs by up to 40 percent. For every $1 in net costs to the government, only about $0.28 is estimated to go to households in hardship. Options for reform, as well as alternative programs, are under consideration. Institutional and policy reforms that reduce inefficiencies and leakage along the supply chain would help reduce the fiscal burden of the program. Reducing the guaranteed price, which was increased 33 percent between 2008 and 2011, would also provide substantial savings.1 But, more fundamentally, a change in perspective has been vital to progress. Although the scheme was initially designed to smooth agricultural production prices within a viable industry, it has now become a de facto social protection system for many of those living in outer islands. Explicit recognition that the scheme serves this purpose has opened opportunities to reconsider the organization of the scheme and the extent to which it represents the most efficient mechanism of using fiscal resources to achieve social protection outcomes. The geographical targeting of the scheme is being reconsidered, with the possibility of limiting copra purchasing to poorer islands and communities, where alternative economic opportunities are most scarce. Alternative social protection mechanisms, potentially including workfare or cash transfers, are being considered to replace the scheme in some areas, including in South Tarawa where copra cannot be grown and poverty and vulnerability is concentrated. 1. These increases in the guaranteed price were made during a time of rising world prices for copra. However, since 2011, world prices have declined substantially. Source: World Bank 2013. Kiribati regulate the prices that retailers charge to lead to deterioration of the quality or availability of consumers on a wide variety of commodities, and goods. In some countries, enforcement capacity is other countries, including Solomon Islands, Tonga, minimal, while in other countries low-quality versions and Tuvalu, have price controls on a small number of goods are sold at controlled prices while better of basic goods, mostly petroleum and staple food versions are not. imports. Around the world, price monitoring and regulation is common for certain goods, such as petroleum products, and can be an appropriate means of public intervention in markets where there Key Messages is significant monopoly power. The regulation of a Traditional systems do not eliminate hardship and can small number of commodities, such as imported provide only partial insurance. Although traditional food staples that households in hardship rely on, systems of resource sharing and self-subsistence are can be justified on these grounds, because they important to the well-being of many Pacific islanders, can prevent price spikes driven by firms’ pricing hardship and vulnerability are still major challenges. behavior, particularly in rural areas where there is Traditional systems do not reach everyone, and little competition. However, monitoring of prices is evidence from household surveys suggest that those complex and costly to effectively implement and can in deepest hardship may be the least likely to be part 66 | HARDSHIP AND VULNERABILITY IN THE PACIFIC ISLAND COUNTRIES of gift-giving networks. In addition, cultural and social responses to economic shocks in the Pacific have pressures seem to require greater generosity than proven to be effective in reducing the negative many households feel they can truly afford. At the impacts on households while also being fiscally same time, traditional systems cannot insure against sustainable. Some ex-ante measures that provide the many aggregate shocks that are common in the protection or insurance against shocks are being Pacific. Governments therefore have a role to play explored, but shocks will continue to be part of in complementing traditional systems with hardship the Pacific landscape. Therefore, some of the most reduction and risk management efforts. important actions governments can take are to pursue prudent macroeconomic policy, including building up Households have limited access to market instruments savings in good times and actively mobilizing revenue that can help them manage risks. In particular, access to have resources to deploy during shocks. to formal financial instruments is limited in most PICs. Evidence from household surveys shows that a minority of households holds savings accounts, loans, References or insurance policies. Without access to these risk Abbott, David, and Steve Pollard. 2004. Hardship and management tools, households are likely relying too Poverty in the Pacific. Manila: Asian Development much on coping. At the same time, growth in financial Bank. access without effective regulation and consumer education can lead households into excessive debt, Chong, Alberto, Marco Gonzalez-Navarro, Dean which is a concern in some countries including Fiji Karlan, and Martin Valdivia. 2013. “Effectiveness (Karan 2012). and Spillovers of Online Sex Education: Evidence from Randomized Evaluation in Colombian Public Governments provide little social insurance, but some Schools.” Working Paper 18776. National Bureau of programs show promise within country constraints. Economic Research, Cambridge, MA. Across the Pacific, Fiji is the only country with a Feeny, Simon, May Miller-Dawkins, Alberto Posso, hardship-targeted cash transfer program. However, Lachlan McDonald, Jaclyn Donahue, Kate Eccles, many other countries provide transfers or subsidies Matthew Clarke, Manoranjan Mohanty, and Vijay to small groups of people identified to be in need. Naidu. 2012. “Vulnerability and Resilience to Broader measures to support those experiencing Shocks in Melanesia.” Report prepared for AusAID, hardship face fiscal and capacity constraints, as well Canberra. as data limitations, particularly in the smaller islands. FSM (Federated States of Micronesia) Department of Two programs that show promise are elderly funds Health and Social Affairs. 2010. Federated States of and cash for work schemes, which have lower data Micronesia National Health Expenditure 2005–2008. requirements and costs that can be managed through Department of Health and Social Affairs, Palikir, straightforward participation requirements. Federated States of Micronesia. Fukuyama, Francis. 2008. “State Building in Solomon Government funding of basic services is under fiscal Islands.” Pacific Economic Bulletin 23 (3): 18–34. pressure from the rapidly rising costs of coping with Gibson, John. 2006. “Are There Holes in the Safety NCDs. Health care expenditures in the Pacific largely Net? Remittances and Inter-household Transfers in go to coping with health shocks: curative, palliative, Pacific Island Economies.” Working Paper 1. Pasifika and rehabilitation care absorbs 80 to 90 percent of Interactions Project, Wellington, New Zealand. national health expenditures. This focus on coping is fiscally unsustainable because of NCDs, which are Gibson, John, and David McKenzie. 2014. “The spreading quickly and are costly to treat. Greater Development Impact of a Best Practice Seasonal emphasis is needed on knowledge and protection Worker Policy.” Review of Economics and Statistics measures to slow their rise, but changing people’s 96 (2): 229–43. behavior is difficult. In addition, funding both ex-ante Gibson, John, David McKenzie, and Steven Stillman. knowledge and protection measures for the future, 2011. “The Impacts of International Migration on while dealing with the present costs of coping, is a Remaining Household Members: Omnibus Results major financial challenge. from a Migration Lottery Program.” Review of Economics and Statistics 93 (4): 1297–1318. Managing aggregate economic shocks through Gibson, John, Susan Olivia, and Scott Rozelle. 2006. coping actions has limited effectiveness, and more “How Widespread are Non-linear Crowding Out protective measures hold promise. Few ex-post Effects? The Response of Private Transfers to Current Approaches to Reducing Hardship and Vulnerability | 67 Income in Four Developing Countries.” Working Sviridova, Tatiana. 2013. “Review of Traditional Safety Paper in Economics. University of Waikato Nets in Pacific Island Countries.” Background paper Department of Economics, Hamilton, New Zealand. for this report. Halstead, Imogen. 2013. “Health, Hardship, and Tonga Ministry of Finance and National Planning. Vulnerability in the Pacific.” Background paper for 2012. Social Protection Issues Paper. Nuku‘alofa. this report. Udry, Christopher. 1994. “Risk and Insurance in a Huffman, Kirk. 2005. Traditional Money Banks in Rural Credit Market: An Empirical Investigation in Vanuatu: Project Survey Report. Port Vila, Vanuatu: Northern Nigeria.” Review of Economic Studies 61 Vanuatu National Cultural Council. (3): 495–526. Hyoung-sun, Jeong, and Ravi Rannan-Eliya. 2010. World Bank. 2010. Emergency Project Paper on a “SHA-Based Health Accounts in Twelve Asia/Pacific Proposed Grant to Solomon Islands for a Rapid Economies: A Comparative Analysis.” OECD/Korea Employment Project. Washington, DC: World Bank. Policy Centre Health and Social Policy Program— ———. 2011. Poverty Trends, Profiles and Small System of Health Accounts Technical Paper No. 10. Area Estimation (Poverty Maps) in Republic of Fiji Seoul, Korea. (2003–2009). Washington, DC: World Bank. IMF (International Monetary Fund). 2013. World ———. 2012a. The Economic Costs of Non- Economic Outlook April 2013. Washington, DC: IMF. Communicable Diseases in the Pacific Islands. Karan, Ram. 2012. Review of Consumer Credit Act and Washington, DC: World Bank. Regulations . . . From Consumers’ Perspective. Suva, ———. 2012b. Health Financing Options for Samoa. Fiji: Consumer Council of Fiji. Washington, DC: World Bank. McKay, Julienne. 2009. Poverty Housing in the ———. 2013. “Review of Government of Kiribati Developing Nations of the Pacific Islands. Bangkok: Copra Subsidy Scheme.” World Bank, Sydney. Habitat for Humanity. ———. n.d. World Development Indicators (database). Morauta, Louise. 1983. “Preliminary Field Report, World Bank, Washington, DC. http://data.worldbank Madang Phase of IASER Project on Inter-household .org/data-catalog/world-development-indicators. Transfers in PNG Towns.” Institute of Applied Social Yoong, Joanne, Lila Rabinovich, and Stephanie and Economic Research, Port Moresby, Papua New Diepeveen. 2012. “The Impact of Economic Guinea. Resource Transfers to Women versus Men: A Nanau, Gordon Leua. 2011. “The Wantok System as a Systematic Review.” London: EPPI-Centre, Social Socioeconomic and Political Network in Melanesia.” Science Research Unit, Institute of Education, OMNES 2 (1): 31–55. University of London. Ratuva, Steven. 2010. “Back to Basics: Towards Integrated Social Protection for Vulnerable Groups in Vanuatu.” Pacific Economic Bulletin 25 (3): 40–63. Implications for Policy: The Way Forward 69 Chapter 5 Implications for Policy: The Way Forward H ardship in meeting basic needs is a reality in Good government policy in all areas should factor many parts of the Pacific, and Pacific islanders in risks. For example, urbanization is changing risk are uniquely vulnerable to aggregate shocks. profiles and presenting new challenges, as well as Chapter 2 showed that more than 20 percent opportunities. Forward planning can help to address of households in most PICs are unable to meet their the challenges, and maximize the opportunities, by basic needs. At the same time, Pacific islanders are identifying and actively managing the risks associated very vulnerable to aggregate economic and natural with it. For example, strengthening the ties between shocks, and, increasingly, to the epidemic of NCDs, government bodies responsible for infrastructure as discussed in chapter 3. This is in addition to the provision and building codes with those responsible many idiosyncratic and localized shocks that occur for disaster risk management and climate change but are not easily captured in the existing data. adaptation is one important step in this area (see Households and communities actively manage risks box 5.1). In addition, given the importance of land in and seek opportunities to increase well-being, but the social, cultural, and economic life of people in the their efforts can only go so far, as detailed in chapter 4. Pacific, governments should aim to support good land Government measures have focused on ex-post risk management, including enabling communities to use management, at high cost, and are hampered by it as a resource to manage the risks they face. limited capacity, financing, and data. Development partner activities should also factor in This chapter presents the implications for risks and seek to reduce rather than inadvertently add governments and development partners that to volatility (see box 5.2). Aid flows comprise a large can be drawn from the results presented. These proportion of public expenditure in many PICs. With implications can be summarized under three broad donor spending often providing an important source headings: govern prudently, invest in data, and enable of employment and demand for goods and services households and communities. from private businesses, changes in donor spending can have important impacts on economic activity and household incomes. Donors could work to ensure Govern Prudently and that changes in expenditure are timed to counteract volatility arising from other sources. Flexible funding Proactively Manage arrangements that are responsive to changed Aggregate Shocks expenditure priorities in the light of major shocks can help. More generally, greater use of budget Governments and development partners have support would increase the capacity of governments important roles to play in managing risk through to align donor support with fiscal policy objectives sound policy, particularly in managing risks from of offsetting the impact of shocks—supporting aggregate shocks. Broadly speaking, government accumulation of reserves that could be utilized to policy should aim to avoid being a source of instability support higher expenditure, economic activity, and itself and should actively recognize and account for employment during downturns. As recent experience risks in all areas. For the aggregate economic and has shown, PIC governments with stretched natural risks described in the preceding chapters, implementation capacity benefit from a focus on the households and communities cannot fully manage on practicalities of implementation support following their own and need systematic support. natural disasters, and this could be strengthened. 70 | HARDSHIP AND VULNERABILITY IN THE PACIFIC ISLAND COUNTRIES Box 5.1 Managing Risks from Natural Disasters in the Pacific The East Asia and Pacific region as a whole is the most disaster-stricken region in the world, and PICs in particular are among the most vulnerable countries globally to natural hazards. They combine high exposure to frequent and damaging natural shocks with low capacity to manage the resulting risks. Their vulnerability is exacerbated by poorly planned development, which has increased exposure and economic losses, and by climate change, which has increased the magnitude of cyclones, droughts, and flooding. Two recent World Bank reports send a clear message about the importance of active risk management for natural shocks. If country governments and their development partners do not act now to reduce Pacific countries’ extremely high vulnerability, the consequences are likely to be serious. Simply put, a “business as usual” approach focused on immediate disaster relief rather than long-term disaster risk management and climate change adaptation will increase economic and human losses, slow economic growth, and delay or even reverse progress toward development goals. Progress in systematic risk management has been made in recent years, but much more remains to be done. Disaster risk management frameworks need to be developed, and where they exist, integrated into existing planning and regulation, including but not limited to land management, building regulations, and crisis management operational guidelines. Investments in disaster risk management and climate change adaptation have increased substantially at the national and local levels and have also begun to be integrated into social and economic planning. However, efforts have mostly been at the project level and fragmented, limiting their impact. Greater coordination and clear leadership are necessary going forward, as the risks from natural disasters continue to increase. Source: World Bank 2012c, 2013a. Box 5.2 The Role of Development Partners in Managing Aggregate Shocks Aid flows are an important part of most Pacific economies. Many PICs benefit from substantial aid flows (for example, in 3000 of GDP in the Solomon Islands, and 9 percent 2012 donor grants were equal to 50.5 percent of GDP in Kiribati, 19 percent Vatu (millions, real 2000) of GDP in Tonga). Aid, whether delivered through projects or—less often—budget support, plays an important role in 2000 financing social services and infrastructure. Further, aid-financed public spending provides a vital source of employment and demand for the private sector where domestic markets are small1000 and export opportunities limited by formidable geographical constraints. The amount of aid that Pacific countries receive can vary substantially between years, sometimes 0 exacerbating and sometimes mitigating economic volatility arising from external economic shocks. –1000 Figures B5.2.1, B5.2.2, and B5.2.3 show changes in grants and domestic revenue for Vanuatu, Samoa, and Tonga since 2001 and suggest the following points. –2000 Figure B5.2.1 Change in Grants and Revenues, Vanuatu Change in Grants Change in Revenue 3000 3000 Vatu (millions, real 2000) Vatu (millions, real 2000) 2000 2000 1000 1000 0 0 –1000 –1000 –2000 –2000 Change in Grants Change in Revenue   Change in Grants Change in Revenue 3000 Vatu (millions, real 2000) 2000 1000 0 –1000 –2000 20 Tala (millions, r 0 –20 –40 Implications for Policy: The Way Forward | 71 –60 Figure B5.2.2 Change in Grants and Revenues, Samoa Change in Grants Change in Revenues 60 60 real 2000) Tala (millions, real 2000) 40 40 30 real 2000) 20 20 20 (millions, 0 0 10 –20 –20 (millions, 0 –40 –40 Pa’angaTala –10 –60 –60 –20 Change in Grants Change in Revenues  Change in Grants Change in Revenues –30 Figure B5.2.3 60 Changes in Grants and Revenues, Tonga Change in Grants Change in Revenues Tala (millions, real 2000) 40 30 30 Pa’anga (millions, real 2000) Pa’anga (millions, real 2000) 20 20 20 0 10 10 –20 0 0 –40 –10 –10 –60 –20 –20 –30 Change in Grants Change in Revenues –30 Change in Grants Change in Revenues  Change in Grants Change in Revenues do not exhibit any consistent relationship with changes in government revenues. This is unsurprising because Grant flows 30 Pa’anga (millions, real 2000) project planning and implementation cycles are typically linked with longer-term development strategies, rather than 20 seeking to coordinate with short-term macrofiscal swings. The macroeconomic impacts of donor projects are seldom 10 explicitly considered. 0 Grant flows sometimes help offset external shocks and negative economic consequences of natural disasters. In Samoa, –10 aid flows increased by around 2.5 percent of GDP per year in 2009 and 2010, helping to counteract the negative growth and revenue–20consequences of the 2009 tsunami, while also financing vital recovery and reconstruction work. In Tonga, donors made a deliberate effort to mobilize budget support to help the government deal with substantial revenue –30 declines resulting from the global economic crisis. Following revenue declines of around 8 percent of GDP in 2010, donors Change mobilized additional grantin Change of Grants of 2 percent support in Revenues GDP by 2012. Unintended procyclicality can sometimes contribute to macroeconomic management challenges and broader economic volatility. In Vanuatu, for example, grants for road construction from 2008 tapered off following project completion during 2010–2012, while the economy was facing lower tourism arrivals and weak revenue performance due to the global economic crisis. The aggregate impact was the withdrawal of donor-financed fiscal stimulus coinciding with a negative external shock. Donor projects are provided for many reasons and offer a broad range of benefits to recipient countries. Donors face many planning constraints and often do not have the flexibility to optimize project spending based on broader economic factors. But because these projects are often large relative to Pacificeconomies, donor spending does have important macroeconomic impacts. There may be scope for donor agencies to work with governments to achieve greater coordination between fiscal policy objectives and the timing of donor spending, especially on major infrastructure projects. Increased reliance on budget support modalities, which allow greater flexibility in scale and timing of flows, may also be useful to improve consistency between aid flows and fiscal policy goals. 72 | HARDSHIP AND VULNERABILITY IN THE PACIFIC ISLAND COUNTRIES Table 5.1: Policy Measures to Manage Aggregate Economic Shocks Type of Aggregate Economic Shock Potential Policy Measures Stage of Risk Management All Fiscal buffers Insurance Fuel price Strong regulatory environment Protection Lifeline tariffs Protection Diversification of energy sources Protection Food or fuel price Price hedging Insurance Supply chain management Insurance Manage Aggregate Economic shocks. Some protection and insurance measures are Shocks costly, long-term undertakings that are challenging to implement. However, interest in measures that can Leaders should maintain macroeconomic and reduce the impacts of price shocks on households fiscal buffers at relatively high levels to counteract and also save costs is increasing. Some efforts are aggregate economic shocks (see table 5.1). Most already underway, but the potential exists for more Pacific countries have a very high dependence on widespread use of measures described below: imports of goods and services and on external strategic fuel reserves, lifeline tariffs, price hedging, financial flows. To ensure stability in the immediate joint purchasing, and diversification. aftermath of drops in external earnings, countries would ideally maintain a higher than usual level of foreign exchange reserves. In addition, governments PICs should review fuel supply chain efficiency and play a major role in assisting people in need, restoring ensure they have appropriately sized and maintained services, and repairing infrastructure after negative storage facilities, either on or off shore. Some shocks. To be able to do so effectively requires an countries have inadequately maintained excessively escalation of spending and a likely deterioration of costly storage facilities that contribute to heightened the fiscal position. Fiscal buffers are essential to be costs and risks of breakdown. Others lack adequate able to absorb these costs without an ensuing fiscal facilities to support fuel security. The Federated States crisis, which can lead to further instability. Low public of Micronesia is an example of a country that has sector debt, accumulation of assets in trust funds or reformed supply chains and passed cost savings on sovereign wealth funds where possible, appropriate to consumers. Investment in fuel storage facilitates revenue policies and strong compliance mechanisms, can support larger fuel deliveries, but in making such and predictable, long-term aid commitments can help investments the unit cost savings need to be carefully build these fiscal buffers. It is, of course, possible to weighed against the upfront capital costs. accumulate assets only if there is a surplus available to save. In some PICs where natural resource rents Establishing lifeline tariffs for utilities can help protect are high, this may be the case. But otherwise, the households from fuel price shocks and reduce costs revenues and grants that PICs receive are needed for vulnerable households. Utilities providers in some for immediate current and capital expenditure, and countries are already using lifeline tariffs, which are the opportunity cost of redirecting these funds to lower rates for the first block of usage of (typically) the accumulation of assets may be prohibitive. In electricity. This benefits small consumers, who are such circumstances, it might be more feasible for often less well-off households. Establishing lifeline development partners to provide additional resources tariffs not only reduces burdens on households in to help PICs build their assets, while governments hardship but also defines a policy lever that can focus on maintaining low public sector debt and good be used to protect these users from price shocks revenue performance. in the future. The costs of shielding households from increased prices could be met through direct Selectively invest in protection and insurance policies subsidization by government, potentially supported to reduce vulnerability to aggregate economic by development partners, or by cross-subsidization Implications for Policy: The Way Forward | 73 involving higher rates for commercial and less- capital costs associated with renewable generation, vulnerable users. the challenge for PICs is to identify projects in partnerships with private sector or development Price hedging of major commodity imports has the partners that can bring positive economic and social potential to reduce risks in the Pacific. Governments returns, including via reducing risk to households and can help to reduce the high degree of uncertainty businesses. around future prices of major commodity imports that are volatile by supporting the use of commodity Manage Aggregate Health Shocks price risk management. This type of risk management Increased knowledge and protection measures uses derivative products to hedge against the risk of are needed to improve health and slow the rise of an unexpected price shock. Appropriate strategies NCDs. As discussed in chapter 3, behaviors related should focus on the simplest approach to hedging the to NCDs such as smoking are widespread in the most volatile commodity prices, such as oil. Although Pacific. Research from around the world finds that not a World Bank member state, Guam has used behavior change is very challenging but is critical to hedging strategies for some time. In the aftermath slowing the rise of NCDs and their associated costs. of the fuel price crisis, Tonga’s electricity utility has PICs are now working on this challenge, through begun experimenting with using financial derivatives multifaceted strategies that aim to increase people’s to hedge against variations in oil prices. knowledge about behavioral choices and disease and that increase protection using multiple tools Effective market regulation, public procurement (World Bank 2012a). These tools include tax policy to policies, and market structures can also support discourage the consumption of unhealthy goods and supply chain efficiencies. A study of PIC procurement building facilities to encourage exercise. At the same practices (Pacific Island Forum Secretariat 2010) noted time, investments in maternal and child care, as well that significant savings might be made in procurement as protection against communicable diseases, are of major commodities, such as fuel. One option that needed in many countries where these are still major has been proposed for some time is joint regional threats. The need for evidence-based, cost-effective, purchasing, but this has gained little traction in and multisectoral approaches to addressing these the Pacific. Although maintaining sovereignty over challenges was recently affirmed by Pacific Health strategically important procurement, seeking to Ministers in the Apia Communique (2013). reduce costs via competitive and open tendering, as well as effective regional and national regulatory Development partners have an important role to oversight, may bring comparable cost savings. Savings play in providing financial support and aligning their could be used by governments, firms, and households policies. Given the fiscal constraints faced by most for other purposes, such as self-insurance against PICs, there will continue to be a tradeoff between future shocks. addressing immediate demand for curative care and investing in knowledge and protection for the The potential for PICs to diversify their economies is future. The difficulty of this tradeoff can be seen in limited, but energy resources may be an important the fact that, in some countries, funds budgeted for exception. The geographic features of most PICs such investments are sometimes diverted during limit the set of viable economic activities that budget execution to meet immediate demands for are available. To the extent that it makes sense, curative care. In addition, testing different approaches governments already do facilitate the development to behavior change and financing the widespread of varied income sources, as well as food resources. implementation of the successful ones is simply High up-front capital costs have limited the scope beyond the fiscal capacity of some Pacific countries. for diversification in energy generation. PICs possess However, development partners should have an considerable renewable energy potential, particularly interest in providing support, in part because they solar and hydro-power, and diversification away from tend to act as de facto insurers, stepping in with the predominantly oil-burning generators that provide funding and relief after shocks have occurred (which is most of their power would offer protection from often more costly than protecting against the shocks future oil price volatility (ADB 2012). Given the high ex ante). In addition, tax and trade policies aimed at 74 | HARDSHIP AND VULNERABILITY IN THE PACIFIC ISLAND COUNTRIES Box 5.3 Turkey Tail Travails in Samoa The threat of NCDs in Samoa is as acute as it is anywhere else in the Pacific. More than 50 percent of adults are obese, and about a quarter have diabetes (World Bank 2012b). Unhealthy diets that include cheap imported meat offcuts is a major contributor to this situation. In an effort to improve the diets of Samoans, the government banned the importation of turkey tails, a popular and fatty offcut, in 2007. However, as part of Samoa’s accession to the World Trade Organization in 2012, the ban was lifted. A temporary 300 percent import duty was placed on turkey tails, giving the Samoan government time to work on changing the eating behavior of its people before the tails become cheap and widely available again. Source: World Trade Organization 2011. protecting against NCDs sometimes face international Governments and their development partners should challenges that need to be addressed (see box 5.3). invest in regularly occurring household surveys for all countries. Some countries have conducted two or more household surveys in the past decade, which is a promising start to regular data collection. The Ten Invest in Data Year Pacific Statistics Strategy (TYPSS) took effect in Timely, good quality data and analysis on the well- 2011 and takes positive steps toward strengthening being of households are valuable for policy making. regional technical support for national statistical They inform the formulation of government policy, offices. However, funding for survey implementation helping to identify communities in need of services or is still lacking in many countries. Additional other investments. Similarly, data can help guide the consideration of surveying methods and the content investments of development partners, whose projects of surveys could help increase data quality and lower are often targeted to specific communities. Data can costs, which would increase the financial feasibility of also be used to assess the effectiveness of government regular surveys. For example, the use of local market policies and programs. Given the resources required price surveys, shorter diary periods, and technology- to collect data, countries should prioritize key data assisted survey methods could all be considered. For sources, ensure their quality and regularity, and keep the countries with more sporadic surveys that rely on systems simple and fit-to-purpose. Box 5.4 describes external financing, funding regular surveys, including an example from Papua New Guinea of how a lack of the technical capacity to implement and analyze data can hamper decision making. them, should be a priority for development partners. Implementation and analytical technical capacity can In many Pacific countries, data are collected be built at the national level for larger PICs and at the sporadically and are not always of high quality. Several regional level to provide support to the smaller PICs. countries have conducted nationally representative household surveys or censuses in the last five years. Conducting surveys is not enough: Public access to However, some have not because they lack the data and analysis is critical to getting the information funding and technical expertise required: Although used by policy makers and partners. When data are these countries are small, their geographically made accessible, researchers and others use the data dispersed populations make the costs of surveying to provide policy-relevant insights and also often help relatively high. Even among the countries that have improve the data quality by pointing out problems conducted surveys, quality can be a challenge, that might not have been recognized. Several PICs because national statistical offices lack the capacity and their regional partners have restricted access to to carefully implement and analyze complex and the collected data and to the analysis conducted with detailed surveys. Out-of-date or doubtful quality the data, including poverty measurement. In general, data lose much of their value, and even if the only summary statistics are available to the public, and provided information is accurate, policy makers and no processes are established for obtaining access. development partners are unlikely to use it in their Consequently, little use is made of Pacific data outside decision making. of the few groups that have access to it, and Pacific Implications for Policy: The Way Forward | 75 Box 5.4 What Happened to Hardship in Papua New Guinea?—Estimating When There Are No Data In Papua New Guinea, the first nationally representative household survey was conducted in 1996, and 15 years elapsed until the second (conducted 2009–10). During that entire period, limited data were available on the well-being of the population, hampering policy makers’ ability to address hardship and other challenges. Now, even with both surveys available, knowledge of changes over time cannot be recovered, because comparing the 2010 and 1996 snapshots of well- being is likely to mask significant variation over this 15-year period. Macroeconomic conditions deteriorated significantly in Papua New Guinea between the early 1990s and the early 2000s, and these trends are likely to have brought worsening living standards and rising hardship rates until around 2002. The country then entered a decade of very strong growth, and hardship likely declined in subsequent years. Without reliable data, it cannot be known with certainty how living standards and well-being changed over this period, and in particular how hardship changed in response to changes in economic growth. However, estimates can be made. Using macroeconomic indicators of GDP growth and inflation, the World Bank estimated that the hardship rate rose from 40 percent in 1996 to just above 50 percent in the early 2000s, before returning to near 40 percent by the time of the 2010 survey (see figure B5.4.1). Consistent with this pattern, formal sector employment declined by 2.8 percent between 1996 and 2002, and cash crop export receipts rose by 88 percent during the same period. Between 2002 and 2010, formal sector employment rose by 46 percent (and a further 14 percent between 2010 and 2012), and nominal cash crop export receipts rose by 173 percent during the same period. Figure B5.4.1 Hardship and Growth in Papua New Guinea Index 350 70% 300 60% Poverty head count 250 50% (basic need; right-hand side) 200 40% GDP per capita (left-hand side) 150 30% Agriculture Exports 100 20% (kina value; left-hand side) 50 10% Employment (left-hand side) 0 0% 1995 1997 1999 2001 2003 2005 2007 2009 2011 Source: World Bank staff estimates. countries are often excluded from external databases its implications about hardship and vulnerability to and analyses. Increased data accessibility is one of the shocks. objectives of the TYPSS, and this accessibility should go beyond summary statistics to ensure researchers within and outside the region have the necessary level Enable Households of detailed microdata to conduct rigorous analysis. and Communities In addition, active communication about the meaning Governments and development partners have and implications of poverty measurement is needed important roles to play in supporting the efforts of to get hardship on the policy agenda. As detailed in households and communities to reduce hardship and chapter 2, poverty measurement has been carried out vulnerability. In particular, selective investment in social using household survey data in many Pacific countries. programs is needed to provide a safety net and relief However, government and the general public are from hardship. Governments in the Pacific are also well often unfamiliar with the analysis and its implications, placed to effectively build human capital by improving so they receive limited consideration in policy making. the education that is already often publically provided Active communication led by national statistics and by creating productive economic opportunities. In offices and supported by development partners is addition, governments have an important role to play needed to help clarify what poverty measurement in increasing people’s access to a broader set of risk actually measures, and to increase awareness of management tools. 76 | HARDSHIP AND VULNERABILITY IN THE PACIFIC ISLAND COUNTRIES A strong case is to be made for expanding the role around the world are finding ways to successfully of government and development partners in social overcome these challenges with help from donors protection. The preceding chapters provide important and international expertise. Important factors to evidence that households and communities are not making social programs work in capacity-constrained able to fully manage the risks they face or to eliminate contexts include sustained, multiyear commitments hardship. In particular, traditional networks do not from donors, limited and straightforward targeting reach many households experiencing the deepest to reach those most in need, and careful, phased hardship and appear to provide only partial insurance implementation (Grosh et al. 2008). to households suffering shocks. In addition, traditional networks cannot manage local or aggregate shocks The quality and accessibility, as well as the portability, that affect most of their members. These findings of basic education should be increased to enable make it clear that Pacific governments, and their people to increase their well-being through development partners, need to consider an expanded increased productivity at home or abroad. High- role in social protection that takes into account quality education equips people to do their work traditional networks as well as fiscal constraints. The more productively, whatever their work may be, preceding analysis makes clear that when carefully and therefore earn more. Given the challenges to designed and managed, the costs of transfer private sector growth in the Pacific, education that programs to protect vulnerable populations from the is “portable” is likely to be the most valuable. For impact of shocks are manageable. With appropriate example, governments can work with development systems in place, adequate protection for those most partners (who often receive migrants as well) to vulnerable to external shocks could be provided adopt education and training qualifications that will through a combination of budgetary resources and be recognized in major migrant-receiving countries aid from development partners. (World Bank 2013b). In addition to making people more productive, education can also increase Investments in social protection programs should knowledge about risks and how to best manage them. be made well in advance of major shocks. Like most policy initiatives, establishing and operating social Leaders should create more opportunities for migration protection programs take time and effort. Design by removing regulatory barriers to increase people’s aspects of these programs—such as targeting and well-being. Currently, overseas migration is restricted delivery mechanisms—take time to fine tune. It also for the majority of people who are not from islands with takes time for administrators to build up experience in special citizenship relationships with large neighbors how to operate social protection programs effectively. (World Bank 2013b). If equipped with the skills needed If social protection programs can be established for success in destination country labor markets, the before a major shock occurs, they have the potential to relatively young populations of many islands could be scaled up more quickly and effectively in the wake greatly benefit from increased migration opportunities. of a shock than if preparations for their establishment begin only after a major shock has struck. Maintaining More opportunities should be fostered for productive these programs with a fairly limited scope on a work at home. Jobs and increased income not only continuous basis enables continuous learning and raise living standards in good times, but also enable the buildup of knowledge and experience, giving the households to better manage the risks they face. programs much greater chances of being successful in Given the unique challenges in the Pacific, realistic mitigating major shocks when they occur. Data, such expectations about the potential for private sector- as those from periodic HIESs, can be used to inform led job growth are needed. Policy priorities for job the operation of these programs on a regular basis. A creation in the Pacific have recently been laid out in strong evidence base built up over time can then be At Work in East Asia Pacific (World Bank 2013b) and readily supplemented by rapid assessment techniques include investment in connective infrastructure, rural in the wake of major shocks. services, agricultural productivity, and public sector productivity. A key message from this work is that Global evidence suggests that the challenges to public sector employment is likely to remain a key social programs in the Pacific can be overcome. As source of good-quality jobs in most PICs, and reform described in chapter 4, lack of data and fiscal space, efforts should focus on ensuring the productivity of as well as concern for disrupting traditional networks, public sector employment, rather than focusing on are major barriers to the development of social reducing the size of the public sector. The report also protection programs. However, several countries recommends a greater focus on job creation through Implications for Policy: The Way Forward | 77 donor support, by means of increasing opportunities Grosh, Margaret, Carlo del Ninno, Emil Tesliuc, for local procurement and hiring of local staff across and Azedine Ouerghi. 2008. For Protection and infrastructure or social service projects. Promotion: The Design and Implementation of Effective Safety Nets. Washington, DC: World Bank. Responsible development of market-based risk McCaffrey, Mike. 2010. “Formulating a Financial management tools should be supported. In many Inclusion Plan for Solomon Islands.” Pacific Financial PICs, few people have access to financial services: Inclusion Program Focus Note No 3. Suava, Fiji. Estimates are that about 45 to 50 percent of Fijians ———. 2011. “Five Catalysts for Financial Inclusion in and 85 to 90 percent of Papua New Guineans do Vanuatu.” Pacific Financial Inclusion Program Focus not have access to formal financial services (Pacific Note No 5. Suava, Fiji. Financial Inclusion Program 2010). Although a formal Pacific Financial Inclusion Program. 2010. “Financial bank account for each household may not be a Inclusion in Practice and the Potential in the realistic goal, increasing access to savings, credit, Pacific.” Paper presented at the Pacific Economic and insurance mechanisms would increase people’s Governance Agencies Seminar Series, Suva, Fiji. ability to effectively manage the risks they face. Alongside increasing access, appropriate regulation Pacific Island Forum Secretariat. 2010. Study on and knowledge dissemination are needed to protect Procurement Policies and Practices in FICs and people from excessive use of credit or other poor Recommendations on the Inclusion of Government choices. Procurement in the Pacific Island Trade Agreement. Suva: Pacific Island Forum Secretariat. The challenges to financial inclusion in the Pacific World Bank. 2012a. The Economic Costs of Non- are beginning to be surmounted. Geographically Communicable Diseases in the Pacific Islands. dispersed populations, limited infrastructure, Washington, DC: World Bank. and restrictive regulations are some of the major ———. 2012b. Health Financing Options for Samoa. challenges to expanding financial services, but Washington, DC: World Bank. mobile technology, regulatory changes, and active ———. 2012c. Acting Today for Tomorrow: A Policy donor support are enabling progress in many PICs and Practice Note for Climate and Disaster (McCaffrey 2010, 2011). The Pacific Microfinance Resilient Development in the Pacific Islands Region. Initiative, launched in 2010, is one effort aimed at Washington, DC: World Bank. increasing access by providing technical assistance ———. 2013a. Strong, Safe, and Resilient: A Strategic and funding to banks and other financial institutions Policy Guide for Disaster Risk Management in East to help them provide sustainable services to Asia and the Pacific. Washington, DC: World Bank. underserved communities in Papua New Guinea, Timor-Leste, and Tonga. ———. 2013b. At Work in East Asia Pacific. Washington, DC: World Bank. World Trade Organization. 2011. Report of the References Working Party on the Accession of Samoa to the World Trade Organization. Geneva: World Trade ADB (Asian Development Bank). 2012. Pacific Energy Organization. Update. Manila, Philippines: Asian Development Bank. Apia Communique on Healthy Islands, NCDs, and the Post-2015 Development Agenda. 2013. Tenth Pacific Health Ministers Meeting. Photo Credits: