83118 DISCUSSION PAPER NO. 1307 Eligibility Thresholds for Minimum Living Guarantee Programs International Practices and Implications for China Nithin Umapathi, Dewen Wang and Philip O’Keefe November 2013 Eligibility Thresholds for Minimum Living Guarantee Programs International Practices and Implications for China Nithin Umapathi, Dewen Wang and Philip O’Keefe* Social Protection and Labor Unit Human Development Sector East Asia and Pacific Region November 2013 JEL Classification: I38, O15 Keywords: Minimum incomes guarantee programs, cash transfers, China ________________________ *numapathi@worldbank.org, dwang2@worldbank.org, pokeefe@worldbank.org Abstract Using a simple framework, this paper discusses the underlying reason of the variation of threshold level in developed countries, from the least generous 20% to around 60% of median wage, with an average of 35 %. The generosity of minimum guarantee social assistance programs is deeply rooted in social values and principles that further underpin the policy objectives. Many OECD countries set their policy targets for minimum living standard programs beyond basic needs and aim to guarantee a minimum socially acceptable level for a decent living. Thresholds are also refined to reflect the differences in family size and demographic structure, difference in regional cost of living and changes in prices and local wages. In some countries the thresholds show some regional variation due to local discretionary powers of sub-national authorities to set the threshold depending on the co-financing mechanisms. These lessons are valuable for China as the Chinese government has made efforts to standardize the implementation and management of its own minimum income guarantee (Di Bao) programs. The policy recommendations for China include accelerating the convergence of localized approaches, raising the administrative level for setting thresholds to higher level, defining the roles of central and local governments in financing and management, and establishing a transparent budgetary management system to transfer and allocate social assistance funds. ii Table of Contents I. Introduction ....................................................................................................................... 1 II. Minimum Living Guarantee: from principles to implementation ..................................... 3 A. Variation of Minimum Income Protection in OECD countries....................................... 3 B. What are the key principles guiding MLG thresholds in developed countries? ............ 6 Box 1 Case studies: low generosity threshold in the US vs. decent living standard in Sweden ................................................................................................................................ 13 C. Typology of basic features among MLG programs in developed countries ................ 14 III. Refinement of thresholds: International experience ...................................................... 15 A. Definition of an eligibility unit and household specific threshold ............................... 16 Box 2 Case study: anti-poverty program Bolsa Familia in Brazil ......................................... 19 B. Indexation and trends of minimum income thresholds across OECD ......................... 20 C. Roles of Central-local responsibilities in delivery and financing of social assistance . 24 IV. Policy implications for China’s Di Bao programs ............................................................. 25 A. Approaches for setting up Di Bao thresholds .............................................................. 26 B. Di Bao thresholds and benefits .................................................................................... 29 C. Refinements and Financing.......................................................................................... 32 V. Conclusion ....................................................................................................................... 33 References .............................................................................................................................. 35 Figures Figure 2.1: Minimum Living Guarantee thresholds for a single parent with one child family as % of average wage in 2009 ........................................................................................... 4 Figure 2.2: Cash benefit (excluding housing benefit) for a single adult family as % of average wage in 2009, OECD countries .......................................................................................... 6 Figure 2.3: Schematic summary from national aspirations to implementation in developed countries ............................................................................................................................ 9 Figure 2.4: Minimum wage relative to median wage of full-time workers in 2009 ............... 11 Figure 3.1: Trends in benefits and CPI .................................................................................... 23 Figure 4.1: Declining Engel’s Coefficients in Rural and Urban China, 1990-2011 ................. 27 Figure 4.2: Comparison between Rural Di Bao Threshold and Rural Poverty Line, 2006-2011 ......................................................................................................................................... 31 Tables Table 2.1: Type of minimum income eligibility threshold ...................................................... 10 iii Table 2.2: Pros and cons of MLS and poverty approaches to setting up the value of eligibility thresholds ........................................................................................................................ 12 Table 2.3: Types of MIG programs .......................................................................................... 15 Table 2.4: To what extent are MIG payments time-limited? ................................................. 15 Table 3.1: Type of household size adjustment by countries .................................................. 16 Table 3.2: Example of variations in threshold due to household size based on OECD equivalence scale ............................................................................................................. 17 Table 3.3: Eligibility thresholds and benefits in Bolsa Familia Program ................................. 19 Table 3.4: Is there a clear mechanism for periodic review of MIG eligibility thresholds? ..... 20 iv LIST OF ACRONYMS AE Adult Equivalent BFP Bolsa Familia Program CPI Consumer Price Index EU European Union MLG Minimum Living Guarantee MLS Minimum Living Standard MOCA Ministry of Civil Affairs NBS National Bureau of Statistics OECD Organization for Economic Cooperation and Development SAMIP Social Assistance and Minimum Income Protection Data-set UK United Kingdom UN United Nations US United States v I. Introduction This note draws lessons from international experience in setting eligibility thresholds for Minimum Living Guarantee (MLG) programs. International experience provides valuable insights for China’s rapidly evolving social assistance policy. These lessons can be helpful to clarify and address a number of key questions about the MLG thresholds in relation to the social policy objectives, specific approaches, threshold generosity and refinement issues. The eligibility threshold is a key factor in the design of the Minimum Living Guarantee (MLG) welfare programs to provide income support for the poor. Most OECD countries have a minimum living guarantee program intended as a short-term and means-tested assistance. The minimum living standards policies aim to fill the gap between actual and minimum income for the low-income households through combinations of cash and in-kind transfers. Such MLG programs are more prevalent in higher income countries and are rare among low and middle income countries. Compared with developing countries, OECD countries have the features of urbanized populations, high proportion of wage based income and strong administrative capacity for social security programs. And within OECD countries, there are large variations of both the level and approaches of the eligibility thresholds due to different socioeconomic contexts, political and institutional arrangements. In a broad framework, the eligibility threshold is determined by the objectives of social policies, while the latter is motivated by broad principles about citizen entitlements and collective responsibility (i.e. social contract). Consequently, specific approaches are adopted to set up the level, refinement and adjustment mechanism to achieve the policy objectives. The OECD practice presents examples to illustrate how broad principles have guided different policy objectives and approaches and consequently treatment of their MLG thresholds. No doubt, any attempt to set eligibility thresholds is closely related to how each country defines the concepts of minimum living or poverty. In some countries we found 1 that broad principles are translated as guarantees for basic living in their policy, while in others the policy objectives relate to more generous levels required for “decent living”. We note that policy objectives may also be guided by other considerations like fiscal and administrative constraints. Ultimately the level of the threshold is an outcome of an iterative process of designing a program that will fit within policy objectives, budgetary, administrative and political constraints, while maximizing the impact on the beneficiaries. The threshold value could vary significantly due to the differences in policy objectives, differences in approaches and refinements. While the country’s policy objective guide a specific approach, within a specific approach the thresholds can be very simple or highly differentiated depending on regional differences, categories of eligibility, household income or other characteristics. The note will review what refinements are used internationally to adjust for different types of households and changes in the cost-of-living across time and space. Such highly customized eligibility thresholds require stronger program administration capacity and communication effort with the program applicants. For instance thresholds may be differentiated by household characteristics such as size, and demographic composition. Safety net programs in middle and higher income countries can usually take advantage of systems developed for other purposes, like national identification and income tax registry data. Thus, complex and highly differentiated eligibility thresholds are more common in the upper-middle income countries than in poorer countries. The structure of this paper is as follows: the next section briefly compares the threshold levels among OECD countries, develops a simple framework linking broad socio-political principles, policy objectives and different approaches of MLG programs from various OECD member countries, and summarizes their main features including two examples from low generosity standard in the US to decent living minimum in Sweden. Sections three discusses key refinements of thresholds and how they are set in practice depending on family size and location, the thinking behind spatial adjustment and how countries account for changes in cost-of-living over time using consumer price index adjustment for inflation. This section also provides an overview of the role of central-local responsibilities and cofunding that can 2 lead to variations in size of the threshold. Finally fourth section draws implications to China’s Di Bao programs, particularly to its approaches, threshold and benefit level, refinements and financing. II. Minimum Living Guarantee: from principles to implementation A. Variation of Minimum Income Protection in OECD countries Almost all OECD1 countries provide minimum living guarantee programs for families unable to earn sufficient income. However, cross-national comparisons of social assistance levels are not straightforward. Different countries have different levels of development, income, currency and what is considered as poverty or as acceptable living standards. In order to present the variation of the levels (or generosity) of thresholds on an internationally comparable basis we calculated the threshold size as a proportion of national average wage. Since these thresholds in practice may vary by size of household, we fixed a reference household to be a single parent household, in later sections will also focus on how thresholds change with family size. There is considerable variation between countries from the least generous income eligibility threshold of 20% of median wages in the United States to around 60% in Japan and Italy2. Figure 2.1 shows a comparison of the threshold across OECD countries as a proportion of median wages. The median OECD MLG threshold for a single parent with a child stands at 35% of median wage. Countries that fall in the middle of this range are United Kingdom and Germany. It is important to note that housing, tax credits and child assistance are significant components of total benefits that contribute towards closing the gap between the minimum living guarantee threshold and actual income. 1 Organization of Economic Cooperation and Development includes Australia, Austria, Belgium, Canada, Chile, Czech Republic, Denmark, Estonia, Finland, France, Germany, Greece, Hungary, Iceland, Ireland, Israel, Italy, Japan, Korea, Luxemburg, Mexico, Netherlands, New Zealand, Norway, Poland, Portugal, Slovak Republic, Spain, Sweden, Switzerland, United Kingdom, and United States. 2 Comparisons are based using thresholds as a proportion of average (median) wages and do not indicate that countries necessarily set their thresholds relative to the median wage. 3 Figure 2.1: Minimum Living Guarantee thresholds for a single parent with one child family as % of average wage in 2009 70% 60% 50% 40% 30% 20% 10% 0% Source: Author’s calculations using social assistance and minimum income protection data-set (SAMip, Nelson, 2009) and information from OECDstat, 2009. Family assistance threshold is for a single parent. Average wages per full-time and full- year equivalent employee in the total economy. The median OECD cash-only-transfer (excluding housing and child benefit) for a single person family without children covers about 15% of median wage. 4 Figure 2.2 shows a comparison of cash-benefits component provided by MLG programs across OECD countries as a proportion of worker’s median wages. The maximum benefit amounts for a single person varies from least generous 5% average wage in the United States to 25% in Netherlands. This excludes housing, tax credits and child assistance transfers. 5 Figure 2.2: Cash benefit (excluding housing benefit) for a single adult family as % of average wage in 2009, OECD countries 0.30 0.25 0.20 0.15 0.10 0.05 0.00 Source: Author’s calculations using Social assistance and minimum income protection data-set (SAMip), OECDstat 2009. The benefit is made of individual amount and excludes in-kind benefits. Social assistance rates for a single adult person below retirement age and without children. Average wages per full-time and full-year equivalent employee in the total economy. For further details on the wage estimates, please see http://www.oecd.org/dataoecd/32/50/43948033.pdf. B. What are the key principles guiding MLG thresholds in developed countries? National aspirations regarding nature of social welfare contract are guided by general principles of citizen’s entitlements, differing views about role of the state, private sector and personal responsibility. This section presents a conceptual framework to review the institutional differences that affect threshold setting approaches across the countries. For example the right to sufficient minimum resources to live with dignity and independently is recognized in various EU institutional texts such as the European Social Charter (1996) and the Charter of Fundamental Rights of the European Union. These socio-political principles provide general guidance for governments to design their social policy objectives aligned with the national expectations of guaranteed minimum living standards. Some countries (e.g. USA) define their policy goals to achieve basic living as the minimum while others (Sweden or Norway) define more generous eligibility thresholds based on what is socially regarded deemed as “decent living”. But even within specific policy goals that may aim to 6 guarantee basic or decent minimum living conditions the exact implementation varies across countries. There are few examples of minimum living guarantee programs in low-income countries since minimum living guarantee programs are administratively challenging when a large proportion of population is agrarian and in non-salaried employment. Therefore this note is focused on learning from experience of developed countries. The design of last-resort programs is simpler in low-income and low-capacity settings involving a flat payment conditional on being poor and it does not entail a “top-up” feature to fill the gap between actual income and guaranteed income. Therefore the primary focus and majority of examples in this note are inspired by experience of OECD countries3. Although in developed countries social assistance constitutes a smaller component of overall social security, it has grown significantly since World War II. European countries see minimum living guarantee programs as short-term assistance of last resort, many of them in order to assure “up to a fixed living standard” on earnings or property. They are all mean -tested, non-contributory, and funded through the tax system. In general, social assistance programs have three types of policy targets: subsistence needs, basic needs, and decent living. The social assistance threshold level tends to get higher when moving from policy objectives of meeting subsistence needs to satisfying decent living conditions. Lower income countries tend to focus on subsistence needs in their social assistance programs, while developed countries choose decent living standard as their policy objective. The specific approaches taken by individual countries may also differ between different policy objectives and even when they have the same broad policy goal. 3 Notably the UN General Assembly adopted a resolution in 2010 that commits to promotion of universal access to social services and a minimum level of income for all. This “social protection floor” is bas ed on the principle that society as a whole accepts the responsibility to provide basic levels of benefits and services to everyone. This system is aimed as a social contract between the state and citizens where each citizen is entitled to essential services and minimum income security (UN, 2010). 7 Figure 2.3 summarizes the different approaches taken by countries. It also indicates that within both basic needs and decent living type of approaches they can focus on 1) absolute needs or 2) relative needs. The methodologies used for setting thresholds have close linkages with poverty methodologies that may be defined in an absolute or a relative manner such as in the US or UK. If the methodology follows an absolute poverty approach such as the cost-of-basic- needs4, the threshold is estimated in two steps: first by estimating the cost of food needs for adequate nutrition, for example at 2,100 Calories per capita per day. The cost of the total needs (including basic food and non-food requirement) is estimated in the second step, and is based on the Engel method where the food poverty line is divided by the share of food to obtain the aggregate poverty line (i.e. food and non-food). The share of food in the total basket can vary depending on which sub-population (richer, middle or poorer households) it is calculated for. Therefore such a threshold will vary depending on how a country defines basic needs. 4 Used to estimate poverty thresholds in Indonesia, Mongolia, Timor-Leste, Kosovo, China, Montenegro, Paraguay, Bolivia, Chile, Colombia, Iraq, Jordan, Egypt, Sri Lanka, Afghanistan, Bangladesh, Kenya, Liberia, Nigeria, Tanzania, Uganda, and other countries. 8 Figure 2.3: Schematic summary from national aspirations to implementation in developed countries Minimum food, cost- Absolute Subsistence Needs of-basic-needs, food requirement share methods Absolute or Relative All of the Above + Broad Principles Basic Needs requirement Percentile methods Relative Minimum Living Decent Living requirements Standards Thresholds may also be defined in relative way. Relative thresholds are defined in relation to the overall distribution of income or consumption in country, in other words relative to some agreed average. The threshold will be higher if the average of a selected indicator is higher. For example a threshold could be set at a given cut-off, such as the bottom 30% of the population in terms of either total expenditure or some other chosen welfare indicators such as income, average wages or other. As a general rule this approach is followed by richer countries. This explains why rich countries have nominally higher poverty thresholds than do poor countries. Many of those counted as relatively poor in the United Kingdom (that sets its threshold at 60% of average income) would be considered comfortably well-off by absolute basic needs standards in a different country like India. In both cases whether the threshold is absolute or relative requires household living condition survey information, especially information about income or food and non-food expenditure. Many OECD countries set their policy targets for guaranteeing income support beyond basic needs. They aim to guarantee a level of what is considered as minimally socially acceptable norms for decent living. In many European countries, these budgets are quite generous relative to the “absolute” poverty needs threshold as they are not meant to just 9 reflect “minimal consumption” but rather what is acceptable to society. In this case many countries with rely on an alternative approach -- Minimum Living Standards (MLS). A minimum living standard approach is different from a cost-of-basic-needs threshold. It is a political and consequently a legal instrument for setting income thresholds for social assistance policy. MLS seeks to provide an answer to the question ‘What level of income is needed to allow a minimum acceptable standard of living?’ The answer to this question can be used in making policy for income maintenance. Such a threshold usually relies on normative judgments about what constitutes sufficient resources to ‘live with dignity’ (e.g. Sweden, France, and Germany). The rationale for this alternative is that basic living thresholds are less relevant in a European setting given the stated social goal not just to reach basic standards of living (as in less developed parts of the world), but to make whole population to share the benefits of high average prosperity. Table 2.1 lists the countries by type of MLG threshold approach used. Table 2.1: Type of minimum income eligibility threshold Decent living (MLS threshold) Basic/Subsistence living (cost of basic needs or poverty Average thresholds) Income Minimum Wage Other MLS methods Spain, Portugal, Bulgaria, Czech republic, Slovak Rep., Austria, Spain, Finland, Romania, Slovenia, Estonia, Hungary, Portugal, USA, Cyprus, Lithuania, Sweden, Switzerland, Hungary, Portugal, Poland. Spain, Netherlands, Germany, Australia, Low income countries with France, Luxembourg, Norway, Denmark, non-MLG cash transfer Bulgaria Slovak Republic Belgium, United Kingdom programs. Source: Casas (2005), Frazer and Marlier (2009), Bavier (2009) An MLS threshold is generally determined relative to a socially acceptable standard of living of the country: two common examples of MLS are minimum wage and decent living budget standard. Examples of MLS are statutory minimum wage provisions or decent living household budgets as defined by government decree. The methods for establishing MLS vary from budget allotments using expert judgment about what is necessary and 10 acceptable, or constructed based on assessment of what the general public considers as acceptable. Studies of family budgets using surveys have also been used to set the MLS. But MLS is not limited to such a methodology and is open to any methodology that creates a political criterion for adequacy of income for a given period of time. For example, the 1992 recommendation of the EU Commission is that thresholds should be set according to appropriate indicators such as average disposable income, legal minimum wage and household consumption. Some countries (e.g. Germany, Poland, Sweden, Netherlands and France) set social assistance eligibility thresholds as a proportion of the minimum wage. The rationale for using the legal minimum wage as a MIS is that it is considered not only as a minimal benchmark for a decent level of work payment but also indirectly represents society’s accepted minimum level of living standards. This very much highlights the relativity of minimum living standards among European countries. Since in several European member states the MLS thresholds are linked to minimum wages it is useful to assess the size of the minimum wage as a share of average wage. Figure 2.4 presents the distribution of minimum wages as a percent of median wage among European countries (and those rely on minimum wage for MLS – Luxembourg, Spain, Netherlands, and France). The size of the threshold varies between 43-60% in these countries. Figure 2.4: Minimum wage relative to median wage of full-time workers in 2009 0.8 0.7 0.601 0.6 0.43 0.454 0.472 0.5 0.441 0.4 0.3 0.2 0.1 Source: OECDstat (2009) 11 There are other standards that represent a “decent living standard” specific to each country set individually and arbitrarily based on local perception of what is considered as “decent living”. For instance Sweden’s National Board of Consumer Affairs maintains a MLS (budget) representing a “reasonable” standard of living that is updated and approved annually by the parliament. This is used by the National Board of Health and Welfare to advise local authorities on setting social assistance levels; local authorities can set it higher if their budget allows but not lower. In Switzerland, thresholds are set for each canton as recommended by the Federal Social Welfare Agency which uses MLS where the threshold includes amounts for “maintenance, pocket money, rent, clothes, electricity, radio, television and telephone fees, and transport”. There are both advantages and disadvantages of these two types of approaches, which are summarized in Table 2.2. Review pf international experience suggests that eligibility thresholds based on minimum income standard are higher than the basic needs or national poverty thresholds. This underlines the fundamental divergence between how societies and nations regard the minimum that is indecent for people to be without, and poverty thresholds as defined by National Statistical Offices to count the number of poor people. The difference is illustrated in Box 1 below, which compares the approaches taken to setting benefit thresholds in the United States and Sweden. Table 2.2: Pros and cons of MLS and poverty approaches to setting up the value of eligibility thresholds Poverty line approach Minimum income standard approach More focused on the bottom poor and more This type of decent living approach is attractive in affordable when properly targeted to low- terms of generosity of social and financial income groups. protection but may be fiscally demanding. Less likely to suffer from work disincentive The higher generosity of this approach may create effects poverty trap situations by encouraging dependency. Easier to communicate and depending on the May require more consensus building as to what methodology may be more transparent to the constitutes as a socially acceptable minimum public. income. May be more politically acceptable. 12 Box 1 Case studies: low generosity threshold in the US vs. decent living standard in Sweden Sweden and US are interesting examples since they take quite different social assistance approaches despite being high income countries, especially in terms of how developed their respective welfare systems. Sweden is regarded as one of the best developed and most generous welfare states co-existing with one of the lowest rates of poverty and inequality among high income countries. In contrast, the US has the least generous and fragmented welfare system and one of the highest rates of poverty among high income countries. This is unsurprising in light of general principles that guide policy in two countries. While the US social contract emphasizes personal responsibility, in Sweden minimum welfare for all emphasizes social responsibility. In the USA a single national income poverty line is set for statistical analysis by the Census Bureau but poverty thresholds for the social assistance eligibility are set by the Department of Health and Human Services for social assistance administrative purposes to determine whether a family is eligible for assistance for federal programs. There are 27 different federal assistance programs linking their eligibility criteria to the US poverty guidelines which are derived from the official poverty thresholds. As one example, the Food Stamp Program is the largest U.S. anti-poverty food assistance program. The program is available to all citizens who meet income and asset tests. It is administered by local government but Federal Government is responsible for all the rules such as the threshold, which is nationally uniform and does not vary by state. An eligibility criterion for the needy families is that household net income must be equal to or less than the national poverty line (the poverty line varies with family size). The threshold is higher if there is an elderly or disabled member in a family. The Census Bureau issues the poverty thresholds that are updated annually for changes in the Consumer Price Index. However, the poverty thresholds do not account for regional variation in cost-of-living. Currently, the official federal poverty thresholds assume that the cost of living is the same over the entire United States. The US National Academy of Sciences has recommended that regional cost-of-living is taken into account in calculating the regional thresholds. Swedish MLG social assistance is regulated via the Social service Act and administered by the local government (municipalities). The Social Service Act states that everyone should be guaranteed a reasonable standard of living. The national social assistance norm (Minimum Living Standard) is approved every year by the parliament; hence, it is not decided locally. The National Board of Health and Welfare in 1985 presented a recommendation for a minimum living standard norm. The standard is meant to cover cost of food, clothing, shoes, sport, leisure, consumable goods, newspapers, telephone, and television license fees, small medical expenses, home insurances costs and utility. Each year it is updated to account for inflation but also takes into account any changes in average expenditure through Family Expenditure Surveys. In order to further diminish the variations between municipalities, a legislated national norm was introduced in 1998 and since then the parliament has taken annual decisions about a nationwide social assistance norm to which municipalities have to adjust. That is, the municipality can raise the national threshold, but not reduce. Estimations indicate that the social assistance norm guarantees an income that is approximately equivalent to 60 per cent of the median income. In a bigger context the struggle between centralization and decentralization can be seen as a clash between two principles that are equally strong. On the one side is local self-governance seen as an important aspect of the political system, on the other side there is a principle of equal treatment of citizens regardless of where in the country they live. There are a number of other type benefits that are linked to means tests. There are for example housing allowance for young people (between 18-24 years) about to establish an independent household and, another housing allowance system, for older people (65+). There are also special pension arrangements for older people with low incomes that are aimed at 13 guarantee an adequate income and thereby lift them above the threshold for social assistance. But these benefits do not mean that a person in principle cannot get social assistance if needed (for example in case of extra expenditures), it just means that these incomes are included in the means test. Sweden has perhaps the most comprehensive means test of any country in the study: in principle all forms of household income are taken into account when calculating social assistance. All earnings of the adults are counted in full net of tax but not taking into account any private pension contributions. Children are not responsible for supporting their parents, but their earnings can reduce social assistance up to the level of the appropriate child rate. There is a strong labor activation element since benefit recipients - at least if they are prime aged and in good health - are obliged to work or, if unemployed then to search for a job. C. Typology of basic features among MLG programs in developed countries Developed countries see minimum living guarantee programs as short-term assistance of last resort, many of them in order to assure “up to a fixed living standard”. While there is variation in the parameters of individual schemes across the European countries, the common thread across them is that the eligibility conditions are determined by age, residence status, access to financial resources, income means tests, and ability and willingness to work. They are all mean-tested, non-contributory, and funded through the tax system. European minimum living guarantee programs are guided by the 1992 EU Council recommendation concerning sufficient resources for social assistance programs. This is a non-binding recommendation that invites member states to recognize the fundamental individual right to sufficient and reliable resources in line with the following common principles: 1) The amount of resources is sufficient to cover essential needs and is consistent with current living standards, and accounts for different types and sizes of households. 2) These amounts are fixed with reference to appropriate indicators, in particular average disposable income in the member state, and statistical data on household consumption, the legal minimum wage, and price levels. 3) Arrangements for periodic review and recertification. 4) Financial aid is individually differentiated and must bridge the gap between the applicant’s household resources and the amounts deemed to be sufficient resources. 14 5) Measures are taken to ensure that the most disadvantaged people are properly informed of their right. None of the countries falls squarely into one category but, rather, in the intersection between eligibility criteria, means testing, and time limitations. This sub-section of the paper will review the key institutional features of MIG programs among selected OECD schemes and classify them into different typologies. The classification sorts different European programs according to the type of eligibility, determination of the minimum income, and benefit time limits. Some countries (See Table 2.3) such as Finland apply both means test and categorical criteria, with time limits on benefits. Table 2.3: Types of MIG programs Means-tested and non-categorical Means tested and partially categorical Belgium, Czech Republic, Germany, Estonia, Austria, Bulgaria, Cyprus, Denmark, Spain, Finland, Lithuania, Luxemburg, Latvia, Netherlands, France, Ireland, Poland, Romania, Slovenia, United Portugal, Sweden, Slovak Republic Kingdom Source: Frazer and Marlier (2009) Although the MIG assistance is seen primarily as short-term support they are generally not time limited, although other conditions may apply such as willingness to work or permanent disability (Table 2.4). In many countries (e.g. Belgium, Ireland, Finland, Poland and UK) beneficiaries of MIG also receive assistance for other needs, such as housing, children and education. Emergency assistance is also commonly provided, like one-time benefit to cover the most basic needs in Poland. Table 2.4: To what extent are MIG payments time-limited? No time limit but periodic reassessment Time limited Sweden, Czech republic, Germany, Denmark, Estonia, Greece, Finland, Bulgaria, Slovenia France, Hungary, Ireland, Luxembourg, Netherlands, UK, Slovak Republic, Austria, Belgium, Estonia, Poland, Portugal, Romania, USA III. Refinement of thresholds: International experience 15 The level of MLS thresholds may vary across households and across areas within a country, according to factors such as household size and characteristics, differences in regional cost-of-living. And over time to account for inflation. For example the 1992 EU council recommendation concerning sufficient resources for social assistance programs sets out the principles of differentiated benefit amounts, and formal indexation procedures. International experience shows that generally national poverty eligibility thresholds for MLG programs are estimated for different families and use a consistent method for annual indexation in line with inflation. There is wide conceptual agreement for adjusting for regional cost-of-living differences; however, this is not followed in many countries as a practice due to limited availability of regional price information. A. Definition of an eligibility unit and household specific threshold In international practice, minimum living threshold eligibility is estimated for a household and thresholds increase with each additional member, but not always proportionately. Typically, OECD countries adjust the benefits and thresholds depending on the number of people in the household. Estimating individual income makes little sense since almost always resources are shared within a household. One way to consider family thresholds is to simply multiply per capita threshold times the household size. But such an approach will ignore the differences in needs between adults and children and that the cost of reaching the same welfare for each additional person may be decreasing with household size as they share common resources such as electricity or heating. Intuitively speaking, the per capita measure of income gives equal measure of needs for every member regardless of their age. In contrast, the Adult Equivalence Scale (AE) is a formula used for comparing different households and measures the relative costs of living for families of different sizes and composition; for instance, if a child needs half as much food as an adult, then a two adult - two children household will consist of three equivalent adults. Table 3.1 illustrates that a number of countries use the equivalence scales to determine the size of the minimum income eligibility thresholds. Table 3.1: Type of household size adjustment by countries 16 Per capita benefit (no Per capita with Adult equivalence scale adjustment adjustment for economies of scale composition of adjustment household) Latvia, Lithuania Estonia Canada, Sri Lanka, Germany, US, UK, South Africa, Thailand, India, Tanzania, Australia, Netherlands, Sweden, Switzerland, Source: Buhmann, Rainwater, Schmaus, Smeeding (2006) The choice of a particular equivalence scale depends on technical assumptions about economies of scale in consumption as well as on value judgments about the priority assigned to the needs of different individuals such as children or the elderly. These judgments will affect results. There is no scientific consensus on what is the correct equivalence formula and there are many possibilities. The needs of a household grow with size but not in a proportional way therefore the choice of the formula depends on the functional form that is appropriately capturing economies of scale in consumption for different household sizes. Two main types of formula essentially reflect whether relationship between household size and consumption needs is treated linearly or non- linearly. The most common formula from OECD (linear growth in size) and World Bank experience (non-linear change) are presented below. Since late 1990’s the Statistical Office of the European Union (EUROSTAT) adopted the following formula for OECD: ( – ) … (1) This assigns a value of 1 to the first household, and 0.5 to each additional adult and of 0.3 to each child. For administrative purposes, and to fit into the budgetary needs, the formula can be capped so that the benefit has a maximum regardless of how large is household size. For an illustration of how this formula will look to the implementation office, suppose the maximum family benefits are capped to a level corresponding to 2 adults and 2 children per household. Suppose that the single adult threshold is 1000 Yuan per year. Using the OECD formula (1) the menu of family MLG thresholds is represented in Table 3.2. Table 3.2: Example of variations in threshold due to household size based on OECD equivalence scale 17 No of adults in the No of children in the Number of Adult Threshold for the total household household Equivalents in the household income household 1 0 1 1000 1 1 1.3 1300 1 2 1.6 1600 2 1 1.8 1800 2 2 2.1 2100 2 or more 2 or more 2.1 2100 Alternatively World Bank uses the following formula when advising low-income countries on poverty thresholds. This formula incorporates economies of scale that enters the function non-linearly: … (2) Where Nadults and Nchildren are number of adults and children living in the household, α is the parameter for children’s share that usually ranges from 0 <α<1, θ is the parameter of economies of scale which is typically chosen from 0.5 < θ <0.95 range. If θ =1 the formula begins to resemble the OECD formula. This simple differentiation between adults and children may be refined further; this formula can easily be adapted not just for two groups (adults and children) but also to incorporate different needs for multiple age groups like elderly and disabled. For a specific example of the values attached to these parameters, the U.S. National Academy of Science Poverty Study (Michael and Citro, 1995) in the USA recommended fixing the parameters at: α = 0.7 and θ = 0.762. But in practice according to the last equivalence scale estimates published by the US government (Jackson, 1968), a typical adult living alone requires 36% of the income of a typical family of four to attain the same standard of living or welfare level as the family. So, a family of four spending $10,000 a year is as well off as a single adult with an expenditure level of $3,600 per year. 18 Box 2 Case study: anti-poverty program Bolsa Familia in Brazil The Bolsa Família Program (BFP) was launched in 2003, consolidating four existing national programs into a single means-tested cash transfer that would reach 24% of Brazil’s population (44 million people) in 2007, making it one of the world’s largest public cash transfers to the poor. The BFP program targets poor and extreme poor families throughout the country. To date, the BFP has achieved some exceptional targeting results (Lindert et al, 2007). The program designers opted for a pragmatic set of benefits that (a) was simple to administer; (b) favored the extreme poor; (c) favored families with children – but with limits to avoid promoting fertility; and (d) prevented eligible beneficiaries of the pre-reform programs from losing from the reforms. The BFP provides two types of benefits: basic and variable according to family composition and income. The BFP provides a base benefit of R$50 to all families in extreme poverty, regardless of their demographic composition (moderately poor families do not receive this base benefit). Both extreme poor (monthly per capita income up to R$50) and moderately poor (income between R$50 – R$100 per capita) families receive a variable benefit set according to the number of children in the family (capped at three for variable benefits purposes) and whether the mother is pregnant or breast-feeding. The original income eligibility thresholds for the BFP program were set at a fixed monthly per capita family income of R$100 (US$48) for moderately poor families and R$50 (US$25) for extremely poor families. With this benefits menu, income transfers range from R$15-95 (US$7-48) per family per month. The average value of benefits paid during the period January – May 2006 was about R$62 (US$30). BFP benefit values are not automatically indexed to inflation or minimum wage increases. But benefits are increased according to law periodically. This is not considered a good practice due to erosion of the benefit value in the interim period between increases by legislative decree. Table 3.3: Eligibility thresholds and benefits in Bolsa Familia Program Level of Monthly per Number of children 0- Bolsa Familia poverty capita income 15, pregnant or breast- Benefit eligibility feeding mothers Received threshold Poor $27 -60 1 $7 2 $14 3 or more $23 Extreme poor Upto $27 0 25 1 33 2 40 3 or more 48 Source: Lindert et al, 2007 While the assistance unit is defined as the family as a whole, payments are made preferentially to the woman in each family as the legally-responsible beneficiary, as established by the BFP law. Currently, 93% of legally responsible beneficiaries are women. This preference for payments to women reflects international experience (and early tests in the municipal BE programs) that suggest that women are more likely than men to invest additional income in improving the education, health and welfare of their family, particularly their children. 19 B. Indexation and trends of minimum income thresholds across OECD An eligibility threshold requires periodic adjustment to account for changes in living cost over time to preserve the real value of the benefit over time. There are two types of temporal adjustments. First, updating of the thresholds based on changes in the relevant basket of goods, and in growth of general levels of consumption or wages. This type of update is not frequent and occurs every 3-5 years. The second one is the more frequent indexation to the changes in the consumer price index, other indicators like wages, or hybrid. Among the OECD countries this is followed automatically and systematically as per the social security legislation. Table 3.4 categorizes the European member states by presence of a formal indexation procedure. Most European member states and OECD states use systematic indexation methods. To understand what type of indexation procedures are used among the OECD countries, trends in MIG threshold levels are estimated using the Social Assistance and Minimum Income Protection Data-set (SAMIP, 2010). Since European benefits vary by family composition, for consistent comparisons the trends are presented for lone parent with two children and compared with the national consumer price index. Table 3.4: Is there a clear mechanism for periodic review of MIG eligibility thresholds? Regular “hybrid” indexation linked to Regular indexation using a fixed procedure linked to CPI combination of indicators: average earnings, minimum wage and CPI Annual Every 6 months Quarterly Annual Austria, Belgium, Canada, Ireland, Luxembourg, Australia, Portugal, Spain, Sweden, Germany, Denmark, Finland, Japan, Netherlands, UK, USA France, Italy Turkey New Zealand, Switzerland, No systematic indexation procedure used, arbitrary updating of thresholds Greece, Iceland, Norway Source: Frazer and Marlier (2009) and Eardley et al (1996) 20 The automatic adjustment procedures in OECD practice are mostly designed to keep up with annual price movement only. Figure 3.1 shows the trend in the average benefit5 level paid to single parents with 2 children in selected OECD countries and compares it with the trend in Consumer Price index (CPI) during 2000 and 2010. Estimates clearly indicate that some national governments (Austria, Australia, France, and since 2003 the UK) used their respective CPIs to adjust their benefit levels for inflation. In other countries the benefits are not linked exactly to CPI changes but do keep up with the general trends in consumer price index. Adjusting thresholds using a national CPI may not capture differences in regional inflation and also in periods of high growth may result in indexation that is not keeping up with indicators such as average wages. In this case, over time, the MIG eligibility threshold becomes less redistributive. In order to address this concern, in some countries the benefit threshold is linked to growth in wages or consumption (Finland, Japan, Denmark, the Netherlands and New Zealand). Other variations can be seen in indexation rules across the OECD and in some cases within individual countries. In Canada and Germany for example, thresholds are set by local governments and may depend on a variety of factors in addition to inflation. Even in countries where thresholds are linked to growth in minimum wages such as in Austria, Denmark and the Netherlands, the benefit levels tend to track inflation. There is a wide conceptual agreement that differences in costs-of-living across regions need to be taken into account. Individuals living in different locations may face different prices for similar goods. Ignoring the differences in prices across regions and using only one national threshold level threshold can lead to under-estimation of the eligible households in areas with higher prices. While there is wide agreement on the desirability of adjusting poverty eligibility thresholds regional differences in prices, there are some practical difficulties. Few countries practice spatial price adjustment as geographic area CPI or cost- of-living indexes are not common. Most European countries are relatively small, and 5 Normalized to the value in 2005 to create a comparable scale to the CPI 21 regional differences may not play a strong role for eligibility determination. This is more relevant for countries like USA. 22 Figure 3.1: Trends in benefits and CPI Austria Japan Australia Denmark 100 110 120 100 105 110 90 95 90 80 National CPI National CPI 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 Sweden United States Netherlands 100 105 110 100 110 120 95 90 90 80 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 year year cpi Single parent 2 child benefit cpi Single parent 2 child benefit Graphs by Time Graphs by Time Canada United Kingdom France Germany 110 250 100 200 National CPI National CPI 150 90 100 80 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 year year cpi Single parent 2 child benefit cpi Single parent 2 child benefit ,labsize(small) Graphs by Time Graphs by Time Source: Author’s calculations based on OECDstat and SAMIP databases 23 C. Roles of Central-local responsibilities in delivery and financing of social assistance Social assistance administration across countries is a product of constitutional and political arrangements reflecting fundamental cultural and historical initial conditions, and to a lesser extent driven by efficiency and expenditure control considerations. Arguably the debate in the developed countries about the distribution of costs of social assistance between the province and the municipalities is not orientated towards actual problems or towards achieving greater efficiency. More often the decisive question is the distribution of political power between the individual territorial bodies. But when countries devolve the responsibility for both financing and administration of social assistance to the local level, this can create spatial inequity in levels of thresholds and benefits due to variations in the rules of eligibility. The financial burden on local authorities means that in some locations there is more rigid enforcement of eligibility, applying local discretion to ration social assistance. This can constrain the potential of social assistance to promote social cohesion and reduces efficient management of social assistance arrangements. The potential for coordination of central schemes, such as the development of unified eligibility criteria means that these schemes are simpler and cheaper to administer. Standards of administrative justice are commonly perceived to be higher in such schemes, mainly owing to improved consistency of decision making which can prevail over the benefit of local decision making arising from improved access to information about local conditions. In sum, Australia, Belgium, Denmark, France, Ireland, Netherlands, New Zealand, UK, and the US have national regulated social assistance minimum income eligibility thresholds. On the other hand, in countries such as Germany and Canada local government at province level establish their own social assistance thresholds. In a number of OECD countries, social assistance schemes are administered and partially cofunded by local, regional or provincial tiers of government rather than by national-level departments. In at least half of OECD countries, the thresholds are set at the national level whether or not the program is locally administered, but in some countries the thresholds show some regional variation due to discretionary powers to set the threshold at the local level. Countries with locally administered schemes but national guidelines for eligibility thresholds include Germany and Finland which have national eligibility threshold (but which varies locally 24 according to differences in the cost-of-living not due to discretionary power of local government); Japan which has six geographical cost-of-living bands and Iceland which uses a single national threshold level. In Sweden thresholds can be set higher locally according to fiscal resources available to local governments (see Box 1). The most extreme case is of Switzerland where there are national guidelines but wide local variation in actual threshold since in theory all 3000 cantons can set their own MLS threshold. Switzerland, Austria, Canada, Norway, Spain and Italy seem to be the only countries where both administration and the setting of threshold rates are entirely the responsibility of the provinces or municipalities. In all other countries there is a national social assistance threshold and benefit rate that may only vary due to cost- of-living differences. Similarly there are variations in funding principles and procedures between central departments and local government. The main examples of large-scale social assistance schemes which are both organized and funded nationally are those of Australia, New Zealand and UK. The share of central financing varies from 50% in Denmark and close to 100% in France and Australia; In the Netherlands central funding share is 80%. In Austria it is on average 60% but may vary across provinces. In Japan and Bulgaria 75% is funded by central government. IV. Policy implications for China’s Di Bao programs International lessons and experience provide valuable insights for China to inform its policy development and improvement of social assistance programs. In developed countries, the generosity of social assistance programs is deeply rooted in social values and principles that underpin their policy objectives. Countries providing decent living for the poor tend to have higher thresholds than those guaranteeing basic needs. Those practices and experience will be useful to China’s social assistance programs to set up a sound threshold level, make adjustments and to define the responsibilities between central and local governments. 25 The Chinese government has made policy initiatives6 to improve Di Bao programs and social assistance system. Urban and rural Di Bao programs have become the backbone of social assistance and social welfare systems in China. In 2012, they covered 21.4 million urban beneficiaries and 53.4 rural beneficiaries, accounting for 3.0 percent of urban population and 8.3 percent of rural population. Although remarkable progress has been made, there is space to improve China’s social assistance programs7. The policy initiatives reflect the efforts that the Chinese government wants to make avoid the fragmentation and towards standardized administration of Di Bao programs. A. Approaches for setting up Di Bao thresholds Unlike developed countries, in China, localized approaches have been used to develop Di Bao thresholds, under broad national guidelines which permit substantial discretion. While some provinces (or cities) apply standardized budget approaches to measure subsistence food and non-food expenditures to identify the threshold levels based on household survey data, such as in Beijing, Shanghai, Hebei, and Chongqing for urban Di Bao programs. Other provinces (cities) refer to minimum wage to develop local urban Di Bao thresholds (usually 30-40 percent of average wage), and use 60 percent of local urban Di Bao thresholds as local rural Di Bao thresholds, as in Fujiang and Hanzhou. The majority of provinces (cities) adopted the proportion of food expenditures to total expenditures (called Engel’s coefficient approach) and/or the proportion of per capita income (called income ratio approach) to set up the thresholds, such as in Inner Mongolia, Anhui, Shandong. 6 They include: (i) the Guidance on Further Formulating and Adjusting the Thresholds of Rural and Urban Minimum Living Guarantee Programs (2011); (ii) A Notice on Establishing the Linkage Mechanism of Social Assistance and Benefit Level with the Increasing of Prices (2011); (iii) A Notice on Strengthening the Budgetary Management of Social Assistance Funds for Rural and Urban Minimum Living Guarantee Programs (2011); (iv) A Notice on the Fund Management Approach for Rural and Urban Di Bao Programs (2012); (v) A Notice on Verification and Approval Approaches (Trial) for Di Bao Prgorams (2012); (vi) A Notice on Further Strengthening Rural Poverty Monitoring Work (2012). 7 Empirical studies show that Di Bao programs had limited impacts on poverty reduction in its early stage due to large under-coverage and exclusion error (Chen, et al, 2006; World Bank, 2007), but have improved over time in performance (World Bank, 2010; Li and Yang, 2009; Gustafsson and Deng, 2011). If the broad policy objectives of Di Bao programs are considered, the targeting performance needs to be assessed against wider measures than simply income or consumption (Ravallion, 2008), suggesting the targeting performance is good. 26 The choice of different approaches is largely driven by local economic conditions and fiscal capacities. In developed areas or large cities such as Beijing, Shanghai and Jiangsu, local governments respond more generously to the needs of poor households by increasing the threshold level and are less strained in their fiscal capacities, but in medium and small cities or poorer areas, the threshold standards are almost determined by local fiscal capacities (Tang and Ren, 2002). In recent years, the central government has contributed more than 60 percent of total fiscal resources for both rural and urban Di Bao programs, enabling local governments to reduce the burden on the local fiscal system, especially for medium and small cities, and inland provinces. Choosing income (expenditure) ratio approaches would underestimate the basic needs of the poor although they are less demanding for implementation. As shown in Figure 4.1, the proportion of food expenditures to total expenditures follows a downward trend in both rural and urban areas, so the approaches could not capture the non-food needs well over time. Moreover, it would be difficult to address the disparity between rural and urban areas in setting up the threshold level based on the existing data. In addition, the average income or expenditures are used, not by income (expenditure) group; this would not consider a higher proportion of food consumption for the poor. Figure 4.1: Declining Engel’s Coefficients in Rural and Urban China, 1990-2012 65.0 Urban Rural 60.0 55.0 Engel's coefficient 50.0 45.0 40.0 35.0 30.0 2001 1990 1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 2002 2003 2004 2005 2006 2007 2008 2009 2010 2011 2012 Source: National Bureau of Statistics (NBS), China Statistical Yearbook (2012), China Statistics Press, Beijing. 27 The fragmentation of localized approaches has caused problems in addressing policy objectives: (i) big gaps of threshold levels suggest the inadequacy of financial protection for the poor in the areas with low thresholds due to the weakness of fiscal capabilities; (ii) different actual proportions with the same approach probably magnified the gaps because poor areas tend to choose low proportions to reduce local fiscal burdens; (iii) it would be difficult for comparison on the incidence of the poor in terms of different approaches, causing problems to standardize the implementation and management of Di Bao programs. The China’s Ministry of Civil Affairs (MOCA) has fully recognized the issues of Di Bao thresholds and initiated the process of convergence of localized approaches8. In 2011, MOCA proposed three approaches for local governments to choose: basic living expenditure approach, Engel’s coefficient approach and consumption expenditure approach. In reality, the last one was widely accepted and applied, because it requires less knowledge and capacity for local Di Bao bureaus to implement it. The proposed three approaches are in fact two. The basic living expenditure approach is equivalent to the standardized budget approach. The Engel’s coefficient approach formulates the threshold as the basic expenditures for food divided by the Engel’s coefficient of the lowest income group. The consumption expenditure approach is an adjusted Di Threshold by using information of Di Bao threshold level in the year and previous and current consumption. The initial Di Bao threshold would be estimated by either the basic living expenditure approach or by the Engel’s coefficient approach. China may consider a strategy toward the convergence of the approaches. At this stage, it is practical to ask local governments to adapt localized approaches toward one of the proposed three approaches. For the next step, MOCA should collaborate with the National Bureau of 8 A few preliminary studies tried to estimate the reasonable Di Bao threshold level by using models similar to the standardized budget approach, Engel’s coefficient approach and income ratio approach. Some used the Extended Linear Expenditure System (ELES) model to estimate the subsistence level by using aggregate household expenditure data and then constructed the thresholds (Mi and Ye, 2007; Yan, 2007; Peng, 2011; Hu, et.al, 2011). The availability of aggregate household data is probably one of the reasons that scholars chose the ELES model. Some used the proportion of per capita expenditures or income to estimate the levels of thresholds and did some simulations for fiscal policy implications (Yang, 2011; Yao; 2011). They usually selected the cutoff point of expenditures where the Engel’s coefficient is 60 percent; or 1/4-1/3 of average income as the threshold levels. 28 Statistics and strengthen its information system to decide the approach that is workable and better to address the policy objectives. Finally, the selected approach will be applied to all provinces to standardize the process of setting up the threshold level. Another option is to set up the minimum threshold level nationwide by using a suitable approach and ask local governments to take it as the bottom line. Local governments are encouraged to top up this line. This would have no need to pay attention to localized approaches, but this may raise bigger responsibility for the central government. B. Di Bao thresholds and benefits Both Di Bao thresholds and benefits have increased since its initiation (See Table 4.1). For urban Di Bao programs, the threshold level rose from 1788 Yuan in 1999 to 3960 Yuan in 2012. However, relative to average urban incomes, the urban dibao threshold fell from over 30% in 1999 to under 16% by 2011. Interestingly though, the average urban dibao benefit size increased relative to average incomes and expenditures during the period as benefits increased sharply as a proportion of the threshold level. For rural Di Bao program, the threshold level increased from 851 Yuan in 2006 to 2068 Yuan in 2012. Its ratio to average rural income was relatively table, while the proportion of the benefits to rural threshold level went up from 48.7 percent to 60.3 percent in the same period. Table 4.1: Di Bao Threshold and Benefit Levels in Rural and Urban China Threshold % of Level %of Benefits Level %of threshold (Yuan/Year) % of income expenditures (Yuan/Year) % of income expenditures level Urban Di Bao Program 1999 1788 30.5 38.7 538 9.2 11.6 30.1 2000 1884 30.0 37.7 544 8.7 10.9 28.9 2001 1764 25.7 33.2 355 5.2 6.7 20.1 2002 1776 23.1 29.5 527 6.8 8.7 29.7 2003 1788 21.1 27.5 696 8.2 10.7 38.9 2004 1824 19.4 25.4 780 8.3 10.9 42.8 2005 1872 17.8 23.6 868 8.3 10.9 46.3 2006 2035 17.3 23.4 1003 8.5 11.5 49.3 2007 2189 15.9 21.9 1232 8.9 12.3 56.3 2008 2464 15.6 21.9 1724 10.9 15.3 70.0 2009 2734 15.9 22.3 2064 12.0 16.8 75.5 2010 3014 15.8 22.4 2268 11.9 16.8 75.2 2011 3451 15.8 22.3 2698 12.4 17.4 78.2 29 2012 3960 16.1 23.7 2869 11.7 17.2 72.5 Rural Di Bao Program 2006 851 23.7 30.1 414 11.5 14.6 48.7 2007 840 20.3 26.1 466 11.2 14.4 55.4 2008 988 20.7 27.0 605 12.7 16.5 61.2 2009 1210 23.5 30.3 816 15.8 20.4 67.5 2010 1404 23.7 32.0 888 15.0 20.3 63.2 2011 1718 24.6 32.5 1157 16.6 21.9 67.3 2012 2068 26.1 35.0 1248 15.8 21.1 60.3 Source: National Bureau of Statistics (NBS), China Statistical Yearbook (2013), China Statistics Press, Beijing; Ministry of Civil Affairs, China Civil Affairs’ Statistical Yearbook (2013), China Statistics Press, Beijing. Di Bao thresholds are low in China compared with rural and urban income level and by international standards of developed countries. A number of studies have pointed out it cannot adequately meet the needs of poor households9. The objective of Di Bao programs is to provide basic needs for the poor10 through an unconditional cash transfer. The gap between poor households’ income and threshold level will be topped up through public income transfer. However, the threshold level has not kept up with income growth. The ratio of urban Di Bao threshold level to per capita income has declined sharply11 (See Table 3.4). Although the ratio of rural Di Bao threshold level to rural income has been relatively stable, the gap with the urban Di Bao threshold level remains large by considering the difference in the cost of living between rural and urban areas. In 2012, the urban-rural gap of Di Bao threshold level was 1.9:1 (rural=1), reflecting urban-rural disparity. Internationally, the threshold level is about one-third of wage, also suggesting low Di Bao thresholds in China. Compared with the official rural poverty line, rural Di Bao threshold level is also low. From 2006 to 2010, rural Di Bao threshold levels were slightly higher than rural poverty line in four 9 See Tang and Ren, 2002; Li, et.al, 2007; Liu, 2009; Feng and Zheng, 2009; Cheng and Wu, 2011. 10 According to the Regulation of Minimum Living Guarantee for Urban Residents in 1997, the urban Di Bao program provides income support to cover clothing, food, accommodation cost, and certain expenses for water and gas, and compulsory education for children to guarantee the basic living needs for urban poor households. Similarly, the State Council in 2007 issued a Notice on establishing the Minimum Living Guarantee System in Rural Areas, asking for the provision of income support to cover food, clothing, and expenses for water and electricity to guarantee the basic living needs for rural poor households. 11 The benefits may be not low if non-cash transfer is considered. Di Bao beneficiaries often received temporary assistance and other cash/non-cash benefits such as educational and medical assistance, housing subsidies, free training, etc, which caused a so-called “cliff effects”, referring that the near poor would be worse off compared with Di Bao beneficiaries. 30 years, but lower than the latter in 2008. China’s rural poverty line has until recently been very low compared with international standards (Ravallion et al, 2008). Recognizing this issue, the Chinese government significantly increased the official line to 2300 Yuan in 2011 when it launched a 10-year poverty reduction plan (See Figure 4.2). This raises a question about how the Di Bao threshold will be adjusted to keep in line with the change in the increasing of rural poverty line. It is worth noting that the average also disguises the low threshold level in a number of provinces, especially in central and western areas. Figure 4.2: Comparison between Rural Di Bao Threshold and Rural Poverty Line, 2006-2011 2500.0 Rural dibao threshold 2000.0 Rural poverty line 1500.0 Yuan/year 1000.0 500.0 0.0 2006 2007 2008 2009 2010 2011 Source: National Bureau of Statistics (NBS), China Rural Poverty Monitoring Report (2011), China Statistics Press, Beijing; Ministry of Civil Affairs, China Civil Affairs’ Statistical Yearbook (2012), China Statistics Press, Beijing. In its 12th 5-year plan, the Chinese government aims to increase Di Bao threshold level by more than 10 percent annually and achieve a 7 percent annual growth of per capita income in both rural and urban areas. The increasing of Di Bao threshold at a relatively faster speed can improve the living conditions for the poor, but its level would be still lower than the official rural poverty line. Assuming rural Di Bao threshold keep an annual growth of 10 percent from 2011 to 2015, its level would reach 2261 Yuan in 2015, lower than rural poverty line in 2011. Increasing Di Bao threshold also indicates a change in direction of China’s social assistance policies from the subsistence protection from prevention and promotion for the poor. 31 China may consider a gradual way to narrow the big gap of threshold level across region and between rural and urban areas. Developed countries often apply a unified formula to set up the threshold level for social assistance programs. This does not mean they choose a unified absolute level nationwide, while there is a variation due to regional cost-of-living differences. Together with the convergence of the approaches, China could consider a gradual move up in setting of thresholds from county (city) to prefecture, from prefecture to province, and finally from province to nationwide. This implies greater equalization within first prefectures and then provinces over time. Both national and sub-national government can play active roles in the equalization of thresholds. C. Refinements and Financing In developed countries thresholds are refined following a standardized mechanism to reflect a number of demographic, spatial and price factors. The threshold level varies in terms of household size and demographic structure. It is indexed over time to reflect changes in consumption, average wage, inflation and cost of living. This means a unified formula is adopted, but the absolute value varies between households and across regions. An adjustment mechanism for Di Bao programs has been introduced, but is at an early stage. Policy guidance was issued in 2011 to request local governments to establish an adjustment mechanism for Di Bao programs, but it is flexible on which formula they can choose. In practice, some provinces use the increase in CPI over consecutive 3 months, some use quarterly averages, while others are based on year-to-year price changes. The adjustment for benefits also varies. Some introduce above values, some use proportions. The next step could consider a standardized adjustment formula and applying it to all provinces. The remaining big issue is how to construct a suitable dynamic adjustment mechanism to keep in line with income/consumption growth, wage growth and price change. MOCA is now collaborating with NBS to collect a special CPI for the poor in several cities. This is an appropriate direction for establishing a sensible adjustment mechanism for Di Bao programs in China. In developed countries, the financing responsibilities are clearly defined between central and local governments, and between sub-national governments, and managed transparently 32 using the budgetary and legislative system. Some countries set up the minimum thresholds at the national level and local governments have the flexibility to top up the minimum thresholds based on local resources. Countries codify both thresholds and financing responsibilities to guarantee a decent life for the poor. Those experiences are useful for China to define the financing responsibilities and strengthen the legislation for its social assistance system. At present, the increasing role of the central government in China provides a channel to standardize and equalize Di Bao programs. In recent years, the central government has taken a major role in financing both rural and urban Di Bao programs and made efforts to strengthen the budgetary management of social assistance funds. Given such changes, China should consider amending social assistance regulations and policies to promote greater system standardization. It could also consider moving ahead with its draft social assistance law to more clearly define the policy objectives and scope of Di Bao programs and the relationship with other social policy programs, and to define the roles of central and local governments in financing and management. In the meantime, the central government should upgrade the authority of setting up Di Bao thresholds from county (city) to province to facilitate the convergence of the approaches and the equalization of Di Bao thresholds and benefits. For the management of social assistance funds, a transparent budgetary management system can be established to transfer and allocate social assistance funds through a formula that includes key factors such as local needs based on poverty, number of beneficiaries, local fiscal capacity, local Di Bao funds, and poverty reduction performance. V. Conclusion This policy note has summarized the international practice and experiences on setting up MLG thresholds. Using a simple framework, it discusses the underlying reason of a large variation of threshold level in developed countries, from the least generous 20% to around 60% of median wage, with an average of 35 %. The generosity of social assistance programs is deeply rooted in social values and principles that underpin their policy objectives. Many OECD countries set their policy targets for MLSs beyond basic needs and aim to guarantee a minimum socially acceptable level for decent living. Different approaches are used to set up the eligibility 33 thresholds across countries, but the same formula is applied within a country. The thresholds are also refined to reflect the difference in family size and demographic structure, difference in regional cost of living and changes in prices and wage. Social assistance programs are managed largely by local government, and mostly co-financed by central and local governments. The responsibilities of financing and management are legislatively defined. China can learn from the international experience and improve its social assistance system. These lessons can be helpful to clarify and address a number of key questions in relation to the policy objectives, specific approaches, threshold generosity and refinement issues as the Chinese government continues its efforts to standardize the implementation and management of Di Bao programs. The recent policy initiatives have signaled the desire for convergence of local approaches in setting and adjustment of thresholds and benefits, and for strengthening budgetary management of social assistance funds. Therefore, the issues of low thresholds, fragmented local approaches, and gaps of thresholds between rural and urban areas and across regions should be properly addressed through a strategic approach. This includes: (i) accelerating the convergence of localized approaches, through choosing a suitable approach to achieve the objectives of Di Bao programs that are clearly defined and applying it to all provinces; (ii) gradually raising the administrative level for setting thresholds from county (city) to prefecture, from prefecture to province, and finally from province to national level; (iii) clearly defining the roles of central and local governments in financing and management through legislation; and (iv) establishing a transparent budgetary management system to transfer and allocate social assistance funds. 34 References English Literature Bastagli F. (2008) “The design, implementation and impact of conditional cash transfers targeted on the poor: An evaluation of Brazil’s Bolsa Fimilia”, PhD thesis, London School of Economics Brandt, L, & Holz, C 2006, 'Spatial Price Differences in China: Estimates and Implications', Economic Development and Cultural Change, 55, 1, pp. 43-86 Buhmann, B., L. Rainwater, G. Schmaus, and T. M. Smeeding. 1988. “Equivalence Scales, Well - Being, Inequality, and Poverty: Sensitivity Estimates Across Ten Countries Using the Luxembourg Income Study (LIS) Database” Review of Income and Wealth (June): 115–142 Casas, R (2005) “Minimum income standards in enlarged EU: Guaranteed Minimum Income Schemes”, EC Working paper Citro F., and R. Michael (1995) Measuring Poverty: A New Approach. Washington DC. Eardley, T., Jonathan B, Ditch J., Gough I., and P. Whiteford (1996) “Social Assistance in OECD countries”, Department of Social Security Research Report 46 Frazer, H and E. Marlier. 2009 “Minimum Income Schemes Across EU Member States”, EC Report Grosh, M., Del Ninno, C., Tesliuc, E., and A. Ouerghi. 2008. For protection and promotion. The design and implementation of effective safety nets. The World Bank, Washington, DC. Gustafsson, Bjorn A. and Quheng, Deng (2011) "Di Bao Receipt and Its Importance for Combating Poverty in Urban China," Poverty & Public Policy: Vol. 3: Iss. 1, Article 10. Income in China’s Cities,” Policy Research Working Paper, WPS 3805, Washington DC, World Bank. Jackson, Carolyn A., (1968), "Revised Equivalence Scale for Estimating Equivalent Incomes or Budget Cost by Family Type," U.S. Bureau of Labor Statistics Bulletin 1570-2, November 1968. Lindert et al, 2007 “"The Nuts and Bolts of Brazil's Bolsa Familia Program: Implementing Conditional Cash Transfers in a Decentralized Context." World Bank SP Discussion Paper. Nelson, K. 2007 “Minimum Income Protection and European Integration: Trends and Levels of Minimum Benefits in Comparative Perspective 1990-2005”, International Journal of Health Services, Vol. 38(1). Nelson, K. 2010 "Social assistance and minimum income benefits in old and new EU democracies". International Journal of Social Welfare, Vol. 19(4). OECD Stat (2009) http://www.oecd.org/dataoecd/32/50/43948033.pdf. 35 Smeeding, T. M. 2009 “Poor People in Rich Nations: The United States in Comparative Perspective” Journal of Economic Perspectives 20(1): 69–90 Ravallion, Martin, 2007, “Geographic Inequity in a Decentralized Anti-Poverty Program: A Case Study of China”, Policy Research Working Paper 4303, Washington DC, World Bank. Ravallion, Martin, 2008, “Miss-targeted or Miss-measured?” Economics Letters, Economics Letters 100 9–12 Ravallion, Martin, Shaohua Chen, and Prem Sangraula (2008). “Dollar a Day Revisited.” Policy Research Working Paper 4620. The World Bank Development Research Group. Ravallion, M, Chen, S., and Y. Wang 2007 “Does the Di Bao Program Guarantee a Minimum Income in China‟s Cities?” (in Public Finance in China: Reform and Growth for a Harmonious Society, edited by Jiwei Lou and Shuilin Wang, Washington DC: World Bank. World Bank, (2007) Urban Di Bao in China: Building on Success, World Bank, Human Development Unity, South Asia, World Bank. World Bank (2010) Social Assistance in Rural China： Tackling Poverty through Rural Di Bao, Social Protection Group Human Development Unit, EASHD. Chinese Literature Chen, Xianshen, and Tianhong Wu, 2011 “Policy Recommendation for Improving Rural and Urban Minimum Living Guarantee System (Di Bao),” Fiscal Studies, No. 1, pp 31-35. Feng, Yuanyuan, and Laibing Zheng, 2009 “Bottlenecks of Implementing Rural Minimum Living Guanranttee System,” Rural Economy, No.7, pp57-60. Hu, Fangxiao, Cui Huang, Tingting Zhang and Chunyan Guo, 2011 “The Measurement of Minimum Living Guarantee Thresholds in Urban China,” China’s Second Social Assistance Conference Proceedings “Contemporary China’s Social Assistance System: Improvement and Innovation,” pp437-447. Li, Wenjin, Xiangyuan Li and Xiaofeng Zhang, 2007 “A Study on Establishing the Regular Growth Mechanism of Rural and Urban Minimum Living Guarantee Thresholds,” Northern Economy, No.7, pp. 26-28. Li, Xiangzhe, and Kunpeng Lai, 2010 “Problems of Implementing Rural Minimum Living Guarantee System and Suggestions,” Hebei Agricultural Science, No. 12, pp 107-108, and 142. Li, Shi, and Sui Yang, 2009 “Impacts of China’s Urban Di Bao on Income Distribution and Poverty,” Chinese Journal of Population Sciences, No.5, pp21-29, and 113. Liu, Xitang, 2009 “Outstanding Issues and Policy Suggestions of Current Urban Di Bao System in China,” No.4, pp55-60. Liu, Bin, and Xiaoyi Zhang, 2011 “The Categorical Approach for Urban Di Bao Households and Its Implementation—An Empirical Study by Using Household Survey Data in Shanghai,” China’s 36 Second Social Assistance Conference Proceedings “Contemporary China’s Social Assistance System: Improvement and Innovation,” pp428-436. Mi, Hong, and Lan Ye, 2010 “Model Development and Empirical Study of China’s Rural Minimum Living Guarantee System,”Zhejiang Social Sciences, No.5, pp55 -61, 67 and 128-129. Ministry of Civil Affairs (MOCA), and Asian Development Bank (ADB), 2006, A Study on China’s Rural Minimum Living Guarantee System. Peng, Xiang, 2011 “Measurement and Assessment of Rural and Urban Minimum Living Guarantee Thresholds from the Perspectives of Consumption, ” China’s Second Social Assistance Conference Proceedings “Contemporary China’s Social Assistance System: Improvement and Innovation,” pp483-489. Tang, Jun, and Zhenxing Ren, 2002 “Policy Suggestions for Improving China’s Social Assistance System,” Economic Reference, No. 76, pp 20-30. Tang, Jun, and Weiwei Liu, 2012 “ 2011 Social Assistance in Urban and Rural China,” Xing Lv, Xueyi Lu, and Peilin Li (ed), Social Bluebook: 2012 Social Trend Analysis and Projection in China, China Social Sciences Academic Press, Beijing. Yang, Lixiong, 2011 “Empirical Investigation and Testing of Measurement Approac hes for the Minimum Living Guarantee Thresholds,” China’s Second Social Assistance Conference Proceedings “Contemporary China’s Social Assistance System: Improvement and Innovation,” pp395-427. Yao, Jianping, 2011 “An Investigation on Integrating Urban Minimum Living Guaranttee Thresholds from the Perspective of Consumption Proportion Approach,” China’s Second Social Assistance Conference Proceedings “Contemporary China’s Social Assistance System: Improvement and Innovation,” pp448-458. Yan, Hongmei, 2007 “A Study on the Measurement Approaches of China’s Rural Minimum Living Guarantee Thresholds,” Ph. D. Dissertation, Nanjing Agricultural University. 37 Social Protection & Labor Discussion Paper Series Titles 2011-2013 No. Title 1307 Eligibility Thresholds for Minimum Living Guarantee Programs: International Practices and Implications for China by Nithin Umapathi, Dewen Wang and Philip O’Keefe, November 2013 1306 Tailoring Social Protection to Small Island Developing States: Lessons Learned from the Caribbean by Asha Williams, Timothy Cheston, Aline Coudouela and Ludovic Subran, August 2013 1305 Improving Payment Mechanisms in Cash-Based Safety Net Programs by Carlo del Ninno, Kalanidhi Subbarao, Annika Kjellgren and Rodrigo Quintana, August 2013 1304 The Nuts and Bolts of Designing and Implementing Training Programs in Developing Countries by Maddalena Honorati and Thomas P. McArdle, June 2013 1303 Designing and Implementing Unemployment Benefit Systems in Middle and Low Income Countries: Key Choices between Insurance and Savings Accounts by David A. Robalino and Michael Weber, May 2013 1302 Entrepreneurship Programs in Developing Countries: A Meta Regression Analysis by Yoonyoung Cho and Maddalena Honorati, April 2013 1301 Skilled Labor Flows: Lessons from the European Union by Martin Kahanec, February 2013 1220 Evaluating the Efficacy of Mass Media and Social Marketing Campaigns in Changing Consumer Financial Behavior by Florentina Mulaj and William Jack, November 2012 1219 Do Social Benefits Respond to Crises? Evidence from Europe & Central Asia During the Global Crisis by Aylin Isik-Dikmelik, November 2012 1218 Building Results Frameworks for Safety Nets Projects by Gloria M. Rubio, October 2012 1217 Pension Coverage in Latin America: Trends and Determinants by Rafael Rofman and Maria Laura Oliveri, June 2012 1216 Cash for Work in Sierra Leone: A Case Study on the Design and Implementation of a Safety Net in Response to a Crisis by Colin Andrews, Mirey Ovadiya, Christophe Ribes Ros and Quentin Wodon, November 2012 1215 Public Employment Services, and Activation Policies by Arvo Kuddo, May 2012 1214 Private Pension Systems: Cross-Country Investment Performance by Alberto R. Musalem and Ricardo Pasquini, May 2012 1213 Global Pension Systems and Their Reform: Worldwide Drivers, Trends, and Challenges by Robert Holzmann, May 2012 1212 Towards Smarter Worker Protection Systems: Improving Labor Regulations and Social Insurance Systems while Creating (Good) Jobs by David A. Robalino, Michael Weber, Arvo Kuddo, Friederike Rother, Aleksandra Posarac and Kwabena Otoo 1211 International Patterns of Pension Provision II: A Worldwide Overview of Facts and Figures by Montserrat Pallares-Miralles, Carolina Romero and Edward Whitehouse, June 2012 1210 Climate-Responsive Social Protection by Anne T. Kuriakose, Rasmus Heltberg, William Wiseman, Cecilia Costella, Rachel Cipryk and Sabine Cornelius, March 2012 1209 Social Protection in Low Income Countries and Fragile Situations: Challenges and Future Directions by Colin Andrews, Maitreyi Das, John Elder, Mirey Ovadiya and Giuseppe Zampaglione, March 2012 1208 World Bank Support for Pensions and Social Security by Mark Dorfman and Robert Palacios, March 2012 1207 Labor Markets in Middle and Low Income Countries: Trends and Implications for Social Protection and Labor Policies by Yoonyoung Cho, David Margolis, David Newhouse and David Robalino, March 2012 1206 Rules, Roles and Controls: Governance in Social Protection with an Application to Social Assistance by Lucy Bassett, Sara Giannozzi, Lucian Pop and Dena Ringold, March 2012 1205 Crisis Response in Social Protection by Federica Marzo and Hideki Mori, March 2012 1204 Improving Access to Jobs and Earnings Opportunities: The Role of Activation and Graduation Policies in Developing Countries by Rita Almeida, Juliana Arbelaez, Maddalena Honorati, Arvo Kuddo, Tanja Lohmann, Mirey Ovadiya, Lucian Pop, Maria Laura Sanchez Puerta and Michael Weber, March 2012 1203 Productive Role of Safety Nets by Harold Alderman and Ruslan Yemtsov, March 2012 1202 Building Social Protection and Labor Systems: Concepts and Operational Implications by David A. Robalino, Laura Rawlings and Ian Walker, March 2012 1201 MicroDeterminants of Informal Employment in the Middle East and North Africa Region by Diego F. Angel-Urdinola and Kimie Tanabe, January 2012 1120 Employment Generation in Rural Africa: Mid-term Results from an Experimental Evaluation of the Youth Opportunities Program in Northern Uganda by Christopher Blattman, Nathan Fiala and Sebastian Martinez, December 2011 1119 Measuring Governance and Service Delivery in Safety Net Programs by Gloria M. Rubio, September 2011 1118 Assessing Safety Net Readiness in Response to Food Price Volatility by Margaret Grosh, Colin Andrews, Rodrigo Quintana, Claudia Rodriguez-Alas, September 2011 1117 Social Safety Nets in Fragile States: A Community-Based School Feeding Program in Togo, August 2011 (also available in French) by Colin Andrews, Elena Galliano, Carolyn Turk and Giuseppe Zampaglione, August 2011 1116 Strengthening Governance of Social Safety Nets in East Asia by Sara Giannozzi and Asmeen Khan, August 2011 1115 International Portability of Health-Cost Coverage: Concepts and Experience by Martin Werding and Stuart McLennan, July 2011 1114 Liberia’s Cash For Work Temporary Employment Project: Responding to Crisis in Low Income, Fragile Countries by Colin Andrews, Prospère Backiny-Yetna, Emily Garin, Emily Weedon, Quentin Wodon and Giuseppe Zampaglione, July 2011 1113 Employability and Productivity among Older Workers: A Policy Framework and Evidence from Latin America by Edmundo Murrugarra, July 2011 1112 Cash Transfers, Children and the Crisis: Protecting Current and Future Investments by Ariel Fiszbein, Dena Ringold, Santhosh Srinivasan, June 2011 1111 Severance Pay Programs around the World: History, Rationale, Status, and Reforms by Robert Holzmann, Yann Pouget, Milan Vodopivec and Michael Weber, May 2011 1110 Portability of Pension, Health, and other Social Benefits: Facts, Concepts, Issues by Robert Holzmann and Johannes Koettl, May 2011 1109 Disability and Poverty in Developing Countries: A Snapshot from the World Health Survey by Sophie Mitra, Aleksandra Posarac and Brandon Vick, April 2011 1108 Advancing Adult Learning in Eastern Europe and Central Asia by Christian Bodewig and Sarojini Hirshleifer, April 2011 1107 Results Readiness in Social Protection & Labor Operations by Laura Rawlings, Maddalena Honorati, Gloria Rubio and Julie Van Domelen, February 2011 1106 Results Readiness in Social Protection & Labor Operations: Technical Guidance Notes for Social Service Delivery Projects by Julie Van Domelen, February 2011 1105 Results Readiness in Social Protection & Labor Operations: Technical Guidance Notes for Social Safety Nets Task Teams by Gloria Rubio, February 2011 1104 Results Readiness in Social Protection & Labor Operations: Technical Guidance Notes for Social Funds Task Teams by Julie Van Domelen, February 2011 1103 Results Readiness in Social Protection & Labor Operations: Technical Guidance Notes for Labor Markets Task Teams by Maddalena Honorati, February 2011 1102 Natural Disasters: What is the Role for Social Safety Nets? by Larissa Pelham, Edward Clay and Tim Braunholz, February 2011 1101 North-South Knowledge Sharing on Incentive-based Conditional Cash Transfer Programs by Lawrence Aber and Laura B. Rawlings, January 2011 To view Social Protection Discussion papers published prior to 2011, please visit www.worldbank.org/sp Abstract Using a simple framework, this paper discusses the underlying reason of the variation of threshold level in developed countries, from the least generous 20% to around 60% of median wage, with an average of 35%. The generosity of minimum guarantee social assistance programs is deeply rooted in social values and principles that further underpin the policy objectives. Many OECD countries set their policy targets for minimum living standard programs beyond basic needs and aim to guarantee a minimum socially acceptable level for a decent living. Thresholds are also refined to reflect the differences in family size and demographic structure, difference in regional cost of living and changes in prices and local wages. In some countries the thresholds show some regional variation due to local discretionary powers of sub-national authorities to set the threshold depending on the co-financing mechanisms. These lessons are valuable for China as the Chinese government has made efforts to standardize the implementation and management of its own minimum income guarantee (Di Bao) programs. The policy recommendations for China include accelerating the convergence of localized approaches, raising the administrative level for setting thresholds to higher level, defining the roles of central and local governments in financing and management, and establishing a transparent budgetary management system to transfer and allocate social assistance funds. About this series... Social Protection & Labor Discussion Papers are published to communicate the results of The World Bank’s work to the development community with the least possible delay. The typescript manuscript of this paper therefore has not been prepared in accordance with the procedures appropriate to formally edited texts. The findings, interpretations, and conclusions expressed herein are those of the author(s), and do not necessarily reflect the views of the International Bank for Reconstruction and Development / The World Bank and its affiliated organizations, or those of the Executive Directors of The World Bank or the governments they represent. The World Bank does not guarantee the accuracy of the data included in this work. The author(s) attest(s) that the paper represents original work. It fully references and describes all relevant prior work on the same subject. For more information, please contact the Social Protection Advisory Service, The World Bank, 1818 H Street, N.W., Room G7-803, Washington, DC 20433 USA. Telephone: (202) 458-5267, Fax: (202) 614-0471, E-mail: socialprotection@worldbank.org or visit us on-line at www.worldbank.org/sp. © 2013 International Bank for Reconstruction and Development / The World Bank Printed on recycled paper