Safety Nets Social Protection & Labor Technical Note June 2014 | Number 6 Highlight Effective Targeting for the Poor The choice of targeting methods for a particular and Vulnerable1 social program depends Phillippe Leite on the objective of Targeting social programs can improve effectiveness of poverty alleviation policies. the program and the Allocating benefits to the poor and vulnerable can help reduce chronic poverty, ensure particular country’s uninterrupted access to social services, and address social exclusion and disparities. circumstances and Recent international experience has shown that effective targeting is achievable and that a combination of targeting methods can be most effective in achieving poverty experience. Literature reduction and promoting shared growth. suggests that three methods —geographical Targeting must be cost-effective and be useable by policymakers and communities in a way that can be used to generate lists of potential beneficiaries. Moreover, procedures targeting, community- must be put in place to ensure that beneficiary selection is objective, transparent, and based targeting, consistent across geographical areas. Several methods and approaches are known as and (proxy) means- means tests; proxy means tests; community-based targeting; geographic targeting and self-targeting. One way to achieve consistency across different approaches is by making testing—can be used proxy means testing (PMT) an integral part of targeting in combination with geographic in combination to targeting, community-based targeting (CBT), or both. In most situations, PMT methods identify households can provide a common definition of poverty and the basis for data collection efforts, that is sometimes lacking is only CBT methods are used. Coordinating and institutionalizing that are chronically procedures, establishing a uniform understanding of poverty and vulnerability, poor or food-insecure, establishing common processes across country assistance programs, and creating the or that are vulnerable basis for creating a database of potential beneficiaries (a social registry) can reduce fragmentation, improve targeting effectiveness, and generate substantial efficiency gains. to these conditions.2 These methods must be Reasons for Targeting Social Benefits3 flexible enough to be From an ethical or human rights perspective, equity is better achieved with policies that used across safety net treat every member of a society equally. Such a universalist approach may look simple in programs of harmonized principle, but its practical application is far from easy.4 Literature shows that in general interventions. the neediest, most marginalized and most discriminated-against groups may have difficulty accessing universally provided programs and may face barriers, such as lack on information, transaction costs, etc. that would exclude them. 1 Synthesis note based on author’s work and on del Ninno, C. and Mills, B. Eds. 2014. Safety Nets in Africa: Effective Mechanisms to Reach the Poor and Most Vulnerable. Washington, DC: World Bank. 2 Del Ninno, C., and Mills, B; Coady, D., Grosh, M. and Hoddinott, J. 2004. Targeting of Transfers in Developing Countries: Review of Lessons and Experience (2004). Washington, DC: World Bank. 3 See more reasons in Leite, Yemtsov and Brixi. 2012. Improving Lives of Poor and Vulnerable Population: Targeting and Universal Approaches (mimeo) prepared as background paper for the Thailand Economic Monitor December 2012. 4 See examples in Guven and Leite (2014). The Slippery Slope: Explaining the Challenges and Effectiveness of Social Pensions to Fight Poverty in Sub-Saharan Africa. Technical Note June 2014 | Number 6 Moreover, in a world of high levels of poverty and limited testing, geographic targeting, and self-selection based on a resources there may be a need to ration or limit access to work requirement are all associated with increasing the share programs. Poverty imposes costs on communities in several of benefits directed to the two lowest expenditure quintiles. ways. Vulnerability to shocks forces families with limited Further, proxy means testing, community-based selection of resources to use negative coping strategies that may cause individuals, and demographic targeting of children achieve further negative consequences on long term well-being.5 For good results on average, but vary considerably depending on example, to maintain minimum level consumption, the poor implementation strategy. Demographic targeting of the elderly might be forced to sell their assets or forego investment in their and self-selection based on consumption do not appear to be children’s human capital by cutting back on care, nutrition or effective in targeting people in the lowest expenditure quintiles. education. As a consequence, delays in children’s development In addition, rapid-assessment methods have also been used to among the poor and vulnerable may be irreversible and can identify vulnerable households affected by shocks. The World undermine their future prospects, locking families in poverty.6 Food Program (WFP), for example, has used indicators like Since resources are limited too, targeting specific groups or food consumption scores (FCS), dietary diversity indexes, and populations such as the poor and vulnerable, who should be food frequency indexes to identify food-insecure households.9 given priority access to limited public resources, would be important to improve program effectiveness. Poverty and food insecurity indicators are correlated, but methods for targeting poor and food-insecure households are Universal vs. Poverty-Targeted Policies often implemented independently. Poverty analysis focusing on household budget surveys (HBS) and food-insecurity work Universal access and targeted approaches are policies are not employing rapid assessments may identify the same group of in contradiction. What matters most for governments and households depending on the time or year. A key challenge for policy makers is not addressing an intellectual question— policymakers has been to find which methods best use data for whether to poverty-targeted programs or not—rather how targeting in specific country contexts. to design and implement a mix of coordinated and well harmonized both universal and poverty-targeted policies Weakness and strengths of five common methods of targeting and programs, which make up a social protection and labor methods, including means testing, proxy means testing system which covers all populations. (PMT), community-based targeting (CBT), geographic targeting, and self-targeting are summarized in Table 1. Social protection and labor policies and programs are universal because they are meant for all population. Targeting some policies and programs to the poor and vulnerable will make them more effective and efficient.7 By allocating benefits to selected population groups, governments can ensure that the poor have access to social programs and services. Brazil, Mexico, the Philippines and Indonesia, among other countries, use targeting to improve the effectiveness of programs in combating poverty and inequality and developing human capital. Therefore, in the words of the Economic Commission for Latin America and the Caribbean, “targeting, or selectivity, 5 is not a social policy as such, but rather a method, which, A third concept treats vulnerability as a categorical status, such as elderly, disabled, widows, children and orphans, because they need special treatment if properly applied, enhances the effectiveness of universal and attention. Thus, children have needs that are markedly different from social programs. The fact that the principle of universality adults and require age-appropriate care; the elderly have unique needs due not only to the increased incidence of illness and disability, but also to the multiple is translated into priority access to basic protection for the and complex interactions of other types of physical and social consequences poor does not mean that selectivity, as a tool, represents the of aging; and orphans can be too young to manage an inheritance or may not even know they have access to an inheritance. social policy. Targeting does not work against the principle of 6 The need to help the poor and destitute imposes costs on communities and universality.”8 strains neighborhood and family support networks, often at times when everyone is struggling. Poverty may also result in socially harmful behavior (crime, begging, etc.) that imposes costs on all members of a society. Common Targeting Methods 7 Hanlon, J., Hulme, D. and Barrientos, A. 2010. Just Give Money to the Poor: The Development Revolution from the Global South. Sterling, Virginia: Kumarian Press. Well-established methods exist for identifying potential safety 8 Economic Commission for Latin America and the Caribbean. 2001. Equity, net program beneficiaries. Grosh et al. (2008) and Coady, Development and Citizenship, pages 78-79. Santiago, Chile: United Nations. Grosh and Hoddinott (2004) present the pros and cons 9 FCS gives a score for each household that is a linear combination of weights of targeting methods such as means testing, proxy means based on the nutrient density of food groups and the level of consumption within groups. The dietary diversity index represents the number of different testing, community-based targeting, geographical targeting, foods or food groups that households have consumer over a predetermined demographic targeting, and self-selection targeting. Coady, time period, varying from one to 30 days (usually seven days). The food frequency index considers the frequency of consumption of foods or food Grosh and Hoddinott (2004), after analyzing numerous social groups over a defined period (usually seven days). protection programs, find that interventions that use means Social Protection & Labor | World Bank Group Table 1: Strengths and Weaknesses of Five Targeting Methods Method Description Strength Weakness Means tests Actual consumption or income is Very accurate with good income or Expensive to collect income compared to eligibility threshold consumption data or consumption data for all potential beneficiaries Proxy means Consumption is proxied though Can accurately and cost effectively Does not address the impact of test readily observable and verifiable target the chronic poor short-term shocks variables and compared to eligibility threshold Community- Groups of community leaders Incorporates local knowledge and Vulnerable to elite capture and based targeting and members determine is responsive to short-term shocks. eligibility decisions can lack household eligibility Can generate community support. transparency Geographic Targets by location, including all Easy to implement and transparent. Does not account for differences targeting residents within a location Can rapidly target in response to in household well-being in area natural disasters and other large covariate shocks Self-targeting Benefits and transaction costs are Easy to implement and low Stigma and lack of program set so that only needy households implementation cost knowledge may discourage enroll participation Source: del Ninno and Mills Eds. (2014) Multiple Methods (b) Proxy-Means-Test-Plus The literature suggests that the use of a combination of Proxy-means-test-plus (PMTplus) is simply a variation of PMT targeting methods within a single program can produce that incorporates the impact that major shocks (e.g., droughts, better targeting results than reliance on a single method floods, incapacitation or death of adult family member) may (Grosh et al. 2008; Coady et al. 2004; Handa et al. 2012, and have on households in the selection to identify short-term poor Leite 2014). For example, Mexico’s Oportunidades program households. and Kenya’s Orphans and Vulnerable Children program combine geographical targeting and PMT; Brazil’s Bolsa PMTplus can be seen as a method that allows small shock- Familia uses geographic targeting and means testing; and related adjustments to the PMT to reduce inclusion errors at Tanzania uses geographic targeting combined with CBT and times of shocks. By applying the value of the estimated impact PMT. In a well-designed process, multiple methods can bring of the shock on welfare to correct the cutoff point for the PMT, complementary strengths in order to minimize errors of households in need can be identified quickly, and geographic exclusion and inclusion. targeting using vulnerability and or shock maps makes it possible to identify households that are or are likely to be (a) CBT and PMT affected by the shocks.11 Some countries are exploring a combination of community 10 Cambodia’s secondary scholarship program applies such a model: (i) geographic targeting: each school in priority areas receives 30 or 45 slots; (ii) based targeting and proxy means testing. Countries as students fill out forms in class, teachers read them aloud and classmates verify Tanzania, Ghana, Niger, Cambodia,10 Indonesia and that information is correct; (iii) selection formula is based on typical statistical analysis, but simple variables and integers; (iv) a school committee scores and Afghanistan are exploring this combination of methods, where ranks forms by hand and awards scholarships. the community is engaged at the project preparation stage to 11 Yemen used this strategy in the aftermath of last year’s political turmoil, provide a pre-list of potential beneficiaries of the program, where the Social Fund for Development (SFD): (i) redefined its priorities by emphasizing cash-for-work programs; (ii) selected areas of intervention based while the PMT is applied over the pre-identified households on current maps of malnutrition and vulnerability; (iii) enlisted the help of either to select beneficiaries or to verify them in order to communities to identify food-insecure households; and (iv) added a means test to the then-current targeting strategy that combined community-based reduce inclusion errors. targeting, self-selection, and a proxy. As a result, SFD was able to increase the number of people benefitting from cash-for-work programs in a short time. Technical Note June 2014 | Number 6 Figure 1: Targeting Food Insecure Populations Affected by Shocks Based on the PMTplus Model After the Shock Notes: B represents chronically Areas Affected by the Shock food-insecure households and regular Areas not Affected by beneficiaries of safety nets. A12 Total the Shock represents households vulnerable Food Secure Food Insecure to food insecurity due to the shock, that is, those who need immediate Before the Shock assistance because of the shock. A22 represents households that are not Food Secure A20 A21 A22 chronically poor but are food insecure A as a result of the shock, so they may be A10 A11 A12 in need of a shorter-term intervention. A10 and A20 represent households Food poverty line living in areas that are not affected by Food Insecure B B10 B11 B12 the shock so are not eligible for the shock-related intervention Source: Leite (2014) at del Ninno and Mills Eds. (2014) The Social Registry: Potential Beneficiaries’ chronic poor. Then scalability will depend on the occurrence Database of a shock. If short-term shocks are a dominant driver of safety net needs, countries may wish to focus on methods One important element of an efficient system that is the need that enhance the identification of vulnerable households for good household and individual-level data on potential and target short-term needs. Conversely, if chronic poverty beneficiaries: that is, a social registry of all potential program generates the greatest need for safety net programs, then the beneficiaries. Different programs can have different eligibility method(s) should focus on the identification of chronically criteria based on their own objective. Regardless of these poor households. different criteria, however, programs should be designed to benefit the neediest (poor and vulnerable). Harmonizing the In terms of a country’s capacity to implement methods, data and information of this population across programs will enable human resources are key constraints. The method(s) selected governments to identify potential beneficiaries of multiple social must be compatible with existing human resources available and assistance programs. In this regard, a social registry would be a training possibilities. Identification of the population of interest major step forward in improving the efficiency of social policies. always requires access to household or individual-level datasets, regardless of the method used, and the information can be used The information investment required to identify program later to inform program monitoring and evaluation. Constraints beneficiaries could be significantly reduced over time if systematic on implementation of more complex targeting methods rather than program-by-program information on household well- become greater as social safety net systems move towards more being is collected as part of a country’s investment in its safety net. structured, quantitative methods. A well-designed social registry can improve coordination across safety net programs immensely, no matter which targeting method Political economy may play a role in the choice of appropriate is chosen. In addition, a social registry can help disseminating targeting methods. Implementation of more quantitative important information, lower beneficiary transaction costs, targeting methods may reduce direct elite capture of assistance generate savings, and improve overall efficiency. A functional resources, but might also lead elites to seek compensation social registry also allows assistance programs to be rapidly through other means. In summary, no matter the targeting scaled up in response to shocks. The cost of collecting data for method or combination of methods, implementation and targeting, particularly for means testing and proxy means testing, coordination are key. A good targeting system should ensure is dramatically lowered if the information is already in the registry. transparency and consistency above all. Personnel must Thus, a social registry can remove major impediments to the ensure clear, consistent application of common criteria in use of these methods and should be considered as an important collecting and processing data, with little political interference component in deciding where to invest time and resources in and manipulation by public officials and beneficiaries alike. order to improve targeting procedures and performance. Good systems are designed to achieve maximum inclusion of the poor and reduce leakage to the non-poor (but allow for Conclusion the inclusion of the near-poor). Technical or administrative targeting errors are understandable, but fraud, corruption and Appropriate targeting methods will depend on country-specific manipulation should be combated. needs and capacity. A social program should always start at the 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 original had problem with text extraction. pdftotext Unable to extract text.