INTERNATIONALBANK FOR WORLD BANK R E T C N O E N STRUCTION PM AND DEVELO December 2005 No. 83 A regular series of notes highlighting recent lessons emerging from the operational and analytical program of the World Bank`s Latin America and Caribbean Region VOICE AND ACCOUNTABILITY IN TRANSFER PROGRAMS IN LATIN AMERICA AND THE CARIBBEAN Katherine Bain, Franka Braun and Jaime Saavedra Background effective in the long run by offering feedback to service providers of potential obstacles and long-term benefits, A two-day workshop, organized by the World Banks beyond the immediate and obvious benefits that these Poverty Reduction and Economic Management (PREM) programs have on human development indicators. and Human Development (HD) departments, was held in June 2005 in the Dominican Republic with the objective to Accountability in Transfer Programs explore whether voice and accountability can improve the effectiveness of Conditional Transfer Programs. Partici- A conceptual framework for accountability in transfer pants included government and civil society representa- programs, developed byAlain de Janvry from the University tives from ten countries in the region, as well as World of California at Berkeley, was presented at the workshop Bank staff. (See Figure 1). The main features include: In recent years, many countries in Latin America and the (i) Main actors of the accountability framework: Caribbean region (LAC) have begun reforming their government, service providers (central government, mu- transfer systems to make them more effective, inclusive nicipalities, or contracted NGO), and beneficiaries of and less burdensome. The central idea of Transfer programs. Programs (TPs) is to condition monetary payments to the (ii) Objectives of accountability: To assure that cash poor for investment in human capital by its beneficiaries. transfer programs are used for poverty reduction in the Some programs, such as the Mexican Oportunidades short and long run; to avoid exclusion errors (people that Program, target individuals or households, and imply qualify who are not covered by program) and inclusion giving cash to poor parents under the condition that they errors (elite capture, corruption, rent-seeking, clientelism). send their children to school and for health visits. Other (iii) Types of accountability: Upstream accountability TPs, such as the Brazil Community Driven Development of service providers towards the central government; (CDD) project, target communities through the transfer of downstream accountability of service providers towards funds for approved community investments directly to the beneficiaries and potential beneficiaries. bank account of the respective associations. (iv) Requirements for accountability: Transparency (updated information on program criteria, implementation Initial evaluations suggest that TPs can be an effective and arrangement, and beneficiaries should be made accessible feasible instrument in both reducing current poverty, as to the central government and interested parties); enforce- well as improving the future of children and communities ment (appeals systems and sanctions for non-compliance through increased investment in human capital. However, or poor performance) these same evaluations suggest that TPs are costly and, (v) Channels to exercise accountability: Referring though "well-targeted", the targeting mechanisms them- to the World Development Report 2004, de Janvry pre- selves can cause tensions within communities. There is a sented the "long-route" of accountability and the "short strong hypothesis that improved accountability through route" of accountability. Via the long route, accountability increased transparency and monitoring mechanisms by can be exercised through local or national elections of citizens might provide better targeting and make TPs more politicians who define budget allocations, tax collection and 1 redistribution systems. Through the "short route," citizens Presented with these facts, participants discussed various demand accountability from service providers in terms of possibilities for LAC countries to improve quantity and implementation, targeting and performance of programs. quality of social spending: Improve tax collection to make the total envelope for The workshop results social spending larger: improving transparency around the revenue process and black listing tax The two-day program was designed around three dimen- evaders can help to improve the collection process. sions: (i) the share of social expenditure in public spending Enhance transparency and accountability: there is in LAC, (ii) the structure and quality of social expenditures, evidence that countries with good governance and and (iii) the design of transfer programs, especially the accountability and low levels of corruption spend evaluation and accountability mechanisms, which were more on social policy. explored in plenary sessions and case studies from Mexico, Improve quality of social spending: governments can, Ecuador, Brazil andArgentina. on the one hand, improve targeting of transfer programs and, on the other, restructure social spend- 1. Public social expenditure as a share of total public ing in order to provide universal rights granted by expenditure in LAC constitutions. The workshop opened with a presentation of the results of Make social programs, including social assistance, a research program analyzing the effectiveness of public pro-cyclical: the creation of stabilization funds help transfer programs in Latin America1. In LAC, social poor households to better absorb adverse economic expenditures, as a share of GDP, rose from 8 percent shocks, but are often difficult to finance. (1970) to 13 percent (2002), with variations from below 2 percent (Nicaragua) to over 20 percent (Uruguay). This 2. Structure of social expenditures: Social Security vs. represents 40 percent (1970) as opposed to 48 percent Social Assistance (2002) of total public expenditure. The increase in expendi- In a second step, participants were introduced to the main ture for social security has been mainly explained by the types of social programs. Transfer programs are divided increase of expenditures for pensions from 2 percent of into: (i) social security that includes risk pooling in the form GDP (1970) to 5.5 percent of GDP (2002), which has of contributions and are linked to the formal labor market primarily benefited the richer 20 percent of the population. (pensions, insurances), and (ii) social assistance, including This, in addition to the fact that transfer programs have conditional transfer programs, school feeding programs been counter ­cyclical, and thus have been cut in times of and others. Whereas evidence derived from the above economic crises, represents a double burden on the mentioned study suggests that redistribution through social poorest two quintiles of the population, who are most security is regressive, targeting of social assistance pro- affected by economic shocks. grams is progressive and tends to reduce inequality. In addition, as opposed to pensions or insurances, cash 2 transfer programs are pro-cyclical and their coverage is mechanisms, such as the costs and the need for organiza- high. However, monthly per capita amounts received tional capacity of beneficiaries, were discussed. However, through most cash transfer programs are relatively low presenters confirmed that social accountability has a (between USD 7 and USD 36) and thus have a limited proven positive impact on poverty reduction, but that few impact on poverty. studies measuring the impact of social accountability on redistribution and equality exist. There is a need, thus, for LAC countries with high social spending (~13.1 percent of further evaluation of these mechanisms. GDP) tend to spend more on social security (10.8 percent of GDP) than on social assistance (2.3 percent of GDP). LAC countries with low social expenditures (~1.6 percent of GDP) spend a larger share on social assistance (1.2 percent of GDP), and a smaller share on social security (0.4 percent of GDP). In countries with high social spending, social security tends to reduce poverty, but not inequality. In light of these facts, the following recommendations and comments emerged: Social assistance programs have an impact on poverty Lessons Learned in the long run when they contribute to the creation of human capital. In plenary sessions and working groups discussing the Reducing social policy to compensatory, palliative four case studies presented at the event (Mexicos measures should be avoided and transfer programs Oportunidades, Ecuadors Bono de Desarrollo Humano, should become a permanent rather than a temporary Brazils Community Driven Development Project (CDD), feature of social policy. In addition and in order to be and Argentinas Jefes y Jefas del Hogar), participants sustainable, transfer programs should be comple- elaborated the following recommendations that should be mented by sound exit strategies, e.g. through linkages taken into account when designing transfer programs: to employment generation programs. Shifting from social security to social assistance or Targeting within social assistance programs with different Apply clear, transparent targeting rules as opposed to target beneficiaries has implications for the political secret proxy-mean formulas, as in Oportunidades. economy, since strong interest groups exist around Secret formulas limit stakeholders right of appeal, traditional social security mechanisms. In addition, third party verification and create inter-personal ten- possibilities to increase tax income and restructuring sions in communities. within social spending are often limited. The political Locate sites for interviews for information collection attitudes about redistribution and equality of opportu- in schools or health centers, away from the Mayors nity versus equality of outcomes vary among coun- office (to decrease clientelism and patronage) and tries leading to a different set of program options away from the home (to allow for witnesses). likely to be politically sustainable. Consider using qualitative information to make target- Social security is an important instrument for redistri- ing formulas fit local realities. This could be done by bution and equality in a society and the question is having a central formula that provides allocations to whether they should or could be replaced by transfer regions or states, and thus, ensure equity. But in programs that target different groups. Reducing addition, authorities could use qualitative information pension payments, for example, could mean that more to make sure that the centrally run program fits local people fall into poverty. realities. This is particularly important for programs The informal sector, the poor and needy, are often not that are urban and rural, and include indigenous and covered, either by social security or transfer pro- non-indigenous groups as beneficiaries. grams. Programs should consider how to include Allow for verification of self-reported income infor- them. mation. This is easier when the identification of eligible beneficiaries is decentralized. 3. Social Accountability in Transfer Programs Define successful targeting not only by reaching Based on the hypothesis that the quality and quantity of intended beneficiaries, but also by achieving accep- social spending, and transfer programs in particular, could tance and comprehension of program goals among be improved, the question of if and how social accountabil- the general population. This is key for the ity might have a positive impact on improved redistribution sustainability and ownership of the program, and through transfer programs was at the centerpiece of the affects the underlying program design. conference. The prerequisites for social accountability 3 Information and Oversight decentralized programs, it seems easier to achieve Provide clear information to providers and beneficia- enforcement and transparency, whereas centralized ries about program rules. For many transfer pro- programs might help to prevent clientelism and elite grams, there exists confusion about who/how ben- capture at the local level. eficiaries are selected. Public perceptions about the purpose and usefulness Post lists of beneficiaries in public places, if there are of social programs influence whether they receive no confidentiality constraints, providing households political support. Therefore, regular evaluations of with benchmark comparisons in claiming their rights. programs, including qualitative methodologies to as- Use civil society organizations (CSOs) to facilitate sess them along with effective communication pro- the flow of information, as they have the advantage grams, are important. Accountability mechanisms that of access to local information. CSOs can use this facilitate continuous evaluation can be catalysts for information to inform and educate providers and improved public policies. Such mechanisms should be government officials about how to better manage well linked to broader monitoring and tracking systems programs and to expose deviations from norms and within the government. corruption. Provide access to citizen oversight committees to Enforcement relevant and timely information (beneficiary lists, Provide appeals mechanisms. In Brazil, appeals can be selection criteria, compliance with conditions by directed to the federal government through telephone beneficiaries, etc.). hotlines; the principal difficulty has been weak follow- up to these complaints. Appeals can also be directed to the local council, although the council then acts as judge and jury. Access to courts or conflict resolution mechanisms, like the ombudsman, are generally lack- ing. Give purchasing power to beneficiaries to help them hold local providers accountable. This includes use of development coupons and targeted vouchers with choice among competing providers. Require beneficiary endorsement before payment is made to providers. Report Cards can be used to evaluate providers as done with Ecuadors Bono de Desarrollo Humano. Conclusion Revise incentive systems for public functionaries to Voice and accountability can strengthen sustainability of provide better information and responsiveness to conditional transfer programs through improved oversight users. over program implementation as well as the use of funds Complement traditional quantitative evaluations with and transparent targeting of beneficiaries. Also, account- qualitative evaluations (such as the one carried out by ability mechanisms for cash transfer programs, especially the International Food Policy Research Institute- when linked to monitoring mechanisms and tracking IFPRI around the Oportunidades program in systems within the State, can help to correct the design of Mexico). Qualitative information can provide impor- cash transfer programs since providers are informed early tant insights into the perceptions and behavior of on about potential obstacles in program implementation. beneficiaries, obstacles to effective implementation, the impact of the program on social relations, the Notes effectiveness of the program in terms of the overall 1 well-being (social capital) of the community. Since Lindert, K.; Skoufias, E.; Shapiro, J. (2005): How social capital is considered to be an important effectively do Public Transfers in Latin America Redistrib- strategy for poverty reduction and inclusive develop- ute Income ?. The World Bank, Washington D.C. ment, impacts at the community level should also be Photos taken from Mexico's Oportunidades program. considered during evaluations. About the Authors With a central provider (as in Oportunidades), there is more downstream accountability, but possibly less Katherine Bain is the Team Leader and Franka Braun is a upstream accountability, whereas a decentralized Junior Professional Associate in the Civil Society Team. provider (Argentina Jefes y Jefas) provides the Jaime Saavedra is the Sector Manager of the Poverty and opportunity for more upstream accountability. In Gender group in the Latin America and Caribbean region. 4