«DIREC TIONS IN DE VELOPMENT Human Development Public Disclosure Authorized The Cash Dividend The Rise of Cash Transfer Programs in Sub-Saharan Africa ...»
Although traditional impact evaluations (box 3.6) are important to understand the effectiveness of CCTs and UCTs on specific outcomes, other studies are also important. An example of such a study has been completed for Zambia’s Monze District SCT (Schüring 2010b). This study combines multiple methods, including secondary data, beneficiary and nonbeneficiary surveys, in-depth interviews, community games, and administrative data, to address important issues regarding the use of conditions in Zambia. It covers many soft issues related to conditioning, including beneficiary and public perceptions of conditions, the challenges in applying conditions, the potential exclusionary role of conditions, and the political economy dynamics that ultimately affect the usefulness of conditions. It also addresses beneficiaries’ preferences with respect to altruism and risk aversion, time discounting, and potential information asymmetries, to determine whether those factors played a role in suboptimal human capital investments.
Although this study found that many Zambians supported conditioning transfers, it ultimately concluded that applying conditions was not optimal at that time. The conclusion was driven by the conditions’ potential tendency to exclude the worst-off beneficiaries, the setting’s low implementation capacity, and the conclusion that conditions had a relatively small potential to significantly change beneficiaries’ health and education behaviors. This type of analysis, combined with rigorous quantitative evaluations, will prove useful for other countries that are trying to determine whether and how conditions should be applied in their own programs.
Another area of study that may prove fruitful is the question of which conditions should be applied—and how often conditions should be monitored—in CCT programs. Many of the vanguard CCT programs condition benefits on school attendance, which must be monitored relatively frequently, and Sub-Saharan African CCT programs are following in their footsteps. However, if a program can encourage children to continue Design and Implementation of Cash Transfers in Sub-Saharan Africa 127
To Condition or Not to Condition? Discussion and Additional References Fiszbein and Schady (2009) provide a thorough discussion on when placing conditions on human capital investments may be appropriate for cash transfer programs. They suggest that conditions may be useful when households underinvest in human capital or when adding conditions increases the program’s political acceptability and sustainability.
Households may underinvest in human capital for several major reasons.
Investments may be privately suboptimal as a result of imperfect information, myopic decision making, or incomplete altruism of parents toward their children.
Incomplete altruism may result from inefficient intrahousehold bargaining outcomes or excessively high parental discount rates in the presence of credit market failures. Even if households invest in privately optimal levels of human capital, this level may still be socially suboptimal because of the presence of positive externalities in human capital investment (Fiszbein and Schady 2009).
Despite their benefits, conditional transfers present potential problems. These problems include the possibility that conditions will be too costly for the most destitute households to comply with, in essence excluding them from program benefits, or that the quality of education or health services will be too low to provide significant benefits to households. Other concerns are that schools and health centers will be unable to support increased demand created by the CCTs, and that the CCT program will not have adequate capacity to monitor and enforce conditions. Additional opposition to conditions stems from the belief that the government should not withhold benefits from citizens who are entitled to a state transfer (Fiszbein and Schady 2009).
Those interested in learning more about the rationale for, implementation of, and known impacts of CCTs throughout the world are referred to Fiszbein and Schady (2009). Fiszbein and Schady also highlight how CCTs should work in tandem with other social protection programs. Readers interested in the issues involved in deciding whether to condition CTs are also referred to Schüring (2010a). Other key references on CCTs include Das, Do, and Özler (2005), who discuss the tension between equity and efficiency objectives in CCTs as they balance redistributing resources with increasing investments in human capital, and Rawlings and Rubio (2005), who discuss programs and evaluation results for some of the earliest Latin American CCTs.
128 The Cash Dividend
Impact Evaluations Although many programs in Sub-Saharan Africa have tried to determine the impact of CTs on key outcomes, some of the evaluations have lacked the credible counterfactual necessary to determine the program’s causal effect on beneficiaries. To have an impact evaluation in the truest sense, a program must be able to identify treatment and control groups. Experimental evaluations randomly assign beneficiaries or households into these groups, often exploiting the program’s need to conduct a phased rollout (because of fiscal, capacity, or political constraints) to ensure that potential beneficiaries are not unfairly excluded from the program.
A quasi-experimental or nonexperimental method does not work from intentionally randomized assignment of treatment and control groups. Instead, these evaluations use econometric methods, including matching, difference-in-difference regressions, and instrumental variables, to try to isolate program impacts. For more information on program evaluations, see Ravallion (1999, 2005).
in school by monitoring grade progression or final exam grades, rather than by monitoring daily attendance, this knowledge would be useful.
Similarly, conditions may be most important at key junctures, such as the transition to the second cycle of primary school or to junior secondary school. If a program has difficulty applying conditions that require frequent monitoring, it could instead learn what value can be added by monitoring only the groups with the highest risk for suboptimal behaviors (for example, school desertion during major transitions). This analysis should consider which conditions have the greatest effect on beneficiary behavior and what unintended consequences monitoring these conditions might have, such as the potential for increased automatic promotion in the case of a grade progression condition.
Finally, additional case studies and knowledge sharing will offer useful information on how Sub-Saharan African CCTs have helped encourage the improvement of condition-related service delivery. Such findings will be particularly relevant for low-income countries trying to implement CCTs in areas that face severe capacity constraints.
Graduation from Cash Transfer Programs The concept of beneficiaries’ graduation from CT programs has received considerable attention in Sub-Saharan Africa. In some cases, beneficiaries Design and Implementation of Cash Transfers in Sub-Saharan Africa 129 graduate only after they have also received support in income-generating opportunities. Beneficiaries may be connected to a public works or microfinance program, or they may have completed a vocational training course and received a lump sum for capital investment. Other beneficiaries may graduate after receiving important psychosocial support that should help them in the future. Households that graduate from a CT program may continue to benefit from other social support at a level commensurate with their need and ability. Rwanda’s VUP and Ethiopia’s PSNP, for example, have strategies that encourage capable households to graduate from their CT components into public works or other incomegenerating activities.
Graduation is not a requirement of social assistance. In Sub-Saharan Africa, where skipped-generation households are increasingly common, program graduation is not always feasible as a short- or even mediumterm goal. Governments that have clear vision about vulnerability among different groups in the country and that fully understand the purpose of their CTs will maintain realistic expectations regarding graduation. In Rwanda, which has looked into social protection issues in great depth, the government recognizes that some households, such as those composed only of the elderly or of the elderly and young children, should not be expected to quickly graduate from its state-sponsored CT program. In some cases, graduation is not expected to occur. Either the transfers or some other form of social protection will be needed indefinitely to sustain individuals who cannot maintain their own livelihood. The status of households is closely monitored in Rwanda’s VUP, with targeting reviewed every six months at the local level, thereby ensuring that households are correctly categorized with respect to social protection issues (Republic of Rwanda 2009).
Ethiopia’s PSNP has struggled to encourage graduation when possible, while still covering beneficiaries who should remain in the program.
Initially, some households were graduated after acquiring significant assets, even if the assets were purchased on credit (Devereux and others 2006). The practice pushed beneficiaries from the program before they were able to survive without the transfers. This dilemma highlighted the program’s need to clarify its role in development (for those able to work) and in welfare (for PSNP-DS beneficiaries, who were unable to participate in public works). Over time, expectations have become more realistic about the time required for households to be able to graduate.
Increasingly, the government understands that graduation depends not just on the success of the CT program, but also on the program’s ability to link households to other support that will help increase their asset 130 The Cash Dividend base, mitigate the effect of shocks, and link them into well-functioning markets (World Bank 2010a).
Monitoring and Accountability Monitoring plays an important role in all CTs, providing officials with information on how funds are spent, how well the program is carrying out its duties, and whether and how many beneficiaries are being reached.
The Management Information System: A Building Block to Improve Program Credibility and Enhance Impact Establishing a strong management information system (MIS) early in a CT’s life is critical to the program’s success and its potential to scale up.
Programs that use manual information systems or other relatively inefficient systems are unable to grow in the way that a program with an appropriate electronic system can. For instance, Kenya’s CT for OVC was slowed down by its MIS for a time because the system was not designed to handle data for a program as large as Kenya’s became. Its MIS eventually began to create implementation bottlenecks that had to be addressed.
Many of the small CTs in Sub-Saharan Africa have monitoring systems with at least some manual components. For instance, at payment distributions of the Lesotho CGP, a payment coupon and receipt are stamped in a Child Grant coupon book, which identifies the household through a unique number (Blank 2008). Similarly, most documentation of Zambia’s pilot SCTs has been filed manually. However, a beneficiary database and payment registrar are maintained on computers. This information is transported to and from districts and headquarters on compact discs (Ministry of Community Development and Social Services 2008). Other programs rely on computer databases that may or may not be able to communicate with other monitoring systems.
In an ideal world, programs will establish and properly implement a comprehensive monitoring system that is wholly assimilated into program processes. However, given limited capacity and resources, Africa’s CT programs will often begin and work with limited monitoring capacity.