«DIREC TIONS IN DE VELOPMENT Human Development Public Disclosure Authorized The Cash Dividend The Rise of Cash Transfer Programs in Sub-Saharan Africa ...»
Other community targeting in Sub-Saharan Africa is conducted in a more indirect manner. Some programs encourage key community members, such as health workers, to identify potential beneficiaries on a caseby-case basis (Mozambique’s PSA). Others target OVC—children who may not identify themselves because of the stigma associated with HIV/ AIDS—by relying on local organizations that already are in contact with them (Senegal’s CCT for OVC).
Other issues to consider in targeting. In some of Sub-Saharan Africa’s CT programs, a major concern is that beneficiaries’ incomes will “leapfrog” over those of nonbeneficiaries. This issue is especially important when CTs are implemented in areas with a relatively flat income distribution. The size of a poverty-targeted transfer must be chosen carefully to ensure that it meets program goals but does not suddenly make beneficiaries substantially better off than nonbeneficiaries who had a similar standard of living before the CT began. If the program fails to address this possibility, leapfrogging may generate significant social tension (Ellis 2008, as cited in Slater and Farrington 2010).
This issue will not affect every country equally. In an analysis of Ghana and Malawi, leapfrogging was found to be a major concern in Malawi, where the distribution of income was very flat (especially in rural areas), whereas it was not as problematic in Ghana, where the income distribution was more unequal (White and others 2009). In Malawi, only US$9 or US$10 per capita monthly divided the lowest income decile from the sixth income decile, highlighting the potential of transfers to cause significant leapfrogging (Ellis 2008, as cited in Slater and Farrington 2010).
Care must also be taken when poverty targeting is accompanied by categorical targeting. For instance, although Malawi and Zambia target Design and Implementation of Cash Transfers in Sub-Saharan Africa 89 ultrapoor, labor-constrained households in their major CTs, ultrapoor households with available labor may be even worse off than laborconstrained households if employment is unavailable, given that ultrapoor households may need more calories per person than do labor-constrained households (Ellis 2008, as cited in Slater and Farrington 2010). Those issues suggest the importance of using analytical work to drive CT designs.
Some programs in Sub-Saharan Africa use quotas to restrict the number of beneficiaries targeted by geographic location. In Malawi and Zambia, surveys determined that 10 percent of households were ultrapoor and labor constrained. On the basis of this determination, the 10 percent worst-off households that fit the labor-constrained definition were selected as beneficiaries in each given locality. However, even in areas with a relatively equal spatial income distribution, such quotas can be problematic, excluding eligible households in areas with greater than 10 percent eligible households, and including ineligible households in areas with eligible populations of less than 10 percent (White and others 2009).
Also important is understanding who may benefit indirectly from a CT based on the targeting scheme. For instance, old-age pensions in Lesotho and South Africa are known to provide significant support to OVC who live with their grandparents. This support is especially helpful in Lesotho, which is still in the process of trying to put more significant support in place for OVC. Samson (2007) calculated that 65 percent of the pension money in Lesotho is actually used by the elderly to care for children. The significant size of the transfer provides a major boost to household income, and it may benefit children. However, this scheme does not negate the need for programs directed to children, because any pension obviously targets the elderly first and will miss many OVC.
Finally, targeting in Sub-Saharan Africa may also need to deal with local cultural and social traditions, such as targeting of polygamous households. This issue arose in Ethiopia’s PSNP. In that case, the government decided that the best approach was to use a standard procedure in which wives and their children should each be registered as separate households (Devereux, Sabates-Wheeler, and others 2008), although implementation across locations has varied in practice (World Bank 2010a).
Data Collection for Targeting Often household surveys, censuses, administrative data, and birth and death records are out of date or nonexistent in Sub-Saharan African 90 The Cash Dividend countries. Institutional capacity and financial backing for data collection may be weak, and access to some communities may be difficult at best. This lack of data can be a significant obstacle in the targeting process. Beyond leaving knowledge gaps about households’ program eligibility, the lack of data can put CTs in danger of overlooking some vulnerable households entirely. Various programs have confronted such data limitations differently as they attempt to collect data for program purposes. The collection methods often are limited by time and financial constraints that affect data accuracy and, consequently, targeting precision. The constraints also make it difficult to maintain up-to-date information about households that are excluded from programs, another key piece of information for targeting and program implementation.
Even information allowing for broad-scale geographic targeting is limited in some Sub-Saharan African countries. Despite the relative dearth of detailed information by region, general geographic targeting is possible in most cases. What is more, the usefulness of data collection exercises, such as detailed poverty mapping in parts of the country where a potential CT program may function, can provide an impetus for a country to develop and maintain more current data.
Community help in collecting data. Programs in Sub-Saharan Africa have typically relied on ad hoc community-level data collection methods to acquire necessary information for targeting households and individuals.
For instance, community members identify potential beneficiaries of Botswana’s Orphan Care Program, and social workers must then assess each case (BFTU 2007). In lieu of relying on an up-to-date census, Malawi’s SCT has collected necessary data for targeting by using community members’ knowledge of where households reside (Miller, Tsoka, and Reichart 2010). Similarly, Eritrea relies on village health committees to identify potential RBF beneficiaries and invite them to enrollment meetings (Ayala Consulting 2009). In program areas of Kenya’s CT for OVC, local committees identify potential beneficiary households, which are then visited by enumerators, who collect additional information about the households to help determine whether they are eligible for the CT (Government of Kenya 2006).
Malawi’s experience illustrates some of the potential difficulties encountered when using community members to collect data for the program’s targeting system. Community-level knowledge of the presence of households within a given area was often inaccurate, with leaders Design and Implementation of Cash Transfers in Sub-Saharan Africa 91 estimating as much as 43 percent more or 44 percent fewer households than those encountered in a systematic canvassing activity. Estimating household numbers incorrectly affected how many beneficiaries would be included in the program in a given location. Listings were sometimes affected by nonrandom exclusion of households because of their remote or relatively inaccessible location or their lack of community ties and by purposeful inclusion of ghost households (Miller, Tsoka, and Reichert 2010).
Data collection through other agencies or means. Some programs identify potential beneficiaries—and, therefore, collect data about them—as the individuals initiate contact with other official support systems.
Potential beneficiaries are identified through their contact with a local official in health, education, or social services who then helps to enroll the individual in the program. This method of identification occurs in Mozambique’s PSA, where health centers may identify potential beneficiaries and local program officials verify their information. The officials receive small payments as an incentive for recommending or enrolling beneficiaries (Datt and others 1997). Similarly, nongovernmental organizations (NGOs) and other groups that provide support to OVC identify potential beneficiaries for Senegal’s CT for OVC. In some countries, such targeting mechanisms have the tendency to overlook individuals who hide because of the stigma associated with a condition such as HIV. In Senegal’s case, this concern is lessened; OVC should be more easily identified by NGOs, because they are used to identifying and assisting them.
Still other programs (including multiple programs in the uppermiddle-income countries of Mauritius, Namibia, and South Africa) leave the onus of providing data on potential beneficiaries, who are required to submit information to program offices to receive benefits. With this type of design, beneficiaries who are not connected to relevant support systems will be excluded from the CT program. Given that some of those individuals will typically be the ones who most need the program’s assistance, this method of identifying eligible beneficiaries can lead to undercoverage of the eligible population. Awareness campaigns have been important in increasing coverage for many of these programs.
Although data collection using community members and incidental data collection are not a first-best practice in many cases, they may be the most appropriate methods, given program constraints. Over time, CT programs may be able to generate increased support and a rationale for 92 The Cash Dividend regularly collecting microlevel data as programs expand and are concerned with households’ ongoing eligibility and recertification.
Knowledge Gaps in Targeting Systems Information about the execution and effectiveness of targeting systems in CT programs throughout Sub-Saharan Africa is still fairly limited. The difficulties associated with effective targeting suggest that case studies analyzing targeting and its effectiveness—in terms of success in reaching intended beneficiaries, implementation successes and failures, inclusion and exclusion outcomes, and costs—will be helpful to other programs throughout the region. Specific areas where knowledge will be helpful include the following: collecting data in limited financial and human resource capacity settings; targeting individuals who may not be easily identified because of stigma or inaccessibility; improving the implementation of community-based targeting, particularly with respect to effectively training communities in targeting practices and successfully navigating local political and cultural dynamics; and effectively communicating targeting criteria to program beneficiaries and nonbeneficiaries.
Targeting in Sub-Saharan Africa faces a set of particular challenges in data collection in terms of weaknesses in available population data and statistics, weak institutional capacities for collecting and analyzing data, and problems with access and cooperation from populations to be surveyed. Balancing the financial and social costs of targeting with the desire to achieve targeting accuracy can be a delicate issue, which further analysis and experience will be able to inform.
Client Registration: A Key Issue for Cash Transfers A validated registration system needs to be in place to enroll transfer recipients once eligible beneficiaries have been identified for inclusion in a CT. A key issue in low-income countries is how to reliably identify eligible beneficiaries for enrollment purposes. Problems arise where procedures are not in place to correctly identify beneficiaries. Individuals may register for multiple grants (when they are not supposed to do so), or someone may wrongly receive grants in another person’s name. For example, Swaziland has found that its Old Age Grant has more recipients than the number of eligible Swazis, a sign that large-scale fraud is occurring. The government blames this corruption on early implementation errors that allowed multiple proofs to be used to identify an individual in the initial enrollment period (RHVP 2007).
Design and Implementation of Cash Transfers in Sub-Saharan Africa 93 Challenges That Arise When Requiring Official Documentation Requiring beneficiaries to register for CTs using a national identification card or a birth certificate is one way to reduce the incidence of fraud.