Chapter 22

Conditional Cash Transfer Programs: A “Magic Bullet” for Reducing Poverty?
Michelle Adato and John Hoddinott

I

n 1997, the government of Mexico introduced a conditional cash transfer (CCT) program called Programa de Educación, Salud, y Alimentación (PROGRESA) to provide assistance to about 300,000 extremely poor households. The essential premise of a CCT program is a cash transfer to households conditioned on their participation in health, nutrition, and education services. Ten years later, PROGRESA, now Oportunidades, covers more than 5 million households in all 31 Mexican states. Approximately 20 countries have adopted a pilot or full-scale CCT program, and another 20 countries have expressed interest in starting one. Most current programs are in Latin America, but others can be found in Asia, Africa, and the Caribbean, and interest is increasing among African countries struggling with extreme poverty and low human capital. CCT programs are increasingly perceived as being “a magic bullet in development.” Are they? This chapter presents a review of the rationale for and operation of CCT programs, an assessment of their effectiveness, and a discussion of key issues facing countries considering these interventions or reforming existing programs.

How Do CCT Programs Work? CCT programs have the following characteristics: they are targeted to poor households, and the cash transfers are usually paid to mothers. Some programs also include transfers such as nutritional supplements or school supplies for children. Cash transfers may be made as a lump sum or determined based on the number

300

mIchelle adato and john hoddInott

of children, with the amount varying by the children’s ages and genders. In some countries, higher transfer amounts are paid for girls’ school attendance and for secondary school attendance. In return for these transfers, recipients commit to undertake certain actions, such as enrolling children in school and maintaining adequate attendance levels, attending pre- and postnatal health care appointments, and seeing that preschool children receive vaccinations, growth monitoring, and regular checkups. Some programs require women to attend regular health and nutrition training workshops. Some provide resources that improve the supply and quality of the schools and health care facilities used by beneficiaries. Through these requirements CCT programs aim to reduce current poverty while also seeking to improve human capital formation and, in so doing, help prevent the intergenerational transmission of poverty.
Impacts of CCT Programs Rigorous evaluations—often built into the programs themselves—show that many, but not all, CCT programs have been successful in improving human capital outcomes. In Mexico, PROGRESA increased enrollment in secondary school by 6 percentage points for boys and 9 percentage points for girls. For girls—who often drop out before secondary school—those making the transition to secondary school increased by 15 percentage points. Children in the program also entered school at an earlier age and repeated fewer grades. PROGRESA had relatively little impact, however, on school attendance rates, on achievement on standardized tests, or in bringing dropouts back to school. These objectives must thus be pursued through design improvements or complementary programs. CCT programs in Colombia, Mexico, and Turkey all improved secondary school enrollment but had little impact on primary school enrollment rates because these were already high. Where preprogram enrollment rates are extremely low, the effects of CCT programs can be very high: in Cambodia, for example, secondary school enrollment increased by 30 percentage points and attendance by 43 points. In Bangladesh, where 3 million children are still not enrolled in primary school, a small CCT program targeting the hardest-to-reach children (including street children) increased primary school enrollment by 9 percentage points, though this occurred only in schools where grants were also provided to improve school quality. In Nicaragua, where primary school enrollment was also low, the CCT program increased overall enrollment by 13 percentage points, enrollment of children from the very poorest households by 25 points, and regular primary school attendance by 20 points. Two years after households stopped receiving benefits, however, enrollment had dropped by 12.5 percentage points, but this was still 8 points higher than before the program, implying some sustainability of impact. This experience points

condItIonal cash transfer programs: a “magIc bullet”?

301

to the important but little understood question of how to maintain the effects of CCT programs after households graduate from a program or after it ends. CCT programs have also had significant impacts on health and nutrition. In Mexico, health visits increased by 18 percent in PROGRESA localities, and illnesses among PROGRESA children 0–5 years old were reduced by 12 percent. Young children in Honduras increased their use of health services by 15–21 percentage points, though there, as in Brazil, no effects on children’s illness rates were found. Some of the largest increases were found in the regular monitoring of children’s growth in CCT programs in rural Colombia, Honduras, Mexico, and Nicaragua. A number of CCT programs are also associated with increased child height, which is an important measure of long-term nutritional status. Stunting was reduced in Colombia by 7 percentage points, in Mexico by 10 points, and in Nicaragua by 5.5 points. Although the exact mechanism that triggers improvement is not known for certain, it may result from one or several program characteristics, such as the provision of higher incomes that permit increased expenditure on food, growth monitoring and information about nutrition and childcare, or nutritional supplements. In both Mexico and Nicaragua, for example, calorie intake increased, as did the consumption of fruits, vegetables, meat, and dairy products. In Honduras, no positive nutritional impacts were found because of implementation problems, while in Brazil, the program was initially associated with a slightly reduced weight gain, but this phenomenon was subsequently reversed. According to anecdotal evidence, mothers may have kept their children underweight under the mistaken belief that they would lose their benefits if the children gained too much. This misunderstanding points to the importance of well-functioning mechanisms to ensure clear and regular communication between the program and beneficiaries so that program conditionalities do not create perverse incentives. In Turkey, insufficient or incorrect information about the program also reduced impacts. Some programs address micronutrient deficiencies. In Mexico, program beneficiaries had anemia rates substantially lower than nonparticipants, though the rates remained high. In Nicaragua, although mothers reported receiving the iron supplement offered, anemia rates were not affected, in part because they did not give the supplement to children, believing it was bad for their stomachs and teeth. Both cases point to the continuing challenge of addressing nutritional deficiencies and the need for multidimensional approaches rather than cash transfers or supplements alone. CCT programs have a sharp gender focus. They have been successful in significantly increasing rates of school enrollment for girls, who have historically faced discrimination because educating them is not considered as important as educating boys. Research in Mexico and Nicaragua has found that CCT programs are associated with improved attitudes toward educating girls, as well as a heightened profile for women more generally. Although there has been concern and some evidence

302

mIchelle adato and john hoddInott

that women’s program responsibilities can lead to conflicts with men, in both countries there is more evidence that the program’s infusion of financial resources has reduced intrahousehold tensions. Where CCT programs organize collective activities for beneficiaries, such as meetings, committee participation, and workshops, women report increases in their knowledge, social awareness, and self-confidence. Nevertheless, not all CCT programs provide these opportunities, and those that do not overlook potential for increasing women’s status. Research in eastern Turkey found that sociocultural biases against schooling for girls were more powerful than cash incentives, indicating the need for complementary approaches to overcome these constraints. Even in this region, however, the program provides opportunities for women to spend time out of their homes and to engage with institutions such as banks and government offices. CCT programs tend to be administratively centralized because their complexity requires standardization; hence, they offer fewer avenues for community participation than do many other development interventions. Nevertheless, the programs still affect communities—positively or negatively—depending on their design and implementation. Programs in Brazil, Colombia, Honduras, and Mexico have all found ways to integrate varying types of local input, from a beneficiary feedback system in Mexico to local input into targeting in Brazil, to school-based parents’ organizations and quality improvement teams for the health services in Honduras to mothers’ assemblies in Colombia. Although data-based centralized targeting has generally been successful in reaching the poor and avoiding political manipulation at the local level, it has also frequently bred discontent in communities when people have not understood the targeting criteria, have perceived it as unfair, or have not had access to a functioning appeals mechanism. Exploring country-specific options for participation could lead to programs that are even more effective in achieving their primary goals while increasing collective and individual empowerment.
Development and Implementation Issues
Are CCT Programs Too Expensive?

The concern that governments in poor countries cannot afford CCT programs should be considered within the context of the large sums spent by many governments on programs directed to the nonpoor. Energy subsidies, for example, are typically highly regressive and often more costly than CCT programs: Egypt spent 8 percent of its gross domestic product (GDP) on energy subsidies in 2004, and Indonesia spent 5 percent in 2005. Bailouts of insolvent contributory pension funds are another example. The expansion of Brazil’s well-targeted CCT program, Bolsa Família, to cover the bottom quintile of the population would cost about 0.4 percent of its GDP, whereas the Brazilian government now spends nearly 10 times that

condItIonal cash transfer programs: a “magIc bullet”?

303

amount covering the deficit in the main federal pension programs, which deliver more than 50 percent of their benefits to the richest quintile. These are not isolated examples: many other countries spend considerable amounts of money on industry subsidies and military expenditures. In some very poor countries, particularly in Sub-Saharan Africa, donors and nongovernmental organizations (NGOs) have stepped in as partners with governments considering or implementing CCT programs, viewing them as potentially cost-effective approaches to increasing human capital—for example, by protecting children in households affected by AIDS. Even if a country can afford a CCT program, it is sometimes argued that, relative to other types of social safety nets, such programs are expensive to operate. For example, in its first year of operation (1998), PROGRESA spent US$1.34 on administrative costs for every dollar spent on transfers to beneficiaries. Statistics like these shape a common perception that CCT programs are too expensive. Closer scrutiny, however, reveals a different picture. First, as with any program, fixed establishment costs, such as those of buying computers, identifying beneficiaries, and so on, are comparatively high. By PROGRESA’s third full year of operation, its administrative costs had fallen to only 5 cents for every dollar spent on transfers. Second, many administrative costs—such as identifying beneficiaries, establishing mechanisms for delivering the benefit, and monitoring and evaluating the program —are common to all social protection programs. Further, some of these costs are incurred to improve the program’s effectiveness. Reducing expenditures on targeting, for example, might reduce administrative costs, but if targeting performance is severely weakened as a result, the cost savings will prove counterproductive. An important cost issue is whether conditionality significantly increases program costs. The existing evidence is mixed, largely because the intensity of monitoring conditionalities is a choice made by program designers and implementers. In the case of the CCT program in Honduras, monitoring accounted for about 9 percent of administrative costs.
Is Conditionality Necessary?

An important question being debated as new countries consider cash transfer programs is whether to impose conditionalities. Are conditional cash transfers in fact better than unconditional ones when it comes to achieving objectives, and if so, for what objectives and under what conditions? Three broad arguments support conditionality. The first relates to the externalities associated with certain types of human capital investments. For example, when making decisions about their children’s care—say decisions about girls’ schooling—parents may not take into account the benefits that society derives from educating girls, and, as a result, they underinvest in girls’ schooling relative to optimal levels from a societal perspective. Conditionality can be an effective means of increasing these investments. Second,

304

mIchelle adato and john hoddInott

sociocultural biases against schooling may be imposed by more powerful groups (for example, men) on the less powerful (for example, their daughters), and conditionality provides state legitimation of social change. Third, conditionality may overcome the possible stigma associated with welfare payments if conditions are seen as part of a social contract between beneficiaries and the state. Finally, conditionality may be required for reasons of political economy. Politicians and policymakers are often evaluated by means of performance indicators, such as changes in school enrollment or use of health clinics, and the impacts of CCT programs provide a basis for sustaining public support. Conditionality has also increased the credibility of programs where, historically, the public has often been suspicious of anti-poverty efforts that were deemed ineffectual. Conditionality also raises several concerns: first, there must be reasonable access to schools and clinics. Second, governments and NGOs must be able to handle the costs and administrative requirements or adapt them to local circumstances. These two concerns are particularly relevant for African countries establishing CCT programs; in Kenya, for example, an evaluation comparing the impacts of conditional and unconditional transfer programs is under way. Third, if poor people’s preferences differ sufficiently from the conditions placed on their behavior by the government, the restrictions that conditionality imposes may actually reduce total welfare gains. Another way to look at this is to consider that conditional transfers can be perceived as demeaning, implying that the poor do not know what is good for them and need to be told by the government. Fourth, poverty, culture, social exclusion, discrimination, and other historical processes may prevent people from participating in activities regardless of the benefits, so the people in most need can actually be punished rather than helped by conditionality. Balancing these arguments and comparing the benefits of programs against the costs associated with conditionality is important but difficult; the sparse evidence to date makes rigorous comparisons even harder. In the case of PROGRESA, monitoring conditionality represented approximately 2 percent of total program costs. Monitoring attendance had no effect on primary school enrollment but appears to have had a major impact in increasing the likelihood that students would continue school after completing their primary education, with conditionality increasing enrollment in the first grade of lower secondary school by approximately 20 percentage points.
Are CCT Programs Sufficient as a Poverty Reduction Strategy?

CCT programs as currently designed are important parts of a poverty reduction strategy that aims to improve the health, nutrition, and education of young children in the short term and their income-earning potential in the future, ultimately reducing the likelihood they will remain poor as adults. Other complementary

condItIonal cash transfer programs: a “magIc bullet”?

305

strategies are needed, however, for people at other stages of the life cycle. Mexico’s Oportunidades is partially addressing this issue by offering (1) benefits throughout high school; (2) a cash incentive for high school graduation conditioned on beneficiaries’ investment in higher education, a productive activity, health insurance, housing, or continued savings; and (3) a cash transfer for beneficiaries 70 years of age or older. Of course, poverty reduction also requires other approaches to promote economic development and job creation.
Would All Developing Countries Benefit from a CCT Program? With a proven track record, CCT programs are a powerful approach not only to reducing poverty but also to improving various educational, health-related, nutritional, and other welfare-related outcomes. That said, not all CCT programs have functioned as well as their designers had hoped. CCT programs are not for every country, and no two countries should necessarily adopt identical programs. In assessing whether a CCT program is appropriate, four main questions need to be answered:

1. What are the current levels of specific human capital outcomes? If the current enrollment rates of primary school children are nearly 100 percent, it makes little sense to condition transfers on primary school enrollment. However, if enrollment rates are only 70 percent, there will be greater scope for a CCT program —although the extent of this scope will be revealed only through further disaggregation of the enrollment rates. For example, are the rates uniform across rural regions or for boys and girls? If pronounced regional, gender, or ethnic differences are present, a CCT program targeted to those lagging groups would be more effective than a countrywide program. 2. Why are specific human capital outcomes currently too low? Do they reflect an income constraint, such as parents’ needing the income that children bring in when not in school? Or are schools nonexistent, too far away, or considered unsafe for children to travel to or attend? Whether low school enrollment rates or poor nutrition outcomes reflect constraints at the household level or the absence of adequate service provision needs to be determined prior to initiating a program. CCT programs are ideal where the supply of supporting services is good but underutilized; they are much less effective when supporting services are limited. In such cases, improvements in the supply of schools, clinics, and so on should precede or accompany the launch of a program. In some countries CCT programs have provided a strong impetus for improving services.

306

mIchelle adato and john hoddInott

3. Is there high-level political support for a CCT program? By design, CCT programs require coordination across different sectors, most notably social welfare, education, and health. This implies the need for interministerial coordination, which is difficult to achieve. An influential political champion of the program is needed to ensure that this coordination occurs. In Mexico, for example, strong support from inside the Ministry of Finance was an important factor in PROGRESA’s success. 4. What administrative resources are available? Is the necessary intersectoral coordination feasible, particularly when both transfers and supply-side interventions are envisaged? The level of complexity of program design should reflect administrative capacity. CCT programs—while not a magic bullet—are worth serious consideration as part of an integrated poverty alleviation strategy.
For Further Reading
Das, J., Q. Do, and B. Özler. Reassessing conditional cash transfer programs. World Bank Research Observer 20, no. 1 (2005): 57–80. Levy, S. Progress against poverty: Sustaining Mexico’s Progresa–Oportunidades program. Washington, DC: Brookings Institution Press, 2006. Maluccio, J., and R. Flores. Impact evaluation of a conditional cash transfer program: The Nicaraguan Red de Protecciòn Social. IFPRI Research Report 141. Washington, DC: International Food Policy Research Institute, 2005. Skoufias, E. PROGRESA and its impacts on the welfare of rural households in Mexico. IFPRI Research Report 139. Washington, DC: International Food Policy Research Institute, 2005.