Investments in human capital in childhood are generally believed to be critical for adult well-being. Children who have higher educational attainment are more productive as adults, earn higher wages, and have better health status than children with less education. In country after country, governments have sought to devise effective policies to increase school enrollment. In this context, cash transfer programs have expanded dramatically in many developing countries, especially in Latin America. These are often conditional cash transfer programs, in which eligible households are given transfers conditional on compliance with certain requirements. In most countries, households are required to send school-aged children to school and to take younger children for regular visits to health centers, where they receive nutritional supplements and growth monitoring. The best known of these programs is Oportunidades (formerly Progresa) in Mexico, although similar programs have also been implemented in a number of other Latin American countries.
This paper evaluates the impact of a cash transfer program in Ecuador, the Bono de Desarrollo Humano (BDH), on school enrollment. This is a large program—in 2004, the BDH budget was approximately 0.7 percent of gross domestic product (GDP). Eligibility for BDH transfers is determined by a proxy means test known as the Beneficiary Identification and Selection [End Page 43] System, or Selben. In theory, the 40 percent of households with the lowest Selben score are eligible for monthly transfers.1
The analysis in this paper is divided into two parts. The first part assesses the overall impact of the BDH on school enrollment. During an expansion of the BDH coverage, program administrators undertook an experimental evaluation of the program’s impact. A sample of households that had not previously received transfers was selected and randomly divided into treatment and control groups. Data on both groups were collected at baseline, before households started to receive BDH transfers, and follow-up, approximately one and a half years later. We use this experimental study design to show that BDH transfers significantly increased school enrollment. These results complement earlier findings of positive effects of cash transfer programs on enrollment in Mexico, Colombia, Nicaragua, and Honduras.2
The second part of the paper provides evidence on the relative importance of transfers and conditions (or perceived conditions) in explaining the positive BDH program effects on enrollment. Conditions attached to transfers introduce a kink in the budget constraint, and their effect depends on whether this compels households to behave differently than they would have done otherwise.3
The BDH was originally modeled on Oportunidades. Program administrators intended to condition transfers on school enrollment and attendance, among other things. Local elected leaders (the heads of the Juntas Parroquiales) were encouraged to hold town-hall-style meetings in which the BDH was presented as a compact between the state and beneficiaries—the state agreed to transfer resources to poor households, and these households, in turn, agreed to send their children to school. The BDH program also briefly aired a series of radio and television spots that explicitly linked transfers with school enrollment, and some BDH administrators appear to have stressed the enrollment requirements when they signed up households for transfers. In practice, however, the BDH did not monitor the schooling condition because of administrative constraints, and it thus did not penalize households whose children were out of school. Nevertheless, the BDH information campaign had an [End Page 44] effect: in the follow-up survey used in this paper, approximately a quarter of respondents stated that they believed that sending children to school was a BDH program requirement.
We compare the impact of the program among conditioned households (that is, those who told survey enumerators that school enrollment was a BDH requirement) and unconditioned households (those who told enumerators that there was no enrollment requirement attached to transfers). Our estimates show that program effects on enrollment are only significant among conditioned households. Because exposure to the information campaign was not assigned randomly, these comparisons are not experimental. However, the effects we estimate are insensitive to adding a large number of controls, trimming the data, and sweeping out fixed differences between conditioned and unconditioned households. We therefore argue that the larger...