Jere R. Behrman: This is a nice paper on an important program. I organize my comments around four topics: some critical aspects of PROGRESA from the point of view of evaluation; the Skoufias-Parker study; questions for future research on PROGRESA; and lessons for future evaluations of similar programs.
Critical Aspects of PROGRESA
PROGRESA is a well-known Mexican rural antipoverty and human resource investment program that was initiated under the Zedillo government. The program has several striking and unusual features that are relevant for social scientists and policy analysts. First, PROGRESA based some of its essential components on the outcomes of social science research in the scholarly literature. Transfers were given to women, for example, because previous research on intrahousehold allocations suggests that income received by mothers has larger associations with human resource investments in children than income received by fathers.
Second, PROGRESA used modern social science tools to guide its decisions. For instance, discriminant analysis was used on census data in the initial stages of identifying target communities and households within those communities, and geographic information systems (GIS) were used to systematize information on the location of schooling and health services relative to the communities.
Third, PROGRESA recognized that baseline data and longitudinal household and service-provider data with treatment and control groups were essential to serious evaluation of the program. The program coordinators implemented the collection of baseline data prior to the introduction of the program (which is always the intent but often not the reality), together with longitudinal follow-up and random assignment to treatment and control communities. [End Page 87]
Fourth, PROGRESA contracted an outside research agency, the International Food Policy Research Institute (IFPRI), to undertake an extensive evaluation of the program. This Washington-based organization is a member of the Consultative Group on International Agricultural Research (CGIAR), and it has a strong history of data analysis and evaluation in developing countries. In addition to IFPRI staff, the evaluation team consisted of myself and three other academic economists, namely, T. Paul Schultz of Yale University, Paul Gertler of the University of California at Berkeley, and Petra Todd of the University of Pennsylvania.
These are important features of PROGRESA. The program and the key individuals behind it, in particular the program's first director, the late José Gómez de León, should be commended for incorporating such elements in their program plans and for giving them sufficient priority that they were carried out reasonably well during the incredibly difficult period of program development, implementation, and rapid expansion, all within an environment characterized by considerable political pressures.
PROGRESA has already served as a model for the development of related programs elsewhere. The hope is that these and other programs will not only build on the substantive aspects of PROGRESA, but also learn from and improve the evaluation process that has been carefully designed and given such high priority by PROGRESA.
The Skoufias-Parker Study
Skoufias headed the IFPRI evaluation team, while Parker first worked for PROGRESA and then joined the IFPRI evaluation team. Together, they bring considerable knowledge of and expertise on PROGRESA to this study. While other complementary works examine aspects of the program's impact on schooling, this is the most satisfactory study available on child time use in both school and work. This is an important topic because investments in schooling are widely thought to shape options over the life cycle and because time spent in labor activities competes with time invested in schooling. The paper blends related aspects of PROGRESA, institutional knowledge, modeling of behavior, use of distinctive data, and estimation. Of particular interest is the impact of the nonlinear incentives created by the PROGRESA cash transfers that are conditional on [End Page 88] schooling and how those incentives differ depending on what household decisionmaking would have implied for the schooling-work choice in the absence of the program. The empirical estimates consider the program's impact on the probability of working in market, domestic, and farm activities; the hours worked in market, domestic, and farm activities; the probability of being enrolled in school; and the hours spent in leisure. Differences are found by age...