Comments on the research proposal “School Attendance, Labor Supply and Formality. An Impact Evaluation of PANES” (Team leader: Veronica Amarante) Nguyen Viet Cuong
Summary The research aims to evaluate the impacts of a cash transfer program (called Ingreso Ciudadano of PANES) on child labor, child school attendance, household income and labor: Regression discontinuity. Difference-in-differences. The research questions are well defined and very policy relevant: Inform policy makers on the impact of an important antipoverty program. Be helpful for the program improvement.
Program description The program selection should be discussed more detailed. It is not clear what criteria and the threshold are used to select the beneficiaries. Is income or score/index (estimated from probit model) used? In addition, it is not very necessary to discuss other components of PANES.
Data The date of administrative records of PANES is not mentioned in the proposal. Which data are used as the baseline? If the administrative data are combined with the follow-up surveys for difference-in-differences estimation, they should be consistent in terms of questionnaires, interview instruments and sampling design.
Methodology Economic theory on impact of transfers on income and labor? Transfers increase investment => higher long-run impact. Transfer can reduce working incentive => lower long- run/short-run impact.
Methodology (cont.) Regression discontinuity: Endogeneity of the transfer program. Endogeneity of score? Difference-in-differences: Attrition problem. Estimation of the impact of the program at the threshold like the regression discontinuity. Other methods: Matching. Fixed-effect regression using panel data.