Income and Bargaining Effects on Education and Health in Brazil

Publication Details

Journal of Development Economics, March 2012, vol.94, iss.2, pp.242-253. Available From:

Link to Source
Vladimir Ponczek
Institutional affiliations
None specified
Grant-holding institution
None specified
Latin America and the Caribbean
Education, Health Nutrition and Population, Social Protection
Girl's Education, Pensions & Social Insurance
Girl's Education, Pensions & Social Insurance
Equity Focus
Evaluation design
Difference-in Difference (DID)
Journal Article


This paper assesses the effects of a Brazilian pension system reform aimed at improving access to old-age pension plans in rural areas. Passed in 1991, the reform made three main modifications to the former old-age benefit system: (a) eligibility was extended to household members other than the head of the household, (b) the minimum benefit payment was set to one minimum monthly salary, and (c) the minimum age for eligibility was reduced to 60 years for men (from 65) and to 55 years for women (from 60). Higher household incomes could lead to increased investments in younger household members, notably in health and education. The present study investigates this hypothesis by estimating the indirect effects of the reform on education and health outcomes of children aged 6–14 years old living with one or various pensioners.
The author uses a difference-in-differences approach to conduct the impact assessment. The specification includes a variable indicating whether the observed child lives with an eligible individual, a binary variable indicating the individual's gender, a dummy variable indicating the period of data collection (that is, before or after the reform), and a vector of household and personal characteristics, such as age, number of family members and number of children. The data originate from the Pesquisa Nacional por Amostra de Domicilios (PNAD) database, an annual household survey gathering information on the national population’s socioeconomic conditions, notably data on school attendance, literacy, labour participation, retirement and income sources. For schooling outcomes, the researcher uses PNAD data for the years 1988, 1989 and 1990 (before the reform), and 1992, 1993 and 1995 (after the reform). For health outcomes, the author uses data for the years 1986 and 1998. The analysis measures intention-to-treat estimates and includes approximately 100,000 observations.

Main findings

Overall, the findings suggest that the reform had no effect on the health and educational outcomes assessed in the analysis. However, disaggregated results show significant effects on various outcomes. Specifically, children living with an eligible man were 4.8 percent more likely to attend school and 7.1 percent more likely to be literate than children living with a non-eligible man. In contrast, the presence of an eligible woman in the household appears to have no effect on any outcome. When disaggregating the estimates by child's gender, the analysis reveals that only girls living with an eligible man were more likely to be literate. Consistent with these results, the effects on attendance for both girls and boys living with an eligible woman lack statistical significance.
To understand the uneven effects observed between children living with eligible men and those living with eligible women, the author investigates whether literacy rate differentials among children could be the result of differences in the bargaining power of men and women within Brazilian households. To do so, the researcher estimates whether child literacy was affected by the family-head status of eligible women. Child literacy improved only when eligible women were head of household, whereas the impact of eligibility on literacy remained similar, independent of the family status of eligible men. This finding suggests that the family status of women affects the allocation of benefits toward educational expenses. Finally, the study finds no significant impact (aggregated or disaggregated) on either of the two health outcomes assessed (the probability of being ill and the probability of seeking health care services).

Scroll to Top