Filhos e renda familiar: uma aplicação do efeito quantílico de tratamento

Abstract

Research on the consequences of childbearing on income and labor-supply is complicated by the fact th at fertility is an endogenous variable, which may bias the conventional estimator. In this study we use the Quantile Treatment Effect (QTE) approach recently proposed by Abadie, Angrist e Imbens (1998 and 2002) to estimate that effect more adequately. It uses the parental preferences for a mixed sibling sex composition to construct an instrumental variable estimator, allowing it to vary along the quantiles of the family income distribution. The QTE estimator can be used to determine how an intervention affects the distribution of the response variable for individuals whose treatment status is changed by a binary instrument. It also has the advantage of collapsing to the conventional quantile regression when the treatment is exogenous. The QTE estimates with PNAD data for Brazil reveal that there are important differences, between the several quantiles, in the effect in family income of childbearing. It is a reduction that varies from14% to 18% throughout the quantiles, and is larger in those at the extremes of the income distribution, in particular, for those that correspond to the families with smaller income.