Your browser does not support JavaScript, You should enable JavaScript to use this website.

Work in progress


Not just a drop in the bucket: the spillover effects of affirmative action.

ABSTRACT: Encouraging low-SES students to apply more to selective colleges is key to bridging the social gap in college attendance that persists at all levels of academic achievements. This paper focuses on affirmative action, a popular policy to address inequalities in higher education, and study the impact it can have on low-income students' application behavior.

The affirmative action program I study was introduced by a prestigious, albeit small, French college. It created a separate admission track for students enrolled in partner high schools, targeting schools with a large proportion of low-SES students.  This program was the first of its kind to be introduced in France.

Ex-ante, one may be quite sceptical about its ability to be a game-changer. Indeed, a model of optimal college application behavior would predict that the introduction of affirmative action would only increase applications and enrollment in the college that started it, at the detriment of other colleges.

However, I present a model in which the overall impact of the affirmative action becomes positive if I assume that low-income students initially under-estimate the returns to a college application. If the program includes an information component which corrects this under-estimation, it can also boost applications and enrollment at other selective universities.

Combining several administrative datasets, my empirical analysis shows that a narrow focus on the college that introduced the affirmative action program finds a significant increase in low-income students' enrollment, but that this increase is small in magnitude. However, widening the analysis to include applications and enrollment at other selective colleges finds that the affirmative action program had long-lasting and positive spill-over effects. Students enrolled in high schools that had access to the program apply more to other selective colleges, and, as a result, enroll more.   This effect is particularly strong among non-high-achievers, who represent the large majority of low-income students but are not typically targeted by affirmative action programs.

New version coming soon.


----------------------------------------
Boosting patience for poverty reduction: a field experiment in Malawi.

Is it possible to develop patience in parents? If so, how does that translate to investments in children? I run a large randomized-control-trial with 2,413 mothers in Malawi to answer those questions. In collaboration with a Malawi NGO, I designed a set of trainings to increase the participants' future-orientation, through a mixture of visualization and goal-setting exercises.

Status:  Intervention and endline data collection completed. 

---------------------------------------- 
Parent-bias (with Guilherme Lichand).

ABSTRACT: Lots of attention has been devoted to present-bias, particularly within Development Economics, as a leading explanation for why parents often fail to invest in children’s human capital. A different source of preference reversals has been, however, not investigated to date: asymmetric geometric discounting – when subjects discount their future consumption more than that of their children. Such time preferences imply that parents’ generosity towards their children increases with the time horizon of the decision and leads to reversals away from planned children’s allocation every period.

This paper is the first to document this time-increasing altruism, through a lab-in-the-field experiment  involving real consumption decisions in rural Malawi. We find that a large share of parents are asymmetric geometric discounters. This leads to large reallocations away from planned children's consumption towards parents’ consumption. Those parent-biased reversals are of the same order of magnitude as the preference reversals induced by present-bias over a 30-day horizon.

When offered commitment against within-household reallocation, parent-biased subjects demand it significantly more than other subjects, including involving children in future decision-making. Our findings provide a new explanation for under-investments in children and inform a range of new instruments to address it -- from school meals to earmarked savings accounts.


Status: Field work completed. Working Paper coming soon.