> Slides

Working Papers



Policy Reports



Work In Progress


  • "Explaining Aspiration Failures: Evidence from French Middle School Students" (with Nina Guyon)

This paper provides evidence on the reasons why low-SES students have more modest educational aspirations than their high-SES classmates who have the same academic performance. We use lab-in-the-field experiments with about 6,000 students from 59 French middle schools to test differences in knowledge of the existing academic tracks and in academic self-esteem, as well as stereotype susceptibility and susceptibility to peers. We find that aspiration failures can come from the fact that low-SES students lack knowledge on higher education, that they perceive themselves less performant than their high-SES counterparts, and that they tend to conform to their peers.

  • "Can Additional Managerial Capital Make Microcredit Work? Evidence from a Randomized Experiment in Morocco" (with Bruno Crépon, Dean Karlan and William Pariente)

This paper provides evidence on the impact of an intervention providing managerial capital to small entrepreneurs in Morocco. Among 1,600 firms which benefitted from access to microloans ex-ante, 700 were randomly selected to receive business training sessions (marketing, accountability, networking, etc., for a total of 40 hours more than in the control group). We find that small enterprises in urban areas did not benefit from this intervention, while agricultural cooperatives benefitted in terms of sales, profits, and personal income. The return to business training may thus be larger in contexts where initial levels of education and training are limited.

  • "Managerial Capital and Entrepreneurship: Evidence from a Randomized Experiment with Young Disadvantaged People" (with Bruno Crépon, Esther Duflo, William Pariente and Juliette Seban)

Chronic unemployment is a considerable problem in France, particularly in the poor neighborhoods. Across the so-called "Sensitive Urban Zones", 36.2 percent of men and 40.8 precent of women aged 15-24 years are unemployed. While self-employment may be a positive alternative, entrepreneurship is low in France. This paper tests the impact of a national program directed at supporting business creation for young people between ages 18 and 30 from disadvantaged neighborhoods. The program study involved 1500 individuals randomly assigned to the business training. We find no impact on business creation, nor on employment and income, suggesting that providing managerial capital and business information is not enough to encourage self-employment.

  • “Mindset Change and Employment: Evidence from a Randomized Experiment with Young Disadvantaged People" (with Yann Algan, Bruno Crépon and William Parienté)

The unemployment rate in France is nearly 10 percent, but among young people aged 15-24 this rate is more than doubled. Although many programs exist to provide support labor market participation, in particular  those who have dropped out of school and live in disadvantaged areas, the best ways of supporting youth in creating or finding a job are not clear. Will young people benefit more from general information on the job market and business creation, or development of the autonomy and decision-making skills? We test a national program in which 452 randomly selected young people (out of 902 candidates) explore their motivations, strengths, and weaknesses, and to develop their initiative and autonomy. We find that the program reduced business creation from 8% down to 3%, but increased employment contracts from 31% to 38%. Personal income increased by 22%. Also, the beneficiaries depend less on social and interpersonal transfers, show higher sense of self-efficacy, higher trust in others, and higher levels of wellbeing. Mindset approaches to support labor market participation may thus be a promising avenue for public action.

  • "The Impact of Violence at School on Educational Performances" (with Yann Algan and Nina Guyon)

This paper explores the causal impact of bullying and victimization at school on school outcomes. It uses a field experiment aimed at reducing the prevalence and intensity of violence between students in middle schools. 80 middle schools in 13 French regions took part of the study, half of which were randomly assigned to an anti-bullying program. The program was effective at reducing bullying and victimization, which provides a powerful first-stage to analyse the causal impact of violence on school outcomes. Results expected in 2016. 

  • “Toward an Equal Start: Closing the Early Learning Gap for Roma Children in Eastern Europe: Evidence from a Randomized Evaluation in Bulgaria“ (with Paul Gertler and Joost de Laat)

Since early learning and socialization provide a critical foundation for all children, this project provides policy relevant information on how to most cost-effectively address the low pre-school participation of poor Roma children in Bulgaria. We test different interventions designed to raise the expected benefit and/or lower the costs of early education : 1) the information component is intended to help parents better understand the returns to preschool education and increase their perception of expected benefits and trust in the school to treat their children well and fairly ; 2) the free access component is a natural way to lower the monetary costs of early education ; 3) finally, the financial incentive component increases the direct utility from having a child at school since it provides an additional disposable income to the household. Since the financial transfer is conditional on attendance, it can help parents to overcome a potential strong preference to raise the young children at home, or their procrastination problems facing a present cost of education for a future benefit. 240 Roma settlements over the whole country were assigned to 8 conditions, one control and seven combinations of these three componants. Results expected in 2016

  • “The Impact of a Large-Scale Mindset Intervention on School Outcomes and Dropouts: Experimental Evidence from France”

This experiment provides evidence on the role of mindset in school behaviour, outcomes and dropouts, using student level data from 97 middle schools all over the country. Randomization took place at the cohort-school level in September 2014. The intervention started in 2014-2015 and consists in 12 sessions aimed at developing the “growth” mindset from grade 6 to grade 9. In total, 25,000 students participate in the study, half in the treatment group and half in the control group. The impact evaluation is planned for four years from the beginning to the end of middle school. Final results expected in 2018.

  • “The Cost of Colonial Empire for French Taxpayers and the Benefit for African Taxpayers” (with Denis Cogneau and Sandrine Mesple-Somps)

This project aims at assessing the importance of public transfers from France to its colonies both for French and for African taxpayers: was colonization costly for France? To what extend did French taxpayers contribute to colonies’ development? To answer these questions, we collected original public finance data on all French colonies. The objective is to expand the analysis that exists for French West Africa (Huillery 2014) to the other colonies: French Equatorial Africa, Indochina, Madagascar, Algeria, Tunisia and Morocco. Results expected in 2016. 

  • “Colonial Leaders, State Capacity, and Long-Term Development: Evidence from French Colonization in West Africa” (with Quoc-Anh Do and Jean-Louis Keen)

This paper examines the role of colonial leaders in determining state formation and development paths in West Africa. We first explore the interaction between the quality of colonial political leaders, the contemporaneous hostility of the population towards colonial political leaders, and the resulting state capacity (taxation and public investments) in colonial times. Then, we examine the long term effects of state capacity in colonial times on current political attitudes, economic and political inequalities, and current intrastate conflicts, using characteristics of the colonial leaders as instruments for past-state capacity. We are currently collecting personal data on 809 colonial leaders at the district level between 1890 and 1920 in the colonial archives in Aix-en-Provence. Results expected in 2016.