Working Papers/Submitted Papers

If I Were a Boy: Gender Stereotypes at School.


This paper aims at quantifying gender stereotypes at school. How does self-perceptions and gendered- oriented-expectations of others influence children’s self -perception of their capabilities? Using PISA data with a specific focus on mathematics,I first construct an index of self-confidence using multiple correspondence analysis. In a second step, I regress this index over PISA scores and other control variables and apply a decomposition method in order to see which part of the difference between boys and girls self-confidence is explained by characteristics and which part is not. Despite the high number of explanatory variables, the decomposition highlights a substantial part in differences between girls and boys that can be attributed to stereotypes and not to characteristics.

 A Bayesian Look at American Academic Wages: Where are the Superstars?

With Michel Lubrano (AMSE-GREQAM), (Forthcoming: Journal of Economic Inequality)


The paper investigates academic wage formation and develops tools in order to detect the presence of superstars. We model wage distributions using a hybrid mixture formed by a lognormal distribution for regular wages and a Pareto distributions for higher wages, using a Bayesian approach, particularly well adapted for inference in hybrid mixtures. The presence of superstars is detected by studying the shape of the Pareto tail. Contrary to usual expectations, we did found some evidence of superstars, but only when recruiting assistant professors. When climbing up the wage ladder, superstars disappear. For full professors, we found a phenomenon of wage compression as if there were an upper bound, which is just the contrary of a superstar phenomenon. Moreover, a dynamic analysis shows that many recruited superstars did not fulfil the university expectations as either they were not promoted or left for lower ranked universities.

Education Politics, Schooling Choice and Public School Quality: Does Income Polarisation Matter?

With Paolo Melindi-Ghidi (BETA, Université de Strasbourg & GREQAM) and Michel Lubrano (AMSE-GREQAM). (R&R European Journal of Political Economy)


Do communities with the same level of inequality but a different level of income polarisation perform differently in terms of public schooling? To answer this question, we extend the theoretical model of schooling choice and voting developed by de la Croix and Doepke (2009), introducing a more general income distribution characterised by a three-member mixture instead of a single uniform distribution. We show that not only income inequality, but also income polarisation, matters in explaining disparities in public education spending across communities. Public schooling is an important issue for the middle class, which is more inclined to pay higher taxes in return for better public schools. Contrastingly, poorer households may be less concerned about public education, while rich parents are more willing to opt out from the public system, sending their children to private schools. Using micro-data covering 724 school districts of California and introducing a new measure of income polarisation, we find that school quality in low-income districts depends mainly on income polarisation, while in richer districts it depends mainly on income inequality.