Housing and discrimination in economics : an empirical approach using Big Data and natural experiments.

Summary The first chapter documents a key parameter for understanding the housing market: the elasticity of housing supply in French urban areas. We show that this elasticity can be understood in two ways by considering intensive and extensive housing supply. Thanks to a large amount of new data collected and an original estimation strategy, this first chapter estimates and decomposes the two elasticities. The second chapter is devoted to the possibilities offered by Big Data to study the French rental housing market. By exploiting online data from December 2015 to June 2017 and comparing these data to traditional administrative data, we show that the internet provides data to accurately track local housing markets. The third chapter focuses on the discrimination of women in politics. It exploits a natural experiment, the 2015 French departmental elections in which, for the first time in the history of French elections, candidates had to run in mandatorily mixed pairs. Using the fact that the order of appearance of candidates on a ballot was determined by alphabetical order and showing that this rule does not seem to have been used strategically by the parties, we show on the one hand that the position of women on the ballot is random, and on the other hand, that right-wing pairings for which the name of the female candidate is in the first position on the ballot receive an average of 1.5 percentage points less votes.