Search

This dissertation consists of four self-contained chapters, each based on an independent research paper. The first two address the role of expert judgments on consumer behavior and welfare by focusing on the Goncourt Prize and the Booker Prize. The third and fourth chapters study the role of demonstrations and the determinants of corruption. All four chapters are empirical and rely on observational data. The challenge in applied work when randomized control experiments cannot be used is to draw causal inferences, which is the core message of the “identification revolution” in economics. The common feature of the chapters of this dissertation is that each one addresses causality by using a specific identification strategy taking advantage of a natural experiment. Specifically, Chapter 1 rests on a regression discontinuity design allowed by the decision process of the Goncourt to assess the effect of the prize on sales. Chapter 2 relies on a difference-in-differences strategy to investigate the impact of the Booker on consumer welfare. Chapter 3 estimates the efffect of demonstrations on voting outcomes in the 2002 French presidential election by using weather conditions as an instrumental variable. Finally, Chapter 4 implements a regression discontinuity design to study the effect of electoral rules on political corruption by leveraging the fact that the French law for municipal elections conditions the electoral rules on the population size of a municipality. I provide below a succinct overview of the method and findings of each chapter., Doctorat en Sciences économiques et de gestion, info:eu-repo/semantics/nonPublished

Expert judgments may increase or decrease consumer welfare depending on experts’ ability to redirect consumers toward goods they enjoy. Leveraging the discontinuity created by the attribution of the Booker Prize, a leading literary award, we confirm that the prize attracts readers to consumption. We then investigate how it affects consumer surplus. We measure consumer ex post satisfaction from reading a book by the sentiment and the rating of the reviews posted on Amazon. We show that the Booker reduces satisfaction and that this negative effect is driven by a misalignment between the tastes of the jury and those of consumers. We quantify the associated loss in welfare by calibrating a structural model of demand. We find that the prize reduces consumer surplus by USD135,000 annually, meaning that a consumer buying a Booker Prize-winning book experiences a loss in surplus of 4% of the average price of a book., info:eu-repo/semantics/published

Using a randomized controlled trial, we study whether showing German respondents a graph plotting the European Central Bank’s inflation target alongside inflation in the euro area from 1999 to 2017 affects respondents’ trust in the ECB. The treatment has, on average, no significant effect on the level of trust in the ECB respondents report, but trust increases among respondents who report no preference for any political party. Within this group, the information about the actual development of the inflation rate, and not information about the inflation target itself, appears to be the main driving force., info:eu-repo/semantics/published

This paper studies the relationship between the size of a jurisdiction and how corrupt its citizens perceive officials to be. The relationship may a priori be driven by four distinct mechanisms: (i) larger communities have more officials, thereby making it more likely at least one official is corrupt; (ii) larger communities have a larger budget, thereby offering more opportunity to be corrupt; (iii) monitoring officials is costlier in larger communities; and (iv) the public is less likely to have contact with officials in larger communities, which raises citizen’s suspicion. Using cross-country analysis, we first establish that corruption is perceived as larger in countries with larger populations. We then test this stylized fact using French survey data on the perception of the municipal government corruption. We again observe that perceived corruption increases with population size. This result holds through a series of robustness checks and many confounding factors. Moreover, our results hold across two distinct periods and for another administrative unit, departments. Finally, we report suggestive evidence that the stylized fact is driven by mechanisms (i) and (ii), but not by (iii) and (iv)., info:eu-repo/semantics/published

This paper surveys the recent literature on the relation between social norms and electoral outcomes. It argues that the relation goes both ways: social norms can affect electoral outcomes and vice versa., info:eu-repo/semantics/published

The literature has pointed out the negative aspects of political dynasties. But can political dynasties help prevent autocratic reversals? We argue that political dynasties differ according to their ideological origin and that those whose founder was a defender of democratic ideals, for simplicity labelled “pro-democratic dynasties”, show stronger support for democracy. We analyze the vote by the French parliament on July 10, 1940 of an enabling act that granted full power to Marshall Philippe Pétain, thereby ending the Third French Republic and aligning France with Nazi Germany. Using data collected from the biographies of parliamentarians and information on their voting behavior, we find that members of a pro-democratic dynasty were 9.6 to 15.1 percentage points more likely to oppose the act than other parliamentarians. We report evidence that socialization inside and outside parliament shaped the vote of parliamentarians., info:eu-repo/semantics/inPress

We examine the effect of studying business on the beliefs and values of students and study the role of economics in influencing this effect. We observe significant differences between business students and students from other disciplines, among which economics, at the start of their first year. We also discover that some of these differences persist or are reinforced at the end of the year. Furthermore, we find changes in beliefs and values of business students that take only one year to manifest. Above all, while we observe that some values and beliefs of business students change over time, we observe no such changes for economics students. This suggests that the effect of studying business is not entirely driven by exposure to economics in business studies., SCOPUS: ar.j, info:eu-repo/semantics/published

info:eu-repo/semantics/published

This paper studies a new mechanism that allows political elites from a nondemocratic regime to survive a democratic transition: connections. Wedocument this mechanism in the transition from the Vichy regime to democracy in post-World War II France. The parliamentarians who had supported the Vichy regime were purged in a two-stage process where each case was judged twice by two different courts. Using a difference-in-differences strategy, we show that Law graduates, a powerful social group in French politics with strong connections to one of the two courts, had a clearance rate that was 10 percentage points higher than others. This facilitated the persistence of that elite group. A systematic analysis of 17,589 documents from the defendants' dossiers is consistent with the hypothesis that the connections of Law graduates to one of the two courts were a major driver of their ability to avoid the purge. We consider and rule out alternative mechanisms., info:eu-repo/semantics/published

We examine the effect of studying business on the beliefs and values of students and study the role of economics in influencing this effect. We observe significant differences between business students and students from other disciplines, among which economics, at the start of their first year. We also discover that some of these differences persist or are reinforced at the end of the year. Furthermore, we find changes in beliefs and values of business students that take only one year to manifest. Above all, while we observe that some values and beliefs of business students change over time, we observe no such changes for economics students. This suggests that the effect of studying business is not entirely driven by exposure to economics in business studies.

Increasing the employment rate of the working age population has been at the top of policy and research agendas alike. In this thesis, I study the impacts of three flagship Active Labour Market programs – a job search monitoring program, a wage subsidy, and a training program – targeted at vulnerable populations. In the first chapter, I study the effects of a job search monitoring policy targeted at the long-term unemployed, focusing not only on transitions from unemployment to employment, but also on participation in other social welfare programs. Using a Regression Discontinuity Design, I show that the program decreased the probability of remaining unemployed, but only through an increase in the probability of receiving Disability Insurance benefits. Employment and participation in other types of social welfare programs are, on the other hand, unaffected. These findings highlight that, for individuals who are on the fringes of the labour market, job search monitoring might not be sufficient to foster a return to employment. In the second chapter, I use a Difference-in-Differences approach to study the effects of implementing domestic service subsidies on employment and work disability in the targeted sectors. I then complement this analysis with an event study Difference-in-Differences approach that allows to uncover how working in these subsidized jobs affects the labour market outcomes of subsidized workers. I show that domestic service subsidies can be an effective means of creating employment opportunities for low-skilled workers in the sector of domestic services, by both reducing unemployment and increasing labour market participation of the target group. I also show that PHS subsidies can entail important unintended costs for the Disability Insurance and possibly deteriorate the health of subsidized workers. In the last chapter, I explore whether jobseekers can be encouraged to follow trainings for shortage occupations through a low-cost information interve, Doctorat en Sciences économiques et de gestion, info:eu-repo/semantics/nonPublished

We study how demonstrating against a far-right candidate changes the behavior of voters and ultimately impacts election results. To do so, we focus on the 2002 French runoff presidential elections which pitted far-right candidate Jean-Marie Le Pen against the incumbent, Jacques Chirac. Between the two rounds of the election, demonstrators protested Le Pen’s quest for power at roughly 300 demonstrations. Using rainfall as an exogenous source of variation in demonstration attendance across municipalities, we find that larger protests reduced the number of votes for Le Pen and the number of abstentions and blank or invalid ballots, and increased the number of votes for Chirac. We show that this positive effect on voting for Chirac results from left-wing voters who did not cast a blank or invalid ballot and right-wing voters who switched from Le Pen to Chirac. Next, we focus on the mechanisms behind these results to find that the 2002 demonstrations both reduced support for the policies advocated by Le Pen and signaled that voting for him was socially undesirable. Finally, we provide evidence that demonstrations affected voting mainly through local media coverage and spread out beyond the municipalities that hosted the demonstrations., info:eu-repo/semantics/published

We study the spatial distribution of the effect of oil and gas revenues on Brazilian municipalities, using variations in international prices to establish causality. Those revenues increase night-time light emissions in oil-producing municipalities but impose negative spillovers on neighbouring municipalities, especially those located more than 150 kilometers away from oil activities. The two effects cancel out at the level of micro-regions. In oil-producing municipalities, oil and gas revenues increase royalties, population, prices, crime, and real wages, particularly in manufacturing and service sectors. Spillovers are negative on wages and prices, but positive on royalties and crime., info:eu-repo/semantics/inPress