Résumé : The advent of experimental methodologies have led to decisive progress in the study of corrupt behaviour in the last two decades. Since they can complement survey data and perception indexes with controlled experimental data, scholars and policy makers have reached a better understanding of decision-making in bribery situations and are able to design innovative anticorruption policies.In this thesis, I use experimental and empirical data to contribute to the field of the economics of corruption. The first two chapters of this PhD dissertation use experimental methodologies to study decision-making in a bribery scenario. The first chapter tests an anti-corruption strategy with a lab in the field experiment in Burundi. The second chapter studies the fairness concerns that might arise when dealing with redistribution in a bribery situation. The third chapter uses an empirical analysis to explore the controversial issue of political moonlighting, i.e. having outside activities while holding public office. It investigates "double-hat politicians", who combine mayor and parliamentary positions in Wallonia.In a first chapter, written with Jean-Benoit Falisse, we explore the effect of anti- corruption messages on corrupt behavior and public service delivery. In a novel lab-in-the-field experiment, 527 public servants from Burundi were asked to allocate rationed vouchers between anonymous citizens; some of these citizens attempted to bribe the public servants to obtain more vouchers than they were entitled to. Two groups of public servants were randomly exposed to short messages about good governance or professional identity reminders. Participants in these two groups behaved in a fairer manner than those of a third group who were not exposed to any message. The result is more robust in the case of the group exposed to the professional identity reminder. The underlying mechanisms seem to be that when a public servant reflects upon governance values and her professional identity, the moral cost increases, prompting more equal service delivery. Bribe-taking was not impacted by the messages. The experiment provides new insights into the design of anti-corruption strategies.The second chapter, written with Lena Epp, investigates the impact of a public officials’ fairness considerations towards citizens in a petty corruption situation. Other-regarding preferences, and, more particularly, fairness concerns are widely acknowledged as crucial elements of individual economic decision-making. In petty corruption contexts, public officials are to a large extent aware of differences between citizens. Here, we experimentally investigate how fairness considerations may impact on corrupt behaviour. Our novel bribery game reveals that bribes are less frequently accepted when bribers are unequal in terms of endowments. These results suggest that fairness considerations can influence corrupt behaviour.In the last chapter, I focus on political moonlighting in Wallonia. Activities outside of public office or combining specific public offices simultaneously is a topic of ongoing heated debates. An element crucial to these discussions is whether moonlighting is detrimental for politicians’ performance. In Belgium, the combination of local executive and regional legislative offices, i.e. double hat politicians, is a frequent habit for a majority of politicians. This accumulation of activities might lead to (un-)desirable outcomes in terms of political achievements. This chapter investigates the impact of holding several remunerated and honorary positions on regional MPs parliamentary activities and mayor’s municipality performance in Wallonia. I use a database of all public and private positions held by Belgian politicians in Wallonia since the disclosure of positions became compulsory for those holding at least one public position, i.e. from 2004 to 2016. For members of Parliament, wearing a double hat reduces global parliamentary activity. For mayors, it seems that holding more remunerated positions is associated with less efficient municipality management.