Annual Report 2018

• • The experiment was implemented from November 3 - 6, 2014 in Itajaí (Santa Catari- na), with one session each day. It involved 112 participants in the four sessions, with 28 participants in each one (20 voters and 8 politicians), and each participant took part in only one session. Each session included two treatments, with and without a rule, and the order of im- plementation was reversed to minimize the effects of ordering. The participants were not informed about the content of the second treatment until arriving at this part of the session. The participants were recruited by announcements, e-mails and posters placed in the campus of FGV in the preceding weeks. The composition of the group of participants was very representative of the popula- tion in the region where the experiment was conducted, although there was predomi- nance of participants previously educated in private schools. The participants received a symbolic payment (about eight dollars) and each session lasted about one hour. RESULTS . The results of the study suggest that morality and social rules are indeed crucial in elections with heterogeneous politicians. The results reveal, in the first place, that voters overwhelmingly condemn corrupt pol- iticians, almost totally eliminating their chances of winning, but often tolerate politi- cians who only create waste of public funds. In other words, voters frequently accept politicians that impose the same loss of wel- fare that corruption imposes, but without directly pocketing the public money for their own benefit. Finally, a rule that is accepted by the majority of citizens permits voters to coordinate their actions and encourages them to consider the interests of society as a whole in- stead of just their districts. The rule is able to reduce the number of votes cast for pol- iticians that cause waste even when voting for those candidates would benefit the dis- trict they represent, and hence is a strictly dominant strategy for the individual voter. CONTRIBUTIONS OF THE STUDY . • Unlike previous laboratory experiments investigating corruption by applying bribery games, and that have found little evidence of the moral costs of corruption, the game implemented in this experiment includes the negative externality of corruption in elec- tions and confirms that corruption is strongly punished. Nevertheless, when politicians were described as not being corrupt, but instead likely to cause wasted public resources, nearly half of voters voted for them, resulting in the same loss of welfare as would have been caused by corruption. The supply of infor- mation to voters about the amounts wasted was not sufficient to minimize the lost welfare designed in the game's equilibrium. For this to occur, it is necessary to have a social rule specifying that waste of public funds will not be tolerated. When considered together with the results of voters' deci- sion to adopt a rule against corruption or waste, the results provide support for the hy- pothesis that although voters may reject corruption but tolerate waste, social rules are able to reduce not only corruption, but also inefficiency in representative democracies. APPLICATIONS OF THE RESULTS AND POSSIBLE EXTENSIONS OF THE STUDY New studies can help to clarify the effect of political ideology or party identification on the willingness of voters to punish politicians who misbehave, but that are members of the party preferred by the voter, and those of other parties. Future research can also investigate to what extent corruption and waste influence voter turnout in elections. There is evidence that corruption induces voters to participate in elections to prevent parties and candidates associated with corruption from winning. Al- though waste directly benefits neither politicians nor voters, specific individuals or firms can indirectly benefit from it, and in this case it is expected that waste can increase the campaign financing of politicians responsible for it. AUTHORS: Paulo Arvate and Sergio Mittlaender. ORGANIZATION: Sao Paulo School of Business Administration (EAESP). SUPPORT: National Council for Scientific and Technological Development (CNPq) and Max Planck Institute. 32 Annual Report 2018 33 RESEARCH

View entire presentation