This thesis studies the strategic behavior of two key components of democracy: political parties and voters.
Three scenarios regarding political competition are presented, coauthored with Socorro Puy (Universidad de Málaga).
The first scenario presents a model of political competition between a traditional party (office-seeking), and an assembly party. An assembly party takes decisions according to what the affiliate members vote in an assembly prior to an election. Results show that the location of the assembly induces the traditional party to locate at different platforms. When the assembly is centrist, the traditional party moves away from the median voter location. However, when the assembly is extremist, the traditional party locates at a moderate position. As a result, the assembly party generates divergence between the platforms of the parties, in contrast to the Downsian model (Downs, 1957).
The second scenario considers a party composed of two internal factions: an ideological and an opportunist. Following Roemer (1999), the ideological faction is in charge of keeping the ideological principles of the founders of the party (i.e., the "party label"), whilst the opportunistic faction is in charge of maximizing the number of votes achieved by the party. Under a sequential setting, results show that a party that gives equal weights to its two factions needs considerable valence advantage to win the elections, even with a first mover advantage. Interestingly, despite the multiplicity of equilibria, the Downsian equilibrium does not hold since parties have incentives to offer divergent platforms.
The third scenario studies a theoretical setting characterized by two parties with a two-fold objective: representing the electorate and winning the elections. Political competition is held considering several cases. If office-seeking incentives are very rewarding, parties sacrifice ambiguity and converge to the ideal policy of the median voter. If ambiguity is very rewarding, no equilibrium exists since parties have incentives to deviate to more ambiguous platforms even when this means losing the elections. We conclude that with risk-neutral voters when the distribution of voters is known, parties cannot choose ambiguous platforms if they want to win the elections.
Second, two scenarios are analyzed regarding voters. The fourth scenario is a joint work with Bernardo Moreno (Universidad de Málaga) that focuses on how the presence of conformity may alter decisions taken by agents under a complete information setting. When all agents are conformist and vote simultaneously, there are equilibria in which the decision adopted does not coincide with the option obtained if all agents vote truthfully. A solution to deal with this problem is replacing conformists by agents who vote truthfully (independents). In this case, the effect of conformity disappears with a sufficient number of conformists replaced by independents. Under sequential voting, conformity does not affect the results.
The final scenario is a joint work with Ismael Rodríguez-Lara (Middesex University) and Bernardo Moreno, in which conformity is induced in a laboratory experiment. A theoretical model is developed considering groups of five agents. Three cases are studied: when agents vote without inducing conformity in their preferences; when conformity is induced by an extra payoff if conforming to some other agent; and when conformity is induced and agents vote knowing the vote of two agents in advance. Results show that inducing conformity reduces the likelihood of voting truthfully. Besides, when agents vote knowing others' agents vote, this leads them to vote more truthfully. Results also show the effect of the voting rule on agents' decisions, being less sincere the less pivotal they are.