The thesis addresses three questions related to the main goals of elections, preference aggregation and politicians’ accountability. In the first chapter, we explore the effect of adopting a closed or open list proportional representation system on the representation of minorities in Parliaments. The two systems differ in the possibility for the electorate to express preferential votes for candidates within party-lists. We show that minorities are proportionally represented only under open list, if voters can vote for a limited number of candidates. In all other cases, minorities are poorly represented, or not represented at all. The second chapter provides an explanation for sub-optimal policy making decisions by incumbents. We show that the possibility to stand for re-election can induce them to address less relevant issues, when more relevant ones are too difficult to solve. As failing to solve an issue constitutes a bad signal about their quality, politicians prefer to address easier ones, even if they are less important for the electorate. The goal of the third chapter is to axiomatically characterize a measure of ethnic voting, defined as the extent to which voting decisions are affected by ethnic motivations. The main property we impose on the measure is neutrality with respect to size of ethnic groups and total support of parties. We identify the class of size neutral measures of ethnic voting and investigate monotonicity axioms that are compatible with this property.