Choosing the type of electoral system in new democracies has become a contested issue for social scientists as well as for political actors. Contrary to the state of public debate on the issue, the article advances the position from a multidisciplinary standpoint (political science, historical sociology, and economics) that proportional representation with large districts and closed lists performs better on a variety of key indicators. We review recent literature on the performance of electoral systems especially in the post-communist and Latin American democracies. The article identifies the centripetal theory of democracy as a normative basis for our institutional prescriptions and discusses how distinct types of political representation relate to the debate on electoral systems. We focus especially on four main concerns commonly associated with proportional representation (the rise of “extremist” parties, government instability, party system deconsolidation, and corruption and clientelism). Contrary to much of the public debate on electoral systems, we conclude that further steps towards personalization (by opening lists or reducing district magnitude) are not advisable.