This article examines why some democracies and semi-democracies develop relatively stable party systems, while others continue to be roiled by high levels of electoral volatility. It is the first broadly cross-regional analysis of electoral volatility, and it is based on the most extensive data assembled on electoral volatility. Our most original finding is that competitive regimes inaugurated in earlier periods have much lower electoral volatility than regimes inaugurated more recently, even controlling for a variety of other factors that have been hypothesized to affect electoral volatility. Parties had very different functions according to when democracy was inaugurated, and these congenital differences had longterm effects on the stabilization of party competition. What matters for the stabilization of party competition is when democracy was born, not how old it is. Our results support social science approaches that emphasize historical sequences and path dependence.

This article is a first attempt to explore the characteristics of political competition in Central America, a subject that has been the focus of little research to date. The article provides a detailed spatial analysis of Central American legislators and parties, as well as an analysis of the number and nature of the political dimensions dividing legislators and major parties for the period 1998–2002. This multi-level analysis illustrates the significant role of agency in the translation of cleavages to the arena of political competition. The article explores the relationship between party system polarization and the number and substance of politically relevant dimensions. It uses data from the Parliamentary Elites of Latin America Survey Project, and the statistical analysis involves factor and discriminant analysis of Central American legislatures.

RESUMEN: Este artículo examina por qué algunas democracias y semidemocracias desarrollan sistemas de partidos relativamente estables, mientras que otras continúan envueltas en altos niveles de volatilidad electoral. Éste es el primer análisis amplio a nivel comparado regional cross-regional de volatilidad electoral, basado en la información más extensa recolectada sobre el tema. Nuestro hallazgo más original es que los regímenes competitivos inaugurados en períodos tempranos tienen mucha menos volatilidad electoral que aquellos inaugurados más recientemente, incluso controlando una variedad de otros factores que se ha considerado que afectan la volatilidad electoral. Los partidos tenían funciones muy distintas según el momento en que la democracia fue inaugurada y estas diferencias congénitas tuvieron efectos de largo plazo en la estabilización de la competencia partidista. Lo que importa en cuanto a la estabilización de la competencia es cuándo la democracia nació, no cuán antigua es. Nuestros resultados apoyan las aproximaciones de las ciencias sociales que enfatizan las secuencias históricas y el path dependency.ABSTRACT: This article examines why some democracies and semi-democracies develop relatively stable party systems, while others continue to be roiled by high levels of electoral volatility. It is the first broadly cross-regional analysis of electoral volatility, and it is based on the most extensive data assembled on electoral volatility. Our most original finding is that competitive regimes inaugurated in earlier periods have much lower electoral volatility than regimes inaugurated more recently, even controlling for a variety of other factors that have been hypothesized to affect electoral volatility. Parties had very different functions according to when democracy was inaugurated, and these congenital differences had longterm effects on the stabilization of party competition. What matters for the stabilization of party competition is when democracy was born, not how old it is. Our results support social science approaches that emphasize historical sequences and path dependence.