Testing Duvergerʼs law: strategic voting in Mongolian elections, 1996–2004

Recent empirical research on voting in single-member districts, based on extensive data-sets of election results, has demonstrated the general (although not universal) validity of Duverger’s law (i.e. that the average outcome under plurality rule is generally consistent with two-party competition). This article tests Duverger’s law through analysis of a data-set covering Mongolian parliamentary elections in the period of 1996–2004. The results show consistent, but not linear, movement towards the Duvergerian equilibrium in Mongolia, with large part of the districts conforming to the Duvergerian norm of two-party competition. Duverger treated his law merely as an important tendency but insisted that social forces are the main determinants of the number of political parties. The main factor that limited Mongolian voters’ rationality, and created problems with their strategic ability to distinguish and abandon hopeless candidates, was weak institutionalization of the Mongolian party system. Finally, I prove that the emergence of bipolar party politics was not an immediate process and will continue over a series of elections, supporting the so-called “learning hypothesis.”