Election forecasting in Lithuania: The case of municipal elections

Abstract

Abstract This article explores the possibilities for election forecasting in Lithuania, a post-communist country that has a party system which is characterized by high levels of electoral volatility and fragmentation. The main argument of the article is that despite these unfavorable conditions, election forecasting in Lithuania has potential. Since the sample of national parliamentary elections is too limited for statistical modeling, the possibility of forecasting at the level of municipalities in the municipal council elections is discussed. Four factors (local unemployment change, party’s belonging to the national government, population size and lagged vote share) are integrated into a model that strives to predict the vote share of the party that holds the mayor’s post (the dominant political power in the municipal council). The model presented explains more than 70% of the variance in the dependent variable. The case diagnostics reveal that the model predicts municipal election outcomes very accurately.