Why in particular election some parties win and others lose votes (if compared to the last election)? One of the dominant theoretical explanations for the dynamics of electoral support is economic voting. Considering weak party identification and cleavage-based voting, it may appear easily applicable for the post-communist democracies: if not economy, then what can explain the dynamics of voting for the incumbent and non-incumbent parties? However, macro-level analysis of national elections in these countries faces the problem of very small samples (considering the quite small quantity of democratic elections). Therefore, research of (economic) voting in the post-communist democracies usually takes a path of regional voting analysis. However, this type of research was largely ignored by the Lithuanian researchers: there is still a lack of systematic study of regional voting, where it would be explained by the economic (and other, supplementary) factors.
This article strives to at least partly fill this gap and to make contribution into research on regional economic voting, resting on strategy quite rarely applied: analysis of local elections (into municipal councils of Lithuania). Hence first part of the article presents relevant (for the presented investigation) aspects of the economic voting theory and also discusses the possibilities of finding this phenomenon at the level of local elections. Potential impact for the vote change of other relevant political-institutional factors is also discussed. Resting on presented theory and existing empirical findings, two groups of hypotheses are discerned: the ones in first group aim for the search of economic voting, and others enunciate link between the vote and politicalinstitutional factors. Two kinds of responsibility objects are selected: first – dominant parties in national government coalition, second – dominant parties in municipal council (hypotheses aim to explain particularly change of their vote share).
Second part of the article presents statistical analysis of vote change in two periods between municipal elections in Lithuania: year 1997–2000 and 2002–2007. The main reason behind this particular selection is quite different nature of both periods, which enables to test the impact of economic voting (and other factors) under different conditions. To be more specific, first period is characterized by the dominance of Homeland Union-Lithuanian Christian Democrats (HU-LCD) in the national government and deteriorating state of economy (aspect of rising unemployment). In second period, Social Democratic Party of Lithuania (SDPL) was in power (dominant party in national government coalitions) and state of economy was improving.
The main findings of empirical analysis (methods of regression analysis and independent samples t-test were applied) are in line with the logic of conditional economic voting: economy matters, but sometimes it matters more and sometimes it doesn’t. Firstly, unemployment change is interestingly important in the explanation of the vote change of dominant party in the municipal council, when economy is improving: more the unemployment falls, the bigger is possibility of reward (or lesser punishment). However, when state of economy is deteriorating, different levels of unemployment rise doesn’t have an effect on the vote change: mere participation in the national government is enough for the punishment (but it is interestingly not important for the reward, if economy is rising). Moreover, in case of rising unemployment, dominant parties in the municipalities that are also dominant in the NATIONAL coalition are clearly punished if compared to other parties: but in case of improving economy, there are no differences between two samples.
Research also confirms that turnout change is an important explanation factor of local election results: it was statistically significant in three of four regression model. It is important for the vote change of both analyzed dominant-in-the-national-government parties, HU-LCD and SDPL: these parties tend to lose votes if turnout is rising (and vice versa). Fragmentation variable has impact in two models: this confirms that party system format is important for the electoral change not just at the national level, but also at the local elections.