ABSTRACT
Literature provides four basic theories to explain regional election results and how they differ from national patterns: authority of regional governments, ethnic or linguistic cleavages, congruence of national and regional electoral systems, and second-order election effects. The second-order national election theory explains why regional elections exhibit lower turnout levels, why government parties lose voter support, and why opposition, minor, and new parties gain support. While second-order election theory provides the dominant explanation for countries with low regional power, we argue in favour of an additional explanation based on incumbency effects on parties’ electoral support. We test the explanations on Czech regional and national election data for the years 2000–2020. The results attest to a strong effect of regional governorship, with a bonus of 5 percentage points for parties whose governors run for re-election. Parties also receive another bonus when national-level MPs and local mayors are present on the ballot.
Disclosure statement
No potential conflict of interest was reported by the author(s).
Notes
1 With some simplification, the governor incumbency situation can be described in terms of US incumbency advantage studies: a sophomore surge occurs when the governor runs for reelection and a retirement slump when they do not.
2 For parties forming coalitions before the regional election, the 1% vote share conditions had to be met by at least one of the parties in that regional coalition.
3 This was the 2016 Karlovy Vary Region list of Mayors and Independents with support from the Karlovy Vary Civic Initiative, KDU-ČSL, and TOP 09. At that time, KDU-ČSL was part of the national government of PM Sobotka while the remaining parties were in the opposition.
4 For more on the specific context of the 2016 regional elections, see Pink and Eibl (Citation2018).
5 The only interaction term related to the state of economy used in our models is that between government party and the economic indicator. In contrast, Schakel (Citation2015), for example, uses the interaction term between the economic indicator and government, opposition, and new party. There are two reasons to believe that our specification is more suitable. First, voters primarily evaluate government parties because they are unable to follow the hypothetical counterfactual situation of opposition parties involved in the government (Bechtel Citation2012, 173). Second, there is no need to differentiate because similar interaction effects are exhibited by opposition and other non-government parties; this is also the case of Schakel’s (Citation2015, 648) models.