半總統制蜜月期選舉時程與政府型態：體制投票（regime voting）的分析觀點

受到美國分立政府逐漸常態化現象的發展，導致選民進行體制性的分裂投票（splite voting）越來越受到學界的重視，甚至認為體制性投票（regime voting）是影響分裂投票的主要因素。本文主旨在於探討下蜜月期（honeymoon election）選舉時程的政治影響分析，特別著重於選民針對政治制度選擇而改變其投票行為的體制性投票之影響，以檢驗總統選舉與國會選舉之間的關係，以及蜜月期選舉時程對政府體制型態的影響。本文以半總統制國家法國為主要分析個案，同時進行美國，法國以及台灣的跨國比較分析。本研究發現法國自2000年修憲將總統任期修改為五年制（Quinquennat），並進行總統選舉時程的調整，將總統任期縮短與國會一致，並採取「先總統、後國會」的蜜月期選舉時程改革後至今歷經三次的蜜月期選舉結果（2002年，2007年以及2012年），皆呈現總統多數與國會多數相同的一致性政府型態，蜜月期選舉時程確實有利於一致性政府的形成；第二、法國個案顯示法國的確存有偏好一致性政府或偏好分立政府的體制性選民，這些選民會依據其體制偏好進行投票，並改變其政黨偏好；第三、透過法國、美國以及台灣跨國比較分析發現，法國與美國體制性選民皆為中間選民，但不同於美國經驗，法國體制性選民傾向支持一致性政府，美國中間選民的體制性投票則傾向支持分立政府，而台灣體制性選民則傾向支持一致性政府。

關鍵字

半總統制；蜜月期選舉；分裂投票；體制性投票；法國第五共和體制

並列摘要

Due to the gradual normalization of the divided government in the U.S., split voting based on regime choice has attracted increasing attention from the academia. Regime voting is even regarded as the key factor affecting split voting. The main purpose of this study is to analyze the political influence of honeymoon election, specially focuses on how the voters' choice of political system affects their regime voting behavior, in order to examine the relationship between presidential election and parliamentary election, as well as the effect of honeymoon election on regime type. This study selected France as the major case for analysis, and conducted a transnational comparative analysis on the U.S., France and Taiwan.This study found that since the reforms of honeymoon election timing in France, including changing the presidential term to five years (Quinquennat) in the constitutional amendment in 2000, shortening the presidential term to be the same as that of the parliamentary term and changing the presidential electoral timing (presidential election first, followed by parliamentary election), all three honeymoon elections until now (in 2002, 2007 and 2012) result in a unified government, in which the presidential majority and parliamentary majority are the same. This confirms that the honeymoon election indeed is favorable to the formation of a unified government. Second, the case of France showed that there are voters preferring unified government or divided government, and these voters will cast their votes according to their regime preference and change their party preference. Third, the transnational comparative analysis on France, the U.S. and Taiwan found that the voters of regime voting in France and the U.S. are both swing voters. However, unlike the voters in the U.S., the voters of regime voting in France prefer supporting unified government, the swing voters in the U.S. prefer mainly supporting divided government. The voters of regime voting in Taiwan tend to support unified government.