不中間的中間選民：以質性方法初探有政黨傾向選民隱藏政黨傾向的原因

＂Pure Independents＂ Reconsidered: Using the Qualitative Approach to Explore Why Independent Leaners Hide Their Partisan Orientation

既有研究對於選舉時程的影響分析，基本觀點認為蜜月期選舉時程有利於新任總統所屬政黨的國會選舉，並有利於大黨，且會降低有效政黨數目，主要關鍵在於總統選舉對國會選舉具有衣尾效應，既有文獻對於衣尾效應如何產生之分析與實際驗證相對較為缺乏。本文主旨在於探討半總統制下蜜月期選舉時程的政治影響，嘗試分析衣尾效應的選票結構，以檢驗總統選舉與國會選舉之間的關係。本文將蜜月期選舉時程視為主要制度變因，認為蜜月期選舉時程會影響選民的投票行為，進而影響政黨結構，選民投票結構會導致政黨得票結構的兩大黨化，進而型塑一致性政府。並以法國為主要個案，區分法國四種選民（大黨選民，中間游離選民，體制性投票選民以及小黨選民）的投票結構，個別檢驗分析總統選舉如何影響國會選舉的關鍵因素，以及對總體選舉結果的影響，以建構衣尾效應產生的內在結構與政治影響。本文發現法國蜜月期衣尾效應主要來自於幾種選票結構的共同組合結果：（1）兩大政黨選民：忠誠度較高，在總統選舉後會因政黨認同因素繼續支持原先政黨的國會選舉，且投票率高，進而導致有利於兩大黨的選舉結果；（2）中間及游離選民：總統選舉後舉行國會選舉，會大幅降低國會投票率，國會選舉時選民棄權比例相當高，特別是中間選民及游離選民會參與總統選舉但在國會選舉時棄權不投票，且在國會選舉時個人認同因素增加，國會選舉往兩大黨候選人集中；（3）體制性投票選民：法國存有偏好一致性政府或偏好分立政府的選民，這些選民會依據其體制偏好持續地進行投票，並依據總統結果而進行策略性投票，法國體制性選民亦為中間選民但傾向支持一致性政府；（4）蜜月期選舉時程不利第三黨：第三黨總統的衣尾效應非常低，總統第一輪得票率與國會得票率落差極大，且國會當選席次大幅下滑。蜜月期選舉制度對法國政黨體制的影響主要呈現於政黨兩大黨化與集中化，政黨得票率與當選席次落差大，特別是兩大政黨當選席次往往高於當選率，特別是有利於總統獲勝之國會政黨。相對而言，第三政黨則呈現低度代表，當選席次往往低於當選率，應驗蜜月期選舉時程不利第三大黨之推論，此一結論與Shugart和Carey認為蜜月期選舉時程有利第三大黨之論證有所不同。

According to studies on the effects of electoral cycles, the fundamental perspective argues that honeymoon elections benefit the parliamentary election to the party of the newly elected president, while lowering the effective number of parties. The key is the presidential election has a coattail effect on parliamentary elections. However, in the literature, this coattail effect has lacked analysis and empirical validation. This paper aims to probe into the political effect of electoral cycles of honeymoon election in the semi-presidential system, and analyze the ballot structure of the coattail effect, in order to validate the relationship between a presidential election and a parliamentary election. This paper uses France as a case study. The case study reveals that the coattail effect of a honeymoon election in France is mainly from the combination of several ballot structures: (1) Voters of two major parties: These voters are more loyal. After a presidential election, due to party identification, voters tend to continue supporting the parliamentary election of the original party with a high voting rate, thus positively influencing the electoral outcomes of two parties. (2) Median and swing voters: A parliamentary election held after a presidential election will significantly lower the parliamentary voting rate. Median and swing voters will participate in the presidential election, however, they abstain from parliamentary voting in parliamentary election. Individuals' identification factors increase and voting in the parliamentary election is centralized to the candidates of two major parties. (3) Regime voters: Currently, in France, there are two regime voters, voters who prefer a unified government or a divided government. They continue voting according to their preference for the regime type and undertake strategic voting according to presidential outcomes. The result indicates that French regime voters tend to support unified government. (4) Electoral cycles of a honeymoon election do not benefit minor parties: The coattail effect of the president of a minor party is extremely low. The gap between voting rates of the first round of a presidential election and a parliamentary election is significant. Parliamentary seats decrease significantly. The impacts of honeymoon elections on the French political system are mainly presented in the bipartisation and the centralization of political parties. There is a large gap between the percentage of vote obtained and elected seats rate, especially the elected seats rate of the two major political parties are often higher than the other parties. In contrast, the third largest party shows a low degree of representatives of elected seats. The honeymoon election is relatively unfavorable to the third largest party. This conclusion is also different from the argument of Shugart and Carey who think honeymoon elections are favorable to the third largest party.