Multilevel Analysis of Voting Behavior under the Mixed-Member Majoritarian Systems
Ⅱ. Electoral Incentives
The configuration of the political system provides various in-centives for parties and candidates; electoral rules such as sin-gle-member districts encourage collective incentives and
mul-98 선거연구_제1권2호(2011년 가을호)
ti-member districts encourage individual incentives (Shvetsova 1995; Katz 1997; Samuels 2002). According to Duverger (1954), the SMD district with simple majority voting system tends to have two-party systems. The simple-majority rule, however, encourages candidates to pursue individualistic campaigns. Each candidate can easily deviate from the equilibrium policy position to pursue vote gains, which makes parties into loose voting coalitions (Shvetsova 1995). Both the United Kingdom and United States have two-party systems, while British parties control assess to nomination and American parties use primary elections to decide their candidates. The different nomination mechanisms in the same party systems partly explain why British parties have stron-ger party discipline. In return, voters chose candidates based on candidate evaluation in the U.S. and party labels in UK respectively.
SNTV allows voters to choose only one candidate among their partisans, so that candidates are encouraged to pursue a personal vote outside party organizations to win a plurality. Because of the multi-member district, political parties tend to nominate more than one candidate in a typical district (Reed and Bolland, 1999). In this regard, particularistic interests outweigh party labels in structuring voting choices. Individual candidates receive incentives from SNTV to conduct candidate-centric campaigns because of the sin-gle-plurality rule (Shvetsova 1995). The SNTV system is arguably responsible for candidate-centered campaigns and factional poli-tics in Japan during the 1955 party system (Lin, 1996). In Tsai’s
Multilevel Analysis of Voting Behavior under the Mixed-Member Majoritarian Systems 99 (2003) analysis of voting behavior of Japan, Taiwan, and the United States, candidate evaluations outweigh partisanship in vot-ing in the congressional election. However, he found mixed evi-dences about the effect of SNTV on voting behavior in Japan and Taiwan.
Carey and Shugart’s overarching research argues that we should be able to compare election systems based on candidates’
incentives to cultivate a personal vote. Because of this two-vote structure, mixed-member systems present parties with coordina-tion problems. On the one hand, the SMD tier encourages candi-dates to cultivate a personal vote, especially when candicandi-dates are nominated by the primary election. On the other hand, the PR tier requires higher party reputation, especially when the closed-list PR system is adopted. They suggest that the conflict between in-dividual politicians and party leaders is critical to party organ-izations, legislative votes, and public policies.
Grofman (1999) agreed with Carey and Shugart’s classifying scheme, yet he suggests that different approaches to compare electoral systems may lead to different classification. According to the criterion of candidate focus, he categorized the MMM system as “intermediate” with close-list PR and SMD plurality on the ex-tremes, which confirms Carey and Shugart’s scoring results.
Bawn and Thies (2003) applied Carey and Shugart’s frame-work to representatives’ behavior. They argued that representa-tives elected in PR tier tend to respond to unorganized interests.
Crisp, Jensen, and Shomer (2007) suggested that the number of
100 선거연구_제1권2호(2011년 가을호)
candidates may not be the function of district magnitude; the number of “copartisans” that candidates compete with could de-crease, not inde-crease, as the district magnitude increases. They ar-gued that candidates cultivate personal votes to get seats that their party is likely to win. According to their theory, therefore, the close-list PR system has the highest incentive to cultivate personal vote, which is contrary to the prediction of Carey and Shugart among other scholars. Their idea of “party seats” may need more empirical evidences.
Despite that scholars disagree with the scoring system, Moser and Scheiner (2005) followed the framework of personal vote and investigated strategic voting in five different MM electoral systems.
They found that the extent to which voters split their votes be-tween the SMD and PR tiers largely depends on the difference in the vote share won by the candidates in the first and second place. Their evidences show that in Japan, New Zealand, Lithuania, and Russia the competitiveness in the district level pos-itively affects the gap between PR and SMD voting results; voters would not cast their two votes to the same party in the PR and SMD districts when their parties are not likely to win the race in the SMD district. Moser and Scheiner’s finding suggests that in mixed-member (MM) systems candidates may need substantial personal votes to beat their opponents while parties may not ben-efit from personal votes.
Kawado (2004) also examined the difference between PR and SMD districts in Japan’s 2000 general election. He not only found
Multilevel Analysis of Voting Behavior under the Mixed-Member Majoritarian Systems 101 that candidates in the 1st place on average won 18 percent of SMD votes more than the PR votes, but also found that the num-ber of PR parties decreases the difference. That means Japanese voters concentrate their votes on two major parties (LDP and Democratic Party) when they sense that small parties may try to split their votes by running party lists.
Based on Carey and Shugart’s scoring system, we expect that the SMD and PR districts emphasize personal vote and party repu-tation respectively. The MMM systems that combine both in-centives of personal vote and party vote provide an arena of ob-serving how two competing incentives affect voting behavior, in addition to candidates. Following Moser and Scheiner, and Kawado’s method, we assume that the difference between the SMD and PR districts measures the level of personal vote; the greater the difference between the two tiers, the greater person-alistic politics is. In particular, the MMM system with two parallel tiers in which a PR vote cast for a small party do not affect the SMD results encourages voters to cast their votes strategically;
strong candidates would win more SMD votes than their parties’
PR votes (Reed 1999).
To sum up, we assume that the PR vote shares are the baseline of party reputation, thus the difference between the vote shares in the SMD and PR districts measures the level of personal voting.
The larger the difference, the more the candidates appeal to the electorate. As for the electorate, they will be more likely to vote for the party’s candidates.
102 선거연구_제1권2호(2011년 가을호)