Electoral Systems and Elections of Korea, Taiwan, and Japan

Japan adopted the MMM system in 1994 and held the first low-er-house election under the new system in 1996. Since then, Japan has held five general elections. Under the new electoral system, the total seat shares of the two largest parties have been increasing. A two-party system seems to take shape in Japan (Lin 2006; Reed 2007). Electoral system change also encourages parties to emphasize party platforms and centralize power (Mori 2009).

However, some scholars argued that electoral system change did not solve the corruption problem among others. They doubted that other factors other than electoral systems are responsible for weak issue voting and persistent vote-subsidy mechanism (Kobayashi and Nagatomi 2009). Natori (2007) also pointed out that, after examining various elections between 1997 and 2003, the LDP vote shares are highly correlated with prefectural subsidies.

One of the reasons that Japan adopted the MM electoral sys-tem is correcting the individualistic interests induced by the SNTV system. Because of the multi-member districts, the SNTV system encouraged candidates to cultivate personal votes. First of all, can-didates who share the same party label must compete with each other. Each candidate’s votes are not transferable to other party members, thus candidates are not motivated to appeal to support-ers beyond their own parties. Instead, they tend to concentrate on

Multilevel Analysis of Voting Behavior under the Mixed-Member Majoritarian Systems 103 small proportion of constituencies in terms of geography or demography. Moreover, factions within parties instead of party or-ganizations control nomination. That not only weakened party leader’s control over ballots, but also created the particularistic politics and patron-client networks. By controlling for specific public policies, factions can secure support of interest groups for campaign spending on faction members. Consequently, the long-term dominant party, Liberal Democratic Party (LDP), can easily extend their dominance not only because it is able to coor-dinate nomination by allocating government resources, also be-cause public policies are fragmented. In the long run, a strong op-position party can hardly emerge and challenged the incumbent party.

As more and more scandals erupted, Japanese blamed the LDP for money politics. The LDP gradually lost the majority in the par-liament, and thus it did not win the simple majority of seats in the general election of 1976, 1979, and 1983. In the 1993 lower-house election, the LDP split and won only 223 seats (256 needed for majority). Eight opposition parties led by the Japan New Party de-cided to form the government without the LDP for the first time since 1955. To address mounting pressure on electoral reform, the new parliament chose the MM system. On the one hand, it was believed that under the new electoral system the LDP would not be able to take advantage of patron-client networks in small constituencies. On the other hand, minor parties may benefit from party-list districts (Lin 2006).¹⁾ Therefore, Japan adopted the

paral-104 󰡔선거연구󰡕_제1권2호(2011년 가을호)

lel MM system that combines 300 SMD districts and 200 (reduced to 180 in 2000) PR seats in 11 regional blocs. The district magni-tude of the PR regional blocs ranges from 7 to 13.

As in Japan, parties in Taiwan managed to coordinate candi-dates because the SNTV system made the intra-party competition higher than inter-party one. Major political parties, the Kuomintang (KMT) and Democratic Progressive Party (DPP) struggled to overcome the nomination problem, while minor par-ties survived in large constituencies. The KMT maintained a he-gemonic regime by delivering resources to their candidates and dividing “responsibility zones” within each districts. Besides, the KMT suppressed the emergence of class, region, or other social cleavages. The opposition forces thus can hardly find a national issue and challenge the KMT until they turned to the ethnicity and national identity issues. Although the DPP won the presidential election in 2000 and 2004, the SNTV system still favored the KMT and its allies due to their massive patron-client connections estab-lished for more than forty years.

Public resentment on corruption and inefficiency of the Legislative Yuan, the legislative body, has grown to an undeniable call for reform. Following Japan’s lead, major parties in Taiwan found the MM system acceptable in 2005. Amid the protest of mi-nor parties, the KMT and DPP passed the reform proposal that

1) Kohno (2007) has reviewed the history of electoral reform, suggesting that the mass public and mass media gradually supported the idea of electoral system re-form and pushed parties to compromise on the measure of rere-form in 1994.

Multilevel Analysis of Voting Behavior under the Mixed-Member Majoritarian Systems 105 some DPP members described as “suicide.” According to the pro-posal, the SNTV system was replaced with the parallel MM system in which 73 legislators are to be elected in single-member districts and 6 are to be elected in aboriginal constituencies. A second bal-lot was introduced for the PR tier that comprises of 34 seats in a nation-wide constituency. Political parties that won more than 5 percent of PR votes can be allocated seats. In addition, the size of the legislature was cut in half, from 225 seats to 113.

In the 2008 legislative election, the Taiwan Solidarity Union (TSU) and other minor parties neither won a seat in the sin-gle-member districts nor received more than 5 percent of votes in the PR tier. While some independent candidates won in certain districts, the trend that Taiwan evolves into two-party system seems inevitable. Although both the KMT and DPP use primary elections to decide nominees in the districts of more than one can-didate, party leaders control access to the party lists. The closed-list PR system and SMD plurality with partial party endorse-ment seemed to make party reputation more important than per-sonal one. That confirms Carey and Shugart’s theory.

Korea used SNTV between 1987 and 1996. Voters cast a single fused vote for a candidate in one of 92 SNTV districts, each of which had two seats. Seats in the party-list district, which had 46 seats, were awarded based on the SNTV districts. Moreover, the party wins the most seats in the SNTV districts will receive enough seats from the party-list tier to form a majority before other seats are divided proportionally among other parties. In 1996, Korea

106 󰡔선거연구󰡕_제1권2호(2011년 가을호)

changed the electoral system; voters still cast a single fused vote in one of the 253 SMD districts but the two tiers are not linked (Hicken and Kasuya, 2003). In a month before the 2004 Assembly election, the National Assembly passed the measure of the new two-vote electoral system. Each voter cast two ballots, one of which is for the SMD district and the other one is for the party list.

The adoption of the MMM system was expected to produce a bi-party system, since the percentage of party-list seats accounts for fewer than 20 percent of total seats (Kim and Park, 2009). In their analysis of the 2000 and 2004 National Assembly elections, Kim, Choi, and Cho (2008) argued that the influence of region-alism weakened while generational cleavage appeared. They found that the influence of liberal-conservative ideology remained strong in both elections; the respondent’s probability of voting for the Grand National Party (GNP) instead of the New Millennium Democratic Party (NMDP) in 2000 and Woori Party (UP) in 2004.

Kim (2007) analyzed the 2004 National Assembly election and his descriptive analysis revealed that UP received large vote shares in Choongshung, Kangwon, and Cholla providence areas. He con-cluded that the regional-based voting pattern remains strong after the introduction of “one man, tow votes” system. Kim and Park (2009) also found that UP and NMDP gained a large vote share from the Honam region because they originated from that area.

Korea and Taiwan’s MMM systems allow voters to cast two bal-lots, and one of which is counted for seats in national party lists.

However, Japan has multiple PR districts and PR votes are

count-Multilevel Analysis of Voting Behavior under the Mixed-Member Majoritarian Systems 107 ed separately. When PR votes are counted separately in several areas, voters may be aware that their PR votes would only influ-ence a portion of PR seats. In this case, they may not cast straight tickets to their most preferred parties in both SMD and PR districts. In Japan’s case, moreover, candidates can be nominated in the SMD district and listed in the PR district. The double candi-dacy system may reinforce the candidates’ incentives to cultivate personal votes even in the PR tier. We therefore assume that party reputation would be more important in countries that pool all of the PR votes together than in countries that have more than one party list.

Provided that Japan has multiple PR districts, our research will calculate the vote shares in SMD and PR gap by PR districts in Japan. As for Korea and Taiwan, each of which has single PR dis-trict, we calculate the difference in vote shares in each SMD district. The comparison of these three countries would have im-plications for the research on the mixed-member systems.