The Dynamics of Electoral Stability and Change of LDP in the Japanese House Elections in 1990's: A Panel Data Exploration
| Autor: | Ming-Feng Kuo, 郭銘峰 |
|---|---|
| Rok vydání: | 2004 |
| Druh dokumentu: | 學位論文 ; thesis |
| Popis: | 92 The post-war party politics in Japan, called “1955 System”, has been dominated by the Liberal Democratic Party (LDP) since 1955. This system is not changed until 1993 when the LDP lost its majority in the House election and the multi-party competitions became to challenge its domination. But only three years later, the LDP staged a comeback and played the major role in forming a new coalition government, and again, it managed to maintain this control in the 2000 House of Representatives election. The purpose of this study, based on the panel data collected by the Japanese Election Study (JES) in 1990’s, is to analyze the dynamics of electoral stability and change of the LDP supporters during the 1993, 1996, and 2000 House elections. Given LDP had confronted the crisis of votes outflow, the data shows that a substantial proportion of LDP supporters remains loyalty during these period. Besides, more elaborate panel data analyses of Nested Logit Model indicates that, the constituent’s residences, ages, and party identification have a significant influence on their stability and change of supporting LDP, which is consistent with the previous studies in this discipline. Specifically, this thesis finds Japanese voters in the urban are more likely to support non-LDP candidates, while the voters from the rural areas prefer the LDP candidates. In addition, comparing to the younger generation, the older Japanese voters are more likely to support LDP. Of those reasons influencing the electoral stability and change, voter’s party identification without doubt plays a very important role. Given the fact that Japanese party system has become a multi-party competition in 1990’s, voter’s strong LDP identification keeps them support LDP candidates consistently and solidly. Most importantly, by exploring several Markov chain models, this study finds the “Black and White Model”, proposed by Converse (1964, 1970), is the most appropriate one to explain the dynamics of electoral stability and change in Japan during 1993, 1996, and 2000 House elections. By splitting the population into two latent subgroups with maximum intragroup homogeneity and intergroup heterogeneity, Black and White Model estimates each group’s transition probability. The results show the first subgroup “stayers”, always maintaining their LDP or non-LDP preference without any change, consists of about 48.28% of the panel samples. While another subgroup of “random movers”, who change their support randomly, consists of about 51.72%. |
| Databáze: | Networked Digital Library of Theses & Dissertations |
| Externí odkaz: |
|