The paper aims to contribute to the workshop by providing single-case narratives of party system changes in Japan and by proposing a measure of party system changes different from those available in the literature. Japan has undergone a significant transformation of its party system since the 1990s as the dominant and governmental party of Liberal Democratic Party (LDP) was split into several parties in the early 1990s. The long-term opposition party of Social Democratic Party of Japan (SDPJ) has shrunk its size throughout these decades while a new party of Democratic Party of Japan (DPJ) was established in the middle of 1990s, which eventually seized power in the end of 2000s. Including these splinter parties, there have been more than a dozen of new parties at national level in Japan in the past two decades. Despite variations in its origin and ideological profile, all of these parties are either a split, splinter, or merged party of the established political forces. In light of these party system changes, the existing measures such as net volatility (Pedersen, 1979) could at best capture the extent to which votes moved between parties across two elections. As the unit of analysis is political party, such measures cannot distinguish these new parties in an old bottle from ‘genuinely new parties’ (Sikk, 2005). The paper instead proposes to set parliamentary members as the unit of analysis to calculate the level of volatility as it would better reflect the extent to which new parties are composed of new members in parliament. The measure could also account whether reelected candidates have been party-switchers or not and this can be weighted based on split, splinter or merger categories. Such elite volatility could also be alternatively used to adjust the existing measure of net volatility.
