Organizer: Ethan Scheiner, University of California, Davis

Chair: Matthew Shugart, University of California, San Diego

Discussant: Sadafumi Kawato, Tohoku University, Japan

When Japan enacted electoral reform in 1994, scholars, journalists and pundits alike put forward a seemingly endless list of predictions about what sort of outcomes would emerge from the new system. Now, with more than a decade having passed and three elections having been held under the new system, this panel examines the extent to which such predictions were borne out.

The papers in this panel re-evaluate many predictions, asking, among other questions: Did, as predicted, the introduction of the new electoral system indeed make Japanese politics more national party policy oriented and less focused on local, particularistic issues? Did the new electoral system bring about an increase in party and candidate competition? Did the new system bring about a decline in or disappearance of many of the accused ills of the '1955 system' such as factions, zoku giin, and the personal vote?

The panel includes scholars who published many of the first prognostications of the system's likely effects and who have published on the specific effects of the new system since its advent. Putting together the different papers in the panel, the authors hope to offer a thorough analysis of what predictions proved correct, which were wrong, and, most important, WHY outcomes that many thought were likely did not bear out in reality. The panel thus hopes to develop a greater understanding of both Japanese politics and the political consequences of electoral laws more generally.

Margaret A. McKean, Duke University

In 2000, the authors of this paper published an article in the journal ELECTORAL STUDIES that argued that Japan’s post-1994 electoral rules would do the following things: (1) It would transform proportional representation seat holders – politicians who in most systems tend to act with non-geographically specific, broad-based policy interests in mind – into locally-based politicians who will rely on personalistic rather than party-based or programmatic campaigning. (2) It would effectively convert the new single member districts back into the multi-member districts of the past. (3) It would enhance incumbency advantage. (4) It would push the ratio of candidates-to-seats down as low or even lower than before.

Since publishing the original article, the authors have been asked on numerous occasions whether their predictions bore out as expected. This paper here re-examines the original analysis. Based on an examination of the three elections held under the new system, the paper evaluates the accuracy of the predictions and considers the reasons behind any failure in the predictions.

Ray Christensen, Brigham Young University

One of the prime claims of pre-reform analyses of Japanese elections was that multi-seat districts contributed to personalistic campaigns and all of their supposed attendant ills (factions, money politics, personal support organization, pork barrel politics). Ignored in this analysis was the role that Japan’s strict campaign laws (which largely remained unchanged in Japan’s election reforms) played in encouraging and even requiring personalistic campaigns to be elected.

I examine the extent to which post-reform voting in Japan has continued to be personal based rather than party based.  I also examine the extent to which the support of other political parties is essential for victory, another characteristic of the previous electoral system.  I accomplish this analysis using a data base of the three post reform elections which compare the baseline vote for political parties (as recorded in the simultaneous voting for political parties), account for endorsements of specific candidates by political parties, and also account for personal attributes of the specific candidates (incumbency status, prior occupation, age).  To the extent that personal attributes account for a significant share of a candidate’s vote, then personalistic campaigns still have a strong appeal with the Japanese electorate, suggesting other factors (such as campaign rules) have a strong independent effect on the party vs. personalistic orientation of Japanese candidates and politicians.

Ellis Krauss, University of California, San Diego

Predictions about the consequences of the 1994 electoral reform, at least at the macro-level of the political party system, the consequences of a mixed-member system of SMD and PR, and policymaking have generally been borne out. However, several predications have fared less well at the more 'meso' level of organizational change in the LDP. As we have pointed out previously, deductive predictions from deductive electoral theory about the disappearance of factions, koenkai, and PARC  have generally not come about. This paper explores why that is the case in terms of how political organizations crucial to the LDP under the SNTV system actually evolved, why their prior historical development helps explain their post-reform fates, and therefore why such electoral determinist theories can go awry in predicting organizational political outcomes.

Steven R. Reed, Chuo University, Japan

I arrived in Japan in 1993, the year before reform were enacted that changed the electoral system. As a student of electoral systems, I immediately began making predictions about how the system would work out in practice. My predictions ranged from the very broad (Japanese politics will be closer to a two-party system after the election than it was before) to very specific (turnout will go up in the next election). Some of these predictions were published in such places as Asian Survey and the Journal of Japanese Studies but many were only made on the Japan Forum chat group. I will gather all of the predictions I made between 1993 and 2000 and ask other experts in the field to evaluate my accuracy. I will then analyze my successes and failures.

Although I have yet to receive my grades, I expect positive results. I will argue that prediction in the social sciences is a great deal easier than is normally assumed. One need only follow two simple rules. (1) Predict those phenomena that are predictable, not those of greatest interest to either the participants or the mass media. (2) Base your predictions on the data, not on the opinions of the participants. Participants base their predictions based on biased data. A researcher who analyzes appropriate data knows more than do the participants.