«by Dyron Keith Dabney A dissertation submitted in partial fulfillment of the requirements for the degree of Doctor of Philosophy (Political Science) ...»
Individual candidates will learn about what works and doesn’t from their second campaign and make necessary adjustments. One point is worth making: the task of developing Type A organizational resources takes a lot of time. Once established they take less time to maintain. Relatively speaking, that would leave more time and energy for Type B strategies in a third campaign and thereafter, although the necessity to depend on the Type B should diminish in subsequent campaigns.
Note that our survey data, reported in the Tables above, do not support an interpretation of a movement from Type B to Type A followed by some movement back to Type B. That is true as well if the scores for each election and individual candidate are examined. However, we believe that this pattern can be found in the way Matsuzawa and Genba explained each campaign, and in our qualitative observations of how they behaved.
Hypothesis 3: The Type A prevails because an effective Type B strategy is precluded by tight election laws and regulations.
The laws on how political candidates can raise and spend money, and in general on what is permissible campaign behavior, are probably stricter in Japan than in any other
half of a broad proposition that it is structural factors rather than cultural factors that accounts for the prevalence of the Type A strategy in Japanese campaigns.
It is true that the rules are often violated—witness the large number of electoral staff, and sometimes the candidates themselves, who are arrested after each election.
However, breaking the law in the direction of the Type B strategy—say, buying TV commercials—would be too obvious even to consider. The violations are nearly always for intensifying the Type A strategy. For example, it is not legal to throw an intimate party with food and drink for local politicians, but it is difficult to detect (and in fact is common). The same would be true of outright bribery of community leaders although we have no information on its prevalence.
Unfortunately, there is no way to test this proposition in our study. Campaign law is a constant, not a variable, in this period, and of course they were the same across the three districts. Readers familiar with election campaigning in other countries will have been struck by the limited strategic repertory available to Japanese candidates. It needs to be mentioned in this conclusion since it is so important, but there is no reason to go into it more deeply.
Hypothesis 4: The Type A strategy is the product of the multimember electoral district system.
Everyone who has written about Japanese electoral behavior has stressed the importance of the medium-sized, single-ballot multimember electoral district that was used in Lower House elections from the early postwar period until it was reformed in
1994. The electoral system’s effects on campaign strategies—multimember districts
is structural rather than cultural factors that make Japanese election campaigns so distinctive. The structural explanation is why it was so widely predicted—or at least hoped—that moving to a single-member district (SMD) system would push candidates more toward Type B.78 This is a question that our study can address directly and more precisely. We look at one election under the old system and two under the new system. We follow the same candidates across all three. They are from an urban, a suburban and a rural district—as noted above, this reflects the most powerful explanation of cross-sectional variation in electoral behavior. All three were in five-member districts in 1993, the most extreme form of the MMD. And we observed and talked with these candidates and their staffs as they themselves tried to figure out the strategic implications of the institutional change for them.
Chapters 3 and 4 have told this story and we need not repeat the details. The most significant finding is that at least in the first instance, the reform led them away from the Type B toward Type A strategy, at least for Matsuzawa and Genba (the Usami case is indeterminate). In 1996 compared to 1993, Matsuzawa and Genba devoted much more of their energy and time to building up organizational resources aimed at reaching voters indirectly, and less to tactics to appeal to voters directly. To understand why this should be so, we need to look analytically at why SMDs should favor Type B strategies, and then why that didn’t happen in 1996 for our candidates, and more generally.
78 We do not consider the implications of having a two-tiered system including proportional representation because it was not an important factor in our candidate’s strategies. McKean and Scheiner (2000) have a good logical argument to the contrary but their considerations did not seem to apply in our districts.
Incidentally, Genba won in the PR race in 1996. His strategy in both that campaign and in 2000, however, was oriented entirely around the SMD race.
district will be M+1, so a five member district is expected to have six serious candidates dividing up the vote. Any individual candidate thus needs only a small percentage of the vote to capture a seat—in 1993, the last seat in our three districts was won with 11.5%, 12.8%, and 10.9% of the votes cast. With so low a threshold to meet to win a seat under the MMD electoral system, how did candidates attract this small slice of the voter? The Type B strategy would be to send some message, or project some image to all voters, hoping it would attract something over a tenth of them. This approach is particularly practical for picking up uncommitted, “floating” voters.79 The Type A strategy would be to concentrate on a specific group of voters, trying to induce some sense of connection, often indirectly via local influentials or organizations capable of insuring voter commitment.
Evidence over decades of MMD voting in Japan indicates that organizational efforts can bring in a substantial percentage of the needed vote. That means that the candidates who appeal to non-partisan, or “floating” voters have to pick up only a few to get over the line. Most candidates most of the time have decided that their marginal investment of time, energy and money will be most productive in trying to enlarge and solidify their connected vote rather than improving their appeal to “floating” voters.
The logic behind the popularity of the Type A strategy for many candidates competing in MMDs is clear, but what about the logic for the single-member district?
According to “Duverger’s Law” fewer candidates compete in SMD competitions. In most cases the SMD races are fought between just two serious candidates under 79 The logic of “floating voter” is germane in a multi-seat district electoral system. The expression describes partisan voters who are uncommitted to a party’s candidates competing in the same election district. Floating voters, unlike independent voters, identify with a certain political party.
Unlike a campaign operating for a MMD seat, it would be very difficult to build and maintain an organizational network to encompass a large proportion of the needed votes for a SMD seat. The marginal investment of a candidate is likely to be more productive if directed to the Type B strategy to pick up “floating” voters.
The logic is so straightforward and compelling that it is understandable why reform-minded academics and politicians would focus on moving from MMD to SMD as the key to changing campaign behavior. And it is understandable that so many were chagrined when it didn’t work and nothing seemed to change, as well expressed in the title of Hideo Otake’s (1998) book, How Electoral Reform Boomeranged: Continuity in Japanese Campaigning. An excellent summary of similar reactions to the 1996 election is McKean and Scheiner (2000).
The explanations of various analysts were summarized in Chapter 2. Here we can take a more analytical look at why Matsuzawa and Genda moved more toward Type A than toward Type B. First, four possible explanations are specific either to their particular situations or to the year 1996: a) political life cycle; b) “Duverger’s Law;” c) Change in district size and organizational rebuilding prerequisites; and d) “Hometown advantage.” a. Political life cycle As noted above, it is logical for first-time candidates from an opposition party to rely on Type B, and incumbents more on Type A. Moreover, for various reasons our candidates were much more Type B than average in Japan. A greater emphasis on Type
b. Duverger’s Law Although a single-member district should tend toward two serious candidates, potential candidates do not always recognize the point, particularly in the first instance.
Here, Matsuzawa and Genba offer an interesting contrast. Duverger’s Law did not work in Kanagawa 9th district. As discussed in Chapter 3, Matsuzawa, backed by the NFP in 1996, competed with representatives of the LDP, the DPJ, and the JCP (a minor party in many districts, but drawing over 15 percent of the vote in the Kanagawa 9th district).
These parties divided up the vote sufficiently for Matsuzawa to win with 36 percent of the vote, far from the half needed according to Duverger’s Law.
In Genba’s case, as seen in Chapter 4, Duverger’s Law did work in the Fukushima 3rd district—the district seemed so stable that some parties in fact were not interested in running candidates. As usual, the JCP did run a candidate, but the candidate was never in serious contention for the SMD seat since he attracted under 6 percent of the vote (there was also an independent who was even less successful). In this two-person-dominated race between Arai, the LDP candidate, and Genba representing the DPJ, it was very close, with Genba losing 46 to 44 percent (easily a good enough effort to win him a PR seat).
Why did both candidates pursue the Type A strategy so enthusiastically in the 1996 general election? In Matsuzawa’s case, his behavior likely was predicated on the size of the competition. He might well have thought that with so many opposing candidates running in the 9th district he was still running a race very similar to a multi
of the votes he needed. In Genba’s case, the logical explanation is simply, as stated above, that his district is rural enough to overwhelm the logic of the new electoral system.
c. Change in district size and organizational rebuilding prerequisites A one-time-only factor in all districts in 1996 was that they suddenly got much smaller. For incumbents, that meant that all their Type A organizational infrastructure outside the boundaries of their new district were of no use to them—except perhaps to trade to other candidates, a sensible tactic but one that requires time and energy to negotiate. If the remaining organizational infrastructure in the new district was to be significant at all, it would have to be expanded and strengthened to draw a substantially higher proportion of voters. Such organization-building takes quite a lot of time, energy, and money, particularly since the “low hanging fruit” of people who were easily accessible were already enrolled.
The importance of this point was amply documented in Chapters 3 and 4.
Matsuzawa devoted much more attention to Type A activities than he had in 1993, and they were much more important to him. He arranged an organizational trade with a former rival, he worked to secure the Soka Gakkai vote made available to him with his NFP affiliation, and in particular he gained the support of an important local politician.
He continued his trademark Type B activities, such as his tiring use of outdoor speeches, but relatively speaking he had switched to the Type A strategy. In the case of rural Fukushima, Genba did not have as much distance to make up because his 1993 campaign had already been quite Type A, but he too devoted himself to making an organizational
women’s organization led by his wife.
Otake’s book and the other literature acknowledging disappointment after the reform offer ample evidence that this effect was quite common nationwide, although to our knowledge its transitory nature has not been emphasized—most analysts took it as an indication that moving to SMDs would not have the expected impact on campaigning even over the long term.
d. “Hometown advantage” The new smaller districts actually give greater weight to incumbents’ “natural” base of support in the district. This final point is not transitory, and it has not drawn systematic attention. When the old big districts were broken up, politicians naturally wanted to run in the portion of their own district where they were strongest, typically the area where they grew up and/or currently live. Indeed, some of the knottiest problems for the LDP in 1996 were cases when two of its incumbents were from the same area. Since both fiercely contested the official nomination for the SMD, the party was forced to make many “Costa Rico” deals so they would trade the SMD and PR seats. Not a few of these deals backfire, which, as described in Chapter 4, is how Genba got his SMD seat back in 2000.
Most incumbent candidates in 1996came to in districts in which their home towns account for a much higher proportion of the electorate than in their old, large districts.
The “hometown boy” and “friends and neighbors” effects (Key, 1949; and Black and Black, 1973) are particularly salient in the new election district in light of the higher proportion of voters who attended the same schools as the candidate, have other
important is that any connection whatsoever makes the voter more accessible to being drawn into the candidate’s kouenkai. Similarly, leaders of groups and local politicians in that area are more likely to have had contact with the candidate or his family.
The implication of this point on localism goes beyond the importance to individual candidates and applies to the nation as a whole. After all, as an extreme thought experiment, imagine an electoral system where districts averaged about 1000 voters each.