How do I read the New Komeito loss in the Upper House election? That's the latest assignment from my favorite Japan analyst. As per request, I am posting my homework here.
Somewhat overshadowed by the coalition's loss of an Upper House majority, as well as the LDP's first ever loss of a plurality since the Great Conservative Merger of 1955, is the heavy setback that junior coalition partner New Komeito also suffered. There seems to have been some defections from the NK's Sokagakkai constituency, and the NK's ability to throw Sokagakkai support behind the LDP where it is not fielding its own candidates (all single- and two-seat districts as well as one three-seat district) has been called into question. In the aftermath of the debacle, the NK and/or the LDP, depending on whom you listen to, was rethinking the value of the coalition.
This is important if anything comes of it because, depending on the results of the next Lower House general election (coming no later than October 2009 and possibly as early as late spring 2008), the NK could hold the casting vote in deciding the new prime minister. However, from an election perspective, the LDP continues to make more sense as the NK coalition partner. Moreover, practical concerns make DPJ-NK cooperation in the next two Upper House elections difficult, if not impossible. Thus, barring a fundamental realignment of the Japanese political parties, the LDP-NK coalition will hold for the foreseeable future. Let me explain…
The coalition brings no obvious benefits in the proportional-seat election, where both parties work to keep their party faithful in line while bringing in as many unaligned and cross-party votes as possible. The NK may have suffered from guilt by association this time around, as polls leading up to the vote showed roughly one-tenth of NK supporters intending to cross over to the opposition, and the actual NK votes did drop from 8.62 million (15.4% of all votes cast) to 7.77 million (13.2%). (It won 8.19 million, or 15.1%, in 2001). The LDP by contrast, managed a much smaller drop, from 16.80 million (30.0%) to 16.54 million (28.1%). I believe that the NK has a core Sokagakkai constituency of around 8 million, and that a small percentage of this religious faithful as well as a very small fraction of non-Sokagakkai floaters are willing to move their votes around. But despite what must have been the drop off in floater and religious faithful vote for the NK, it managed to keep seven seats, down only one from 2004 (and 2001). So the downside for the NK has been limited.
The benefits accrue to the coalition in the local-seat election, where the seats are apportioned to the provinces according to the number of eligible voter residents with each province receiving a minimum of one seat. The value of the LDP-NK coalition for the LDP is in the support the NK gives to the LDP candidates where the NK is not fielding its own candidates. This is most evident in the single-seat or two-seat districts, where the NK has no realistic prospects of winning on its own. (The NK has in the past chosen not to field candidates in some three-seat districts.) The value to the NK is in the support the LDP can throw to the NK candidate in any three-seat district where the LDP fields only one candidate. The LDP does not have realistic prospects of winning two seats on its own in these three-seat districts because they are urban provinces, where the traditional LDP support base of farmers, public works-dependent construction firms, the local post office and others is less effective.
In the July election, one province had five seats, five had three, and twelve had two, while the rest (29) had one each. In the provinces where collaboration occurred (one-, two-, and three-seaters), both the LDP and NK received significantly more votes for the local election than they received for the (national) proportional election. The crossover effect is easier to discern for NK candidates. Since the NK has very limited appeal beyond its institutional Sokagakkai support base, I think that it is fair to assume that most of the additional votes for the local NK candidate came from LDP supporters.
NK help for the LDP candidate in the single- and two-seat elections is harder to discern because the difference between the proportional and local votes varied wildly. The more personal nature of the support for LDP candidates, as well as the resultant potential for defections from the LDP support base, undoubtedly contributed to this variance. However, there is enough of a gain to make me believe that the NK support did substantially help LDP candidates both successful and unsuccessful in receiving more votes than they would have received on their own. The coalition did work, if not to the total satisfaction of the two parties.
But what would the electoral prospects of the NK in a coalition with the DPJ? Not as good as the deal they have now. The DPJ has a much smaller support base than the LDP, and is consequently more reliant on the unaligned "floater" votes. In other words, the DPJ has less control over the number of votes it can throw to the NK. Moreover, there are practical concerns that virtually preclude collaboration in the 2010 and 2013 general elections.
In a perfect world, the NK would be able to keep its coalition benefits by fielding the second coalition candidate in the three-seat districts. But the DPJ will have two incumbents in two of the three-seat districts in the 2010 election in question, while both the DPJ and NK have one incumbent in the other. Chiba will step up from a two-seat to three-seat district. But the NK was not strong enough to field a candidate there in the July election, when the DPJ won two seats. As for the other three-seat district, Osaka, the single LDP and KN candidates will have their hands full in a four-way fight with the DPJ and JCP. In other words, there is little room for DPJ-NK coordination to come into play in the 2010 Upper House election in the all-important (to the NK) three-seat districts. The NK has nothing to gain from leaving the coalition even if the 2010 election were the only general election under consideration, and prospects look even worse in 2013, when the DPJ will be fielding two incumbents in four of the five (and one incumbent in Osaka, where it will likely again face a four-way single-candidate fight, leaving no room for collaboration).
On a substantive policy level, some people think that the DPJ is a more natural fit for the pacifist Buddhism that underlies the NK foreign and security policy. It is true that Ichiro Ozawa's stand against the renewal of the anti-terror law and his emphasis on the importance of the UN is much more than a mere tactical choice to position the DPJ in opposition to an unpopular coalition policy to support the US in Iraq (and, to a lesser extent, in Afghanistan), or to appease the still significant support from the traditionally pacifist, politically anti-American labor unions. However, Mr. Ozawa's shelf life could be fairly limited, if only because of his health problems, and there is reportedly substantial mistrust between Mr. Ozawa and the NK. Moreover, the younger, more libertarian generation of LDP Diet members, represented by people like the previous pro-US DPJ leader Seiji Maehara and current policy head Takeaki Matsumoto, prefer a much more muscular security policy. There is no reason to believe that, over the long-term, the NK will find more ideological comfort in collaboration with the DPJ.
Barring a fundamental realignment of the political parties, the LDP-NK coalition, now in its ninth year, shall continue.
A quantitative analysis should yield a better understanding of the benefits of the coalition. More importantly, my conclusions are based solely on the analysis of the Upper House general elections. I do not expect my conclusions to be significantly altered by an analysis of the Lower House general elections, but someone will have to pay me to do that. I need to use less user-friendly databases for that, and the analysis itself promises to be more complex due to the more complicated nature of the Lower House electoral system.
The NK suffered a big loss in the local-seat election, going in with five seats and escaping with only two. Three of the four NK incumbents standing for election in four of the five three-seat districts lost. (The NK did not field a candidate in Chiba, the fifth three-seat district.) In all three cases, the big winner was the DPJ, which took two seats in each of them, while the LDP managed to keep one apiece. The three-seat districts are all urban districts, where the LDP has not fared well of late. Here, the coalition fielded two candidates, one each from the NK and LDP (the exception being Chiba, where the LDP fielded two candidates and the NK none). But the DPJ won two seats in each of the three three-seat elections that it contested with two candidates. The coalition did a good job of vote allocation - the NK received considerably more votes in the three- and five-seat elections it contested than it did in the same prefectures for the proportional seat election and the LDP seemed to have benefited from NK support in districts where the NK did not field a candidate (though the benefit to the LDP is more difficult to gauge from publicly available data). But in each of the three cases, the JDP outpolled the LDP-NK coalition total by a wide margin, and in only one of these cases could even a perfect 50-50 split of the cumulative LDP-NK vote have lifted the NK candidate above the successful DPJ candidate with the fewer votes. Unfortunately for the coalition, the shift in voter sentiment swamped the positive effects of the vote sharing.