In 2007, I reviewed the twenty-nine single-seat upper house races and offered a prediction (or, rather, a range) that actually proved too optimistic as far as the LDP was concerned.
Rather than review the races in the 300 single-member districts, I've decided that I will look at the state of the races in each of the country's 11 proportional representation blocks, looking at important single-member district races in which bloc and making predictions on a block-by-block basis.
The goal is not necessarily to arrive at a precise prediction in which I have total confidence, but rather to identify important battlegrounds and ultimately address the question of whether the DPJ can win an absolute majority in the House of Representatives. Winning an absolute majority would not spare the DPJ from having to govern along with coalition partners, but it would have practical importance for improving the DPJ's bargaining power with its coalition partners. More importantly, it would have symbolic importance, signifying a clear public mandate for the DPJ to proceed in implementing the policies included in its manifesto — and forcing the DPJ's adversaries in the Diet and the bureaucracy to think carefully about how to oppose the government.
As a prelude to the survey, it is worth recalling just how well the DPJ has to do to win an absolute majority. In 2005, of course, the party suffered a humiliating defeat. Its share of the country's 300 single-member districts fell from 105 to 52, and its proportional representation seats fell from seventy-two to sixty-one. The DPJ's collapse was a result of the vicissitudes of a first-past-the-post electoral system. The DPJ vote share in single-member district voting held roughly steady, falling from 36.66 percent in 2003 to 36.44 percent in 2005. At the same time, however, the party's vote share in proportional voting fell from 37.39 percent to 31.02 percent from 2003 to 2005. Meanwhile, in 2005 the LDP increased its share of SMDs from 168 to 219 on the back of an increase in vote share from 43.85 percent to 47.77 percent. In PR voting, the LDP's share improved from 34.96 percent to 38.18 percent. Komeito's vote share in SMDs fell slightly from 1.49 percent to 1.44 percent, and also fell in PR voting from 14.78 percent to 13.25 percent.
It seems reasonable to treat 2005 as an outlier that tells us little about how the LDP and the DPJ will perform on 30 August. Indeed, I think there is a decent argument from treating its 2003 vote as a baseline for what the DPJ will do in this general election. In 2003, the DPJ wound up with 177 seats to the LDP's 237 seats. In only its second general election, the party gained fifty seats over its 2000 returns.
But it is also worth nothing that the LDP's 296 seats in 2005 was the first parliamentary majority won by the LDP under the new electoral system. In other words, it is hard to win a majority under the current system, although perhaps it is getting easier over time as the two-party-plus system consolidates. Can the DPJ reach the magic number of 241, nearly 130 seats over its 2005 returns?
(Incidentally, for more information on how Japan distributes PR seats, here is the Wikipedia entry on the D'Hondt method, the method used by Japan for allocating seats.)
So with that for an introduction, I will begin the survey in Hokkaido.
Other posts by Tobias Harris: