Saturday, July 25, 2009

What If the DPJ Did as Well as the LDP Did in the 2005 Popular Vote?

What would the overall outcome of the upcoming Lower House election look like if the DPJ do as well in the proportional representation blocs that yields 180 out of the 480 seats as the LDP did in its historic 2005 landslide victory under Koizumi? Not good at all—if you’re rooting for the DPJ. If the DPJ wins only 77 PR seats in the Lower House election, it is hard to see it winning a simple majority. In fact, in that case, I think that there’s a very good chance that the LDP and Komeito combined will end up with more LH seats than the DPJ alone, and that they can even dare to hope for a majority. Here in my view is the reason why:

Let’s assume that all the other parties except the LDP each earn more or less the same proportion of votes (and seats) in the PR blocs give or take a couple of dozen basis points as they did in the previous 2005 election—not an unreasonable assumption if you look at how Komeito, JCP and SDP have fared in the last couple of very different LH elections. The PNP and NPJ only go back as far as the last election, and this time around there will also be a couple of new groups of has-beens and wannabies clustered around Watanabe-Eda and Hiranuma respectively, not to mention a few renegades from the LDP and (surprisingly to me) the DPJ. But let’s assume that all of this will not collectively throw off voting patterns by whole percentage points. If you accept these assumptions, you are also accepting the assumption that the LDP will gain more or less the same proportion of PR votes (and likely seats) as the DPJ did in the 2005 election, which means that we are assuming that the LDP and Komeito will have in the neighborhood of 61 and 23 PR seats respectively, or 84 PR seats between them—7 more than the DPJ by itself.

So what kind of guesses does this lead to regarding the 300 SMDs? Even assuming the kind of significant—say 1.5 million—drop-off in the DP-Komeito coalition vote tally from the PR to AMDs that occurred in the last two elections, the DPJ will be hard put to match the voting tallies of the LDP-Komeito coalition. Of course its two coalition partners will do their best to throw their votes to the DPJ and the JCP will provide some unsolicited help, so the DPJ could very well hold its own against the LDP-Komeito coalition there. Add to that the SMD seats that its coalition partners will win, and the DPJ-led coalition has a good chance of beating its opponent. Still, the DPJ’s path to victory, even with its coalition partners on board, is anything but clear in this scenario.

Of course the DPJ is currently enjoying much larger advantages in public opinion polls than would accompany the kind of voting patterns that would give the DPJ “only” 77 seats. If it can maintain that kind of lead on 30 August, and given the assist from the rest of the opposition, it’s a safe bet that it will win an overall majority on its own.

I’ve only done a rough approximation of the arithmetic in my head, but I’m confident that it will stand up to closer scrutiny. Any serious analysis will of course have to look at the numbers district by district, but hey, I have only one life to live.


Bryce said...

This was covered here in a presentation by Ko Maeda:

He noted, among other things, that the DPJ would need a six percent swing to overtake the LDP. The LDP only had a three percent advantage in terms of the popular vote in 2005.

Jun Okumura said...

Bryce: What I’d like to do is put the 2005 voting results for each of the SMDs on an Excel sheet, fill in the holes where the DPJ did not field candidates in 2005 but will be doing so in this election, and see what the effects of different percentages of across-the-board shifts in the share of votes from the LDP/Komeito candidate to the DPJ candidate would be. A similar way to produce approximations should be available for the PR blocs. Then I would work in some assumptions about the JCP effect. (There must be other important considerations that I’ve missed.) Someone with computer programming skills should be able to create a more dynamic simulation program. I assume that Prof. Maeda’s 3% swing between 2003-2005 is a +3-3=6% swing, not a net 3% one. In that sense, I too find it difficult to find a clear path to a DPJ-led coalition government with just a +6-6=12% swing. However, the polls—unfortunately for the LDP-Komeito coalition—are predicting a much, much larger swing, and it’s hard to see anything short of a tragic ending for Taro Aso’s career (or some reprehensible crime on the part of Yukio Hatoyama just before the vote) that could trigger a sea change in voter sentiment that would significantly narrow the difference.