The electoral system combines single-member district (SMD) seats allocated by plurality rule and proportional representation (PR) seats separately allocated in 11 regional districts. Until this year, there were 300 SMD seats and 180 PR seats. For 2014, five rural SMDs were eliminated to correct for malapportionment. The table below shows the vote and seat totals in both the SMD and PR tiers of the system for 2012 and 2014. We can see that the most substantial shift in votes was away from the Third Force parties, who contested far fewer districts than in 2012. The LDP-Kōmeitō coalition increased its SMD vote share, but won 14 fewer SMD seats. The DPJ took home a slightly smaller share of SMD votes than in 2012 but won 11 more seats because, unlike 2012, its votes were now more concentrated in districts without Third Force spoilers. In the PR tier, the Coalition, the DPJ and the JCP all benefited from shifts away from the Third Force parties, although their combined gains (3.75 million votes) were small compared to the decline in turnout (6.85 million votes). Perhaps most important, in an election in which millions more voters than ever before chose to stay home, the LDP-Kōmeitō coalition held on to its vote base from the previous election, while the non-communist opposition as a group lost votes that it needed to compete with the governing parties.
After the 2012 election, we argued that the LDP’s landslide was not due to a vote swing in its favor. Rather, dissatisfaction with the 2009-2012 DPJ government led to multiple defections and the establishment of new parties, and this fragmented party system spoiled potential DPJ pluralities in numerous districts and handed cheap wins to the LDP. Low turnout also hurt the DPJ, which has a weaker organizational support base than the LDP and Kōmeitō.
In 2014, despite the surprise of the election, opposition parties did a much better job coordinating their nominations in SMDs. The following figure shows the average effective number of candidates per SMD for the seven elections held under the current electoral system, divided into three groups based on population density. After a steady decline through 2009, the effective number of candidates spiked in 2012, as the anti-LDP camp splintered. The 2014 results show a strong correction.
The next table shows the number of SMDs in which the DPJ and the Third Force parties (together, the “Big Opposition”) fielded no candidate, a single candidate or more than one candidate. The 188 single-candidate endorsements would seem to be cases of optimal coordination, while the 60 districts with more than one are coordination failures. But what about those 47 districts in which the Big Opposition didn’t enter the field at all? Were these missed opportunities to take seats from the LDP?
We think not. In seven of those 47 SMDs, the incumbent going into the 2014 election was from a small party or was an independent. Any Big Opposition entrant might have become a spoiler and handed the seat to the LDP. In the remaining SMDs that lacked Big Opposition candidates, on average in 2012 the vote total of the LDP or Kōmeitō candidate was 41 percentage points greater than the combined vote total of all Big Opposition candidates running in the district! It would seem that those 40 Big Opposition absences, then, were reasonably calculated to avoid wasting resources in government strongholds.
In the districts where more than one opposition candidate did run, might better coordination have deprived the LDP/Kōmeitō candidates of victories? How we answer this depends on how we treat the JCP. The JCP has a long history of non-cooperation with other opposition parties. Should we include as coordination failures the districts in which the Big Opposition fielded a single candidate, but the JCP candidate was a spoiler? Or should we take the JCP vote as inevitable and only count as coordination failures those districts in which the Big Opposition vote together exceeded the Coalition vote?
The final table (below) displays the calculations for all scenarios (with the heroic assumption that voters wouldn’t cross the Government/Opposition line or alter their turnout decisions). The most important finding is that unlike 2012, when the failure of the non-communist opposition to coordinate on a single candidate might have handed the LDP/Kōmeitō as many as 115 SMD seats, that failure cost a much less severe 31 seats in 2014. (Just as in 2012, cooperation failure that helped produce LDP/Kōmeitō victories was most frequent in urban districts.) However, the JCP, which won many more votes than in 2012, was the spoiler much more often in 2014: coordination between the Big Opposition and the JCP could have clawed back from the governing parties an additional 62 seats (i.e., 84 seats, rather than just 22) over what the Big Opposition might have done on its own in 2014, compared with only 22 (i.e., 134 seats, rather than 112) in 2012.
The LDP and Kōmeitō won the 2014 election because they were able to hold on to their core supporters, while abstentions most hurt the non-communist opposition. To close the gap in the future, the opposition will have to continue to coordinate as it did in 2014, and it will have to find a way to drive up turnout among the millions of voters who supported it in 2009, but stayed home in 2012 and, especially, 2014.
A more complete analysis of the 2014 poll will appear in Japan Decides 2014, to be published by Palgrave Macmillan next year.