From 1947 through 1993 Japan used what they called the 'medium-sized district system' to elect
the more powerful lower house of the Diet. This system can be best explained as a system of
multi-member-districts (MMD), the simplest extension of single-member districts (SMD). Each
voter gets a single vote and the top M candidates are declared elected, where M is the number of
seats allocated to the district. In precisely the same sense that SMD is a first-past-the-post
system, MMD is an Mth-past-the-post system. M ranged from one through six, with most
districts between three and five. The Japanese system produced campaign styles similar to that
of the single transferable vote (STV) in Ireland and is also called the single non-transferable vote
(SNTV).
Complaints about MMD focused first on the fact that strategic errors often produced anomalous
outcomes. MMD was one of the rare electoral systems in which a party could lose a seat because
of a particularly popular candidate: a party with enough votes to win two seats in a three-member
district might finish first by a large margin but leave only enough votes for the second candidate
to finish fourth, thus winning only one seat. More importantly, parties could lose seats by
running too many candidates: if the party in the previous example ran three candidates, splitting
the vote evenly among them, they could easily finish third, fourth, and fifth, winning only one
seat. Such anomalous outcomes, however, were never common, declined over time, and
occurred primarily when a party's support rose or fell drastically at a single election.
The primary problem with MMD was that it forced candidates from the same party or the same
camp to compete for votes. The dominant party throughout this period, the Liberal Democratic
Party (LDP), ran multiple candidates in most districts. An LDP candidate who needed a few
more votes to gain a seat found it easier to attract voters from other LDP candidates than from
one of the opposition parties. Intra-party competition fostered factionalism inside the LDP.
While the factions of the LDP were held together by power, the opposition parties fragmented as
the competed for the progressive vote. The Japan Socialist Party (JSP), the largest opposition
party, found it increasingly difficult to field multiple candidates. MMD tended to mute political
competition between ideological camps over policies and to promote competition within camps
over constituency service.
Since the party in power possesses the advantage in any competition to perform every greater
constituency service, MMD facilitated one-party dominance. Constituency service also
contributed to serious corruption problems which, in turn, finally led to the defeat of the LDP in
the 1993 election and an electoral reform, which changed the electoral system. Both the process
and the result resemble those in Italy.
The new system is the simplest of the parallel PR/SMD systems, which have proven so popular
recently. 300 seats are elected from SMDs and 200 from PR in eleven blocs ranging in district
magnitude from seven to 33. There is no mixing of SMD and PR votes. The only connection
between the two parts of the system is the double candidacy provision, that a candidate may run
both in an SMD and be included on a PR list. Parties may also rank several candidates equally
on their lists, allowing the tie to be broken by the SMD results.
The double candidacy provision became the focus of criticism after the first election under the
new system in 1996. The original complaint was that candidates defeated in an SMD were
'revived' by PR and making them 'zombie Dietmen.' A more serious problem is that some
districts now have one representative while others have two, the winner of the SMD and the loser
who won in PR. A few districts even have three with two SMD losers winning PR seats.
The process leading to political reform involved defections from the LDP, the decline of old
parties, the birth of new ones, and a rapid succession of unlikely coalition governments. It is still
difficult, after only one election, to distinguish between the effects of the new electoral system
and the effects of this political turmoil.
The clearest difference is that political parties now have only one candidate per district. This
marks a major change for the LDP. In the past, when more than one candidate sought the LDP
nomination, the decision was often left to the voters: let all the candidates run as independents
and whoever wins will join the LDP. This option is now much more dangerous. Candidates will
be tempted to accept the support of other parties and the winner might not join the LDP. This
tendency has long been visible in gubernatorial elections and was clear in several districts in the
1996 general election. The LDP is making rapid progress toward appointing a single person as
the head of the local branch, and thus the prospective nominee for the next election, in every
district. There are also indications that the factions have been weakened. The political turmoil
surrounding political reform shook the factions badly and now that candidates no longer need
electoral support against their internal rivals, there is one less incentive to reform them.
Theoretically, the system should produce incentives for the opposition to coalesce around a
single candidate capable of defeating the LDP but so far we have witnessed further fragmentation
of the opposition parties. Either the incentives toward a two-party competition are weakened by
the 200 PR seats or the incentives are masked by the effects of political turmoil and will become
more visible over time.
Finally, will the new system reduce corruption? On the one hand, incentives for creating a
personal vote based on constituency service have been weakened but little if at all. On the other
hand, the potential for policy competition among parties has increased. We will have to wait to
see how these competing incentives interact but so far the major changes relating to corruption
have not been due to the new electoral system but to the new law that electoral violations may be
used to declare an election invalid even if the candidate was not directly involved in the
violations.
