For those interested, there is a piece by Thomas Rochon "Electoral Systems and
the Basis of the Vote: The Case of Japan" in J. Campbell (ed.), Parties,
candidates and voters in Japan (Univ. of Michigan Papers in Japanese Studies,
no. 2) which argues that even under the multimember constituency system, the
tendency to vote based on the personality of the candidate increased as the size
of the district increased. As a result, it was the urban districts which
displayed the greatest tendency for personal voting because they tended to have
the most seats per district. One would have expected the rural voters to display
more "traditional" modes of personalistic focus, whereas the opposite was shown
in the data analyzed by Rochon.

The comparison with 19th century England and Japan is problematic for a variety
of reasons but the most important is the clarity of lines of cleavage. In
England, the Liberals represented a clear set of socio-economic interests
concentrated in certain areas of the country (esp. the North) and urbanization
created increasing electoral pressure for change. It is an exaggeration to
suggest it was a clear split between aristocrats in the countryside and urban
industrialists, but there was certainly more of a clear division around two sets
of competing interests than can be found in postwar Japan. In fact, the amount
of industry in some nominally rural prefectures in Japan is astounding. In
addition, many Japanese "workers" in these prefectures are also part-time
"farmers", there are locally influential farmers in most urban areas, and even
the purest urban "salaryman" may have elderly parents who are dependent on farm
income (sudsidized) to supplement their retirement income. For most of the
postwar period, increasing urbanization under a multi-member constituency system
has merely meant the fragmentation of the opposition and even provided the LDP
with seats where they would not have had a chance under a single-member system.

When I met with Steve Reed in December last year, he argued that the main
cleavage under the new electoral system will probably be the urban-rural
cleavage. I am not so sure, but it would be interesting to hear his case for it.
Certainly the new districting system will have an impact. Yet, the opposition to
the LDP in the rural areas, the JSP, is doomed to disappear, and the major
conservative opposition party, the New Frontier Party, contains rural voters
supporting former LDP/former Japan Renewal Party MPs, Komeito and DSP semi-urban
voters, and the volatile urban voters who voted for the new parties (Japan New
Party, Japan Renewal Party, etc.) in the 1993 election but may have abandoned
the major parties altogether if the results of the Tokyo and Osaka mayoral
elections are evidence of voting trends. The Upper House election was relatively
good news for the New Frontier Party, but one wonders how long they can continue
to espouse policies (on public works, farm subsidies, taxation, etc.) which try
to please everyone, including both urban and rural interests.