22
statistically insignificant.
21
Turning to the DPJ, I was unable to uncover any statistically
significant relationship between rival expenditures and vote shares. Overall, the
relationship between rival expenditures and vote shares did not appear to be very strong
in either party.
The results of the empirical tests suggest that money continues to be an important
factor in the quest for votes in Japan’s SMD tier. However, it would appear that the actual
correlations between money and votes is relatively small. Thus, one might ask why
politicians raise and spend large sums of money in the first place. One scholar has
ventured an answer related to the value of winning election. Levitt (1994: 796) argues that
“…The opportunity costs of raising funds may be very low compared to the value of
winning an election, so even if there is a small probability that spending affects the
election outcome, it is worthwhile.” In this regard, money in the new system would appear
to be central to the individual efforts of candidates to capture votes and to hopefully win
election or re-election.
21
The most significant correlation uncovered was between rival fund agent spending and LDP vote shares
in Table 2, where each additional spending of ¥100 ($.94) spent by the main rival is related to a loss of 3%
of the LDP district vote share.
Convention