11
6. Types of Campaign Expenditures
Campaign expenditures for Japan’s Lower House are disclosed for both the
official campaign as well as the year-based period. To date, much of the scholarship on
Japanese elections has relied on the official campaign-period figures.
7
Building from the
insights of this scholarship, I utilize newly collected figures for the year-based period.
8
The effects of own and rival expenditures on vote shares should be similar for both types
of figures. However, the strength of the effects may be related to how candidates and
parties have strategically adapted to the specific formal regulations shaping the use of
money in the new system.
First, the figures for the official campaign-period expenditures are limited to a 12-
day period prior to the actual day of the election. Second, unlike the figures for the year-
based period, campaign-period expenditures are subject to campaign spending limits.
9
Both of brevity of the period as well as the spending caps suggests that the effects of
campaign-period data on vote shares may be limited. However, Kawato (1999 & 2002)
empirically shows that the level of campaign-period spending is partly a function of how
much the candidate spends during the year-based period. If one believes there is a
reciprocal link between both types of expenditures, the effect of campaign-period figures
on votes may not only approximate year-based figures, but also respond to the
hypothesized relationships between own and rival expenditures on vote shares.
7
The number of scholars using campaign-period spending data is considerable, including Cox and Thies
(1998, 2000); Horiuchi (1999); Reed (1999); Cox, Rosenbluth and Thies (1998) and Carlson (2001).
8
For specific details about the collection of this data, see Carlson (2003).
9
In the 1996 election, for example, the limit averaged ¥24 million ($226,000) for each single-member
district. Typically, the spending limits are set high enough that many politicians fail to spend the maximum
amount.
Convention