8
proportional representation (PR), systems. Under this legacy, the term mechanical effect has
subsequently become equated with the systematic biases that occur in translating votes into
seats, the degree of which varies significantly between the two electoral systems in affecting
the fate of third parties. However, once we begin disengaging our focus away from
contrasting SMD and PR as the two ideal-type electoral systems, it becomes self-evident that
the (given) vote count and the much-studied vote/seat translation biases are not the only
determinants of the legislative seat distribution under any electoral system. Problems known
as malapportionment, for example, reflect the mechanical distortion inherent in geographical
districting practices, and they directly affect partisan seat share and thus government
formation and distributional policies (Ansolabehere, Snyder and Gerber 2002; Cox and Katz
2002). More generally, in so far as mechanical workings propel voters and other political
actors to behave strategically, a host of institutional properties may affect, at least indirectly,
the distribution of seats via these actors’ strategic responses.
In the case of the SNTV used in postwar Japan, any assessment of its institutional
mechanism would be incomplete unless it takes into account the number of candidates
endorsed by majority-seeking parties, in addition to the actual votes accrued by these parties.
As is now widely known, Japan’s SNTV was a system where a voter cast a single
non-transferable vote to elect typically three to five representatives in each electoral district.
Accordingly, the distinct characterization of this system was that it induced two, seemingly
conflicting, strategic reactions, one that compelled individual candidates to cultivate
personalized votes in order to compete against rivals from the same party, and the other that
forced a majority-seeking party, like the LDP, to centralize its nomination process in order to
Convention