Abstract
Counting the number of `real' candidates in elections is complicated by the frequent presence of individuals who win a small percentage of the vote and have no realistic chance of winning the election. Increasingly, researchers have solved this problem by using the Laakso—Taagepera `effective N', originally devised to count numbers of political parties (usually in the overall electorate). We argue that when counting candidates in individual districts, the focus should be on viability and that the effective N is therefore inappropriate. We discuss the concept and measurement of candidate viability and propose two alternative measures. We illustrate the use of these counting rules by re-analyzing and updating Reed's (1990) analysis of the number of candidates in Japanese elections under the Single Non-Transferable Vote. In contrast to Reed, we find lower numbers of candidates, more rapid learning of strategic behavior, and closer adherence to the theoretical upper limit of one more candidate than the district magnitude (M + 1).
Keywords
Get full access to this article
View all access options for this article.
