Why Vote-seeking Parties May Make Voters Miserable

Abstract

This paper explores an unexpectedly hard question for representative democracy, which is usually thought to be enhanced when political parties compete with each other for the support of voters in free and fair elections. Defining optimal representation in terms of the probability that a voter will find a party to support at election time that promotes a policy position close to his/her ideal point, it transpires that vote-seeking parties do not deliver optimal representation. They tend to locate ‘too close’ to the centroid of voter ideal points to minimize the aggregate distance between party positions and voter ideal points. Instead, optimal representation in this sense will be delivered if parties set policy positions using ‘aggregator’ rules, which seek to represent the views of current party supporters, but not to attract new supporters. One policy implication is that the representativeness of inter-party politics is increased by enhancing the representativeness of intra-party politics.

Keywords:

• representation
• party competition
• elections
• agent-based models

Notes

Named after the Russian mathematician Georgy Voronoi (1868–1908); a student of Andrey Markov.

By ‘very good’ in this context we mean solutions with a route length for the salesman that is close to that of the optimal route for a particular solved case.

Laver and Sergenti (2012) provide formal definitions of these rules, and computer code for implementing them.

Author's calculation.

The possibilities were Hunter, Aggregator and Sticker. Other work by Laver and Sergenti endogenizes rule reproduction probabilities as a function of the past success of each rule, for a much-expanded set of decision rules.

Other party system parameters, discussed by Laver and Sergenti (2012), also have a significant effect on the number of surviving parties. The scale of such effects, however, is nowhere near as large as that of the survival threshold, shown in Figure 4.