Recontesting Elections by Incumbents in Democratic Legislatures in the World

• In Event: The Relationship between the Public and Parliament

Abstract

Theories of democratic accountability rest on the assumption that politicians want to retain public office. This goal makes them susceptible to the fear of losing office, which may occur if their performance proves unsatisfactory to voters. Thus, politicians who do not seek reelection (or who are prevented from seeking reelection by their political party) cannot be held accountable while in office.

Yet surprisingly little is known about patterns of rerunning by incumbent legislators. For accountability take place, politicians need to rerun for reelection. In the best-studied country, the United States, we know that most national legislators rerun and win again. We also know that this pattern emerged in the 20th century, and that most legislators did not rerun (nor achieve reelection) in the 19th century. Evidence on rerunning in other democracies shows that US rates are probably unusually high and that legislators elsewhere often rerun in relatively low frequencies. However, systematic data on rerunning has been assembled for only a few countries.

In this paper, I assemble data on rerunning and reelection from a sample of countries to assess the relationship between rerunning and reelection under different party organizational, electoral system, and informational environments. Pakistan, India, Brazil, Argentina, and France are among the countries for which I have located rerunning and reelection data over a period of elections (ideally at least six). I also include rerunning and reelection data on a selection of approximately a dozen US states over the 20th century.

A preliminary review of the data shows that in many cases, rerunning and reelection closely track one another. This suggests that incumbent legislators have good information about whether they are likely to win again and run only when they foresee likely reelection. Data from US state legislative elections in a dozen states over the course of the 20th century documents this. The goal of the paper is to explore whether the same is true in contemporary less developed countries, especially those with weak party systems, where incumbents are likely to lack much accurate information about their reelection prospects.

Author

• Miriam A. Golden, Stanford University