Changes in the Vote Margins for Congressional Candidates: A Specification of Historical Trends

Abstract

Recent research on House elections has focused on the decline in congressional competition since the mid-1960s. However, this body of research suffers from three major problems: 1) an imprecise specification of the nature of change in the degree of congressional competition, 2) its exclusive reliance on a limited time frame, which calls into question the generalizability and adequacy of its explanations, and 3) its overemphasis on incumbency-oriented explanations. Using a Multiple Interrupted Time Series (MITS) analysis of trends in aggregate electoral data from 1824 to 1980, we find that current low levels of congressional competition are not historically unique, but instead constitute a continuation of a long-term trend established in the mid-1890s. In addition, and in contrast to the general thrust described in the literature, there appears to be a post-1965 trend toward greater competition, although the absolute level of competition is lower after the mid-1960s. Finally, a disaggregation of electoral results by incumbency status suggests a growth in the incumbency advantage since the mid-1890s, but that lower post-1965 levels of congressional competition are the result of a winners' and not an incumbency, advantage. Such findings cast doubt on the adequacy of explanations that focus on post-1965, incumbency-oriented factors.