Made in Congress? Testing the electoral implications of party ideological brand names

We investigate the connection between legislative parties and election outcomes, focusing on ideological party brand names that inform voters. If the source of information conveyed by brand names is the party’s aggregate roll-call record, then changes in legislative party membership should influence election returns. We formalize the argument with an expected utility model of voting and derive district-level hypotheses, which we test on U.S. House elections from 1952 to 2000. We test alternative specifications that vary with respect to the specificity of voter information and find that party positions and heterogeneity both affect vote share independently of incumbents ’ positions. The results provide modest support for the expected utility model but nevertheless suggest that Congress is an important source of the public’s beliefs about the parties, and this effect is clearest for challengers, rather than incumbents, who run under the party’s label. P olitical scientists and democratic theorists have long been concerned with understanding the role of parties as intermediaries between the people and their representatives. Decades ago, Stokes and Miller asked whether a necessary condition in the theory of responsible party government was satis-fied: ‘‘To what extent do public reactions to the legis-lative records of the major parties influence voter’s choices among party candidates...?’ ’ (1962, 531). They concluded that legislative parties and voters are disconnected for two reasons. First, it is difficult to attribute any record of accomplishment to legis-lative parties that are a ‘‘cacophony of blocs and individuals’ ’ (545).1 Second, voters lack the necessary knowledge about political affairs to make informed decisions. While acknowledging the two problems identi-fied by Stokes and Miller, a growing body of rational choice theorizing rooted in Downs ’ (1957) seminal work argues that the importance of parties is as the producers of brand names (Aldrich 1995; Cox and McCubbins 1993; Kiewiet and McCubbins 1991