Partisan Effects of Voter Turnout in Senatorial and Gubernatorial Elections

Conventional wisdom holds that higher turnout favors Democrats. Previous studies of this hypothesis rely on presidential and House elections or on survey data, but senatorial and gubernatorial elections offer better conditions for directly testing turnout effects in U.S. politics. In a comprehensive analysis of these statewide elections since 1928, we find that the conventional theory was true outside the South through 1964, but since 1965 the overall relationship between turnout and partisan outcomes has been insignificant. Even before the mid-1960s, the turnout effect outside the South was strongest in Republican states and insignificant or negative in heavily Democratic states. A similar but weaker pattern obtains after 1964. In the South, which we analyze only since 1966, higher turnout helped Republicans until 1990, but in 1990–94 the effect became pro-Democratic. The conventional theory cannot account for these complex patterns, but they are impressively consistent with DeNardo's (1980) theory.