Yes, structurally low turnout favors the right

Abstract

ABSTRACTIt is well known that socioeconomically disadvantaged citizens tend to vote less than other citizens. Does this cause democratic politics to systematically favor conservative causes? Surprisingly, the conventional wisdom in a very large scholarly literature says “no.” Whether turnout is high or low, according to numerous studies, makes no substantive difference on political outcomes. This article refutes the scholarly conventional wisdom on theoretical and empirical grounds. It points out that the evidence marshaled in the mainstream turnout debate focuses almost exclusively on partisan vote shares as the dependent variable and “nonstructural” variation in turnout. Unfortunately, many studies do not consider the policy consequences of turnout, a more substantively interesting variable for studying patterns of disadvantage that may result from unequal voting rates. Moreover, conventional research designs do not distinguish between structural changes in turnout caused by legal changes to the voting regime and nonstructural changes precipitated by partisan mobilization, candidate valence, or other election-specific factors. This distinction has substantial implications for understanding turnout's impact. The article employs nonparametric meta-analytic methods to synthesize a large number of empirical studies and demonstrates that low structural turnout does indeed favor the right in theoretically predictable ways.