Why Turnout Buying Seems Deadlocked
Political machines traditionally rely on targeted rewards to encourage supporters to vote. Yet meta-analyses of experiments and observational studies find little to no effect of these selective incentives on turnout, creating a puzzle: are machines irrational, or are empirical methods misleading scholars? Kenneth Greene tackles this puzzle in a BJPS article by arguing that common measurement choices obscure the true impact of turnout buying.
Greene’s Argument
Greene shows that many empirical studies use compound measures that collapse three distinct tactics—turnout buying (rewards conditional on turnout), abstention buying (payments to stay home), and vote-choice buying (rewards conditional on casting a particular ballot)—into a single category. Data that mix these strategies produce countervailing and null effects that bias estimates toward zero, making turnout buying appear ineffective even when it matters in practice. He further argues that machines rationally diversify tactics, which makes these compound measures especially misleading.
How the Claim Is Tested
- Proposes straightforward alternative measurement strategies that separate turnout buying from other clientelist practices.
- Evaluates these alternatives with new survey data and a constituency-level analysis of machine strategy in Mexico, comparing results from compound measures versus disaggregated measures.
Key Findings
- Lumping different forms of buying together masks the positive effect of turnout-targeted incentives by averaging out opposing effects.
- Machine incentives to mix strategies make standard compound indicators likely to understate turnout buying’s impact.
- Disaggregated measurement approaches produce clearer evidence that turnout buying is more effective than previous meta-analyses imply.
What This Means For Research and Practice
Greene’s paper narrows the gap between theory and evidence on machine politics: when scholars measure clientelist tactics separately, turnout buying reappears as a rational and effective component of electoral strategy. The study provides concrete, implementable measurement fixes and empirical demonstrations—especially using Mexican constituency data—that should reshape how political scientists quantify clientelism and interpret null results in voter-mobilization research.
