How Widespread is Strategic Partisan Voting in Congress? Revisiting “Backward Induction in the Wild”

“Backward Induction in the Wild” by Spenkuch, Montagnes, and Magleby (BIW) claims that the US Senate’s alphabetical voting procedure affects legislators’ votes. Voting earlier allows senators to defect from their party to vote their district or personal preference. Its findings largely stand alone as evidence of widespread strategic roll call voting. I examine whether roll call voting in other time periods, legislatures, and research designs exhibits the same pattern. I find no evidence in recent congresses; in the US House or three state legislatures; or in a higher-powered research design in the US Senate. Replicating BIW’s original analysis reveals several limitations which raise questions about its findings. Finally, with a new dataset that matches voter preferences to senators’ votes on some of the most salient bills of the past two decades, I find no cases where a senator votes against their constituents and is pivotal. I conclude that ordered voting does not contribute meaningfully to partisan voting by influencing legislators’ perceived pivotality. To the best of our knowledge, our finding that senators’ choices depend on the probability of casting a pivotal vote is one of the first systematic pieces of evidence to suggest that tactical voting is, in fact, widespread. (Spenkuch et al., 2018, p. 1976).