Information asymmetries and simultaneous versus sequential voting

Sequential voting takes place when some voters make choices with knowledge of earlier decisions in the same election. Historically, voting in U.S. president primaries is sequential, but recent primaries have been "front-loaded" into the early weeks of the season. We explore sequential voting in drawn-out primaries and simultaneous voting in front-loaded ones theoretically and use laboratory elections to examine our predictions empirically. We find evidence that in sequential voting later voters can use early outcomes to infer information about candidates and make choices that better reflect their preferences. The ability of later voters to infer information increases with higher levels of risk aversion and information provided on early outcomes. We discover that when a moderate candidate is largely unknown, information aggregation in sequential voting can inrease the probability s/he will win, which supports the contention of policymakers that sequential voting can lead to different electoral outcomes.