Are Caucuses Bad for Democracy?

On 4 March 2008, Texas held both primary elections and caucuses statewide to select delegates to the Democratic National Convention. This unique, hybrid procedure, dubbed the “Texas Two-Step,” took place on the same day and was open to the same universe of voters, but the similarities did not extend much further. Participation in the primary, in which nearly 2.9 million ballots were cast, vastly exceeded turnout in the caucuses, which attracted an estimated 1.1 million voters across the state. This is not atypical for caucuses, which tend to attract fewer participants than primaries. More crucially, the two elections yielded different outcomes. With 50.9 percent of the vote, Hillary Clinton bested Barack Obamaʼs 47.4 percent in the primary, but Obama won the caucuses with support from 56.2 percent of participants, compared to Clintonʼs 43.7 percent. The results in Texas mirrored a more general pattern in the 2008 contest for the Democratic nomination, in which caucus participants favored Obama while primary voters were more favorable to Clinton. In the end, Obama won in 14 out of 16 caucus states, while Clinton was victorious in 22 out of 39 primaries. Are such differential outcomes byproducts of systematic differences between primary elections and caucuses? If so, is one system of preference expression superior to the other? These questions and the results observed in the 2008 cycle highlight the impact of institutional variation on voter preferences and election outcomes. Put more bluntly, the rules of the game matter. Scholars and practitioners alike have acknowledged this reality, and have grappled consistently with evaluating the effects of electoral institutions and implementing reforms accordingly. In