The Origins and Impact of Votes for Third-Party Candidates: A Case Study of the 1998 Minnesota Gubernational Election
Lacy, Dean, Monson, Quin, Political Research Quarterly
We estimate a multinomial probit model of vote choice and turnout to examine the 1998 Minnesota gubernatorial election. Like supporters of recent third-party presidential candidates, voters who elected Jesse Ventura tended to be young, male, lower in education, liberal on social issues, and fiscally conservative. Ventura support was not due to a general dissatisfaction with U.S. government, but it was correlated with voter dissatisfaction with Minnesota state government. Ventura was the Condorcet winner in the election; Hubert H. Humphrey was the Condorcet loser. With Ventura out of the race, Norm Coleman would have beaten Humphrey by approximately ten percentage points. Coleman voters overwhelmingly preferred Ventura to Humphrey, but Humphrey voters preferred Ventura to Coleman by a slim margin. Ventura's candidacy added seven percentage points to the turnout rate. Under full turnout, the vote shares of the candidates would not have changed significantly
Strong third-party candidacies are becoming commonplace in U.S. national elections, particularly at the presidential level. In four of the eight presidential elections between 1968 and 1996, a third-party candidate won at least 5 percent of the popular vote. No other thirty-year period in U.S. history has witnessed such third-party performance.1 Ross Perot's popular vote of close to 20 percent in 1992 was the highest ever recorded by a third party candidate who had no prior political experience, and his nearly 9 percent of the popular vote in 1996 was the highest recorded by a repeat third party candidate. In 2000 the impact of third-party candidates became even more significant: even though Ralph Nader polled barely 2 percent of the popular vote, his candidacy likely took enough potential Gore votes to give George W Bush an Electoral College victory.
Third-party success at the presidential level pales in comparison to third-party success in recent gubernatorial elections. While the 1998 election of Jesse Ventura in Minnesota may seem unusual, three other third-party candidates won gubernatorial elections in the 1990s: Walter Hickel in Alaska (1990), Lowell Weicker in Connecticut (1990), and Angus S. King, Jr., in Maine (1994 and 1998). At the presidential level, only Ross Perot's short-lived lead in the polls during the summer of 1992 approaches the success of third-party candidates in gubernatorial elections.
Third-party candidacies in gubernatorial elections are an understudied phenomenon in American electoral politics. The existing research on third-party candidates in US elections has focused on the presidential level (e.g. Abramson et al. 1995; Alvarez and Nagler 1995, 1998; Gold 1995; Herron 1998; Lacy and Burden 1999, 2000; Rosenstone, Behr, and Lazarus 1996) while recent thirdparty gubernatorial and congressional candidacies have received relatively little attention. Beiler (1999) and Frank and Wagner (1999) examine the candidacy of Jesse Ventura in the 1998 Minnesota gubernatorial election. Reiter and Walch (1995) describe voter support for James Longley (elected Governor of Maine in 1974), Lowell Weicker, and Bernard Sanders (elected to the U.S. House from Vermont in 1990). Donovan, Bowler, and Terrio (2000) study voter support for minor-party candidates in the 1994 California gubernatorial and senatorial elections, and Magleby, Monson, and Walters (2000) examine support for Merrill Cook's 1994 independent run for Congress in Utah. However, none of these studies have examined the impact of the third-party candidates on the vote shares of the other candidates or on turnout.
We examine Jesse Ventura's victory in the 1998 Minnesota gubernatorial election in order to compare the electoral origins and impact of his support to that of third-party presidential candidates. Using recent advances in the study of voting with three or more candidates (Alvarez and Nagler 1995, 1998; Lacy and Burden 1999, 2001), we develop a unified multinomial probit model of vote choice and turnout to answer four questions central to the literature on thirdparty candidacies. …