The 2004 presidential election has been termed the values election. In one widely discussed exit poll, the plurality of voters (22 percent) ranked moral values at the top of their list of concerns, and of that group 80 percent voted for George Bush. In addition, 11 states passed ballot questions that wrote bans on gay marriage into state constitutions. These referenda, according to some analysts, galvanized the Christian right, mobilizing voters who might otherwise have stayed at home. They came to the polls to strike a blow for traditional values, and they cast their ballots for George Bush while they were at it. Were this line of reasoning correct, John Kerrys defeat could be laid directly at the feet of Chief Justice Margaret Marshall of the Massachusetts Supreme Judicial Court, who authored the courts decision legalizing gay marriage in the state.
But a carefulexamination of the 2004 election returns provides no evidence forthis interpretation. Marriage referenda mobilized voters on bothsides, not just the conservatives, and the net result may have been to John Kerrys benefit. If the Democrats want to increase their chances of winning in 2008, they would be well advised to study the results of the recent election with care.
* * *
Did the marriage referenda help Bush? A hard look at where Bushs support came from and where he gained support between 2000 and 2004 reveals that his victory reflected a nearly uniform national swing. The accompanying graph shows the relationship between Bushs share of the two-party vote in 2000 and in 2004 among the 50 states. (The District of Columbia, which gave Bush only ten percent of the vote both years, is excluded.) The diagonal line represents no change between 2000 and2004; in states above the line, Bush drew a larger share of thetwo-party vote in 2004 than he did in 2000. Most states (including Massachusetts) are slightly above the diagonal line, reflecting a1-to-5-percent overall shift in favor of Bush. On the graph, mostof the exceptions to this trend fall near 50, 50the point representing states where Bush won close to 50 percent of the two-party vote in both 2000 and in 2004. These are the battleground states, where the massed political attention on both sides yielded yet another effective stalemate. Another exception is Vermont, which in spite of its backlash against gay rights in 2000 registered the greatest gain for the Democratic candidate between 2000 and2004.
In general, then, voters shifted in a pro-Bush direction in 2004. What happens when we factor in the issue of gay marriage?
Eleven statesArkansas, Georgia, Kentucky, Michigan,Mississippi, Montana, North Dakota, Ohio, Oklahoma, Oregon, andUtahhad measures before the voters that would prohibit gaymarriage, and in some cases civil unions. Nine of the 11all but Michigan and Oregonhad gone for Bush in 2000, and only threeMichigan, Ohio, and Oregonwere battleground states. Theseare hardly the states one would choose if gay marriage were being used as a wedge issue.
Consider the case of Ohio: John Kerrylost Ohio, a state with a ballot initiative and substantial effortsby the Christian right to mobilize voters. But Kerry won a greater percentage of the vote than Gore had (48.9 percent rather than 48.2 percent). Indeed, Bush lost vote share in each of the three battleground states with gay-marriage bans on the ballot, fallingfrom 49.7 percent of the overall two-party vote in these states in 2000 to 49.6 percent in 2004. In contrast, Bush gained vote share inthe battleground states that did not vote on gay marriage: in Florida, Iowa, Minnesota, New Mexico, Pennsylvania, and Wisconsin, Bushs combined 50.4 percent of the vote represented a one-percentage-point increase since 2000.
At the statelevel, then, marriage referenda seem not to have worked to Bushsadvantage. If we move down to the county level, we find even firmer support for this conclusion. In states with gay marriage on the ballot, Bush gained additional support in the counties he carried in2000. But in these same states he also lost votes in Democratic counties generally andperhaps more surprisinglyin evenlydivided counties. The overall result is that the polarization of theelectorate over gay marriage aided Kerry, not Bush.
What was themagnitude of the effect? In states without gay marriage on theballot, Bushs gains were fairly constant across counties. In those states, you get a pretty good estimate of Bushs share of the two-party vote in a given county in 2004 by adding three percent tohis share of the two-party vote in 2000. But in states with gay marriage on the ballot, where counties that were pro-Bush in 2000 were even more pro-Bush in 2004 and counties that were pro-Gore in 2000 were even more pro-Kerry in 2004, there was an overall net shift of 2.6 percentage points away from Bush from the first election to the second. In other words, in states where gay marriage was on theballot, partisan voting patterns became more pronounced, with a netadvantage for Kerry.
If you put together the states with and without gay marriage on the ballot, the evidence suggests that the initiatives somewhat hurt Bush. The effect is not enormous, but it is statistically significant and politically meaningful. Overall, Bushdid particularly well when gay marriage was on the ballot in counties where Bushs percent of the two-party vote in 2000 was more than 61 percent, that is, in very strongly pro-Bush counties. Bush ran less well in counties where he received less than 61 percent. In short, in more-closely contested counties, the marriage referenda appear to have worked to Bushs disadvantage.
A skeptic might dismiss this conclusion. But members of the Bush campaign team themselves were uneasy about the anti-gay-marriage movement and had a hard time deciding to take a position on the issue. According to Alan Cooperman and Thomas Edsall of the Washington Post, the idea of a ban on gay marriage initially met with resistance from the WhiteHouse. As one minister put it, It was a good thing we werent coordinating with the Republican Party, because there wasnt anyone to coordinate with.
* * *
If the issue of gay marriage cannot explain George Bushs victory, why did Bush enjoy a nearly uniform swing in his direction? Most likely it was simply politics as usual.George Bush had the advantage in the types of issues that typically determine electionsthe economy, war and peace, leadership, and ideology. First, the economy in 2004 was better than average, and nearly all economic voting models predicted a Bush victory. Professor Ray Fair of Yale University, who devised the leading model forelection forecasting, predicted that Bush would win 57 percent of the two-party vote. Second, concern about terrorism remains widespread inthe United States, and the voters most concerned about terrorism favored Bush. (In a poll taken less than three weeks after theelection, most Americans expressed disapproval of Bushs handlingof the economy, foreign policy, and war with Iraq, but pretty strongapproval for his handling of the war on terror.) Third, Kerry was more liberal than Gore: when the two were in the U.S. Senate together, Gores votes were generally similar to those of other Democrats, while Kerrys were clearly more liberal. Fourth, Bush was the incumbent and enjoyed the advantages of his office, ranging from the bully pulpit to the federal purseit was he, for example,who sent FEMA money to hurricane-ravaged Florida.
Viewedfrom this vantage point, Bush won a surprisingly unimpressive victory. Three factors likely dragged down Bushs support. First, the initial three years of his administration saw very weak economicgrowth. Second, while concerns about terrorism helped Bush, the warin Iraq hurt him; in pre-election and exit polls, those citing Iraqas the most important issue to them tended to favor Kerry. Third,while only 22 percent of Americans identify themselves as liberals,less than half of all Americans think the country is heading in the right direction.
The interpretation of Bushs 2004 victory will surely shape the agenda of his second term. Many commentators have described the election as a triumph of the Christian right, which rallied around moral values using the threat of gay marriage as a catalyst. This interpretation, if it takes hold, will embolden those on the right within the Bush administration. It will also lead Democrats in the wrong direction as they respond to their loss. John Kerrys running mate, Senator John Edwards, has already commented that voters have to believe that our valuesmy values and the values of other Democratic leadersare the same values they believe in. The evidence shows that the Republican victory rests more on fear of terrorism and an election-year uptick in the economy than on the activism of the partys right wing. Responding to the tangible worries of the vast middle of the political spectrum rather than a polarizing moral agenda should be the basis of Democratic strategy over the next four years.
Stephen Ansolabehere is a professor of political science at MIT and the co-author of The Media Game and Going Negative.
Charles Stewart III is a professor of political science at MIT and the author of Budget Reform Politics.