Presidential nominees love to consider home state advantage when choosing their running mates. But a vice president’s state turns out to make no difference in an election.
Few myths of American politics remain as deeply entrenched and blindly accepted as that of the electoral power of the vice presidential pick. Already in 2016, pundits have speculated that Governor John Kasich would deliver Ohio if selected as the Republican vice presidential nominee, or that fellow Ohioan Senator Sherrod Brown could deliver the state to Democrats instead. Likewise, Senator Tim Kaine could deliver Virginia. In 2012, Paul Ryan would do the same for Wisconsin, Rob Portman for Ohio, and Bob McDonnell for Virginia. In 2008, it was Bill Richardson; in 2004, Bob Graham; in 2000, Tom Ridge. You get the idea.
The underlying assumption is this: Vice presidential candidates add votes in their home state. The right VP pick can help carry a competitive state, the thinking goes, or put an uncompetitive state into play. Knowing that, a presidential candidate would be foolish not to use this strategic opportunity to try to pick up a key state in the Electoral College. At a minimum, the list of pros and cons for each vice presidential finalist must include his or her potential to deliver a home-state advantage.
Like all unquestioned shibboleths, it’s come to seem almost a law of nature by now. Analyzing news coverage between 2000 and 2012, we found that journalists invoked geographic strategy in about 50 percent of their profiles on potential veep candidates. But it’s wrong. According to our analysis of election and voter data over the course of a little more than the past century, a vice presidential candidate’s state of residence generally has no effect on how a presidential candidate performs in that state. The vice presidential home state advantage is, essentially, zero.
Pundits who tout the notion of a home-state advantage clearly believe in it. So do the journalists who report on it. But the most eminent constituency for this belief, of course, is the presidential candidates themselves. Despite ultimately selecting Dick Cheney as his running mate in 2000, in his autobiography George W. Bush revealed that he was tempted to pick Tennesseans Lamar Alexander, Bill Frist or Fred Thompson, each of whom seemed sure to deprive Al Gore of a victory in his home state (a feat that Al Gore would later manage by himself). Bush never may have faced Gore in 2000, but instead Bob Graham, had Bill Clinton chosen the Florida senator as his running mate in 1992—a candidate, Clinton later wrote, who “would almost certainly bring Florida into the Democratic column.” Richard Nixon wrote in his memoir that he chose Maryland Governor Spiro Agnew in 1968, in part, to garner votes in mid-Atlantic and border states. And let’s not forget the granddaddy of them all: John F. Kennedy’s selection of Texas Senator Lyndon Johnson in 1960. (The math seemed so simple in this case that campaign manager Robert Kennedy, upon learning of his brother’s selection, directed a campaign adviser to “add up the electoral votes in the states we’re sure of and to add Texas.”)
In our new book, we analyzed state-level election returns from 1884-2012 and individual-level survey data from 1952-2008 to determine whether vice presidential candidates do, in fact, deliver a home-state advantage—and, if so, by how much. If the advantage is real, we should be able to detect and quantify it.
We used three distinct methods. First, if the home-state advantage is real, we should see predictable (and statistically significant) deviations in the voting trends of a running mate’s home state, relative to national voting trends, across decades of elections data. Second, linear regression analysis should show that parties win a higher percentage of the vote from states that are home to the running mate on their ticket in a given election. Finally, survey data from the American National Election Studies should tell us that individuals (not states) respond differently to elections that feature a vice presidential candidate from their home state. Do they vote for a different party? Are they more likely to vote at all, and to become engaged in the campaign?
Our conclusion: While presidential candidates typically enjoy a home-state advantage (approximately 3 points to 7 points), vice presidential candidates generally do not. In each of the three analyses described above, a presidential ticket performs no better in the vice presidential candidate’s home state than we would expect otherwise. Statistically speaking, the effect is zero.
We did find that veep home-state voters are more likely to care who wins an election compared with non-home state voters—but they aren’t more likely to turn out to vote, volunteer for or donate money to a campaign, influence other voters or attend political rallies.
There is one important exception to all of this: In the small handful of cases where a vice presidential home-state advantage did occur, consistently we find that the state in question has a relatively small population, and the candidate in question has a great deal of experience representing the voters of that state. In other words, the candidate who actually delivers a vice presidential home-state advantage truly must be an institution in state politics—an object of intense affection, loyalty and intimate familiarity. Most running mates (indeed, most politicians) do not meet this remarkably high standard. Those who meet the standard—for instance, Joe Biden in 2008 and Edmund Muskie in 1968—do, indeed, improve their ticket’s performance at home. But, of course, the prize is small: By definition, the states that can be “delivered” this way have relatively few electoral votes.
This evidence ought to affect the vice presidential selection process, and the way pundits, journalists and campaign staff discuss its strategic potential. The takeaway is simple: There is very little chance that a vice presidential candidate from Florida, Ohio or any other large battleground state will deliver its electoral votes. In most cases, voters simply vote for the presidential candidate they like best.
In fact, VP picks aren’t even particularly effective on the national scale. Our analysis suggests that presidential candidates have at least three times the influence on vote choice as the vice presidential candidate. In order for a running mate to help a candidate on a national scale, he or she must be exceedingly popular; in order to hurt, the VP must be tremendously unpopular. By and large, neither happens.
From a historical perspective, then, has a vice presidential home-state advantage ever decided a presidential election? According to our analysis, no. But that doesn’t mean it couldn’t happen or that it is irrelevant to understanding presidential campaigns.
In 2000, Al Gore chose then-Senator Joe Lieberman of Connecticut as his vice presidential running mate. That same year, Jeanne Shaheen, a two-term Democratic governor of New Hampshire, was up for reelection in her home state. Shaheen was leaked as a finalist on Al Gore’s vice presidential short list, and would win reelection that fall—the same election in which New Hampshire became the only New England state to cast its electoral votes for Republican George W. Bush. According to estimates from our election forecasting models, a counterfactual Gore-Shaheen ticket would have won the small state of New Hampshire by at least 1 percentage point. Assuming that national political dynamics remained the same, Gore would have secured a majority of Electoral College votes, regardless of the outcome in Florida. The 2000 presidential election is the only election in recent history where a known vice presidential finalist plausibly could have delivered an electorally decisive home state.
But what about 1960? Every political junkie knows the myth that LBJ “delivered” electoral votes in his home state of Texas and throughout the South that year, effectively deciding the election. Even JFK, as president, told an aide: “You’ve got to admit that I was right. We couldn’t have carried the South without Johnson.” However, our data provide no evidence of such an effect. We analyzed LBJ’s geographic impact in a way that (surprisingly) researchers have never done before: using survey data from 1960, including the ANES and internal campaign polls. The results of this analysis are counterintuitive, to say the least. According to ANES data, Texans had a more negative opinion of LBJ than other southerners did; and southerners viewed LBJ more negatively than non-southerners. Asked to explain their opinions, the most common response from Texans was that LBJ lacked integrity and was overly ambitious. Indeed, none of our empirical models indicates that Johnson was particularly popular in Texas or the South, nor do we find any evidence that opinions of LBJ had more influence on vote choice and other political attitudes and behaviors among Texans versus other southerners, or among southerners versus non-southerners. As a final validation of our surprising findings, the ANES data show that only two of 972 respondents, including one northerner and one southerner, said their opinion of Kennedy had ever changed due to selection of the vice president. The kicker: Both voted for Nixon. (A foray into the Kennedy and Johnson presidential archives confirmed our finding. According to internal polls conducted during the 1960 campaign, in Texas, Johnson was no more popular than Kennedy in terms of job approval, and in Louisiana Johnson was less popular than Kennedy.)
The perception of a vice presidential home-state advantage has major consequences for the conduct of presidential campaigns. In 2012, for example, we find that the Obama and Romney campaigns devoted a disproportionate amount of their most precious resources—time and money —on contesting Paul Ryan’s home state of Wisconsin. The campaigns visited Wisconsin only once before Ryan’s selection but 43 times afterward. This made Wisconsin the fifth most-visited state in 2012, when, in fact, it was about the tenth most-competitive state, according to our analysis of state voting in previous elections and state polling in 2012. In terms of money, campaign ad spending in Wisconsin (as a percentage of national spending) increased by nearly 400 percent in the weeks following Ryan’s selection. At one point in September, ad spending in Ryan’s home media market represented one out of every $80 spent nationally. Clearly, both campaigns overestimated the competitiveness of Wisconsin, increasing their activity to anticipate the likelihood of a vice presidential home-state advantage that never materialized. Polls did not show that Romney was closing the gap on Obama after he chose Paul Ryan. And, Paul Ryan, as the VP running mate on the ticket, lost his home state by nearly 7 percentage points that year. Ten states were decided by a smaller margin, yet few got more attention from the campaigns than Wisconsin.
There are higher, more serious costs to the home-state advantage fallacy than just campaign money. The vice president is the country’s second in command—primed to become chief executive in the case of the president’s death, resignation or removal from office. It would be a shame if one of the primary considerations used to select a running mate in the United States continued to be a politician’s address, up there with executive experience or policy expertise—especially now that the evidence is in that, in nearly every case, the “home-state” strategy doesn’t even work.