It's the political equivalent of "tastes great!" vs. "less filling!" among light-beer lovers: the Clinton-Obama battle over who will be a better general-election candidate based on the primary results. The Clinton campaign says she'd be the better fall candidate because she's stronger with her party's core of white working- and middle-class voters in Democratic states. The Obama campaign argues that he'd be better in the fall because he can attract independents, bring new younger voters to the polls, and compete in traditionally red states.
Who's right? Neither side. Why? Because they are both arguing from the false assumption that primary contests can provide a guide to the fall campaign. Look back across recent political history and you'll be hard-pressed to find such a link.
Some of the counterexamples are blatantly obvious: In 1988, the Rev. Jesse Jackson won Democratic primaries in Alabama, Georgia, Louisiana, Mississippi, and Virginia. Did that reveal Jackson's potential strength in the South as a general-election candidate? Of course not; it demonstrated that legions of white Southerners had fled the party in those states, leaving blacks a powerful voting bloc in the primary but insufficient in numbers to carry the general election.
Or look at what happened in 1980 in the Michigan Republican primary. On May 20—well after Ronald Reagan had effectively clinched the GOP nomination—Michigan Republicans voted for George H.W. Bush in a landslide, 57.5 percent to 31.8 percent. Proof that Reagan would be weak in that state? That fall, he beat President Jimmy Carter there by six and a half points, a bigger margin than homeboy Gerald Ford had racked up against Carter four years earlier. Today, when journalists and campaigns set out to find "Reagan Democrats," they head straight for Macomb County, Mich. * There was no sign of enthusiasm for Reagan in the Republican primary of 1980 because Reagan Democrats weren't voting in the primary.
I offer this blindingly obvious point to suggest why it is mostly a fool's errand to find autumn portents in winter and spring primaries. To be even more blindingly obvious, the great majority of voters do not participate in the primaries. As of today, some 27 million people have votes in Democratic primaries and caucuses (counting Florida, where all the candidates were on the ballot, but not Michigan, where only Clinton and Chris Dodd were). In the 2004 general election, more than 122 million votes were cast. Any extrapolation about voting blocs based on primary results has to confront that elemental difference.
Moreover, exit poll results from primaries don't always tell us what we think they tell us. Consider the much-sought-after independent voter. Independents are permitted to participate in primaries and caucuses in such competitive states asOhio, Wisconsin, Iowa, and New Hampshire. But exit polls simply ask respondents to identify themselves, so a registered Democrat or Republican who considers herself an independent thinker might tell an exit pollster she's "independent." In addition, even those voters who don't formally register with a party often have strong leanings one way or the other; the number of genuine "swing" voters is comparatively small.
In the case of the current battle, we're divining, for example, whether Obama can draw white voters based on those who have decided to vote in Democratic primaries. We don't really know how this historic contest between a woman and an African-American is playing with white voters who are not part of the primary process. Maybe race and gender matter a lot less than they would have a few decades ago; maybe such voters are sitting this round out and will flock to the white guy in the fall. We are unlikely to get a persuasive answer to this question until the fall. Nor do we have any real clue about whether Clinton's showing among white working-class voters would mean much of anything should she be the Democrat to confront John McCain … or whether a campaign focused on the economy as opposed to national security would pull such voters to either Democrat. Can we guess? Sure. Can the primaries offer us actionable intelligence? Highly unlikely.
This is not to say that there are no clues at all to be gleaned from the primaries. Michael Barone—who is to political numbers what Bill James is to baseball statistics—offers this take on last week's Ohio primary: "In southeast Ohio, settled originally by Virginians and still Southern-accented today, Clinton carried all-white counties with 70 percent to 80 percent of the vote—more than she was carrying nearly all-white counties in central Texas. That raises doubts that Obama could run well in these counties, which provided critical votes in Bill Clinton's wins in Ohio in the 1990s and Jimmy Carter's narrow win there in 1976." Those findings have to give Obama backers pause.
If you're looking for better news for Obama, the measurable surge of younger voters in the primaries and caucuses suggests that the decadeslong wait for "the youth vote"—a wait that makes Godot look like the most punctual of men—may be over. After splitting their votes almost evenly between Gore and Bush in 2000, the 18- to 29-year-old cohort—some 20.5 million of them, by my exit poll arithmetic—produced a nine-point edge for John Kerry, or a boost of 2 million-plus votes. Greater numbers and a bigger margin for Obama in the fall could be decisive.
There's also one historical example that is heartbreakingly intriguing. When he won the 1968 Indiana primary, Robert Kennedy had the vote of a large number of conservative white working-class voters. (In 1970, two ex-Kennedy aides wrote a book debunking that claim; in his new book on the '68 race, The Last Campaign, historian Thurston Clarke debunks the debunkers.) There is anecdotal and statistical evidence suggesting that a chunk of the RFK primary voters wound up supporting George Wallace's third-party bid that November, when Democrat Hubert Humphrey ran against Nixon. We will never know whether Robert Kennedy could have kept those voters from defecting to Wallace—or whether the huge turnout of Hispanic and black voters for RFK in California that June would have occurred again in November and turned the tide in his favor in what was back then a Republican-tilting state.
Finally, there is one clear and consistent scenario in which primary contests provide telling clues about the fall: when an incumbent president faces a meaningful challenge for re-nomination within his own party. No president in modern times has ever survived such a challenge to win in the fall. Lyndon Johnson abdicated in 1968; Gerald Ford fended off Ronald Reagan in 1976 but lost that fall; Jimmy Carter beat Ted Kennedy but was swept away by Ronald Reagan four years later; and even George H.W. Bush, embarrassed in New Hampshire by Pat Buchanan, though never seriously threatened by him, had to invest so much time shoring up his base that the episode helped lead to his defeat in 1992.
But the main lesson is that in searching the primary terrain for general-election hints, tread very carefully. As a rule, what happens in the primaries stays in the primaries.