So to figure out whether gay marriage will hurt Obama in the fall, you have to figure whether gay marriage alone is likely to block any of these five paths—that is, whether Obama is likely to receive fewer votes from these specific constituencies in these specific states than Kerry received in 2004. For that to occur, Obama would have to suffer a 32-point net loss in Latino support in Nevada; a 27-point net loss in Latino support in New Mexico; a 27-point net loss in Latino support in Florida; a 9-point net loss in black support in Virginia; a 19-point net loss in black support in North Carolina; a 12-point net loss in working-class support in Iowa; and a 5-point net loss in working-class support in Ohio.
In other words, it’s unlikely. Right now, 43 percent of Latinos—a group made up largely of the kind of younger men and women who are driving population growth in key states—approve of same-sex marriage. Among independents that number is up to 52 percent. And opposition among African-Americans has fallen 20 points since 2008. It’s hard to imagine that Obama’s personal opinion about same-sex marriage—remember, he’s not pushing any kind of federal legislation—will be such a turn-off for key demographic groups in key states that their support for the president will plummet to sub-Kerry levels come November.
That said, politics does not occur in a vacuum. Outside organizations may use Obama’s announcement to mobilize evangelicals who would have otherwise been unenthusiastic about voting for Romney; if the president doesn’t match Kerry’s performance among white men, which seems likely, his cushion among minorities will shrink. And so on. But it’s just as likely that these forces will be balanced out by equal and opposite forces: young voters reinspired to volunteer and turn out on Election Day; Latinos appalled by Romney’s far-right immigration stance. The bottom line is that it’s very hard to imagine Obama shedding enough votes on gay marriage to really make a difference where it matters most.