After a year of carnival-like fun, the Republican race has suddenly become boring and predictable. Voting in state after state has just revealed the same basic demographic divide at work. There’s been a great deal of discussion about how Romney and Santorum voters are split by income, education, and ideology, but only one indicator really seems to matter: if you look closely at the exit polls, the single best predictor of whether Romney loses a state is the percentage of voters who describe themselves as evangelical or born-again. Of the seventeen states where exit polls have been conducted, Romney has lost each one where evangelical voters make up a majority of the electorate, and he’s won each one where their share of the vote was forty-nine per cent or less. Here’s what it looks like when charted, courtesy of a great exit-poll feature over at the Washington Post:
With this information in mind, let’s see if we can predict the outcome of the remaining twenty-three states. Naturally, we don’t have 2012 exit-poll data for those states, but we do have some 2008 exit polls, and, courtesy of The Pew Forum on Religion & Public Life, we have some data about the adult population of evangelicals in every state except Maryland and Washington, D.C. Using this information, here’s what the states that have voted look like:
We learn two things from this. First, that Santorum can win in states without big populations of evangelicals if the state uses a caucus system, which generally attracts the most conservative voters, to select its delegates. And second, we learn that, just as the magic evangelical number in exit polling was fifty per cent, the corresponding magic evangelical number for the total adult population of a state seems to be thirty-seven per cent. (Romney has lost every non-caucus state where the adult evangelical population exceeds that percentage.) What does that mean going forward? Bad news for Santorum:
There are only three remaining states where the adult population of evangelicals exceeds thirty-seven per cent. But perhaps that number is deceiving. For instance, some Democratic states might have a small but intensely conservative Republican electorate, and the percentage of evangelicals in the over-all adult population may mask this fact. Separately, the 2008 exit polls for Texas and Louisiana make it clear these will be promising states for Santorum, even though they don’t reach the thirty-seven per cent threshold. While the adult population of evangelicals in Texas is thirty-four per cent, the Republican primary electorate was sixty per cent evangelical in 2008. In Louisiana, the adult population of evangelicals is thirty-one per cent, and the Republican primary electorate was fifty-seven per cent evangelical in 2008. Santorum came very close in Ohio and Michigan; evangelicals make up twenty-six per cent of the adult population in both of those states. So maybe the real magic number for the adult evangelical population is closer to thirty per cent, which brightens Santorum’s prospects quite a bit.
Another bad piece of news for Santorum: there are only two caucus states left, Nebraska and Montana.
With all this in mind, here’s a guide to who is likely to win the remaining states. I’ve listed the states in order of my degree of confidence about the victor. Our two simple rules about evangelicals and caucuses predict eleven states and the District of Columbia for Romney and eleven states for Santorum:
Romney: Rhode Island, Washington, D.C., Utah, Connecticut, New York, New Jersey, Maryland, Delaware, California, Wisconsin, New Mexico, South Dakota.
Santorum: Arkansas, Kentucky, North Carolina, West Virginia, Indiana, Texas, Louisiana, Pennsylvania, Nebraska, Montana, Oregon.
Despite its small evangelical population, Santorum’s home state of Pennsylvania will probably break his way, so I have it in his column. Some smart people I respect would give Santorum Delaware (because of the very right-wing G.O.P. electorate there) and South Dakota (because he won North Dakota). But I’m sticking to my evangelical rule for Delaware and assuming his win in South Dakota was a function of the caucus there. Finally, it’s worth noting that the much discussed Wisconsin primary, which is being touted as a place for Santorum to break Romney’s momentum, does not look promising for the former senator. I put it in Romney’s column.
If these predictions are correct—I’ll revisit them as the campaign goes on and see how they hold up—Romney is on track to lose more states than any G.O.P. nominee since Gerald Ford lost twenty-three to Ronald Reagan, in 1976.
Finally, we all know that the nomination is not secured by winning states but by accumulating 1,144 delegates. The delegate rules are quite complicated, but I’ll dive into the details in an upcoming post and tease out what the almost even split of states between Romney and Santorum states means for Romney’s ability to hit 1,144 after the last contest in Utah, on June 26th. (Spoiler: it’s going to be close.)
Bottom two charts by Andrew Prokop.
Photograph by Stephen Morton/The New York Times/Redux.
