I’m a strong believer in structural explanations for things, but I think Jonathan Bernstein goes a bit too far with this:
Now, the second part of what Yglesias said — how can anyone know what Romney will actually do it elected? I think the answer is, basically, the same way you can know that about anyone. He’ll follow party incentives, and institutional incentives, and other such things that have little or nothing to do with what he “really” thinks. And that’s mostly a good thing!
There’s a lot of truth to that. But still, I do think beliefs matter. One reason is that people engage in “motivated reasoning” all the time. A politician with a visceral loathing of gay people is more likely to interpret the objective incentives facing him as having anti-gay implications than is a politician who has warm personal relationships with gays and lesbians. I don’t think there’s any guarantee that President Tom Ridge would have pursued the 2003 Medicare bill with as much alacrity as George W Bush did, or that President John Corzine would have been as committed to the “cadillac tax” on high-value health insurance plans as Barack Obama was.