Economists don’t know where preferences come from. They might come from god or culture or genes. Whatever. We don’t care. ((Or we do but we’re just not paid to care.))
If preferences are more or less constant within individuals over relatively short periods of time, it doesn’t matter where they come from. If people optimize over preferences today, resulting in some behavior, they will behave the same way tomorrow when optimizing over the same preferences ((Yes, framing matters. But there’s a fix for this: just extend the product space to include frames… I like apples on Tuesday, but oranges on Wednesday; I prefer apple-Tuesdays and orange-Wednesdays.)). The pragmatic case for taking preferences as the primitive objects of analysis is just that they’re so powerful at predicting behavior and doing policy analysis.
Treating morals as preferences over how one wants other people to act sidesteps moral reasoning (i.e. consideration of what are the correct morals) and allows for analysis to focus on predicting behavior or on policy issues.
Moral reasoning is important and I’m not saying it shouldn’t be done. As Haidt points out, though, it is often done by white upper-middle class liberals and as such reflects the moral intuitions of that particular sub-culture. The universe of moral intuitions is much broader and, more importantly, some of those non-white, non-upper middle class, non-liberal moral intuitions are held by people that don’t feel the need to rationalize or vocalize moral intuitions. Why should extra weight be given to those moral intuitions that happen to have been articulated best?
In any case, the point is that moral reasoning is tangential to many of the policy issues we’re interested in such as the optimal size of moral communities. Why not abstract away from the issues moral reasoning raises if given the opportunity?
PS – I know that I’ve been a bit cryptic on some of these topics. I’m a little busy ramping up to teach a course this summer and finishing my first chapter in my dissertation (fingers crossed). I hope to circle back to these issues, fleshing out them when I do.