Ideally, yes, we would like to explain those. As Matt says, without a general model of almost everything, that could be extremely difficult. Consequently, our ideas about what is important may be right without having a lot of predictive power.

I would hazard to guess you use models of multiple equilibrium, and try to make the process of equilibrium selection more transparent.

Of course, this runs into the same sort of issue as assuming that preferences aren’t stable … and as a result, it is something that requires great care. And potentially a recognition that our ability to understand what, where, and why is oft limited.

The interesting question is why – I’d put it down to the same reason why economists are uncomfortable with preference shocks. It feels like “assuming your result” and is hard to truly test or observe in real time 🙂