Sure, I was quoting it more for the point about methodology than Rogeberg’s argument about the plausibility of rational addiction. I think plenty of economists are happy to use as-if justifications for their models to avoid defending the assumptions. That’s fine, as long as they don’t infer welfare consequences from the as-if predictions.

In terms of rational addiction, I’m a bit torn. The model has some empirical support and models augmented with hyperbolic discounting seem to fit the stylised facts reasonably well. At the same time, humans don’t appear to do rational choice all that well, which casts doubt on the assumptions underpinning the model. Maybe we’ll find that there are just some tweaks that need to be made to it to properly represent people’s decision heuristics, but that’s fairly optimistic.