From Aaron Schiff:

A model should not be judged solely by its assumptions (although highly dubious assumptions are not a good thing). Rather we should focus on the model’s ability to teach us something, and its ability to explain the economics of something in a plausible way.

To summarise, a model does:

• Highlight the incentives or tradeoffs that are relevant in a particular economic situation.
• Generate predictions about the behaviour of economic agents in response to controlled changes in conditions.

I took this from a good post by Aaron Schiff.  I agree with it.

Note that the purpose of economic models isn’t prediction (as we have been discussing) – but we do want a testable hypothesis, in order to make our models scientifically valid.  So the model MUST be testable, but this is not a sufficient condition on models.

Furthermore, predicitive accuracy is not part of testability – as the difference could stem from a change from one of our ceteris paribus (CP) assumptions.

The goal is explanation and description.  And trust me the grey line between prediction and testablility is problematic.  But for the purpose of discussing economic models, the fact that our CP assumptions are the things that break unexpectedly does not invalidate the usefulness or purpose of economic models.

Note:  I will stop writing on this soon and go back to NZ economics.  For me this stuff is interesting, and I like to have a record of where my head is at.