I think that they're simply in love with academic models and "model UN" games that simply don't pass the test of reality. Then they try making up excuses why their models don't work, swearing it was yet another one time fluke, instead of creating a better model from scratch. I have no evidence...