The problem with economic rationality as commonly used in orthodox economic analyses is not that it’s a straw man. It’s an initial hypothesis adopted because – as Krugman says – without it economic decisions are very messy and hard to analyse. But it provides cover for a lot of other assumptions which are rarely examined (eg that people are individual utility-maximisers), and it carries it’s own definition of rationality which rarely checks back against what people actually want or do. It was seen as an unrealistic assumption early on, but carried on when it was realized that the alternatives are not mathematically treatable (this was stated explicitly in the 40s). In short, the appearance of rigour won out over the messy reality.

It’s not so much outmoded as radically incomplete, and the key issue is that the profession – and the economics-using public – are reluctant to recognise the limitations this imposes. Worse, the growing realisation that this incompleteness renders hollow many arguments has not led, on the whole, to opening the field, but to hostility to studies which could endanger economics’ status (eg the marginalisation of other schools, and the disbandment of economic history departments).