George Akerlof gave a fascinating talk on this topic to a World Bank forum last month. His earlier Nobel Prize lecture on behavioral macroeconomics is here. The World Bank talk is a gas - I would like to hear a stern neoclassical macroeconomist's views on Akerlof's demolition of Ricardian equivalence, the permanent income hypothesis, the Modigliani-Miller theorem, the natural rate hypothesis and rational expectations - and his enthusiastic support for Keynesian underpinnings for macroeconomics.
Akerlof's main line is that norms regarding what people should do - not only real outcomes - should enter the utility function we regard agents as maximising. He argues that the wrong models we have endorsed stem from a false endorsement of parsimonious 'positive economics' and flawed econometric methodology. Economists should study the 'microscopic' not just the average. Keynes' got it right and neo-classicals who supported the big 5 neutralities in economics got it wrong.
By the way, the discussion at the end of the Akerlof talk identifies major issues. Key issue: Where do the norms come from? Role of evolutionary economics? Neuroeconomics? Should macroeconomics be culture specific? Can you change behaviour by changing norms? Can you foster development by changing norms?
The World Bank talk is a stunning lecture - watch it circulate across the web! My speciality is microeconomics not macroeconomics so I'd be very interested in views of others.