Not only do I disagree with Arnold's views on the self-interest and rationality assumptions; it also looks like Arnold disagrees with Arnold! In the past, Arnold has expressed a lot of sympathy for my view that systematically mistaken economic beliefs lead to a lot of bad policies. Check out his great article "Economic Illiteracy Quadrifecta."
My view, which I almost think Arnold shares, is that the self-interest and rationality assumptions are severely wrong for voters, but pretty accurate for consumers. I basically agree with Arnold that, in markets, "the amount of interesting and empirically supported predictions that you get out of those assumptions far outweighs the occasional clinker." I'm inclined to add that Arnold understates the size of the market's clinkers - but given his continuing sympathy for Keynesian macro, maybe he doesn't. (After all, without money illusion, it's hard to see why Keynesian prescriptions would work).
But in politics, the assumptions of selfishness and rationality are so wrong it's not even funny. Or so I argue here.