notes-econ-risk

http://en.wikipedia.org/wiki/Ellsberg_paradox

the obvious solution, to me, seems to be that risk aversion should not just be captured by switching the utility function but still assuming that expected utility is the quantity being maximized, but rather, that we should explicitly model the (downside) variance over outcome scenarios (similar to how we speak of a 'bias/variance tradeoff')

http://en.wikipedia.org/wiki/Choquet_expected_utility