Risk aversion

The upshot of this chapter is a number of impossibility theorems, demonstrating that given a few reasonable desiderata for strongly risk averse decision rules, there is no rule that can fulfil all the proposed desiderata. Together these theorems give indirect support to the expected utility principle. However, a further point I seek to make is that many of our normative intuitions about risk aversion can be explained by considering a much different, epistemic notion of risk aversion. According to the epistemic (belief-guiding) notion of risk aversion, intuitions about risk aversion should be accounted for in terms of what rational agents are urged to believe.

In Section 8.1 the concept of a fatal outcome is defined and discussed in relation to some prominent examples of strongly risk averse decision rules. In Sections 8.2 and 8.3 a number of desiderata for such decision rules are stated, and proved to be inconsistent. In Sections 8.4 and 8.5 a set of even weaker desiderata are proposed, which appeal to the intuition that in many cases when risk aversion is called for it does not make sense to presuppose that the agent has access to quantitative information about the utility and subjective probability of each outcome.