Abstract

We perform a lab experiment where we measure higher order risk preferences under gains and losses and investigate combinations of higher order risk preferences. We find no evidence that decision makers have a preference for combining good with bad or good with good, a preference which has been suggested as a basis for higher order risk preferences. Such a preference implies mixed risk preferences, where risk averse individuals are temperate and risk lovers are intemperate, and both are prudent. Instead, we  find a reflection effect for higher order risk preferences: When outcomes are presented as losses, the aggregate patterns for prudence and temperance reverse.