Tempering effects of (dependent) background risks: A mean-variance analysis of portfolio selection

In a mean variance framework, we analyse risk taking in the presence of a (possibly) dependent background risk, exemplified in a linear portfolio selection problem. We first characterise the comparative statics of changes in the distribution and dependence structure of the background risk. For unfair, undesirable and loss-aggravating increases in background risks (both dependent and independent), we then present necessary and sufficient restrictions on preferences such that greater background uncertainty leads to reduced risk taking. With mean-variance preferences, these restrictions boil down to simple conditions on the marginal rate of substitution between risk and return. They can be easily related to familiar notions such as risk vulnerability, properness or standardness.