Choice, Background Risk, and Household Behavior

Individuals cannot insure against or avoid all risks. People have to accept some background risk. Risk due to the un-insurability and non-tradability of human capital is normally the most important environmental factor people have to cope with when making decisions. Models of background risk arising from uninsurable idiosyncratic shocks to labor income are therefore the standard workhorses of current research. Yet, their quantitative results hinge critically on the quality of estimated background risk, wh ich so far is doubtful. This proposal presents a rigorous framework to identify components of human capital risk than cannot be avoided nor insured against and study empirically how they impact the individual assets allocation decisions at various stages of the life cycle. Our strategy consists of decomposing variation in wages into three exogenous components: a) A component that is specific to the firm, and that cannot be altered by workers´ actions (other than by moving to another firm, an issue we m odel explicitly); b) An aggregate component which represents background risk but does not vary across individuals and hence is of little help in testing models of financial risk-taking with background risk; c) A worker-specific component, which mixes bo th exogenous, shocks to human capital as well as worker choices. The literature makes no distinction between the first and the third component and thus de facto does not identify background risk. Hence, models calibrated on these measures overstate the i mportance of background risk; tests based on them to assess the effects of background risk on behavior produce inconsistent estimates. We implement our empirical strategy using large-scale Norwegian matched employer-employee data containing high-quality information over a long period of time on firm-specific shocks and data on wages and very detailed data on assets from the Norwegian Tax Registry.