The Value of A Statistical Life in Absence of Panel Data: What can we do?

In this paper I show how reliable estimates of the Value of a Statistical Life (VSL) can be obtained using cross sectional data using Garen's instrumental variable (IV) approach. The increase in the range confidence intervals due to the IV setup can be reduced by a factor of 3 by using a proxy to risk attitude. In order state the "precision" of the cross sectional VSL estimates I estimate the VSL using Chilean panel data and use them as benchmark for different cross sectional specifications. The use of the proxy eliminates need for using hard-to-find instruments for the job risk level and narrows the confidence intervals for the workers in the Chilean labor market for the year 2009.