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

TL;DR

This paper demonstrates how to reliably estimate the Value of a Statistical Life (VSL) using cross-sectional data with an instrumental variable approach, and shows that proxies for risk attitude improve precision.

Contribution

It introduces a method to estimate VSL with cross-sectional data using a proxy for risk attitude, reducing reliance on hard-to-find instruments and improving estimate precision.

Findings

Using a proxy reduces confidence interval range by a factor of 3.

Cross-sectional VSL estimates align with Chilean panel data benchmarks.

Proxy-based approach narrows confidence intervals for 2009 Chilean labor market.

Abstract

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.