Consistent estimation of the value of statistical life

Conventional estimators of the value of statistical life are biased. People differ in risk from each of many health threats, ability to reduce these risks, willingness to pay to reduce risk, and other utility parameters-creating a problem of multi-dimensional heterogeneity existing single-equation methods cannot handle. Herein we propose a general method of moments (GMM) approach that uses functional relationships between underlying parameters and observed data to estimate a person's willingness to pay for mortality risk reduction. This approach yields a consistent estimate of the value of statistical life. We use simulations to show that the GMM estimate of the value of statistical life performs well even when combining data from different sources that are sampled at different, low frequencies.