Working Paper: NBER ID: w4649
Authors: John Mullahy; Jody L. Sindelar
Abstract: The economic costs of adverse health outcomes have typically been evaluated in a context of risk neutrality, an approach that ignores the potential welfare importance of individuals' risk preferences. This paper presents a framework that unifies the research in health capital and earnings with that on risk preferences in the presence of stochastic outcomes. The model is implemented to obtain estimates of the economic damages due both to general health problems as well as to one specific health problem that is of considerable interest from society's perspective: alcoholism. Our empirical findings, based on data from the Epidemiologic Catchment Area survey, indicate that failure to recognize the possibility of risk averse preferences leads to a potentially serious underestimation of the magnitudes of the 'costs' of alcoholism and poor health. In particular, it is shown that while alcoholism problems have negative impacts on the conditional mean of income (consistent with most of the existing literature), they also have positive impacts on the conditional variance of income. Our conclusions are to some degree provisional because our estimates of conditional variances are necessarily biased to the extent that unobserved heterogeneity is an important determinant of the moment structure of income in our sample.
Keywords: health economics; risk aversion; alcoholism; welfare costs
JEL Codes: I12; I18; D81
Edges that are evidenced by causal inference methods are in orange, and the rest are in light blue.
| Cause | Effect |
|---|---|
| poor health (I14) | conditional mean of income (D10) |
| alcoholism (I12) | conditional variance of income (D31) |
| health status (I12) | income outcomes (E25) |