We develop a new revealed preference framework to estimate the value of statistical life (VSL). Our framework starts from a hedonic model of health care in which heterogenous individuals choose how much to spend on medical services that reduce mortality risk. Their choices generate an equilibrium survival function that can be differentiated to recover their marginal willingness to pay for mortality risk reduction. Our IV estimator uses survey data on Americans over age 66, linked to their federal administrative records. The mean VSL is approximately $1 million at age 67 and increasing in health, income, education, and life expectancy.
Authors
- Acknowledgements & Disclosure
- We are grateful for insights and suggestions from Joe Aldy, Kelly Bishop, Glenn Blomquist, Mary Evans, Chad Jones, Alan Krupnick, David McCarthy, Alvin Murphy, Julian Reif, Lisa Robinson, Dan Silverman, Kerry Smith, Jonathan Skinner, David Slusky, Greg Veramendi, Jeffrey Wooldridge, and audiences at Camp Resources, WCERE, AERE, ASHEcon, WEAI, RFF Workshop on Improving Federal Environmental Regulation, Arizona State University, Carnegie Mellon University, Georgetown University, Hong Kong Polytechnic University, Michigan State University, Stanford University, University of Birmingham, University of British Columbia, University of Florida, University of Georgia, University of Hawai'i-Manoa, University of Illinois Urbana-Champaign, University of Kansas, University of Nova School of Business and Economics, and University of Southern California. This research was supported by the National Science Foundation (award #2049902) and the U.S. Environmental Protection Agency (award #84018401). The views expressed herein are those of the authors and do not necessarily reflect the views of the National Bureau of Economic Research.
- DOI
- https://doi.org/10.3386/w33165
- Pages
- 67
- Published in
- United States of America
Table of Contents
- Introduction 3
- Data and Measures 7
- Sample construction 8
- Medical expenditures 8
- Health 10
- Socioeconomic Characteristics 11
- Environmental Exposures 12
- The Demand for Mortality Risk Reduction 13
- Survival 13
- Budget constraint 14
- Preferences and optimization 14
- The Value of Statistical Life 16
- Choice frictions 18
- Interpreting the VSL with endogenous future health 19
- Econometric Model 20
- Survival function 20
- Identification and estimation 21
- Constructing an instrument for medical spending 21
- Control function estimation 23
- Results 25
- First-stage results and instrument validity 25
- Survival functions 26
- VSL measures by age 29
- VSL measures by health, education, and income 31
- Validation Tests and Sensitivity Analysis 33
- Main validation tests 33
- Additional sensitivity analysis 35
- Conclusion 36
- Supplemental Appendices 43
- Data 43
- MCBS Spending Measures 43
- Evolution of Health and Medical Spending 44
- Environmental Exposures 45
- Homicide mortality and automobile mortality 46
- Fine particulate matter 48
- Average temperatures 48
- Share urban 48
- Median household income 48
- High school graduation rate 48
- Hospital compare index 49
- Ambulatory Care Sensitive Conditions 49
- Access to Medical Care 49
- HCC risk adjustment score 49
- References 50
- Additional Results 51
- First Stage Results 51
- Second Stage Results 53
- Model Fit 53
- Heterogeneity in the Return to Medical Spending 55
- Heterogeneity in the Value of Statistical Life 58
- References 61
- Additional Sensitivity Analysis 63
- Modifiable features of the research design 63
- Including or excluding workers 63
- Alternative instruments for medical expenditures 63
- Interacting the instrument with observed health 64
- Interacting medical spending with health, age, or education 64
- Alternative Parametric Forms of the Survival Function 65
- Allowing heterogeneity in agency and information frictions 65
- Results 65
- References 67
