Abstract : I develop and calibrate a dynamic life-cycle model of health investment for people over age 65 that allows me to evaluate the efficiency and equity of policies targeting health and mortality. The model allows for dynamic complementarity between health (quality) and longevity (quantity of life). To account for measurement errors in any given proxy measure for health, I estimate the health production using a dynamic latent factor model. My results imply that returns to an additional $1,000 medical spending is on average 0.3 percentage points and increases with age. The framework allows me to estimate the value of a statistical life (VSL) and the value of disease prevention. I find that the VSL at age 67 is well below the conventional estimates for younger, healthier workers, and that the VSL declines with age. I also find that dynamic complementarity is quantitatively important. For example, my results suggest that a 10% reduction in mortality risk from the baseline increase VSL by about 3% and the willingness to pay to avoid Chronic Obstructive Pulmonary Disease by 4% at age 67. Finally, I use the model to evaluate the morbidity and mortality benefits to seniors from the Clean Air Act and find that it has a benefit-cost ratio of approximately 12:1. I also find that by reducing mortality among seniors, the Clean Air Act increased average VSL by 8%, and total health benefits by 6%.
"Valuing Statistical Life Using Seniors' Medical Spending " joint with Nicolai V. Kuminoff and Jonathan Ketcham (Submitted)
Abstract : This study provides the first revealed preference evidence on the value of statistical life (VSL) for US seniors aged 67 to 97, a crucial input to evaluating polices that affect mortality. We derive VSL measures for seniors from the rates at which they choose to consume medical care relative to other private goods, and by the effects of their choices on their survival probabilities. These effects are estimated from novel data linking survey information to the respondents’ reconciled Medicare records. Instrumental variable estimators provide robust evidence that the mean VSL is below $1 million (year 2010 dollars) and that the value per statistical life year is below $ 100,000. We also find that the VSL decreases with age, and, given age, increases with income, education, and health and is higher for women and people who never smoked.
Abstract : We use the 9/11 terrorist attack as an instrument for depression in order to understand its causal effects on illicit drug consumption among young adults age 18-28. Comparing individuals interviewed before and after 9/11, we find a large and significant jump in symptoms of depression. Our results imply that, depression triggers illicit drug consumption with heterogeneous impact for different drugs. We find substantial heterogeneity in these effects by sub-groups of population such as gender, race, cognitive ability and education. We also partially identify the impact of depression on productivity in school/workplace and find large effects for students. Overall, our findings suggest treating depression will help to curb illicit drug consumption.