October 16, 2020, 11:00–12:30
This paper adopts the sufficient-statistics approach to obtain a measure of the marginal willingness to pay for long-term care insurance. Theoretically, we extend the standard consumption-based framework to account for the presence of committed consumption goods, that is, goods which face adjustment costs. This yields a weighted-average representation of the marginal value of insurance which implies the need to account, separately, for the intensive margin consumption responses of adjusters and non-adjusters of the committed consumption good as well as the extensive margin response in committed consumption. We implement our framework using an event-study design on panel data from the Health and Retirement Survey. This analysis finds that, on average, households experience sharp drops in food consumption of about 37 percent upon encountering a nursing home episode. However, this masks a significant difference between those who simultaneously adjust their housing and those who do not. Indeed, while the latter group would be willing to pay at least a 57 percent markup over the actuarially-fair rate for insurance, households which adjust both food and housing consumption value insurance at around the actuarially-fair rate. Despite this, applying our weighted-sum formula implies that overall, households in our sample would be willing to pay 1.4 times the actuarially-fair rate for coverage against nursing home expenses.