|Essays on Risk and Insurance in the Extended Family
|Year of Publication
|Number of Pages
|Adult children, Medicare/Medicaid/Health Insurance, Older Adults, Risk Factors
This dissertation examines the role of extended family members as 'informal insurance' against risk over the lifecycle. The three chapters contained in this dissertation focus on long-term care risk and income risk between parents and their adult children. The first chapter analyzes the role of informal care by family members in the demand for long-term care insurance. The second chapter quantifies the substitution between informal care and formal long-term care by exploiting variation in Medicaid eligibility rules across states over time. The third chapter presents a framework to quantify a network's ability to share risk and applies it to extended family networks with income risk.
The first chapter demonstrates that informal care plays an important role in the demand for long-term care insurance. A key determinant of this result is the 'in-kind' nature of insurance benefits that creates a lack of coverage of informal care. To quantify the effect of informal care on insurance demand and evaluate potential welfare-enhancing policies, I develop a dynamic model of long-term care decisions between an elderly parent and her adult child. I estimate the model using detailed data from the Health and Retirement Study from 1998-2010. The model replicates the low demand for insurance across the wealth distribution in addition to asset accumulation paths and informal care rates. I find that the availability of family care lowers the demand for insurance by 5 percentage points overall and by 25 percentage points for wealthy individuals. Counterfactual insurance policies that incorporate family care can generate large welfare gains to families.
The second chapter examines the interaction between sources of long-term care more closely. Specifically, while most care in the United States is provided informally by family members, it is not obvious whether informal care acts as a substitute for formal care - such as nursing homes - or whether family and formal care service different health needs. I measure the extent to which family care responds to the price of formal long-term care services by exploiting panel variation in Medicaid 'spend-down' provisions across states.
Using data from the decennial Census, the American Community Survey and the Health and Retirement Study over the last 30 years, the analysis finds that Medicaid generosity decreases the use of family care by 1 to 4 percentage points for single individuals aged 80 and over, with a corresponding increase in the use of nursing home. These findings suggest that individuals substitute family care for nursing homes.
The third chapter, which is joint with Orazio Attanasio and Costas Meghir, presents a framework to quantify a network's potential to share risk and the extent to which the network takes advantage of this opportunity. The standard life cycle model implies that a household will prefer smooth consumption to a variable one. However, the amount of smoothing a household can achieve depends on the instruments at its disposal, such as formal insurance, tax and transfer policies, or informal risk-sharing. The paper first develops a framework of partial insurance and group risk sharing that decomposes income shocks into idiosyncratic (i.e. insurable by the group) and group-aggregate (i.e. uninsurable by the group) components. This allows us to study the differential impact of these shocks on consumption.
The second part of the paper applies this framework to extended family networks in the United States. The analysis uses data from the Panel Study of Income Dynamics, which provides longitudinal information on income and consumption of both parents and adult children. The estimates indicate that over 60% of income shocks are idiosyncratic within extended family networks. We argue that this implies a large potential for the family risk sharing network to have a non-trivial impact on the transmission of income shocks into consumption. However, no such insurance occurs on average. Suggestive evidence shows that some insurance is provided by family networks in narrower geographic areas.