Moral Hazard and Long-Term Care Insurance

The financing of long-term care will become increasingly challenging as the US population ages and the prevalence of chronic conditions grows.  Currently, the Medicaid program serves as the default payer for long-term care when individuals exhaust private resources, but Medicaid-funded long-term care is associated with poor quality and the expense is straining state budgets.  Although only 13% of elderly individuals have private long-term care insurance (LTCI), an expansion of the private market is often posited as one solution to the financing challenge because any increase in private responsibility could offset some of the public burden.  At the same time, significant concerns exist over the sustainability of the private market.

In both public programs and private long-term care insurance markets, estimation of the degree of moral hazard in utilization of covered services is central to pricing and long-run sustainability, yet there is remarkably little evidence on the degree to which moral hazard exists.  Two prior studies found mixed results on moral hazard, one finding no evidence of moral hazard in Medicaid-funded nursing home use and one finding some moral hazard in privately insured nursing home use.  We use US Health and Retirement Study data from 1996-2010 to assess moral hazard in nursing home and home care use in the presence of private LTC insurance, employing bivariate probit models with instrumental variables to estimate consistent effects.  Our paper addresses the limitations of the existing literature by examining utilization of long-term care services using multiple waves of the Health and Retirement Study with geographic identifiers, enabling better instruments to control for selection bias.  Rather than examining the holding of LTCI, we model LTCI purchase, which allows the use of stronger instruments based on changes over time in state tax treatment of LTCI.  We then examine utilization of long-term care services in subsequent waves of the survey as a function of prior LTCI purchase.

We find evidence of significant moral hazard in nursing home use, in contrast with the lack of a moral hazard effect in Medicaid-funded nursing home use from prior studies. Medicaid-funded nursing home care might be considered an inferior good, in which case the lack of a moral hazard effect is less surprising.  In this paper, we examine the private LTCI market and find a significant effect.

Our finding of moral hazard in nursing home use has several possible implications. One is that public policies designed to encourage LTCI purchase and to cover long-term care services should consider the additional spending associated with moral hazard and potentially incorporate disincentives to socially inefficient spending by policyholders.  However, the current moral hazard literature is marked by inadequate attention to the welfare effects of moral hazard, which depend crucially on the extent to which the increased utilization in the presence of insurance is efficient (based on income effects) or inefficient (based on price effects).  Disentangling the welfare effects of increased utilization of long-term care services is an important area of further research.