Evaluating the Allocative Efficiency in Long Term Care Using Variation in Occupancies

Wasteful spending is a major policy concern in many health care systems, especially when financed through public sources. In this paper, we analyze overuse of nursing home care, relative to other forms of long term care, which has come to the forefront of policy debate. Various states have expanded community care programs over the last decade that aim to reduce the demand for institutional care and thereby reduce Medicaid long term care spending (e.g., through Money Follows the Person grants). In this paper, we test whether nursing homes counteract these cost saving initiatives by prolonging nursing home stays of Medicaid beneficiaries whenever it is in the nursing home’s best financial interest.

We examine this possibility by comparing the length of nursing home stays between private residents that pay out-of-pocket and Medicaid beneficiaries at low and high occupancy rates (the fraction of occupied beds). Economic theory predicts that nursing homes have a financial incentive to prolong the nursing home stay of Medicaid beneficiaries whenever they have free capacity and regulated Medicaid reimbursement rates exceed marginal costs. Prolonging the stay of Medicaid beneficiaries is facilitated by moral hazard of Medicaid beneficiaries who do not pay for their nursing home stay out-of-pocket. At high occupancies, the nursing home’s incentive is muted because it is willing to keep beds vacant for more profitable private residents who pay generally higher rates out-of-pocket.

We test these hypotheses using health assessment data from the Minimum Data Set (MDS) on the universe of nursing home residents in California, New Jersey, Ohio, and Pennsylvania. We combine the assessment data with administrative Medicaid and Medicare claims data to assess the payer source for each resident on each day. We focus on residents who initially pay for their stay out of pocket. Variation in payer type changes stems from differences in wealth levels, which determine the point in time when a resident has spent down most of her wealth and becomes eligible for Medicaid. Exogenous variation in occupancy within facility and year comes from changes in the number of filled beds other than the individual resident’s. To control for potential differences in health profiles between payer types, we include week-of-stay fixed effects and observable health measures from the MDS.

Our empirical findings support the theoretical predictions. We find smaller weekly discharge probabilities for Medicaid residents than for private payers at low (less than 85 percent) occupancy rates, which indicates prolonged nursing home stays. We also see a clear increase in weekly discharge probabilities for Medicaid beneficiaries as the nursing home approaches full capacity while discharge probabilities for private payers remain constant as the occupancy increases. The variation in discharge patterns is almost only visible on the discharge to home margin, suggesting that nursing homes prolong stays for relatively healthy Medicaid beneficiaries, who are returning to the community. We find no health benefits from extended nursing home stays, indicating that the extension of nursing home stays for Medicaid beneficiaries at low occupancies is socially wasteful.