The Value of Medicaid to Older Households
Medicaid is an importance source of funding for long-term care for the elderly. While Medicaid is the payer of last resort, contributing only after Medicare and private insurance pay their share and individuals spend their assets down to a relatively low "disregard" amount, the very high cost of nursing home care - on the order of $60,000 to $75,000 per year - means that Medicaid ends up financing care for seventy percent of nursing home residents. While most individuals on Medicaid qualify based on their low income and assets (the "categorically needy"), some states also offer eligibility for those with higher income who face high medical expenses (the "medically needy"). Thus, the reach of the Medicaid program extends beyond the poorest elderly to include many middle-income and some higher-income families as well.
Despite the rising importance of Medicaid to an aging population and the program's rising costs, relatively little is known about how Medicaid benefits are distributed among the elderly and about their value to older households. In Medicaid Insurance in Old Age (NBER Working Paper 19151), researchers Mariacristina De Nardi, Eric French, and John Bailey Jones aim to help fill this void. Specifically, they explore issues such as which elderly households receive Medicaid, what the insurance value of those benefits is, and how much people would lose if benefits were cut.
The authors begin by documenting Medicaid use using the Assets and Health Dynamics of the Oldest Old (AHEAD) data. Among respondents in the lowest quintile of lifetime income, nearly seventy percent are receiving Medicaid at age 74, a fraction that remains roughly constant as the group ages. In the two highest quintiles, eligibility at age 74 is negligible, but grows to twenty percent by age 96. As the authors note, "even high income people become Medicaid recipients if they live long enough and are hit by expensive conditions."
Next, the authors use the AHEAD data to estimate a rich model of individual saving and medical expenses, where medical expenses depend on both the individual's resources and his or her changing health status. In estimating the model, they choose parameters so as to enhance their model's ability to match aggregate out-comes in the data such as out-of-pocket medical expenses and the fraction of people on Medicaid or in nursing homes.
The model generates some interesting findings regarding medical spending. In the top income quintile, out-of-pocket spending rises rapidly from around $4,000 per year at age 74 to $20,000 at age 100; for the lower two income quintiles, out-of-pocket spending remains below $2,000 per year as an individual ages. Once Medicaid spending is added, however, all income quintiles experience rapidly rising medical spending with age, and the gaps in spending across income quintiles narrow dramatically, with the highest income quintile never spending more than twice as much as the lower two quintiles. While the bottom two quintiles are more likely to qualify for Medicaid and thus receive higher average Medicaid transfers than the top quintile, both groups receive substantial benefits at very old ages, with mean Medicaid spending of $10,000 per year in the top income group by age 100 and nearly twice that amount in the low income group.
Finally, the authors conduct several simulations in which they change the Medicaid program rules and compute the wealth transfers that would make each individual as well off as he or she had been prior to the change. One such change is to reduce the consumption floors (allowable consumption for those that are categorically or medically needy) by 25 percent. This drops the value of Medicaid payments to individuals in all income groups, but it is the top income group that most values the lost insurance—at five times its cost, compared to just slightly above cost for the lower two income groups. The insurance value is greater for higher income people because they have more consumption to lose if they begin to incur large medical expenses and are at greater risk of living a long time and facing very high medical needs. When the authors simulate making Medicaid more generous by raising the consumption floors, all income groups except the highest value the expanded benefit at less than its cost.
In concluding, the authors note, "although the lifetime discounted presented value of Medicaid payments does decrease with permanent income, even higher income people can receive sizeable Medicaid payments, as they tend to live longer and face higher medical needs in old age. Furthermore, our compensating differential calculations show that many higher income retirees value Medicaid insurance as much or more than lower-income ones."
Jones acknowledges support from the Social Security Administration through the Michigan Retirement Research Center (MRRC grants UM10-16 and UM12-12).