EVENTS

Date:	Thursday, February 17, 2005
Time:	3:30 PM -- 5:00 PM
Location:	Wohlstetter Conference Center, Twelfth Floor, AEI 1150 Seventeenth Street, N.W., Washington, D.C. 20036

February 2005

Does Medicaid Crowd Out Private Long-Term Care Insurance?

Medicaid finances care for 70 percent of all people in nursing homes and covers half of all the money spent for long-term care. The strain on state budgets is intense, and the National Governors Association has demanded federal relief. Private long-term care insurance could reduce some of those spending pressures, but few people buy such coverage today. Is Medicaid making its own problems worse by discouraging people from investing in long-term care insurance? At a February 17 AEI conference, Jeffrey Brown of the University of Illinois presented a new study of Medicaid crowd-out (co-authored with Amy Finkelstein). A panel of experts offered their critique of the study and assessed the barriers facing private long-term care insurance.

 

Jeffrey R. Brown
University of Illinois

Professors Jeffrey Brown and Amy Finkelstein approach long-term care (LTC) from their perspective as public finance economists. Various (and varying) policy implications can be drawn from their findings, but the authors do not recommend any specific actions in the paper. 

Insurance in the United States is generally provided by a mix of public and private entities. One question addressed by this paper is whether, even when public insurance is incomplete, it can crowd out more comprehensive private insurance, thus potentially increasing the population’s overall exposure to risk.

Long-term care is a major source of risk for seniors, costing $135 billion in aggregate in 2004. Nonetheless, only 10 percent of the elderly buy private coverage, and private insurance pays for only 4 percent of LTC costs. One-third of long-term care expenditures are paid for out-of-pocket--double the share of costs that individuals incur for health care overall.

Professors Brown and Finkelstein have two papers on LTC insurance. The first considers the market for long-term care insurance in general. It concludes that the low penetration of private plans is primarily caused by demand-side problems, even though supply-side issues plague the market as well. The second paper focuses on one particularly important demand-side explanation--the role of Medicaid in influencing the demand for LTC insurance.

Medicare is the primary payer for nursing home care, but that coverage is restricted to post-acute care, it does not pay for long-term care per se. Medicaid is the secondary payer of long-term care expenses, following private insurance. It provides essentially free coverage to individuals meeting strict income and asset tests--to qualify, individuals must first exhaust their personal resources. Consequently, it offers incomplete insurance protection.

To assess demand, the authors estimate how much a sixty-five-year-old would be willing to pay for private LTC coverage. The private policy they model has a $100 daily benefit cap that covers only two-fifths of expected costs. The annual premium is approximately $1,800 for both men and women, despite the fact that women typically live longer and use substantially more LTC.

The authors’ findings are largely consistent with the empirical data on LTC coverage. Their model predicts that individuals at most income levels will not purchase existing LTC policies, which accords with the fact that 90 percent of the elderly today are uninsured for LTC. Individual willingness to pay also rises with wealth; surveys observe this trend in coverage rates across wealth levels. Their model also suggests similar willingness to pay for men and women. Though LTC coverage is a far better deal for women, surveys report similar take-up rates for both sexes.

The model shows that only the wealthiest people are willing to buy private insurance in the presence of Medicaid LTC coverage. That is true in their model even when men are charged actuarially fair premiums. Between two-thirds and 90 percent of people would not buy private insurance because of the existence of Medicaid, even taking other factors into account.  This is because a large share of the benefits that are provided by a private policy are duplicative of what Medicaid would otherwise cover.  

Medicaid is a “secondary payer” for LTC. That is, private insurance pays first, and Medicaid pays only after the private benefit is exhausted. This adds to the implicit tax on private insurance. Accounting for this tax, the “net load,” or difference between the cost of the policy and the expected benefits, is much higher.

People must divest themselves of their financial assets before becoming eligible for Medicaid--effectively an insurance deductible paid by the person before Medicaid pays. As a result, this interferes with consumption smoothing across health states and time and leaves individuals exiting from care with few resources. Most people would be willing to complement, or “top up,” Medicaid with a second policy if it were possible, but such policies do not exist, and indeed are conceptually difficult to implement.

Policy interventions to encourage investment in private long-term care coverage have not been able to eliminate Medicaid’s large implicit tax on private plans. Tax subsidies for the purchase of private insurance have been too small to fully compensate for the implicit tax. Some states have scaled down Medicaid’s asset tests for individuals who buy a minimum amount of private insurance, but these Partnership programs have, so far, not significantly impacted the take-up of private coverage. Even making Medicaid the primary payer would fail to induce most of the population to purchase private insurance.

To effectively stimulate the demand for private LTC insurance, Medicaid payment for LTC services would need to be separated from private coverage status. For instance, the government could offer a refundable tax credit to individuals buying private coverage equal to the costs that Medicaid would have incurred for their long-term care.  This proposal, however, is difficult to implement due to a variety of concerns, including adverse selection.  In addition, such a program might actually increase government outlays. 

Mark R. Meiners
George Mason University

It is debatable whether Medicaid is an attractive alternative to private coverage. On the other hand, insurance products tend not to appeal to individuals with limited resources. Are low-income people, then, more likely to divest than people with higher incomes? Will a more generous Medicaid increase taxes?

The authors find that, despite the questionable quality of Medicaid coverage, it effectively crowds out private alternatives. Their more enigmatic conclusion is that Medicaid imposes an implicit tax on private insurance by forcing private plans to cover services that Medicaid would otherwise have paid for. That finding assumes that divestiture is rampant. The implicit tax raises the net load, which captures both the loss ratio and the asset divestiture necessary to get on Medicaid, and the price of private long-term care insurance. If you can get something for free on one side of the road, you will not cross the street to buy it elsewhere.

Professors Brown and Finkelstein find that state and federal policies to stimulate demand do not work, even though Partnership incentives appear to increase willingness to pay by as much as 30 percent. Medicaid, they conclude, is the real barrier to private insurance, and the incentives people face to seek coverage through Medicaid must be reworked.

Private insurance cannot by itself reach the people most vulnerable to spending down on long-term care. Partnerships for LTC merge Medicaid and private insurance, allowing people to protect some of their assets by substituting private insurance payment for Medicaid payment. They encourage high-quality and comprehensive insurance, though it can be short-term if people spend down and qualify for Medicaid.

Partnerships protect consumers against inflation, preserve the balance between cost and quality, inform consumers, standardize reporting for insurers, and protect individuals’ assets.

LTC Partnerships offer an efficient subsidy to induce middle- and modest-income individuals to get private insurance. That deters seniors in the program from shifting assets solely to become eligible for Medicaid LTC coverage. Asset transfer is a last resort for most people who face impoverishment from long-term care, particularly for middle-income people who finally face viable options to protect against future LTC costs. Partnership programs can bolster consumer confidence because insurance companies are no longer the only ones backing the policy’s quality. This is a public program that involves the middle class to strengthen the safety net.

In short, Partnerships increase value and decrease cost. They can double the size of the potential LTC insurance market and lay out straightforward criteria to help consumers purchase protection.

Partnership programs have sold fewer than 200,000 policies, largely because the Omnibus Budget Reconciliation Act of 1993 (OBRA 1993) effectively restricted their operation to only four states. The modest take-up of Partnership insurance only suggests that the Partnerships were squelched by restrictive policies. This does not mean that those programs cannot work, and lawmakers should not be discouraged from implementing them elsewhere.

Long-term care financing has been debated for years. One option is socializing LTC insurance by adding it to Medicare, for example, but Medicare’s current fiscal realities make this unlikely. Another approach would be to eliminate Medicaid’s means-testing, but many people rely on the richer benefits that means-testing allows. Professors Brown and Finkelstein propose improving Medicaid, perhaps referring to expanding eligibility, but that again is improbable from a fiscal standpoint.

Public-private Partnerships hold the most promise. They establish a social contract in which individuals who protect themselves are assured a social safety net through Medicaid. The Partnership insurance strategy offers workable versions of the concepts of "topping up" Medicaid benefits and "consumption smoothing" that are identified in the paper as important outcomes to be gained with long-term care insurance. Removing the OBRA 1993 restrictions on Partnership insurance growth is therefore an important policy option that this research may well support.  It deserves a closer look from the perspective of this new research.

Stuart Hagen
Congressional Budget Office

Professors Brown and Finkelstein have illuminated many of the problems and inefficiencies in our long-term care financing system. Their application of consumption models to this area is both innovative and useful. Their work could prove very valuable in gauging the impact of different policy proposals on long-term care.

Their analysis raises the question of people’s actual preferences for long-term care protection: do people want comprehensive coverage, or is catastrophic insurance sufficient?

Long-term care expenditure comprised about 1.2 percent of GDP in 2004 and is projected to rise to 1.4 percent in 2040. That estimate assumes that disability rates continue to decline, as recent trends suggest, but LTC spending could be 2 percent of GDP instead if those rates plateau. Federal spending for Medicare and Medicaid--a major payer for LTC--is projected to increase dramatically over the coming decades.

The model used by Professors Brown and Finkelstein provides an alternative to models that borrow simple elasticities of demand from the literature on health insurance demand. Those earlier models assume that the demand for long-term care insurance is similar to the demand for insurance for acute health care. The authors’ measure of willingness to pay can be interpreted as a reservation price, and as such used to model different subsidies and credits for LTC insurance. The model could also be employed to simulate the impacts of non-tax policies, such as offering supplemental Medicaid coverage or changing Medicaid eligibility rules. It could also be useful in modeling new health insurance products, like health savings accounts, that allow for wealth accumulation as well as payment of health care costs.

It may be worthwhile to test the model’s sensitivity across varying utilities to capture the variation of marginal utility of consumption between time spent in a nursing home and healthy years lived at home. The utility could be varied across institutional settings (nursing home, assisted living facility, etc.) or across health statuses. It could also incorporate varying expectations of being nursing-home-bound, which may be higher, for example, in people with a family history of Alzheimer’s or another debilitating disease.

The demand for more comprehensive coverage stems from circumstances imposed by Medicaid’s financial eligibility requirements. Beneficiaries in nursing homes have few resources available for non-care consumption or to return to the community upon recovery. Medicaid also offers no protection for bequests. At the same time, nursing home residents value non-care consumption less than they did when they lived on their own.  Moreover, very few will ever recover and return to the community, where their demand for non-care consumption would probably increase, as would their desire for financial resources to finance that consumption.

The 1999 National Nursing Home Survey found that three-quarters of nursing home residents leave the nursing home alive, but most of those are bound for a hospital or other care facility. Of the one-third that is discharged to home, only 3.5 percent stay home longer than six months. Individuals who stay in the nursing home less than three months are usually not receiving long-term care. They are typically recovering from an acute illness, and their stay is paid for by Medicare. So only individuals staying in nursing homes for longer than six months can have their resources decimated. Of all discharges, only 2 percent are for individuals who have stayed in nursing home care for longer than six months (and spent down on Medicaid) and who are headed home. It may be the case, then, that some people cannot afford to leave the nursing home, even if they are well enough to do so. The authors suggest that more comprehensive insurance could enable more people to reenter the community once they recover. But even with strict spend-down requirements that drain people of their wealth, returning residents still have their incomes, residences, and cars. Comprehensive insurance, then, may not be of as great value as Professors Brown and Finkelstein posit.

Robert B. Friedland
Georgetown University

Many of Professor Brown’s modeling assumptions are obscure, leaving his conclusions difficult to interpret. The model results raise three specific concerns: the statistic on nursing home use resembles the figure for total long-term and post-acute care use, but the latter is covered by Medicare and health insurance and would overstate LTC expenditures and optimal demand. Second, the demand for long-term care insurance compares the lifetime utility with and without insurance and computes the amount of wealth that would match that utility. The authors appear to consider income and financial wealth in the same way, but people are likely not as willing to liquidate their assets to obtain coverage as they would be to dispose of some income. Finally, the decision to purchase LTC insurance is different from other decisions in several ways. For instance, the existence of a spouse matters, and all assets and income are considered jointly. The paper is unclear on how it treats this dynamic.

The policy modeled in the paper is a very expensive choice--a $100-a-day lifetime benefit with no waiting period. It is plausible that the purchase of an inadequate insurance policy decreases expected utility, but the model does not appear to allow for this. Medicaid may indeed crowd out private insurance, but that may only occur at the margin in the same manner that Medicaid crowds out private health insurance. Why would the LTC crowd-out be so much more complete?

Long-term care is not inevitable, and people younger than eighty have a low risk of needing LTC services. The modest lifetime risk of needing long-term care should make LTC insurable, but there was inadequate data available to start-up insurance companies to develop their products and price them accurately. These insurers had little information on family caregiving, which is a significant component of long-term care. As a result, LTC products typically provide fixed dollar payments per day in a nursing home (regardless of the actual daily cost). In effect, LTC coverage insures against the risks that individuals purchase, not the actual risk of using long-term care services.

There are many barriers to private LTC insurance other than Medicaid. The decision to purchase insurance involves several variables--income, marital status, children, and especially the cost of the policy. It is too expensive to purchase more than once. Unlike mortgages or other major investments, they cannot be refinanced if the purchaser finds a better deal.

The risk pool for LTC insurance is relatively limited. The policy covers only the fixed risk that the purchaser decides. Also, the policy must be sold, making it a very labor-intensive process.

We know very little about individuals who do not purchase insurance. Some cannot buy LTC protection because their income is too low; others have chronic conditions that make the policy too expensive to afford. More generally, long-term care has only recently become part of formal retirement planning.

A typical long-term care policy for a couple can cost $400 a month. Combined with Medicare Part B and Medigap premiums, the average cost of health insurance in Montgomery County, Maryland, is $776 a month, or 40 percent of the income of a couple at the poverty level, where most seniors fall. Even for a couple with an income equal to four times the poverty level, that health insurance would consume 20 percent of their monthly budget.

So what options do seniors face? They can purchase LTC insurance, independently save for their long-term care, plan to move to a life care community, or divest to qualify for Medicaid and pay substantial taxes on their assets.

Medicaid may crowd out private insurance at the margin, but it is completely rational to pass on current insurance options, which are too few and too expensive. To an extent, people experience denial in their long-term planning, but they generally save what they can. Who, after all, has planned exactly along their optimal consumption path? If people can scarcely commit to a cell phone provider, we should not expect long-term care planning to be a perfectly rational exercise.

AEI research assistant Ximena Pinell prepared this summary.