No, Long-Term Care Costs Aren’t Going to Cause a Retirement Crisis*

*Unless we eliminate Medicaid, Medicare, state programs, private insurance and charities. Then sure, we’ve got a problem.

Debunking claims that Americans face a retirement crisis is apparently the reason I was put on Earth. Because the sooner I take care of one, another pops up. God doesn’t want me idle.

In my new book, The Real Retirement Crisis, I devoted an entire chapter to critically examining two decades of studies claiming that a crisis-level shortfall of retirement saving is just around the corner. (Note that the “two decades” and “just around the corner” implies that the early generation of such studies has already been shown to be incorrect.) I also recently published a survey of the retirement crisis studies in the Journal of Retirement, which is available ungated here.

But The Real Retirement Crisis came out all the way back in April, meaning there’s a new retirement crisis study on the autopsy table.

The latest is a report by Spencer Look and Jack VanDerhei of Morningstar, who have developed a computer model of retirement income security. The model estimates the probability that households will run short of money in old age. This specific study measures how long-term care costs – often referred to as Long Term Services and Supports (LTSS) – will affect retirement income security. Long-term care can include both in-home assistance and residential nursing homes.

Look and VanDerhei are smart and, to the very best of my knowledge, intellectually honest. Nevertheless, their study doesn’t at all say what you’d think it says.

In a summary article about their report, Look writes:

We calculate the percentage of households simulated to run short of money in retirement under two scenarios: one that includes LTSS costs, and a counterfactual where LTSS costs are excluded.

The results are striking. When LTSS costs are included, 41% of households are projected to run short of money in retirement—compared with just 26% when those costs are excluded.

That’s a big difference: long-term care costs will cause an additional 15% of seniors to run out of money in old age. That sure seems to undermine people like me who think the retirement crisis is largely a myth.

The authors of the report agree:

Our results help expand our earlier discussion on whether there is a retirement crisis in the United States. Notably, the model results excluding LTSS costs are more in line with optimistic assessments of retirement outcomes for the working population, highlighting the significant impact of omitting LTSS costs or assuming another source will cover them.

The picture is clear: long-term care costs pose a grave risk to financial security in old age. The retirement optimists should be taking a hard look in the mirror.

Or not. Because what the Morningstar team describe they are doing – comparing retirement outcomes with and without the inclusion of long-term care – isn’t what’s really happening.

Here’s why: “Including long-term care costs,” as Look and VanDerhei describe it, doesn’t mean calculating the long-term care costs that seniors will actually pay.

Rather, it means assuming that retirees bear nearly all of long-term care costs.

As the report’s background material states,

“These estimates reflect the costs a household would have to pay to avoid spending down to qualify for Medicaid-financed LTSS or otherwise relying on charity.”

In terms of judging the retiree population’s income security, this assumption is just nuts. It makes no sense, because currently, retirees bear only a tiny fraction of total long-term care costs.

The chart below is drawn from a Congressional Research Service analysis of who pays for long-term care. Total long-term care costs in 2021 equaled $468 billion. For context, that’s about 14% of seniors’ $3.3 trillion in total incomes that year. That would be a lot to ignore.

But as the CRS tabulation shows, retirees themselves paid only $64 billion out of pocket, meaning seniors paid 14% of total long-term care costs while 86% came from Medicaid, Medicare, private insurance and other payers.

In reality, less than 2% of seniors’ total incomes went toward out-of-pocket long-term care costs. One dollar out of fifty.

Morningstar’s estimates for the share of seniors who run out of money assumes that all long-term care costs are borne by retirees themselves. That’s just untrue.

The same picture holds when we look at who pays for Americans living in nursing homes. According to the Kaiser Family Foundation, three-quarters of nursing home residents have either Medicaid or Medicare as their primary payer. While Medicare typically does not pay for long-term care as such, the program will pay for a skilled nursing facility for up to 100 days following a hospital stay of at least three days.

Even if you feared that, due to budgetary pressures, Congress would reduce government appropriations for long-care by 50%, Morningstar’s figures would overstate retirees’ costs. Even if government spending on long-term care were cut by 75%, Morningstar’s household costs would still likely be too high.

Moreover, Morningstar’s report wasn’t about the prospect of Congress eliminating or severely cutting back on public support for long-term care. It was ostensibly about recognizing long-term costs that already exist. But that isn’t what the figures describe. Morningstar’s figures simply don’t mean anything. Neither of their simulations corresponds to reality.

But assuming that retirees pay nothing for long-term care is actually closer to the truth than Morningstar’s assumption that seniors pay 100% of costs out of pocket, given that we know seniors actually only pay about 14%.

This would be particularly true for lower-income seniors, where long-term care costs almost certainly would be covered by Medicaid. For instance, a 2022 analysis published by the Department of Health and Human Services projected that 73% of low-income seniors will pay nothing out of pocket for long-term care services. The overall out-of-pocket average for this group is slightly over $6,000. Not per year, but ever.

Sure, there is uncertainty regarding how much assistance government will be able to provide to seniors in the future. For that reason, for instance, projections of Social Security benefits sometimes show both promised level of benefits and the level that would be affordable if the trust funds were exhausted.

But those studies don’t assume the elimination of Social Security benefits and the elimination of households’ personal savings. But when Morningstar assumes seniors receive long-term care assistance neither from Medicaid and Medicare nor from private insurance, that’s more or less what they’re doing.

In short, the Morningstar analysis doesn’t point toward Americans facing a retirement crisis. If anything, it indicates the opposite.

Some wonky stuff

When I see an analysis like this, I look to the sky and ask “Why?” Why set up an analysis that doesn’t reflect reality? (The sky rarely answers.)

The Morningstar model predicts whether a household’s income and wealth will be sufficient to maintain its projected level of retirement spending. Retiree spending in the Morningstar model is based upon the Health and Retirement Study’s (HRS) Consumption and Activities Mail Survey (CAMS). The HRS itself includes demographic information on older households as well as details on their sources of income and wealth, while the CAMS includes data on retiree households’ expenditures in various categories. By itself, that approach is similar that used in studies by RAND Corporation economists Michael Hurd and Susanne Rohwedder, which project that roughly three-quarters of Americans will have sufficient resources in retirement. So far, so good.

But Morningstar adds a second element. Their model simulates retirees’ health status, ranging from good health, to poor health, to in-home healthcare, to nursing home care. Once the model predicts that a retiree requires in-home or nursing home care, the model incorporates costs for those care types from GenWorth’s Cost-of-Care survey. For instance, GenWorth finds the median annual cost of in-home care is about $26,000 while a semi-private room at a nursing home costs over $111,000 per year. Yikes.

But GenWorth’s survey measures the full pay private rates for long-term care, which ignore any discount, government program, or payment from private insurance. Sure, some people pay full freight, but the vast majority do not. That’s how the CRS ended up calculating that seniors themselves pay only 14% of total long-term care costs.

Why does Morningstar do this? I think the answer is that the Morningstar retirement model doesn’t explicitly simulate Medicaid and other third-party payers of long-term care.

But implicitly, it does. The HRS CAMS data that form the foundation of the Morningstar model do incorporate healthcare costs, including out-of-pocket costs for long-term care. And so, a more accurate estimate would be available if Morningstar simply stuck with the CAMS data throughout retirees’ lives, never abandoning it for data from GenWorth’s Cost-of-Care study. The HRS data aren’t perfect, but they’re better than assuming that governmental support for long-term care disappears overnight.