Facebook Twitter Tumblr Close Skip to main content
A Project of The Annenberg Public Policy Center

Clinton’s Shaky Cost-Shifting Claim


Hillary Clinton claimed in the Dec. 19 debate that private insurance premiums have “gone up so much” in some states that didn’t expand Medicaid because hospitals shifted their costs for providing emergency care for the uninsured. But we have found no data to support that claim, and the idea that such cost shifting occurs is debated among experts.

Clinton made her claim about cost shifting in the Democratic debate when she was asked about the increase in private insurance premiums that has occurred since 2010, when the Affordable Care Act was passed. Moderator Martha Raddatz of ABC News asked: “But for Americans who already had health insurance the cost has gone up 27 percent in the last five years while deductibles are up 67 percent, health care costs are rising faster than many Americans can manage. What’s broken in Obamacare that needs to be fixed right now?”

(Employer-sponsored insurance premiums for family plans have gone up 27 percent from 2010 to 2015, according to the Kaiser Family Foundation and the Health Research & Education Trust’s annual survey. That’s the same percentage increase as 2005 to 2010, but much lower than the hike from 2000 to 2005. The figure on deductibles also comes from the survey.)

Clinton put part of the blame for rising private insurance premiums on states that hadn’t expanded Medicaid, nearly all of which are run by Republican governors.

Clinton, Dec. 19: And one of the reasons in some states why the percentage cost has gone up so much is because governors there would not extend Medicaid.

And so people are still going to get health care, thankfully, in emergency rooms, in hospitals. Those costs are then added to the overall cost, which does increase the insurance premiums for people in the private system.

Clinton was explaining a concept called cost shifting, a controversial and debatable idea that hospitals charge higher rates to those with private insurance to cover uncompensated costs for caring for the uninsured. The federal government does provide some payments to hospitals that treat low-income patients. But those payments, called Disproportionate Share Hospital Allotments, or DSH, through the Medicaid program, aren’t enough to cover all of the uncompensated care costs.

Some researchers say this cost shifting occurs, and others say it doesn’t — or doesn’t occur on a large enough scale to make a significant difference in private insurance premiums. But we could find no research that shows cost shifting has led to increased private premiums in states that didn’t expand Medicaid, compared with states that did. (Thirty-one states, including Washington, D.C., have adopted the expansion of Medicaid called for by the Affordable Care Act, while 16 states have not. The expansion, which covers most adults up to 138 percent of the federal poverty level, is under discussion in four others.)

We asked two organizations that have published a lot of research and analysis of the impact of the Affordable Care Act: the Kaiser Family Foundation and the Urban Institute, and neither knew of any research that would address Clinton’s claim. A spokesman for the Kaiser Family Foundation told us KFF’s experts haven’t attempted to study whether private insurance is going up more in nonexpansion states because of any cost shifting, nor were they aware of any such studies.

We looked at KFF data on changes in ACA marketplace premiums from 2015 and 2016 in major cities in each state, and there’s no clear pattern for expansion versus nonexpansion states. Instead, there’s a wide range of increasing and decreasing premiums for each state. And statewide changes also vary: The New York Times reported in October that “premiums for the benchmark plan will decline by 13 percent in Indiana, 8 percent in Mississippi and 1 percent in Ohio and Maine.” Two of the four — Mississippi and Maine — are states that didn’t expand Medicaid.

It would take a lot more analysis, of course, to determine whether there are connections between uncompensated care and premium changes.

It is the case, however, that uncompensated care costs are higher in nonexpansion states than they could have been if those states had expanded Medicaid, a move that would have covered more of the uninsured. In an April report, the Kaiser Family Foundation estimated that states that didn’t expand Medicaid would face an additional $81 billion in uncompensated care costs for the uninsured from 2015 through 2024, compared with the uncompensated care costs they would have incurred if they had expanded Medicaid.

The question, then, is whether Clinton is right to say that those uncompensated care costs are passed on to private insurers and result in higher private insurance premiums. Experts disagree on whether, and to what extent, that happens.

Stephen Zuckerman, senior fellow at the Urban Institute and co-director of the institute’s Health Policy Center, told us that nonexpansion states “are no worse off than they were before,” in terms of the number of uninsured right now. Some of the uninsured in those states have gained coverage through the marketplaces or through Medicaid programs that already existed before the ACA. Zuckerman questioned the idea, therefore, that the burden of the uninsured on hospitals has grown. It “doesn’t seem logical to think it’s gotten worse … and is driving premiums up.”

It’s not clear how uncompensated care is changing, Zuckerman said. “It’s not even clear that hospitals’ financial needs are greater now to seek those higher payments,” he said.

Kenneth Thorpe, chair of the Department of Health Policy & Management in the Rollins School of Public Health of Emory University, has conducted research in the past for Families USA, a nonprofit that works to expand health coverage and calls the cost of uncompensated care a “hidden tax” on Americans with private insurance. Thorpe, who analyzed President Bill Clinton’s health overhaul proposals while working at the Department of Health and Human Services at the time, told us that higher uncompensated care costs in nonexpansion states would put “pressure on private insurance premiums.”

A June 24 Stateline article said that taxpayers could pay more in some states that didn’t expand Medicaid in the future, once the federal government begins reducing payments to help cover uncompensated care in 2018, as part of the Affordable Care Act. In many states, like Texas, local taxes are used to help hospitals that treat the uninsured.

Stateline reported that Texas hospitals spent $5.5 billion on uncompensated care in 2014 — that’s according to the state hospital association — while the federal government reimbursed $1 billion. That federal reimbursement could be reduced by about 20 percent in fiscal 2017, according to an estimate published in the November 2014 issue of the journal Health Affairs. The Texas Hospital Association told Stateline that cost shifting would make up the shortfall.

Stateline, June 24: Lance Lunsford, a vice president with the Texas Hospital Association, said without that federal money, residents would have to foot the bill in the form of higher insurance premiums and, in jurisdictions with public hospitals like Dallas County, higher taxes. “Hospitals have no choice but to limit their exposure,” he said. “That means cost-shifting to consumers and taxpayers.”

But that’s a prediction of what will happen in the future, once federal uncompensated care payments are reduced, not what’s happening now. And, if that does happen, it remains to be seen how much of that cost would shift to taxpayers and how much would result in higher premiums.

How Much Does Cost Shifting Occur?

It’s not clear to what extent cost shifting occurs in general. In 2014, the Kaiser Family Foundation published an analysis of the issue, by Urban Institute researchers, that called cost shifting of uncompensated care “a long-standing, complicated and controversial issue.”

The report cited data that suggested private insurance payments to hospitals were 30 percent above costs, while Medicare and Medicaid payments were under hospital costs. And some care for the uninsured isn’t paid for at all. “However, there is limited evidence to indicate that rising numbers of uninsured people (and thus increases in uncompensated care) have caused hospitals to increase their charges to the privately insured,” said the analysis.

The Urban Institute researchers found that if there was cost shifting to private insurance, it would be “relatively small” in scope. They estimated there was $21.1 billion in 2013 of uncompensated care nationwide that wasn’t reimbursed by government sources or covered by office physicians doing charity work. That represented 2.3 percent of total private health insurance spending for the year.

If hospitals don’t pass on costs, what do they do about uncompensated care costs? Craig Garthwaite, an assistant professor of strategy at Northwestern University’s Kellogg School of Management, found hospitals largely absorbed the costs.

Garthwaite researched years of uncompensated care records at hospitals and the relationship to hospital profits, finding no evidence that the vast majority of hospitals passed uncompensated care costs onto privately insured customers. “We find hospitals absorb nearly all of the increase in uncompensated care through reduced profits,” he told us in a phone interview, estimating that 60 percent to 100 percent of such costs are absorbed by the hospitals.

Garthwaite explained his findings in a June 2015 article on the Kellogg website. If cost shifting were happening, he said, “we wouldn’t see profits fall — but we see profits fall meaningfully following an increase in the share of the population that is uninsured.”

Garthwaite and Zuckerman both questioned whether hospitals could actually raise prices to cover uncompensated care costs. If higher prices increase profits, Garthwaite told us, “they should’ve been charging these higher prices beforehand.” The only way hospitals could shift costs would be if they weren’t charging a profit-maximizing price before, he said.

He said he knew of “no credible research” that would show a change in the insurance premiums in nonexpansion states was linked to uncompensated care costs, a causal relationship he says would be “very hard to tease that out” if it were true. “At best this is a sloppy claim and at worst it’s deliberately misleading,” he told us of Clinton’s comment.

Researchers with MedPAC, an independent agency that advises Congress on Medicare, published a paper in Health Affairs in 2010 that questioned the “common assumption” that hospitals charge higher rates to private health insurance when Medicare payment rates are below hospitals’ costs (the same type of cost shifting that Clinton says is occurring for caring for the uninsured). “We present evidence that contradicts that common assumption,” the report said, explaining that some hospitals had stronger cost controls than others.

“Hospitals with strong market power and higher private-payer and other revenues appear to have less pressure to constrain their costs,” the researchers wrote. “Thus, these hospitals have higher costs per unit of service, which can lead to losses on Medicare patients. Hospitals under more financial pressure — with less market share and less ability to charge higher private rates — often constrain costs and can generate profits on Medicare patients.”

The left-leaning Families USA, however, believes cost shifting does occur, and the group put a price on it: $1,017 for the year for private family health coverage in 2008. That figure, however, comes from taking the national cost of uncompensated care for the uninsured and dividing it among the privately insured population.

Obama cited this figure back in 2009, and we said at the time that it was disputed. In a 2008 report, Urban Institute and KFF researchers took issue with an earlier Families USA estimate, saying: “[W]e are highly skeptical that the high and growing cost of private insurance is strongly related, if at all, to the amount of uncompensated care delivered by private providers or to the growing number of uninsured people.”

The actuarial firm Milliman Inc. conducted the recent analysis for Families USA, and it notes in the technical “limitations” section that this was based on the assumption that uncompensated care is passed on to private insurers. “It is possible that uncompensated cost is absorbed by providers in the form of lower profits or margins, or that uncompensated cost should be considered a marginal cost of doing business,” Milliman says.

Economist Austin Frakt, then with Boston’s Veterans Affairs health care system and Boston University, conducted a review of the available literature in 2010 and concluded that “[p]olicymakers should take hospital and insurance industry claims of inevitable, large scale cost shifting with a grain of salt. Though a modest degree of cost shifting may result from changes in public payment policy, it is just one of many possible effects.”

Frakt wrote that the “[m]ore careful theoretical and empirical examinations” he reviewed “suggest that cost shifting can and has occurred, but usually at a relatively low rate.” He pointed to changes in hospital market power as having “a significant impact on private prices.”

There are differing opinions on whether or not cost shifting occurs and to what extent — and Clinton is entitled to her opinion on that. But there’s little in the way of hard evidence showing that it’s common practice, and again, no analysis that we could find showing that it has put more upward pressure on private insurance premiums in states that haven’t expanded Medicaid under the Affordable Care Act.

Clinton Campaign’s Alternative Explanation

When we asked the Clinton campaign for support for this claim on cost shifting, it first pointed to a separate issue: The idea that the ACA marketplaces in nonexpansion states would include a higher percentage of unhealthy individuals and that would lead to higher marketplace premiums. In expansion states, such low-income individuals would be on Medicaid and therefore not affecting the risk pool for the individual market insurance plans.

Thorpe, at Emory University, also referred us to one such analysis, a September 2012 issue brief by the American Academy of Actuaries that said states’ decisions on the Medicaid expansion would impact private insurance premiums.

Here’s the reasoning: The Medicaid expansion covered low-income adults earning up to 138 percent of the federal poverty level. Subsidies for the state and federal marketplaces begin at 100 percent of the poverty level. So, in states that didn’t expand Medicaid, those earning between 100 percent and 138 percent could get subsidized coverage on the marketplaces, while in expansion states they’d be on Medicaid. This population of low-income individuals is likely to have more health issues than higher-income individuals, causing risk pools to contain more high-cost enrollees, and therefore leading to higher premiums in the individual market.

That issue brief cited a July 2012 Congressional Budget Office report, which said that lower-income individuals “generally have somewhat poorer health. As a result, CBO and [the Joint Committee on Taxation] now estimate that the premiums for health insurance offered through the exchanges, along with premiums in the individual market, will be 2 percent higher than those estimated in March 2012.”

Other researchers have written about this impact of having a greater population of high-health-cost individuals in marketplace risk pools than would be the case if all states expanded Medicaid. The Clinton campaign cited a November 2014 study by researchers with the University of California, San Diego and UC Berkeley that found evidence that not expanding Medicaid led to increases in private plan deductibles, of 10.5 percent on average, but did not coincide with changes in premiums.

The ACA marketplace coverage began in 2014, so it’s still early in terms of having a body of research that documents the impact on premiums, and the impact of the Medicaid expansion. But so far, Clinton’s cost-shifting explanation is debatable and not backed up by the available data.