Testimony before the House Committee on Energy and Commerce (Subcommittee on Oversight and Investigations)

On January 31, the House Committee on Energy and Commerce Subcommittee on Oversight and Investigations, holds a hearing entitled "Medicaid Oversight: Existing Problems and Ways to Strengthen the Program." 

Watch Paul Howard deliver his testimony (beginning at 20:00).

Summary

Medicaid for the 21st Century: Improving Health Outcomes, Accountability, and Efficiency in Partnership with the States

Medicaid is a vital component of our nation’s safety net for low income Americans.  Many physicians, nurses, and hospitals go to extraordinary lengths to provide health care to low-income populations that face many non-health related challenges, including to access to affordable housing, transportation, and employment.

But the current structure of the Medicaid program – particularly the open-ended federal matching formula, or FMAP - has led to vast unintended fiscal consequences, encouraging states (and the federal government) to devote ever larger sums to Medicaid to draw down additional generous federal matching funds not available to other state programs.

Thus, Medicaid has effectively crowded out funding for other state safety net programs that might have a greater likelihood of improving measured health outcomes. Providing a significantly increased earned income tax credit, supportive housing for the seriously mentally ill, or well-designed prisoner re-entry programs might all be better investments in the long-term health and economic mobility of low income Americans than spending more dollars on health care per se.

Congress, in consultation with the states, should reform Medicaid to ensure that federal subsidies don’t encourage states to allocate ever larger sums of money to health care that might be better spent on safety net supports elsewhere, including by beneficiaries themselves. Structural reforms to federal Medicaid funding, along with greater regulatory, financial, and administrative flexibility from Washington in how states manage their Medicaid programs and other sources of safety net spending would better serve the interests of America’s most vulnerable citizens and should be a bipartisan priority of the 115th Congress.

Testimony

Chairman Murphy, Ranking Member DeGette, members of the committee.  I’d like to thank you for the opportunity to testify today about Medicaid Oversight: Existing Problems and Ways to Strengthen the Program. 

Medicaid is a vital component of our nation’s safety net for low income Americans.  Many physicians, nurses, and hospitals go to extraordinary lengths to provide health care to low-income populations that face many non-health related challenges, including access to affordable housing, transportation, and employment. Any one of these factors can make both providing care and patient compliance with recommended care exceptionally difficult. 

But it is critical to understand how the current structure of the Medicaid program – particularly the structure of the open-ended federal matching formula, or FMAP - has led to vast unintended fiscal consequences, crowding out funding for other state level programs that might have a greater impact on improving measured health outcomes for the poor.  Providing a significantly increased earned income tax credit, supportive housing for the seriously mentally ill, or well- designed prisoner re-entry programs might all be better investments in long term health and economic mobility for low income and vulnerable populations than spending more dollars on health care per se. 

As you know, Medicaid is hybrid program, with funding responsibilities shared between the federal government and the states. Each state must meet certain federal guidelines, but each state administers its own Medicaid program. States also establish reimbursement levels for health-care providers and can, with federal approval, add optional benefits to Medicaid coverage and expand eligibility beyond the populations identified by federal baselines.  In fact, even setting aside the ACA’s Medicaid expansion, every state has expanded its program to include optional populations and services.[1]

The matching rate for federal funding varies greatly by state, with poorer states receiving larger shares from the federal government. According to the most recent figures, Kentucky receives the highest federal share (79.6%), with Wyoming and Virginia close to the minimum match of 50%.  On average, about 62% of Medicaid funding comes from the federal government, according to the Kaiser Family Foundation.[2] Thus, for each dollar states commit to their Medicaid program, the federal treasury automatically provides almost two more.

Because each state has an incentive to minimize its own financial responsibility while maximizing the drawdown of federal dollars, states have found ever more "creative" (and sometimes legally questionable) ways to draw down additional federal dollars.[3]  This is a predictable result of the uncapped federal matching formula; it is also responsible for a byzantine web of formulas, cross subsidies, and supplemental payments that makes evaluating Medicaid program integrity and efficiency extremely challenging, and the GAO has identified it as a high risk program since 2003.[4]

What is much less often commented on is the program’s regressive structure.  Because poorer states have fewer resources (from a smaller tax base) available to devote to health care services, the current FMAP matching rate provides much more federal support for wealthy states compared to poorer ones.  In their 2010 book, Medicaid Everyone Can Count On, Thomas Grannemann and Mark Pauly note that:

"In general, Medicaid benefits are considerably higher in higher-income states than in lower-income states, as are Medicaid payments per beneficiary, despite the much higher federal matching percentage share in lower income states. This spending level is about 89 percent greater in the highest quintile of states by income compared to the lowest."

Pauly and Grannemann's observation is confirmed by the experience of New York, one of the highest states in Medicaid per capita spending and total spending.  While the Empire State has only 6%[5] of the nation's population, it accounts for more than 11% of national Medicaid spending (in 2015, the price tag for New York's Medicaid program was estimated to be nearly $60 billion).[6]

According to a 2014 report from the Medicare and CHIP Payment and Access Commission (MACPAC) using the most recent administrative data from 2011, New York spent 44% more per Medicaid enrollee than the national average ($10,426 versus $7,236) and spent more than almost any other state across every Medicaid category:  New York spent 21% more per adult enrollee than the national average ( $5,297 versus $4,368), 68% more per disabled enrollee ( $31,989 versus $19,031), and 56% more per aged enrollee ($25,382 versus $16,236). Spending on children was comparatively modest: New York spent only about 3% more than the national average ($2,961 versus $2,854).[7]

While differences in cost of living undoubtedly explain some fraction of the disparity between New York's figures and those of the rest of the nation, much of the gap results from the program’s financial incentive to draw down additional federal dollars. 

Complex and fragmented funding streams also makes it extremely difficult to provide adequate accounting controls for the program. In 2012, reports from the Congressional Committee on Oversight and Government Reform and the Office of the Inspector General at the U.S. Department of Health and Human Services revealed that New York had systematically overbilled federal taxpayers for Medicaid services for the mentally disabled for nearly 20 years.

New York's state developmental centers — which offer treatment and housing for individuals with severe developmental disabilities — had received $1.5 million annually per resident in 2009, for a total of $2.3 billion. Of that amount, the HHS Inspector General found $1.7 billion to be above actual reported costs. State centers were compensated at Medicaid payment rates ten times higher than the Medicaid rates paid to comparable privately run developmental centers. 

How did these enormous overpayments go unnoticed for nearly two decades? According to the congressional oversight-committee report, the overbilling resulted from a funding formula agreed to by HHS and state Medicaid officials in 1990. Over the course of the following 20 years, however, HHS never audited the payment rate to ensure that it was still in line with actual costs. State officials, of course, had no incentive to bring the overpayment to the attention of federal regulators. As a result, New York benefitted from $15 billion in excessive payments.[8]

Despite its outsized spending, New York's health-care outcomes have historically ranged from poor to average compared with other states'. For example, in a 2009 report by the Commonwealth Fund, New York ranked 50th in avoidable hospital admissions.[9]

In 2011, Governor Andrew Cuomo's Medicaid Redesign Team, tasked with slowing program growth and improving patient outcomes, conceded that the state's Medicaid program offered poor value for both enrollees and state taxpayers — in part because of its excessive focus on institutional and hospital based care – and embarked on a multi-year program to right size state spending, shift spending to prevention and chronic care management, and reduce preventable hospital admissions.  To the state’s credit, spending growth in the program has fallen even faster than the national average in recent years, although it still faces significant challenges.[10]    

Right Sizing Health Care Spending Relative to Other Safety Net Programs 

Medicaid’s challenges are not just programmatic – they are political.  The program has developed deep roots in local and state economies, with hospitals and health care systems representing a large source of employment and revenues.  Today, Medicaid is the single largest payer of the costs of long-term care services for the elderly and nursing-home care, accounting for more than 40% of both markets. It also pays for around 40% of all U.S. births, according to the National Association of Medicaid Directors.[11]

As a result, health-care providers and others who rely on Medicaid for a significant share of their incomes protest vigorously against any federal efforts to slow Medicaid spending growth.  Highly organized and highly motivated stakeholders advocate for increased Medicaid payment rates and program expansion, which in turn increase their political clout.[12]

The simple reality is that U.S. health care policy is only partly about delivering better health – at nearly 20% of U.S. GDP, or $3.2 trillion, taxpayer funded subsidies for health care spending have become industrial policy, largely walled off from the economic forces that drive productivity up and costs down in other industries.  

Medicaid is often defended by pointing out the program’s low reimbursement rates compared to other payers, like Medicare and private insurance.  Even setting aside the access problems[13] that low reimbursements create for patients, this is the wrong way to think about the program (and health care in general). 

Rather than asking whether Medicaid is paying a low, per unit cost for services delivered, we should be asking whether it is delivering value for the populations it serves compared to alternative uses of the same dollars for other safety net supports and services that might be more highly valued and more effective. 

My colleague Oren Cass has written an important paper[14] on how Medicaid distorts spending for America’s safety net programs and populations.  He notes that,

"[from] 1975–2015, government social spending per person in poverty more than doubled, from $11,600 to $23,400. Rising health care expenditures accounted for more than 90 percent of that increase. For 2015–20, White House budget proposals call for 89 percent of additional social spending to target health care." [emphasis added]

Cass goes on to estimate that “over-allocation to Medicaid may exceed $100 billion annually. If states with above-median Medicaid enrollment rates or spending per enrollee in each recipient category (adult, child, disabled, etc.) returned to median levels, more than $100 billion could become available for other antipoverty programs.”

When Medicaid budgets rise inexorably they crowd out other safety net spending (given that state residents have expressed tolerance for a given level of taxation that tends to be fairly “sticky”).  As a share of state budgets, K-12 education spending and other safety net spending have remained relatively stagnant or fallen while Medicaid’s has risen.[15]

Defenders might argue that our national emphasis on health-care spending merely reflects its high-cost, and the importance of providing health care to low-income populations.  But the incentives created by hybrid federal-state system tell a different story: because only a state’s Medicaid spending earns a generous federal match (of at least $1, sometimes $3 or $4), states will quite rationally put marginal dollars toward health care even when that spending produces a far worse return for recipients than alternative programs.

For instance: if New York receives one federal dollar per state Medicaid dollar and faces the choice between putting a budget increase toward education that will return $1.25 of value per dollar spent or putting it toward Medicaid with a return of $0.75 per dollar spent, it will choose Medicaid.

Why?  Its dollar gets matched with a federal dollar and the total of $2 produces $1.50 of value—making a very poor use of funds end up looking like a highly attractive, positive return to the state. Even if the $1 produces almost no tangible return for recipients, the state is still rewarded with an extra federal dollar sloshing through its economy.

Understanding Medicaid’s Value

This not to say that Medicaid has no value. This is certainly not the case.  It is simply that we must remember that its value may vary widely across the populations and services it offers.

Medicaid covers low-income pregnant women and children, and (largely through the ACA’s optional Medicaid expansion) childless adults, the disabled, and the elderly (through long term care or nursing homes).  Studies have found high value for Medicaid funded early interventions for pregnant women and children, where the benefits may persist for decades and are clearly cost effective.[16]

But in rigorous randomized controlled experiments, like the Oregon Health Insurance Experiment, the results have been much less clear cut.  Over two years, researchers found that Medicaid increased the use of both physician and ER services, improved self-reported mental and physical health measures, and reduced the financial impact of illness (including lower out of pocket costs) but did not show a significant impact on objectively measured health outcomes.   

A follow-on analysis by the researchers found that each dollar of Medicaid spending only provided 20-40 cents of value to the recipient based on how much recipients expected to pay for care without Medicaid coverage, and how much that coverage impacted life expectancy. 

On the other hand, Oregon’s FMAP – 73 percent – ensured that the state could effectively collect roughly $3 in federal support for every $1 it spent on Medicaid coverage.  That’s a far greater return to the state compared to beneficiaries, and goes a long way towards explaining why states have “voted” for expanding Medicaid while starving other safety net programs. 

Another study, from the New England Journal of Medicine, compared mortality rates in three states (Arizona, Maine, and New York) that expanded their Medicaid programs in the early 2000s compared to three “control” states that did not expand coverage.  Of the three expansion states, only one (New York) showed a statistically significant improvement in mortality.  Arizona’s increase was not statistically significant, and Maine showed a statistically significant increase in deaths.[17]

Medicaid undoubtedly does influence beneficiaries’’ health status, but we have to keep in mind that health insurance is only one component – one input – involved in producing and sustaining good health.  Many other factors influence measured health outcomes. 

For instance, a 2006 survey from researchers at Harvard University famously found “Eight Americas,” broadly grouped by longevity and health status. Asians lived the longest, with Asian women in Bergen County, NJ, living an impressive three years longer on average than women in Japan, the country with the largest national female life expectancy: 91 years compared to 88. Young and middle-aged African Americans living in high-risk urban areas fared worse, as did rural whites living in Appalachia and the Mississippi.  Lower income whites living in the northern plains states and Dakotas lived longer than middle-income whites.  Other research of has found that descendants of Scandinavians in the U.S. have higher incomes, higher educational attainment, and even live longer than other Americans.[18]

A 2016 study in JAMA by Raj Chetty et al found that “geographic differences in life expectancy for individuals in the lowest income quartile were significantly correlated with health behaviors such as smoking, but were not significantly correlated with access to medical care, physical environmental factors, income inequality, or labor market conditions.”[19]

A May 2016 study published in Health Affairs underscored the importance of the social determinants of health by noting that “states with a higher ratio of social to health spending (calculated as the sum of social service spending and public health spending divided by the sum of Medicare spending and Medicaid spending) had significantly better subsequent health outcomes for the following seven measures: adult obesity; asthma; mentally unhealthy days; days with activity limitations; and mortality rates for lung cancer, acute myocardial infarction, and type 2 diabetes.”

If we’re interested in supporting better health, and better lives, for low income populations, we need to bring much more nuance to our discussions - and greater balance to our safety net spending.

To Cut or Not to Cut is Not the Right Question

I realize that today’s hearing is not about solutions. But I would like to suggest a few guidelines given the programmatic and fiscal realities policymakers confront. 

Federal policy debates revolve around spending levels – which given existing deficit and debt trends is entirely understandable. And I agree with former OMB director Peter Orszag (President Obama’s first budget director) that “health care reform is entitlement reform.”   We cannot address our budget woes without slowing the growth of health care spending.   

But if we were to become more agnostic about the value of an additional dollar in Medicaid spending compared to other uses of the same dollars – especially if we share programmatic savings between states and federal taxpayers - states would have a much greater incentive to consider the relative tradeoffs between competing safety net programs given their relative impact and efficiency in improving the lives and well-being of the poor.  Finding more effective ways to lift people out of poverty can, in turn, help us bend the health care cost curve. 

Setting a predictable national budget framework for all safety net programs – based on the persons served, not the programs funded – should be framed in this way, not as a pure exercise in budget cutting.   

Shifting more spending to the poor directly, through an expanded EITC, would also help to obviate the political obstacles involved in reordering spending priorities, since third party providers often have much greater ability to lobby for their preferred spending priorities (and thus protect and expand them), and little direct accountability to the populations they ostensibly serve.

Expanding direct financial assistance would also give low-income households spending discretion over safety net supports.  They could decide whether to allocate an additional marginal dollar to health care services, or some other higher value priority – like education or housing.  With low-income Americans gaining more market power, we should also expect providers to develop more affordable health care options in more accessible settings so they can compete for those marginal dollars.  

An all funds-on-deck approach to safety net funding would also allow policymakers to better evaluate programs based on their objective performance, or provide more direct financial assistance to low income families.  Targeted programs and interventions, carefully controlled and measured, are more likely to bear the fruit we want – improved health and economic mobility. This can allow us to scale up good programs, and phase out poor ones.

Many more small scale, state-based experiments are preferable than continuing a system where health care spending increases are essentially on autopilot.  To better address the diverse needs of low-income and vulnerable populations, I recommend that:

1. Congress and the states should agree on broad national programmatic goals for safety net services – including economic mobility, attachment to work, and improved health care outcomes – and then hold states accountable for reaching specific goals for specific populations.          

2. Reform efforts should focus on reducing or eliminating incentives to continue to shift vast amounts of federal and state safety net funds to health care compared to other safety net supports and services. Streamlining and consolidating federal programs to support large block grants of federal funds for all safety net services (on a person-centered basis, i.e., based on metrics like poverty rates, number of disabled, etc.) would be one potential approach.  States might still decide to spend a disproportionate share of their funding on health care, but federal structures should not effectively bribe them to do so.

3. If we continue to fund Medicaid separately, CMS should continue to support and accelerate state level reforms by standardizing Medicaid (1115) waivers and allowing states the option of accepting either block grants or per capita caps (or perhaps some mix of both, depending on the population addressed).  States should also be allowed to use Medicaid funds for alternative non-health related purposes, like an expanded state EITC.     CMS can also help states improve their own health care markets by supporting data enclaves that pool Medicare, Medicaid, and private claims data (with appropriate privacy protections) to help identify efficient providers; benchmark, test, and scale up new payment models; identify and address antitrust issues (like provider or insurer consolidation); and even allow entrepreneurs to re-bundle and re-price health care services to more effectively meet the needs of Medicaid beneficiaries.

4. The Centers for Medicaid and Medicare Innovation (CMMI) can also help states identify and design rigorous trials for testing alternative programs designed to address the social determinants of health, monitor the performance of state Medicaid waiver programs and disseminate best practices, or assist in the any development of other tools (like reference pricing, competitive bidding, or value-based insurance design) that states believe can improve measured health outcomes for Medicaid enrollees or meet other programmatic goals.

Any savings generated from program innovations should be shared between federal and state taxpayers, to encourage continued state innovation and experimentation. 

Additional challenges will need to be navigated to address legitimate state concerns about the sustainability of federal support and changing economic conditions.  Federal spending can be made explicitly countercyclical, for instance, to address state budget weakness in the event of a recession. 

But all of these challenges and concerns are tractable. 

Without comprehensive changes to Medicaid’s financing and administrative structure, along with broad programmatic flexibility and clearer goals from Washington, Medicaid reform will continue to be episodic and halfhearted. Costs will continue to climb unchecked without commensurate offsetting improvements for beneficiaries. 

This does not mean that federal policymakers should simply hand states cash and then walk away. Transparency and accountability for health and non-health related outcomes should allow policymakers to work with states – and program beneficiaries – to modernize and strengthen America’s safety net programs.

Of course, there is no silver bullet for Medicaid reform that will work for every population, in every state.

But this is precisely why states need much greater flexibility—and much better incentives—to  experiment with a wide variety of tailored approaches for safety net programs, while simultaneously putting health care spending on a more sustainable trajectory.

Both liberal and conservative policy priorities could be met by such an approach.  That doesn’t guarantee its success, but it should at least guarantee a productive conversation.       

Thank you and I welcome your questions.

• [1] Oren Cass, Over-Medicaid-ed: How Medicaid Distorts and Dilutes America's Safety Net, Manhattan Institute, May 2016, p. 4.
• [2] “Federal and State Share of Medicaid Spending,” Kaiser Family Foundation, 2015.
• [3] “States’ Increased Reliance on Funds from Health Care Providers and Local Governments Warrants Improved CMS Data Collection,” U.S. Government Accountability Office, July 2014.
• [4] “High Risk Series: An Update,” U.S. Government Accountability Office, February 2015.
• [5] World Bank Open Data, The World Bank, 2015
• [6] “Federal and State Share of Medicaid Spending,” Kaiser Family Foundation, 2015.
• [7] “Report to the Congress on Medicaid and CHIP,” MACPAC, June 2014.
• [8] Laura Nahmias, “New York may have to give back some Medicaid funding,” Politico, February 3, 2014.
• [9] The state has improved its performance in recent years, and now ranks 26^th in avoidable hospital admissions and costs, according to a 2015 report from the Commonwealth Fund. 
• [10] Bill Hammond, “Bending NY’s Medicaid Curve,” Empire Center, September 26, 2016.
• [11] “Long-Term Services and Supports,” National Association of Medicaid Directors, 2017.
• [12] Daniel DiSalvo and Stephen Eide, “The Union that Rules New York,” City Journal, Summer 2015.
• [13] Esther Hing, et. al., “Acceptance of New Patients With Public and Private Insurance by Office-based Physicians: United States, 2013,” NCHS data brief, no 195, 2015.
• [14] Cass, Over Medicaid-Ed
• [15] Ibid, 11
• [16] Sarah Miller and Laura R. Wherry, “The Long-Term Effects of Early Life Medicaid Coverage,” University of Michigan, August 2015.
• [17] Benjamin Sommers, et. al., “Mortality and Access to Care among Adults after State Medicaid Expansions,” New England Journal of Medicine 367, no. 11 (September 2012).
• [18] Christopher Murray, “Eight Americas,” Harvard Magazine, January-February 2007.
• [19] Raj Chetty, et. al., “The Association Between Income and Life Expectancy in the United States, 2001-2014,” Journal of American Medicine 315, no. 16 (2016).

______________________

Paul Howard is a senior fellow and director of health policy at the Manhattan Institute. Follow him on Twitter here.