Hatch-Upton Report On Costs to States of Expanding Medicaid Relies On Seriously Flawed Estimates
March 17, 2011
Senate Finance Committee Ranking Member Orrin Hatch (R-UT) and House Energy and Commerce Committee Chairman Fred Upton (R-MI) recently released a report purporting to estimate the net cost to states of expanding Medicaid under the Affordable Care Act.  The report claims it provides, on a state-by-state basis, the best estimates of state costs under the Medicaid expansion.
The report, however, is wholly unsound. It uses non-comparable state estimates that have widely varying scope and use different time periods and makes no effort to standardize them. It picks and chooses among available estimates, selecting only the estimates with the highest costs and ignoring the others (and failing to disclose their existence). In addition, a number of the state estimates it uses are highly unreliable and overstated because they rest on severely flawed assumptions, including the following:
- Inflated participation-rate assumptions for individuals eligible for Medicaid, including assumptions that 100 percent of eligible people will enroll, despite the fact that 100 percent participation has never been achieved in any means-tested program in the decades those programs have operated. Such assumptions result in overblown estimates of Medicaid enrollment increases and costs.
- Estimates of the costs per newly enrolled Medicaid beneficiary that are highly inflated, as can readily be seen by comparing them to current Medicaid cost data; and
- The inclusion of costs that are not required under the Affordable Care Act.
Furthermore, few of the state analyses cited in the Hatch-Upton report account for any of the savings that would accrue to the states as a result of having more people insured or from other provisions of the Affordable Care Act. By significantly shrinking the ranks of the uninsured, the health reform law will produce significant savings for states and localities; it will enable them to reduce state and local spending on uncompensated care and to scale back or end state-funded programs in a number of states that provide health insurance or health services to state residents who otherwise would be uninsured. Urban Institute researchers have noted that such savings may offset much, and possibly all, of the cost to states of the Medicaid expansion.
Despite claims in the Hatch-Upton report to the contrary, the Affordable Care Act's Medicaid expansion is, overall, a beneficial deal for states. According to the non-partisan Congressional Budget Office (CBO), under the Affordable Care Act (ACA), Medicaid and the Children's Health Insurance Program will cover (by 2021) an estimated 18 million more low-income adults and children than they do today, with most of them being people who otherwise would be uninsured. CBO estimates that over this period, the federal government will pay 92 percent of the cost of this expansion, and state costs for Medicaid and CHIP will increase by only 2.6 percent above what they would be in the absence of the health reform law.
Who Pays for the Medicaid Coverage Expansions Under the Affordable Care Act?
Under the Affordable Care Act, state Medicaid programs will be required to cover all non-elderly individuals up to 133 percent of the poverty line ($24,700 for a family of three), starting in 2014. The federal government will pick up the vast majority of the costs of covering the newly eligible individuals. The federal government will pay 100 percent of the cost of covering newly eligible people in the first three years of the Medicaid expansion, with the federal share phasing down to 90 percent by 2020 and remaining at that level thereafter. This is well above the federal government's normal share of state Medicaid costs, which is 57 percent on average.
Because of this generous federal financing, the Medicaid expansion under health reform is expected to increase state costs by only a modest amount. As noted, CBO projects that state Medicaid costs will be 2.6 percent, or $60 billion, higher over the next ten years than they would be in the absence of health reform. (See Figure 1; CBO's estimate of the state share of these costs includes the cost of covering individuals who are already eligible for Medicaid but are not enrolled, as well as the costs of covering people who become newly eligible.)
It should be noted that the $60 billion figure covers Medicaid costs only. It does not reflect the offsetting savings in uncompensated care and other areas that states will realize.
Report Relies on Non-Standard Estimates
To arrive at its conclusion that the Medicaid expansion will cost states at least $118 billion between 2014 and 2023,  the Hatch-Upton report adds up estimates of the costs of the Medicaid expansion in each state that it derives from a number of sources. For 19 states, the report relies on state-by-state cost estimates from an Urban Institute analysis.  (The Urban Institute report includes a "standard" cost estimate and an "enhanced," or higher, cost estimate for each state; Hatch and Upton use only the high estimates and ignore the others, as noted below.)
For the remaining states, the report discards the Urban Institute estimates and instead uses estimates produced by the states themselves, even though these state estimates are based on widely varying methodologies and assumptions, have different scopes of analysis, cover different time periods, and have different degrees of reliability (which, in some cases, are very low). The report makes no effort to standardize these state estimates to make them comparable, which would be necessary to produce any credible estimate of the total state share of the cost of the Medicaid expansion over some period of time. The Hatch-Upton report also makes no effort to assess the reliability or even the plausibility of the widely varying state estimates. And, where a state has produced a range of estimates, Hatch and Upton use only the estimate at the top of the range, and do so without disclosing the existence of the other estimates.
Report Cherry-Picks the Highest Estimates
The Urban Institute has produced state-by-state estimates of the cost of the Medicaid expansion between 2014 and 2019. The Hatch-Upton report uses Urban Institute numbers for 19 states where the states themselves did not produce their own estimates.
The Urban Institute produced two sets of estimates for each state: an estimate of costs under a "standard scenario" that assumes that the percentage of newly eligible and currently eligible individuals who enroll in Medicaid will be at about the program's current participation rate; and an estimate of costs under an "enhanced scenario" that assumes the participation rate will increase sharply. Nevertheless, in every case, the Hatch-Upton report cites only the enhanced scenario estimate. The Hatch-Upton report fails even to acknowledge the existence of the standard estimates, despite the fact that those Urban Institute estimates are the ones most consistent with the estimates produced by the Congressional Budget Office. The costs under the standard estimates are only about half as much as the costs under the higher participation scenarios. 
For the 31 other states, the Hatch-Upton report only shows estimates produced by the states themselves, which in all but three states are higher than the Urban Institute's standard scenario estimates. Moreover, nine of these state studies included a range of costs based on different enrollment assumptions. For these nine states, the Hatch-Upton report presents only the estimate that shows the highest costs and makes no mention of the existence of the lower estimates. For example, the analysis from Mississippi contains three sets of estimates based on the participation-rate assumption that is employed. The Hatch-Upton report includes only the highest figure, despite a statement from the author of the Mississippi analysis that the low and medium estimates are more likely to reflect reality. 
A Number of the State-Based Estimates Rely on Highly Flawed Assumptions
Several of the state estimates that the Hatch-Upton report uses suffer from severe flaws. These state estimates clearly overstate the costs of the Medicaid expansion by relying on assumptions that almost certainly cannot be correct.
Using Enrollment Assumptions Not in Line with Historical Experience
A number of the state-produced estimates cited in the Hatch-Upton report assume rates of Medicaid participation among eligible individuals that are sharply inconsistent with decades of experience in means-tested programs. For example, many state estimates, including those from California, Florida, Mississippi, Indiana, and Nebraska, assume that 100 percent of individuals who are eligible for Medicaid will enroll. Some studies compound the effect of this inflated participation-rate assumption by also assuming that all Medicaid-eligible individuals will enroll on day 1 of the expansion, rather than incorporating some ramp-up effects over a period of two to three years. Studies by other states, including Louisiana and Texas, assume almost-as-dubious enrollment rates of about 95 percent.
These participation rates are flatly inconsistent with experience in Medicaid and other programs. No means-tested public program has ever achieved a 100 percent participation rate. Even Medicare, a universal social insurance program, enjoys a participation rate of only 96 percent, and a variety of other means-tested programs have participation rates from 43 percent to 86 percent.  While a mandate to have health insurance and the publicity and outreach efforts surrounding the expansion should increase enrollment to some degree, the evidence is overwhelming that the participation rate will not reach or come close to 100 percent. The Congressional Budget Office estimates that individuals who are eligible for Medicaid but do not enroll will constitute about one-third of the people who remain uninsured under the Affordable Care Act. This is related to the fact that under the law, many low-income individuals will be exempt from paying a penalty if they do not have health insurance coverage. 
Moreover, a significant number of people who will become newly eligible for Medicaid already have private insurance that they are likely to keep. One factor that makes the participation rate assumptions in some of the state studies especially unrealistic is that they assume that most or all people who have private insurance and become eligible for Medicaid will drop their current coverage and enroll in Medicaid instead. For example, the Virginia study estimates that 426,000 more people in that state will enroll in Medicaid as a result of health reform, even though Census data show there are only 364,000 uninsured individuals in the state with incomes below 133 percent of poverty. Similarly, the Mississippi, Nebraska, and Indiana analyses estimate that between 35 percent and 45 percent of all new Medicaid enrollees will be people who drop private coverage to enroll in Medicaid.
Such assumptions regarding the number of people who will drop private coverage and shift to Medicaid are sharply inconsistent with the evidence from states' actual experience with previous Medicaid expansions. For example, several studies that looked at state expansions of children's eligibility for Medicaid in the 1990s found that only 10 percent to 20 percent of new Medicaid enrollees previously had private coverage, well below the 35 percent to 45 percent rates assumed in some of the cost estimates the Hatch-Upton report uses .
Assumed Costs Per Newly Enrolled Medicaid Beneficiary Are Inflated
Some of the state analyses also include highly problematic assumptions concerning the cost per beneficiary of covering new enrollees. For example, the Indiana analysis based its estimate of the per-person cost of covering newly eligible adults and parents in Medicaid on the state's cost in covering uninsured adults under the Healthy Indiana Plan (HIP), a state waiver program that covers some poor and low-income childless adults and parents. Using the average cost per beneficiary in HIP as the average cost under the forthcoming Medicaid expansion, however, is clearly not valid because current HIP enrollees are likely to be much less healthy, on average, than those who will enroll in Medicaid in 2014.
An analysis of the HIP program found that enrollees were likely to be in poorer health -- and thus to use more health care -- than typical enrollees in commercial health insurance plans. The study found that by and large, the low-income people who have elected to enroll in HIP are people with much greater-than-average medical needs. This apparently has occurred because HIP requires most enrollees to make a substantial monthly premium payment despite their limited incomes, and the program thus is considerably more attractive to sicker people than to people in good health.
The Indiana cost estimate used by Hatch and Upton is based on a series of contradictory assumptions. It assumes that everyone who qualifies for Medicaid will enroll, including all eligible individuals who already have employer-based insurance. As a result, it assumes that all of the eligible people who are in good health will enroll, along with those who are less healthy. With enrollment at such levels, the per person cost would clearly be considerably lower than the cost of covering adults in HIP, as those who would enroll in Medicaid would include a robust mix of healthy and less-healthy individuals. (Moreover, those who now are uninsured tend overall to be in better health than those with coverage.) This means that using the per-person costs under HIP, which has been shown to disproportionately cover sicker people, as the basis for determining the per-person cost of those who would newly enroll in Indiana's Medicaid program is entirely inappropriate.
The estimates of the per-person cost of covering new enrollees in appear much too high in a number of other states as well. The Mississippi analysis estimates that if the expansion were in effect in 2009, the cost would be $4,540 per beneficiary for the adult expansion population and $2,421 for children. But estimates from the Urban Institute based on expenditure data actually reported by the states to the federal government show that per-capita costs for the adult and child populations in Mississippi's Medicaid program in 2009 were $2,612 and $1,798, respectively.  The Mississippi per-capita cost estimates for the cost of the Medicaid expansion should not be 35 to 74 percent higher than the Urban Institute's estimates, which are based on the state's own data. Similarly, for Nebraska, the estimate of the cost per beneficiary of covering newly eligible parents in 2009 is $4,881, some 74 percent higher than the Urban Institute estimate of an actual cost of $2,812.
Estimates Include Spending Not Required Under Health Reform
Another factor causing some of the state estimates to be overstated is that a number of the state-specific analyses include costs that health reform does not require. For example, the Florida and Texas estimates include the cost of raising Medicaid's payment rates for primary care to Medicare levels after 2014, a change the health reform law does not require states to make.
The Affordable Care Act directs states to raise temporarily -- for 2013 and 2014 -- the fee levels they pay for certain primary care services so those fees are on a par with what Medicare pays. The federal government will pay 100 percent of the increased costs associated with these rate increases. After 2014, states are free to maintain or to scale back these increased primary care payment rates; the ACA does not require states to maintain the higher fee levels. Some states may elect to maintain the higher fee levels after 2014, but maintaining the higher provider rates would be inconsistent with past Medicaid practices in most states. In any event, the assumption that states will be mandated to maintain the high fee levels is not valid.
Some of the state estimates go farther. The California estimate assumes the state will raise its Medicaid payment rates for all care that all physicians provide, including specialty services. Yet the health reform law does not raise payment rates at all outside of primary care. Similarly, Louisiana's estimate includes nearly $2 billion in costs related to raising provider payment rates -- by including not only the cost of maintaining enhanced provider rates for primary care beyond 2014 but also the cost of raising payment levels for specialty services to 110 percent of Medicare rates.
Louisiana also includes an increase in hospital rates of $1.5 billion. The state argues that this will be necessary to pay for the cost of the "uncompensated care" that will result from serving a higher proportion of patients on Medicaid, which pays lower rates to health care providers than Medicare or private insurance. But this assumption defies logic: in the absence of health reform, these patients would have been uninsured and the state and the hospitals would have borne a larger portion of the cost of providing health care to them. The costs associated with these provider rate increases -- none of which the ACA requires -- account for 27 percent of Louisiana's estimate of the Medicaid expansion's cost.
None of the Estimates Incorporate State Savings in Uncompensated Care and Other Costs
By dramatically shrinking the ranks of the uninsured, the health reform law will lighten the burden on states of providing health care to their uninsured residents. However, none of the state estimates cited in the Hatch-Upton report (including the Urban Institute estimates) account for the state savings that the health reform law will produce in this and some other areas. (The Urban Institute report did not attempt to estimate such costs on a state-by-state basis, but as noted below, the researchers who conducted the study have indicated that savings in this area are real and substantial.)
In other words, even if the estimates in the state studies all were credible, they would be estimates of gross coverage costs, not of the net costs to states after taking other savings into account. A recent review of various state estimates of Medicaid expansion costs, conducted by the Kaiser Commission on Medicaid and the Uninsured, noted that such savings could be quite substantial and that the state estimates do not account for them.
Many individuals who will become newly eligible for Medicaid currently receive state- and local-funded health services, such as mental health treatment or hospital treatment. The Urban Institute found that in 2008, state and local governments shouldered $10.6 billion, or nearly 20 percent, of the cost of caring for uninsured people in hospitals. State and local governments also provided 47 percent of the funding for state mental health agencies, amounting to $14.7 billion in 2006. Because of health reform, Medicaid now will cover many of those services, so the federal government will pick up much of the costs of a number of services for which states now are bearing all of the cost.
In fact, the same Urban Institute researchers who produced the state-by-state estimates of the costs of the Medicaid expansion have pointed out that the reductions in state costs in caring for the uninsured could more than offset the state costs of expanding Medicaid.  Thus, some or all of the $60 billion that CBO estimates states will incur in additional state Medicaid costs under the new law will be offset by state savings in other areas as a result of many fewer people being uninsured. This means some of the additional state Medicaid spending will be substituting for existing state health care spending, with some of the expenditures that states currently make qualifying for federal Medicaid matching funds for the first time.
Containing Future Cost Growth
The ACA also contains a wide range of measures to start restructuring the way Medicaid and other health care programs pay for services and deliver care, in order to slow the growth of health care costs over time. The ACA begins to advance many of the strategies that health policy experts consider promising ways to reduce the growth of Medicaid, Medicare, and other health care spending.
- The health reform law establishes a new Federal Coordinated Health Care Office to examine how to improve care and lower costs for the dual eligibles -- the elderly and disabled people enrolled in both Medicare and Medicaid -- who constitute 15 percent of Medicaid enrollees but account for 40 percent of program costs. The Federal Coordinated Health Care Office recently announced that it will award contracts to up to 15 states for up to $1 million each to help them design demonstration programs to develop, implement, and evaluate models aimed at improving the quality, coordination, and cost-effectiveness of care for dual-eligible individuals.
- The ACA also sets up a number of demonstration projects that will examine ways to better deliver health care. For example, a Medicaid demonstration project will test whether "bundled payments," in which hospitals and post-acute care providers like nursing homes and rehabilitation facilities receive one payment, improve care and provide savings. By receiving a single Medicaid payment for all hospital and post-acute services rather than being reimbursed separately for each service, as is the case today, hospitals and post-acute care providers would have greater incentives to coordinate care and provide more cost-effective care both during a hospital stay and after the patient is discharged.
The ACA also promotes the establishment of "accountable care organizations" in Medicaid, in which groups of physicians would be paid fixed payments and/or bonuses to provide coordinated and integrated primary and specialty care services to Medicaid beneficiaries. The intent is to encourage different physicians who care for the same patient to better coordinate their care, rather than providing episodic, fragmented, and sometime duplicative health services. The new Center for Medicare and Medicaid Innovation, which the health reform law created, will help fund these state demonstration projects.
These demonstrations hold promise of identifying innovations that could save state Medicaid programs substantial sums over time.
The Hatch-Upton report's estimate of the cost to states of the Medicaid expansion is unreliable and overstated. The estimate is derived in substantial part by cherry-picking the highest estimates from various studies and by relying on studies that use flawed assumptions and generate inflated estimates of enrollment and per beneficiary costs, count costs that the ACA does not require states to bear, and fail to account for uncompensated-care and other savings that states will secure under health reform.
Contrary to the claims the Hatch-Upton report makes, the Medicaid expansion does not pose a substantial financial burden for states. The best estimate -- from the Congressional Budget Office -- is that the additional state expenditures resulting from the expansion will increase state Medicaid costs by 2.6 percent, compared to what those costs would be in the absence of health reform, while covering 18 million people who otherwise would be uninsured.
 Orrin Hatch (R-UT) and Fred Upton (R-MI), "Medicaid Expansion in the New Health Law: Costs to the States," March 1, 2011.
 Douglas Elmendorf, Letter to John Boehner, Congressional Budget Office, February 18, 2011 (on H.R. 2, which repealed the Affordable Care Act and was passed by the House of Representatives on January 19, 2011).
 In comparison, the Urban Institute estimates that the Medicaid expansion would cost states $21.1 billion from 2014 through 2019 (assuming that participation in the program would remain at its current rate) or $43.2 billion over the same period (assuming that participation is substantially higher than the current rate). As noted, the Congressional Budget Office estimates a cost to states of $60 billion from 2014 through 2021.
 John Holahan and Irene Headen, "Medicaid Coverage and Spending in Health Reform: National and State-by-State Results for Adults at or Below 133% FPL," Kaiser Commission on Medicaid and the Uninsured, May 2010.
 CBO estimated last year that the cost to states of the Medicaid expansion would be $20 billion between 2014 and 2019.
 The nine states are Alabama, California, Indiana, Maryland, Mississippi, Nebraska, New Hampshire, Virginia, and Washington.
 John D. Meerschaert, "Financial Impact Review of the Patient Protection and Affordable Care Act As Amended by H.R. 4782, The Reconciliation Act of 2010 On the Mississippi Medicaid Budget," Milliman, Inc., October 1, 2010.
 January Angeles, "Some Recent Reports Overstate the Effect on State Budgets of the Medicaid Expansions in the Health Reform Law," Center on Budget and Policy Priorities, October 21, 2010.
 Dahlia Remler and Sherry Glied, "What Other Programs Can Teach Us: Increasing Participation in Health Insurance Programs," American Journal of Public Health, January 2003.
 Sherry Glied, Jacob Hartz, and Genessa Giorgi, "Consider It Done? The Likely Efficacy of Mandates for Health Insurance," Health Affairs, November/December 2007.
 Those with incomes below the tax filing threshold (in 2010 the threshold for taxpayers under age 65 was $9,350 for singles and $18,700 for couples) would be exempt from paying a penalty if they did not have minimum essential coverage. Exemptions will also be granted for financial hardship, religious objections, American Indians, those without coverage for less than three months, undocumented immigrants, incarcerated individuals, those for whom the lowest cost plan option exceeds 8 percent of an individual's income.
 Center on Budget and Policy Priorities analysis of the ASEC supplements to the 2009 and 2010 Current Population Surveys.
 Matt Broaddus and January Angeles, "Medicaid Expansion Not Likely to 'Crowd-Out' Private Insurance," Center on Budget and Policy Priorities, June 22, 2010.
 Rob Damler, "Experience under the Healthy Indiana Plan: The Short-Term Cost Challenges of Expanding Coverage to the Uninsured," Milliman Inc., August 2009.
 The Urban Institute and Kaiser Commission on Medicaid and the Uninsured estimates are based on data from the Medicaid Statistical Information System (MSIS) and CMS-64 reports from the Centers for Medicare and Medicaid Services (CMS). Data provided were for 2007. Estimates were adjusted to 2009 dollars using the National Health Expenditure Survey's estimate of the per-capita growth in medical expenditures between 2007 and 2009.
 Randall Bovbjerg, Barbara Ormond, and Vicki Chen, "State Budgets Under Federal Health Reform: The Extent and Causes of State Variation in Impact," Kaiser Commission on Medicaid and the Uninsured, February 2011.
 Jack Hadley, et al., "Covering the Uninsured in 2008: Current Costs, Sources of Payment, and Incremental Costs," Health Affairs, August 25, 2008.
 National Association of State Mental Health Program Directors Research Institute, Inc., "SMHA-Controlled Mental Health Revenues, By Revenue Source and by State, FY 2006."
 John Holahan, "Briefing on Medicaid Coverage and Spending in Health Reform: National and State-By-State Results for Adults at or Below 133% FPL," Kaiser Commission on Medicaid and the Uninsured, May 26, 2010.