The US Supreme Court’s ruling on the Affordable Care Act in 2012 allowed states to opt out of the health reform law’s Medicaid expansion. Since that ruling, fourteen governors have announced that their states will not expand their Medicaid programs. We used the RAND COMPARE microsimulation to analyze how opting out of Medicaid expansion would affect coverage and spending, and whether alternative policy options—such as partial expansion of Medicaid—could cover as many people at lower costs to states. With fourteen states opting out, we estimate that 3.6 million fewer people would be insured, federal transfer payments to those states could fall by $8.4 billion, and state spending on uncompensated care could increase by $1 billion in 2016, compared to what would be expected if all states participated in the expansion. These effects were only partially mitigated by alternative options we considered. We conclude that in terms of coverage, cost, and federal payments, states would do best to expand Medicaid.

The Affordable Care Act expands health insurance primarily through three provisions: the expansion of Medicaid to cover the poorest segment of the population (those with annual incomes below 138 percent of the federal poverty level), health insurance subsidies on the new exchanges for low- and medium-income people (those with annual incomes of 100–400 percent of poverty) who lack access to employer coverage or Medicaid, and a mandate for the uninsured to buy coverage.

After the act was signed into law on March 23, 2010, Florida and twenty-five other states sued the federal government to prevent its implementation, claiming that the Medicaid expansion provision was unconstitutionally coercive. The US Supreme Court’s June 28, 2012, decision 1 affirmed the constitutionality of the individual mandate, but it made states’ participation in the Medicaid expansion voluntary.

As of April 25, 2013, the governors of fourteen states—Alabama, Georgia, Idaho, Iowa, Louisiana, Maine, Mississippi, North Carolina, Oklahoma, Pennsylvania, South Carolina, South Dakota, Texas, and Wisconsin—have publicly said that their states will not participate in the expansion of Medicaid, and several other governors have expressed serious reservations about participating. 2 In this article we explore the implications of states’ decisions about expanding Medicaid for coverage expansion through Medicaid or other sources of insurance and for the state and federal costs associated with implementing health reform. Additionally, we investigate policy options that might be used to mitigate the effects on coverage expansion and costs if some states do not expand Medicaid.

Medicaid Expansion Options

The Affordable Care Act expands Medicaid eligibility to nearly everyone with an annual income of less than 138 percent of poverty. Previously, most children with annual family incomes below 200 percent of poverty were covered by the Children’s Health Insurance Program, and their eligibility for that program is minimally affected by health reform. In 2013, 138 percent of poverty would be about $15,400 for a single person and $31,800 for a family of four (Alaska and Hawaii have higher federal poverty levels than the rest of the United States and therefore higher income thresholds).

The federal government will be responsible for paying to cover people who are newly eligible for Medicaid until 2016; after that, each state’s share of the cost for its residents will gradually increase. States will pay 5 percent in 2017, 7 percent in 2018, 8 percent in 2019, and 10 percent in subsequent years, with the federal government paying the balance. This arrangement is more generous for the states than the current federal contributions for Medicaid, known as the Federal Medical Assistance Percentages, 3 which average 57 percent of the total costs.

In states that do not expand their Medicaid programs, people with incomes of 100–138 percent of poverty will be eligible for federal subsidies on the individual insurance exchanges. The Affordable Care Act does not allow people with incomes below 100 percent of poverty to receive subsidies on the exchanges. There is an exception for recent immigrants, who are generally not eligible for Medicaid. 3 But in the absence of expanded Medicaid coverage, others in the poorest segment of the population may lack access to affordable health insurance.

Many states that have said that they will expand Medicaid have concerns about the costs of expansion. These states might prefer to expand Medicaid only to people with incomes below 100 percent of poverty—instead of a ceiling of 138 percent of poverty—because those above the 100 percent threshold would be eligible for federal subsidies on the exchanges. Having potential Medicaid enrollees receive coverage through the exchanges would not save the states money initially. However, it could be advantageous for states after 2016, when federal Medicaid matching rates begin to decrease. The Centers for Medicare and Medicaid Services has said that the Affordable Care Act does not allow the use of federal funds to support partial expansions of Medicaid. However, after 2017 such expansions might be allowed if states offered the same level of coverage as with a full expansion at no additional cost to the federal government. 4

The expansion of Medicaid has the potential to provide substantial benefits to adults with incomes below 138 percent of poverty—a population that is disproportionately uninsured and therefore has limited access to health care services. As of March 2011 about 21 percent of all nonelderly adults were uninsured, but about 42 percent of nonelderly adults with incomes below 138 percent of poverty lacked insurance. 5

In addition to benefiting the new enrollees in Medicaid, providing insurance to this segment of the population should reduce the costs of uncompensated care. Jack Hadley and coauthors found that uncompensated care for the uninsured population cost $56 billion in 2008 6 —with medical inflation, that would be nearly $80 billion in 2016. Hadley and coauthors also found that state and local governments covered 30 percent of this amount. Thus, states now pay substantial amounts to treat the uninsured although they are not formally enrolled in a public (state-sponsored) insurance program.

In addition, the Affordable Care Act decreases the Medicare and Medicaid disproportionate-share hospital (DSH) payments that are made by the federal government to compensate hospitals for the added costs associated with serving Medicare, Medicaid, and uninsured patients. If a state chooses not to expand Medicaid, the decrease in DSH payments could shift the cost for uncompensated care from the federal government to states, localities, and hospitals.

Medicaid’s low compensation rates have led many providers to refuse to treat Medicaid patients. Sandra Decker found that 31 percent of physicians were unwilling to accept new Medicaid patients in 2011, compared to 17 percent and 18 percent who were unwilling to accept new patients with Medicare and private insurance, respectively. 7 The expansion of Medicaid eligibility will increase the number of people seeking health care, but because of physicians’ reluctance to participate in Medicaid, those people might not have the level of access to health services that they expect. The health reform law does allow for temporary increases in Medicaid payments to providers, which could make providers more willing to accept new Medicaid patients initially.

After 2016, as explained above, states will need to cover a portion of the Medicaid costs for the newly eligible. Although these amounts will be small relative to states’ current share of Medicaid costs, they could be substantial in absolute terms. Considering constrained state budgets, even the small increase in payments that states will make beginning after 2016 could be hard for them to bear.

In some states adults with incomes of up to 100 percent of poverty are already eligible for Medicaid through optional Medicaid categories and federal waivers. However, most of the states seeking not to expand eligibility have more restrictive eligibility requirements. Section 1931 of the Social Security Act requires all states to cover parents to some degree, but effective eligibility levels for people in this group are often well below 50 percent of poverty. 8 States are not required to cover childless adults, and the majority of states offer them no coverage.

If states choose not to participate in the current expansion of Medicaid, federal policy makers interested in maximizing access to affordable insurance will need to determine how best to cover poor adults by other means. In this article we analyze two potential policy responses. The first would be to pass a law removing the restriction on exchange subsidies, so that everyone whose income was below 100 percent of poverty would be eligible for such a subsidy, unless he or she was eligible for Medicaid or had affordable health insurance through an employer.

The second potential response would be to allow states to expand Medicaid to a lesser degree than is currently planned: increasing eligibility only for people with incomes of up to 100 percent of poverty, instead of those with incomes of up to 138 percent. As discussed above, current law prohibits this limited expansion, at least until 2017. But such a partial expansion would reduce the costs to states because people whose incomes were 100–138 percent of poverty would be covered through the exchanges instead of Medicaid. However, it could reduce coverage if people preferred to be uninsured instead of enrolling in a subsidized exchange plan with some premium payments.

Study Data And Methods

We used the RAND Comprehensive Assessment of Reform Efforts (COMPARE) microsimulation 9 to model the effects of different scenarios of health reform implementation related to Medicaid expansion. COMPARE models the health insurance decisions made by individuals, families, and firms by allowing them to weigh the costs and benefits of alternative options. These calculations include the premium cost, out-of-pocket cost, financial risk, and inherent benefit of each insurance option or of being uninsured. We present a basic overview of the COMPARE microsimulation in the online Appendix. 10 Christine Eibner and colleagues have provided a more thorough description of it elsewhere. 9

The COMPARE microsimulation uses data from the 2008 Survey of Income and Program Participation 11 for demographic characteristics, the Medical Expenditure Panel Survey 12 Household Component from 2002 and 2003 for information on medical spending, and the 2010 survey of employer health benefits by the Kaiser Family Foundation and the Health Research and Educational Trust 13 for firm characteristics. These data sets are merged using statistical matching to create a synthetic population of individuals and firms.

Total uncompensated care costs are the values from the analysis by Hadley and coauthors, 6 inflated to 2016 dollars. We computed the change in uncompensated care costs using the change in health spending by the uninsured in each scenario (which varied because the number of uninsured people varied by scenario).

Analysis Of Scenarios For Medicaid Expansion

To assess the implications of Medicaid expansion decisions, we analyzed five scenarios related to the implementation of the Affordable Care Act (ACA). The first two, “no ACA” and “full ACA,” were meant as benchmark scenarios to allow us to understand changes in coverage and cost relative to no reform and relative to the implementation of reform as originally intended by Congress. The third, “opt out,” assumed that fourteen states would reject Medicaid expansion. This was our most realistic assessment of how the law will affect outcomes, given the Supreme Court ruling allowing states to opt out. 1

The last two scenarios represent how outcomes of implementing health reform might be further changed with two possible policy modifications, each of which would likely require intervention by Congress. In the fourth scenario, “expand subsidies,” federal exchange subsidies were extended to people with incomes below 100 percent of poverty in states that did not expand Medicaid coverage. This option represented a way of covering low-income adults who would otherwise be uninsured. This policy might make states committed to expanding Medicaid reconsider that decision, but we did not address this possibility.

In the fifth scenario, “partial expansion,” states were allowed to expand their Medicaid programs to cover people with incomes of up to 100 percent of poverty, instead of all those with incomes of up to 138 percent of poverty. This partial expansion of Medicaid would increase federal subsidy spending because people with incomes of 100–138 percent of poverty would become eligible for subsidies. It also might make Medicaid expansion more attractive to states because their contributions to Medicaid would be lower in the long run.

Unless otherwise specified, all dollar values in the results and discussion sections are reported in 2016 dollars.

Limitations

The COMPARE microsimulation has limitations and includes simplifying assumptions. Some of its outputs, such as Medicaid take-up, depend in part on state and federal policy decisions that have not yet been made. For example, Medicaid take-up could be higher if a state mounted a very effective campaign to promote enrollment or lower if the state had policies to discourage enrollment. Also, reporting results at the national level may average out some variations across states resulting from differences in state populations.

An additional limitation is related to the underlying data, which came from 2008 and might have been disproportionately affected by the recent recession and its aftermath. As a result, Medicaid costs might be overstated.

Study Results

Exhibit 1 shows how many nonelderly adults would be covered by Medicaid, the individual health insurance exchanges, or some other form of insurance (such as employer-sponsored insurance), as well as those who would be uninsured, in each of the five scenarios. With full implementation of the Affordable Care Act, compared to no health reform (the “no ACA” scenario), we estimate that 16.2 million more people would be in Medicaid and 20.1 million more would be covered by an exchange, while fewer people would have other types of insurance. The net effect of these changes would be an additional 27.4 million people with insurance.

Exhibit 1 Five Different Scenarios For Medicaid Expansion And Number Of Nonelderly Adults Covered As Of 2016 Scenario No ACA Full ACA Opt out Expand subsidies Partial expansion Medicaid expansion by all states No Yes No No No Expanded subsidy to <100% FPL No No No Yes No FPL threshold in states with expansion — a 138% 138% 138% 100% Populations under age 65 affected (millions) Enrolled in Medicaid 46.7 62.9 58.5 58.4 54.4 Covered by an exchange — a 20.1 20.6 21.8 24.1 Covered by other insurance b 178.5 169.6 169.9 169.9 170.2 Uninsured 51.7 24.3 27.9 26.8 27.9

Exhibit 2 shows the net changes in government costs, compared to no health reform, for the four scenarios that assume implementation of the Affordable Care Act. These costs include those resulting from the act’s provisions related to coverage (involving Medicaid expansion, the exchange subsidies, and penalties). However, they do not account for changes in Medicare reimbursement or various other taxes and fees in the act. Exhibit 2 also presents estimated changes in uncompensated care costs based on the analysis by Hadley and coauthors. 6

Exhibit 2 Net Cost Changes From No Health Reform In Four Scenarios For Medicaid Expansion Scenario Full ACA Opt out Expand subsidies Partial expansion Medicaid expansion by all states Yes No No No Expanded subsidy to <100% FPL No No Yes No FPL threshold in states with expansion 138% 138% 138% 100% Federal spending or savings (billions of 2016 $) New Medicaid costs 77.4 56.1 56.2 32.6 Other ACA net costs a 52.1 63.3 68.7 82.3 Uncompensated care savings −26.9 −25.2 −25.4 −24.8 Net change 102.6 94.2 99.5 90.1 State spending or savings (billions of 2016 $) New Medicaid costs 1.0 0.8 0.8 0.8 Uncompensated care savings −18.1 −16.9 −17.0 −16.6 Net change −17.1 −16.1 −16.2 −15.8

If health reform were fully implemented and no states opted out of Medicaid expansion (the “full ACA” scenario), federal costs would be about $102.6 billion in 2016 (Exhibit 2 ): The $77.4 billion increase in new Medicaid spending and the $52.1 billion net change from other provisions in the Affordable Care Act (subsidy costs less the income from penalties) would be offset by the $26.9 billion reduction in spending for uncompensated care.

A small percentage of the people newly enrolled in Medicaid would be eligible under the rules that were in place before health reform was enacted; they would therefore be covered using the current Federal Medical Assistance Percentages. This coverage would cost states about $1.0 billion. If states reduced their spending on uncompensated care in proportion to the cost of people who were newly insured, reform would result in savings of roughly $18.1 billion for states (Exhibit 2 ).

The ‘Opt-Out’ Scenario

If fourteen states opt out of expanding Medicaid, there are likely to be 27.9 million uninsured people—3.6 million more than if the Affordable Care Act were fully implemented (Exhibit 1 ). In that “opt-out” scenario, federal Medicaid spending would decrease by $21.3 billion (Exhibit 2 ). Other federal spending would increase because people in states that did not expand Medicaid coverage would receive exchange subsidies.

The fourteen states that did not expand Medicaid would see their federal transfer payments decrease by $8.4 billion, relative to full implementation of the act. Additionally, these states would spend about $1 billion more on costs associated with uncompensated care.

The ‘Expand Subsidies’ Scenario

If low-income adults in states that did not expand Medicaid were eligible for federal subsidies for buying insurance through an exchange, 1.1 million fewer people are estimated to be newly insured than in the “opt-out” scenario. This would leave 26.8 million uninsured (Exhibit 1 ).

Relative to the “opt-out” scenario, expanding the subsidies without expanding Medicaid would increase other federal costs by $5.4 billion, from $63.3 billion to $68.7 billion (Exhibit 2 ). The fourteen states that did not expand Medicaid would see much lower net payments from the federal government with the extension of subsidies to the lowest-income population than if they had participated in the expansion. States would also save less on uncompensated care, compared to full implementation of the Affordable Care Act.

The ‘Partial Expansion’ Scenario

If states were permitted to expand Medicaid eligibility only to people with incomes of up to 100 percent of poverty, 54.4 million would be covered under Medicaid—8.5 million fewer than with full implementation of reform (Exhibit 1 ). About half of the people not enrolled in Medicaid in this scenario, or 4.3 million, would be covered through an exchange, and 600,000 would have another source of insurance, but there would be 3.6 million more uninsured.

In the “partial expansion” scenario, federal Medicaid spending would fall by $23.5 billion, compared to the “opt-out” scenario (Exhibit 2 ). However, other Affordable Care Act net costs would increase by $19 billion.

Discussion

States’ decisions about whether to expand Medicaid eligibility will have a sizable impact on health insurance coverage once the Affordable Care Act is fully implemented. If all states expand Medicaid eligibility, reform will mean that in 2016, 27.4 million people will be newly insured and 16.2 million will be newly enrolled in Medicaid. If the fourteen states that have said they will not expand eligibility for Medicaid act on that decision, there will be 3.6 million fewer newly insured people and 4.4 million fewer new Medicaid enrollees.

The decision to expand Medicaid would not have substantial effects on state budgets prior to 2017. Although expanding eligibility would be costly, the majority of the cost would be borne by the federal government. In 2016 the states that expand Medicaid would not contribute to Medicaid costs for the newly eligible, but by 2020 they would be required to pay 10 percent of those costs. At 2016 spending levels, that 10 percent would amount to about $7.8 billion (for purposes of comparison, in 2010 the states spent $125.7 billion on Medicaid).

Furthermore, the cost to states of expanding Medicaid would generally be lower than the cost of uncompensated care borne by states and localities after the implementation of the Affordable Care Act. Increases in insurance coverage from implementation of the act would reduce state and local spending on uncompensated care by $18.1 billion (Exhibit 2 ). Thus, states that expanded Medicaid would have a net savings on funds spent on the uninsured for the first several years after the expansion. Those savings might continue beyond 2020, when the states’ share of Medicaid costs plateaus at 10 percent.

Regardless of the budgetary impact, expanding Medicaid coverage would have health benefits for states’ populations. If fourteen states do not expand eligibility, 3.6 million fewer people will have insurance. In a recent study, Benjamin Sommers and coauthors found that previous expansions of Medicaid led to a substantial decrease in mortality. 14 Applying their mortality estimates, we project that fully expanding Medicaid eligibility could reduce mortality by 90,000 lives per year. The mortality reduction would be only 71,000 lives per year if fourteen states opted out of the expansion.

People who lived in a state that did not expand Medicaid and whose incomes were 100–138 percent of the federal poverty level would not be eligible for Medicaid but would have access to subsidized health insurance through the new exchanges. However, people in those states with incomes below 100 percent of poverty would have no such affordable option because the Affordable Care Act makes them ineligible for subsidies in the exchanges.

We also simulated a scenario in which states were allowed to expand Medicaid coverage to those with incomes of up to 100 percent of poverty—instead of 138 percent—without financial penalty. This approach might be seen as beneficial by states facing budget crises, and its use could encourage more states to participate in the expansion because it would reduce the number of people newly eligible for Medicaid. This partial expansion of Medicaid led to no net change in uninsurance relative to the opt-out scenario, but—relative to full expansion—almost four million fewer people are estimated to become insured. However, states would see lower federal spending.

Finally, we considered expanding subsidies to people with incomes below 100 percent of poverty as a way of enabling them to access affordable coverage even if their state does not expand Medicaid. We found that this option increased coverage by 1.2 million people compared to the “opt-out” scenario, but it did not result in the same level of coverage as the full implementation of the Affordable Care Act.

Policy Implications

To maximize health insurance coverage, our analysis indicates that all states should expand Medicaid eligibility. However, this does not seem likely to happen, at least initially. Federal policy makers could extend subsidy eligibility to mitigate the effects of states’ decisions not to expand Medicaid. This would increase coverage to some extent, but not by as much as if all states expanded Medicaid. If federal policy makers allowed states to expand Medicaid eligibility for people with incomes of up to 100 percent of poverty—instead of 138 percent—states would be able to shift costs to the federal government, but there would be only a minimal effect on total health insurance coverage.

State policy makers should be aware that if they do not expand Medicaid, fewer people will have health insurance, and state and local governments will have to bear higher costs for uncompensated care. Because the federal share of Medicaid costs for the newly eligible is much higher than current levels, we estimated states’ costs for expansion to be less than the reduction in their costs for uncompensated care.

Conclusion

The bottom line is that the expansion of Medicaid eligibility is a key provision of the Affordable Care Act. If the fourteen states that have said they will opt out of Medicaid expansion do so, 3.6 million fewer people will have health insurance than would otherwise be the case. This would save the federal government around $8.4 billion a year compared to the full expansion of Medicaid. However, the states that opted out of Medicaid expansion would see a net increase in spending in the short term because they would spend more on uncompensated care. Furthermore, even states that opt out of the expansion will be subject to the reductions in Medicare payments and disproportionate-share hospital payments, as well as various other taxes and fees in the Affordable Care Act. Thus, there may be large net transfers of federal funds out of the states that do not expand Medicaid.

If some states choose not to expand Medicaid eligibility in spite of the budgetary and economic impact of that decision, the federal government could consider expanding the eligibility for premium and cost-sharing subsidies on the individual exchanges to people with incomes below 100 percent of poverty. However, making people eligible for subsidies would not fully make up for the decline in the coverage rate that would result from failing to expand Medicaid eligibility.

ACKNOWLEDGMENTS

The authors are grateful to RAND’s Health Economics, Finance, and Organization Program for providing internal investment funds to support this analysis.

ABOUT THE AUTHORS: CARTER C. PRICE & CHRISTINE EIBNER In this month’s Health Affairs , Carter Price and Christine Eibner analyze how health insurance coverage and government health care spending could change in states opting out of the Affordable Care Act’s Medicaid expansion. Price is an associate mathematician at the RAND Corporation. He led several studies of the economic effects of the Medicaid expansion component of the Affordable Care Act on states. Price earned a master’s degree and a doctorate in applied mathematics from the University of Maryland, College Park. Eibner is a senior economist at the RAND Corporation and director of the RAND Comprehensive Assessment of Reform Efforts (COMPARE) initiative, which uses economic modeling to predict how individuals and employers will respond to major health care policy changes and which is used in this article. She leads several projects related to the Affordable Care Act, including a study for the Department of Health and Human Services that will assist state Medicaid programs with income counting and Federal Medical Assistance Percentage calculations. Eibner earned a doctorate in economics from the University of Maryland, College Park.

NOTES