Medicaid and the State Children’s Health Insurance Program

State legislators should

• refuse to implement the Patient Protection and Affordable Care Act’s Medicaid expansion;
• conduct randomized, controlled experiments of the effects of Medicaid and the State Children’s Health Insurance Program (SCHIP), along the lines of the Oregon Health Insurance Experiment, with existing populations;
• deregulate medical care and health insurance; and
• demand that the federal government grant states flexibility with the existing Medicaid and SCHIP programs, not additional funds, to provide medical and long-term care to the needy.

Congress should

• repeal the Patient Protection and Affordable Care Act;
• liberalize Medicare and the tax treatment of health insurance;
• deregulate health care and health insurance;
• permit states to conduct randomized, controlled experiments of the effects of Medicaid and SCHIP coverage on existing populations;
• eliminate any federal entitlement to Medicaid or SCHIP benefits;
• freeze each state’s Medicaid and SCHIP funding at current-year levels;
• give states full flexibility to use Medicaid and SCHIP funds to achieve a few broad goals; and
• begin phasing out Medicaid and SCHIP federal funding.

Counterintuitively, the most important thing policymakers can do to improve access to care for the poor is not to provide direct assistance to the poor, but to liberalize the health care sector. The great virtue of a market system is that it uses innovation to fill in the cracks in the health care sector so that fewer vulnerable patients fall through. A market system drives prices for medical care and health insurance downward. It minimizes the problem of preexisting conditions by offering protection against preexisting conditions even to the uninsured. Liberalization would reduce unmet medical need by bringing health care within the reach of those who could not previously afford it, making health care of ever-increasing quality available to an ever-increasing number of people. Liberalization would make the problem of unmet need smaller and leave the rest of society wealthier, so we could better help the shrinking number of patients who still could not help themselves. Government intervention merely causes the cracks in the health care sector to widen.

Even so, no matter how well a market system improves quality and access, there will always be patients who cannot afford the medical care they need, either because they never had the resources or because they chose not to purchase health insurance. This chapter discusses how to address, in a market system, the shrinking number of patients who cannot help themselves.

The Samaritan’s Dilemma

Any effort to help people in need confronts what economists call “the Samaritan’s dilemma” — the idea that, just as it is possible to help too little, it is also possible to help too much. Helping too much induces people who could be self-reliant to take advantage of charitable assistance and do less to help themselves. As the assistance becomes more generous, fewer people provide for themselves. They contribute less to the economy — and to charity. Helping too much perversely increases the amount of “need” and the burden of charity, while simultaneously reducing society’s ability to bear that larger burden.

Ideally, voluntary charity and government assistance would avoid both helping too little and helping too much. Yet there is no obvious “right” place to strike that balance. The optimal amount of charitable assistance depends on actual need, the costs of assistance and dependence, and donor preferences. All of these vary across space and time.

Some approaches are more effective than others. Voluntary charities have the incentive and the ability to ensure their resources assist only the truly needy. For example, voluntary charities can lose funding if donors learn their contributions are going to people who don’t need assistance. In contrast, the government has little incentive or ability to strike an optimal balance. Politicians must craft broad eligibility rules for government programs. Typically, these take the form of a legal entitlement to benefits for anyone who meets certain criteria. When bureaucrats identify beneficiaries who technically meet those criteria, but nevertheless need no assistance, they have little ability or incentive to exclude them. On the contrary, they face incentives to provide assistance to nonneedy applicants, because their careers depend on a thriving program and beneficiaries can sue the government for withholding benefits to which they are legally entitled. Taxpayers might want to cut programs that provide assistance to nonneedy applicants but lack the freedom to withdraw their “contributions” in protest, so programs that provide assistance to nonneedy applicants rarely see their funding reduced. Thus, government welfare programs tend to err on the side of subsidizing lots of people who don’t need assistance.

Instances in which government appears to help too much abound.
• In 2004, budget constraints led Missouri to cut more than 100,000 people from its Medicaid rolls. Though the share of low-income children enrolled in Medicaid fell from 50 percent to 40 percent, one study found that “increases in other types of insurance coverage prevented an increase in the share that were uninsured.” Before the Patient Protection and Affordable Care Act (ACA) was implemented, Medicaid-eligibility expansions in Arizona, Delaware, Maine, and Oregon did not reduce those states’ uninsured rates. They were accompanied by declines in private coverage.
• A study of Medicaid expansions by ACA supporters projected “high rates of crowd-out for Medicaid expansions aimed at working adults (82 percent), suggesting that the Medicaid expansion provisions of [the ACA] will shift workers and their families from private to public insurance without reducing the number of uninsured very much.”
• Economist Casey Mulligan found that the ACA creates larger disincentives to work than any law Congress has enacted in 70 years. In some cases, workers find that working more leaves them with lower incomes.

The most important thing that policymakers can do to help the poor obtain health insurance and medical care is adopt policies that spur cost-saving innovations and lower prices. Falling prices do not involve a Samaritan’s dilemma. Welfare traps the poor in poverty; falling prices help them climb out. The government can help the poor most of all by reforming Medicare (see Chapter 38) and the tax treatment of health care (see Chapter 37), and by deregulating medicine (see Chapter 35) and health insurance (see Chapter 36).

In addition, federal and state governments operate three main programs to provide medical care to low-income Americans: Medicaid, the State Children’s Health Insurance Program, and premium subsidies available through the Patient Protection and Affordable Care Act’s health insurance “exchanges.” Congress should repeal or drastically reform each of these programs.


Medicaid is a $578 billion program that exists ostensibly to provide health care to the poor. The federal government jointly administers Medicaid with state and territorial governments. States that wish to participate in Medicaid must pay a portion of the cost of a federally mandated set of health benefits to a federally mandated population of eligible individuals. In return, each state receives matching federal funds to administer its program. The federal treasury matches any amount a state spends on its Medicaid program. States receive unlimited matching funds when they make their Medicaid benefits more comprehensive or extend eligibility to more people than the federal government requires. Overall, the federal government currently finances between 63 percent and 65 percent of Medicaid spending.

All states participate in the traditional Medicaid program, which primarily serves four low-income groups: mothers and their children, the disabled, the elderly, and those needing long-term care. Specific eligibility criteria vary by state. Historically, 57 percent of traditional Medicaid funding has come from the federal government and 43 percent has come from the states.

The ACA gives states the option to expand their Medicaid programs to all adults with incomes below 138 percent of the federal poverty level ($16,394 for single adults in 2016). The principal beneficiaries are able-bodied adults. The federal government pays 100 percent of the cost of the ACA’s Medicaid expansion from 2014 through 2016. The federal share then phases down to 90 percent by 2020. Thirty-one states have implemented the ACA’s Medicaid expansion.

For beneficiaries, Medicaid is an entitlement. So long as they meet the eligibility criteria, they have a legally enforceable claim to benefits. People tend to cycle on and off Medicaid for various reasons, but the Congressional Budget Office projects the average monthly enrollment for 2016 will be 77 million Americans, and the total number who will enroll at some point during the year will be 98 million.

Medicaid’s Perverse Incentives

The federal government’s method for distributing Medicaid funds to states encourages fraud, creates perverse incentives for state officials, and encourages states to expand their programs to people who don’t need assistance. Because federal and state governments share the burden of Medicaid spending, neither side cares about the drawbacks of the program — induced dependence, waste, and fraud — as much as they should.

The more a state spends on its Medicaid program, the more it receives from the federal government. When a state spends $1, it receives between $1 and $3 from the federal government. States can thus double, triple, or even quadruple their money by spending more on Medicaid. This leads state and federal officials to tolerate stunning amounts of fraud. The Government Accountability Office consistently designates Medicaid as a “high-risk” program: official estimates of improper Medicaid payments suggest the federal share alone was $17.5 billion in 2014.

The system of matching federal funds creates perverse incentives for state officials to spend too much on Medicaid and too little on other priorities. Spending $1 on police buys $1 of police protection. Spending $1 on Medicaid, however, buys $2 to $4 of medical or long-term care. States tend to spend the marginal dollar on Medicaid even when spending it on police, education, or transportation would provide greater benefits.

The perverse incentives are even greater under the ACA’s Medicaid expansion. States that want to reduce state spending by $1 million would have to cut outlays in the “old” Medicaid program by anywhere from $2 million to $5 million. (The additional savings revert to the federal government.) By contrast, since states pay only 10 percent of the cost of the Medicaid expansion, states must cut Medicaid-expansion outlays by $10 million in order to achieve $1 million of budgetary savings. In other words, the Medicaid expansion creates perverse incentives for states to cut health care spending on needier individuals rather than less-needy individuals. Cutting health care for able-bodied adults requires state officials to inflict up to five times more political pain than cutting health care for needier, more-vulnerable enrollees.

Medicaid both pulls and pushes enrollees into dependence. Medicaid makes private health care less affordable — thus pushing people into the program. Economists Mark Duggan and Fiona Scott Morton found that Medicaid’s system of setting drug prices increases prices for private payers by 13 percent. The more federal and state governments expand Medicaid, the more expensive private medical care and insurance become.

Medicaid, historically, has paid health care providers directly, on a fee-for-service basis. However, states often contract with private insurers to provide Medicaid benefits to enrollees in the hope of making the program more efficient. Currently some 60 percent of enrollees receive Medicaid benefits through private insurers. This practice creates some of the same problems as in Medicare (see Chapter 38). Once states determine how much they will pay insurers per enrollee, insurers identify and recruit Medicaid-eligible individuals who will cost them less than that amount — pulling them into the system. These are often healthy people who were eligible for Medicaid but never enrolled. Whatever unnecessary expenditures these “private” Medicaid plans might avoid, the added costs of new enrollees swamp those savings. One study found that when California decided to “switch from fee-for-service to managed care,” there was “a substantial increase in government spending but no corresponding improvement in infant health outcomes. The findings cast doubt on the hypothesis that health maintenance organization (HMO) contracting has reduced the strain on government budgets.” Overall, contracting with private carriers tends not to reduce Medicaid spending in the average state.

The State Children’s Health Insurance Program

Congress created the State Children’s Health Insurance Program in 1997 to expand health insurance coverage among children in families that earn too much to be eligible for Medicaid. The federal government funds each state’s program much as it funds traditional Medicaid, but with two main differences. First, states receive a larger federal match under SCHIP than under traditional Medicaid. In 2017, the federal government will pay for at least 88 percent of the cost of each state’s program because the ACA authorized a temporary increase in the federal share. For every dollar that states invest in SCHIP, they can “pull down” at least $7 from the federal government (i.e., from taxpayers in other states).

Second, the federal government ostensibly limits the amount it will contribute to each state’s program. But the cap is not as binding as it appears. States often burn through their federal SCHIP funds before the end of the fiscal year and then demand additional funds. In effect, states create an emergency situation, daring Congress to throw sick children off the SCHIP rolls. Congress has repeatedly bailed out such states, effectively rewarding them for committing to spend more federal dollars than federal law allows.

As a result of these perverse incentives, states have expanded SCHIP eligibility dramatically. Nineteen states offer SCHIP to families of four with incomes of $73,000 or more. In New York, SCHIP is available to families of four earning $98,000 annually. Because SCHIP targets families higher up the income scale than Medicaid does, and because higher-income families are more likely to have health insurance to begin with, SCHIP leads to even greater “crowd-out” of private insurance than Medicaid.

Are Medicaid and SCHIP Even Helping?

Remarkably, there is little reliable evidence that these programs have a net positive effect on health, and absolutely no evidence they are the best way to improve the health of targeted populations.

Critics argue that despite the expense, Medicaid is lousy coverage. The Government Accountability Office reports that, compared with privately insured individuals, Medicaid enrollees notoriously have “greater difficulty accessing specialty and dental care,” and “over two-thirds of children in Medicaid with a potential mental health need did not receive mental health services.”

A study by John Bates Clark Medal-winning economist Amy Finkelstein and other top health economists examined the effects of Oregon’s decision to expand Medicaid in 2008. The Oregon Health Insurance Experiment randomly assigned applicants to receive Medicaid or nothing, and then compared outcomes for the two groups. As it happens, the study’s participants were drawn from the same vulnerable population targeted by the ACA’s Medicaid expansion. Random assignment made this experiment the most reliable study ever conducted on the effects of health insurance. The authors found that “Medicaid coverage generated no significant improvements in measured physical health outcomes in the first 2 years, but it did increase use of health care services, raise rates of diabetes detection and management, lower rates of depression, and reduce financial strain.”

The authors chose measures of physical health that should have been amenable to treatment over a two-year period. Yet Medicaid produced no improvement in the people it served compared with the people who got no coverage. Like Finkelstein’s study that found no effect of Medicare on elderly mortality in that program’s first decade (see Chapter 38), the lack of any improvement in physical health outcomes among Medicaid enrollees should throw a stop sign in front of Medicaid generally and the ACA’s Medicaid expansion in particular.

Similarly, there is no evidence that Medicaid is cost effective. The Oregon Health Insurance Experiment did find small improvements in self-reported mental health. But not even that study attempted to quantify whether Medicaid is a cost-effective way of achieving those gains — that is, whether state and federal governments could have purchased better health by spending those funds differently.

Whether or not Medicaid, SCHIP, or the ACA’s premium subsidies turn out to improve health for some populations, or to be a cost-effective way of doing so, these programs become increasingly less cost effective the higher up the income scale they reach. Higher-income households have higher baseline access to health insurance and medical care. As these programs move up the income scale, they offer taxpayer-financed coverage to increasing numbers of people who already had private insurance. The 82 percent crowd-out estimate mentioned previously suggests the ACA’s Medicaid expansion could be covering fewer than 2 previously uninsured Americans for the price of 10.

Federal and state governments should not continue to take trillions of dollars from taxpayers to support these programs when they don’t even know what they are buying.

Investigate Whether Medicaid Is Actually Helping

Rather than expand Medicaid, federal and state policymakers should conduct further experiments to determine what benefits Medicaid and SCHIP actually produce and whether other uses of those funds would produce greater gains in health and financial security. Policymakers should model these studies on the Oregon Health Insurance Experiment. The studies should be conducted with existing populations rather than new enrollees, so as not to impose additional burdens on taxpayers.

States should apply for waivers from the federal government to conduct such studies. Where federal law does not provide authority for the Secretary of Health and Human Services to approve such waivers, Congress should create such authority or enact legislation directly approving such studies.

Critics will object to randomly assigning some Medicaid enrollees to receive no coverage. Such criticism makes the mistake of assuming that Medicaid improves health when that is exactly what we do not know and precisely why states need to conduct such studies. It is unethical to preserve or expand Medicaid without knowing whether it even helps its presumed beneficiaries. The ethical course is to determine whether Medicaid is cost effective. That requires random assignment.

Refuse the ACA’s Medicaid Expansion

States that have implemented the Medicaid expansion are buckling under the expense. The program is costing states far more than they or the federal government projected. In 2015, the cost was $6,366 per enrollee, nearly 50 percent more than the federal government projected. The cost to states will be even higher because enrollment in implementing states has exceeded projections by an average 91 percent. Enrollment has even exceeded maximum-enrollment projections by an average 73 percent.

The 19 states that dodged those bullets by refusing to implement the Medicaid expansion should continue to refuse. The 31 states that have implemented it should withdraw from the program.

Repeal the ACA

Without reliable evidence of cost-effectiveness, neither those 31 states nor Congress can justify the Medicaid expansion, particularly when every penny Congress spends on it adds to the federal debt. Congress should repeal the ACA’s Medicaid expansion along with the rest of that act. Repealing the Medicaid expansion alone would reduce federal spending and deficits by $969 billion from 2017 through 2026 and eliminate the low-wage trap that the program creates. Repealing the remainder of the ACA would eliminate the low-wage traps its exchange subsidies create, while reducing the cost of private health insurance for the vast majority of enrollees of those programs. (See Chapter 36.)

The ACA remains an unpopular law. Nineteen states have rejected its Medicaid expansion. Those states have reduced federal spending, federal deficits, and the future tax burden of taxpayers in all states. Projections by the Urban Institute indicate that those 19 states will save taxpayers $349 billion by 2022. It is unfair to force taxpayers in states that have rejected the Medicaid expansion to pay for the expansion in the other states.

Reform Medicaid and SCHIP

Repealing the ACA is not enough, however. It makes little sense for taxpayers to send money to Washington, only for Congress to send those funds back to their state capitols with strings and perverse incentives attached. Congress should devolve control over Medicaid and SCHIP to the states.

In 1996, Congress eliminated the federal entitlement to a welfare check, placed a five-year limit on cash assistance, and froze federal spending on such assistance. It then distributed those funds to the states in the form of block grants with fewer federal restrictions. The results were unquestionably positive. Welfare rolls were cut in half, and poverty reached the lowest point in a generation.

The federal government should emulate that success by eliminating all federal entitlements to Medicaid and SCHIP benefits, freezing federal Medicaid and SCHIP spending at current levels, and distributing those funds to the states as unrestricted block grants. Congressional Budget Office projections indicate that simply freezing federal Medicaid and SCHIP spending at 2016 levels would produce $945 billion in savings and deficit reduction by 2026. Together with repeal of the ACA’s Medicaid expansion, block grants would reduce projected federal deficits from 2017 through 2026 by roughly 20 percent.

With full flexibility and full responsibility for the marginal Medicaid dollar, states could then decide whether and how to navigate the Samaritan’s dilemma. States that want to focus only on their neediest residents could do so and put the savings toward other priorities like police or tax reduction. States that want to spend more on their Medicaid programs would be free to raise taxes to do so, and vice versa, without federal strictures. States would learn from the successes and failures of each other’s experiments. Since states would bear the full marginal cost of their reformed Medicaid programs or successor programs, they would be more likely to conduct randomized, controlled experiments to determine the most cost-effective uses of those funds.

As an alternative to the current system of matching grants, some members of Congress have proposed that the federal government contribute to each state’s Medicaid program through “per capita block grants.” In that case, the federal government would provide states with a fixed amount of funds per Medicaid enrollee. Per capita block grants would eliminate the incentive that the current matching-grant system creates for states to offer more benefits to enrollees. Indeed, they could encourage states to offer less coverage and even worse access to providers. Unfortunately, this proposal would not encourage states to remove from their Medicaid rolls people who could obtain coverage on their own. On the contrary, it would preserve the current incentive for states to add more and more nonneedy people to their Medicaid rolls.

Block grants like those used in welfare reform would eliminate the perverse incentives that induce dependence, favor Medicaid and SCHIP spending over other priorities, lead states to tolerate widespread fraud, and encourage states themselves to defraud federal taxpayers. Over time, the federal government should give the states full responsibility for Medicaid by eliminating federal Medicaid spending while concomitantly cutting federal taxes. States can hasten these reforms by pressuring the federal government for maximum flexibility in administering their Medicaid programs.