Medicaid Empire: Why New York Spends so much on Health Care for the Poor and Near Poor and How the System Can Be Reformed
Monday, March 20, 2006
by John C. Goodman, Michael Bond, Devon M. Herrick, Joe Barnett, and Pamela Villarreal
Table of Contents
- Executive Summary
- Overview of Medicaid
- What Difference Does Medicaid Make?
- How New York Compares to Other States
- Recommendations for New York Medicaid Reform
- How the Federal Government Can Help
- Appendix I: The Federal Medicaid Matching Formula
- Appendix II: Medicaid Regression Methodology
- About the Authors
What Difference Does Medicaid Make?
Medicaid was intended to improve access to medical care and thereby the health of the indigent. Arguably, it has done neither. And as the program has expanded to cover additional populations, including the near poor and even middle class individuals, the evidence suggests that it is displacing, or crowding out, private health insurance coverage.
Impact on Access to Care. It is commonly assumed that Medicaid enrollees have greater access to health care than if they were uninsured. But there is little evidence that is true. Most enrollees rely on the same public and not-for-profit hospitals and clinics that have always provided a health care safety net for the poor.33 Medicaid is the primary social safety net for indigent health care — but not the only one. For example, community health clinics and hospital emergency rooms are often providers of last resort. The existence of this free care (charity care) often makes people eligible for Medicaid feel that enrollment is unnecessary. Free care, in other words, appears to be a substitute for Medicaid.
“Medicaid spending substitutes for free care by doctors and hospitals.”
In fact, where Medicaid spending is low, spending on free care is high — and vice versa. For example, an NCPA analysis of health regions in Texas found that the area with the highest average Medicaid cost per enrollee spent more than twice as much as the area with the lowest average cost.34 But we found that adding spending on free care to Medicaid spending cuts the variation in regional health care spending in half. In other words, free care substitutes for Medicaid spending, as regions that spend less on Medicaid tend to spend more on free care and vice versa. [See sidebar on “Free Care.”]
Impact on Health. An oft-cited argument for Medicaid is that making health care virtually free at the point of consumption encourages preventive care and improves health outcomes. This could potentially reduce overall health care costs. Unfortunately, there is little evidence that this occurs. Studies suggest explicit attempts to encourage the use of preventive care by Medicaid beneficiaries are generally unsuccessful. For example, one study found that outreach programs in North Carolina had a very small impact on utilization.35 Another study found that receiving Medicaid benefits for a year increased the probability children would receive checkups by only 17 percent. The researchers concluded that “factors other than insurance and income, such as the low educational attainment of low-income mothers, explain approximately 80 percent of the gap between low-income and other children in their well-child visits.”36 Additionally, there is no evidence that becoming eligible for Medicaid significantly improves child immunization rates.37
However, analyzing the use of Medicaid services tells us only about inputs, not outcomes such as health improvement. The evidence of Medicaid’s effect on health is conflicting. For example, Medicaid eligibility is somewhat associated with a lower risk of infant mortality.38 And University of Washington researchers found some evidence that Medicaid decreases low-weight births for medically high-risk women.39 Studies in other states, however, have found Medicaid expansion had little effect on prenatal care and outcomes.40
“Medicaid creates disincentives to work and save.”
Impact on Income and Wealth. Like other means-tested government benefits, Medicaid creates disincentives to work and save.41 Income tests discourage work by withdrawing benefits as income rises. Asset tests encourage people to transfer or spend down their assets or avoid saving to become or remain eligible. Knowing their medical needs will be covered by Medicaid, eligible families may boost consumption and save less for emergencies. University of Kentucky economist Aaron Yelowitz and Jonathan Gruber of the Massachusetts Institute of Technology found that Medicaid recipients consumed more and saved less, reducing their average household wealth in 1993 by $1,600 to $2,000 (in today’s dollars).42
Impact on Private Insurance. Many assume Medicaid insures people who otherwise would not have access to private insurance.43 However, Medicaid also induces some people to drop their private coverage in order to take advantage of free health insurance offered by the state. Often this occurs when employers cease offering insurance coverage as large numbers of current and prospective employees become eligible for Medicaid. As a result of such crowding out, the cost of expanding public insurance programs has been high relative to the gain. For example, if for each new enrollee in a public program at least one person loses private insurance, there will be no net reduction in the number of uninsured, despite the higher taxpayer burden. If for every two new enrollees in the public program one person loses private insurance, the cost to the taxpayers for each newly insured person will double.44
David Cutler of Harvard University and Jonathan Gruber of the Massachusetts Institute of Technology found that Medicaid expansions in the early 1990s were substantially offset by reductions in private coverage.45 They found that for every additional dollar spent on Medicaid, private sector health care spending was reduced 50 to 75 cents on the average.46 Thus taxpayers incurred a considerable financial burden, but at least half, and perhaps as much as three-fourths, of the expenditures replaced spending by individuals, rather than buying additional medical services. [See Figure V.]
“Medicaid crowds out private insurance coverage.”
In 1997 Congress expanded access to public health insurance with the State Children’s Health Insurance Program (SCHIP). This program provides additional federal matching funds to states to cover children in families whose incomes are too high to qualify for traditional Medicaid. SCHIP also crowds out private coverage. Take a low-income working family covered by an employer-sponsored health plan. The employer might have covered some or all of the cost of insurance premiums for the employee and family with pretax dollars. But paying wages is more attractive to actual and potential employees if coverage is provided by the state. Thus SCHIP offered some employees the opportunity to increase wages and reduce their out-of-pocket health care costs.
“Increased Medicaid enrollment was offset by a drop in private insurance coverage.”
As a result of expansions of Medicaid and SCHIP, the number of poor children without health insurance fell from 19 percent in 1997 to 11 percent in 2003. During this period enrollment of low-income children in public programs increased from 29 percent to 49 percent.47 Meanwhile, the percentage of children from low-income families covered by private insurance fell from 47 percent to 35 percent, although there was little change in the percentage of privately insured children in households at higher income levels. [See Figure VI.] Thus, SCHIP apparently induced low-income employees to drop their children from company-sponsored health plans.48 The crowd-out of private insurance due to the expansion of public programs was 0.6, meaning that every percentage point increase in public coverage resulted in a reduction of about 0.6 percentage points in private coverage among low-income children.49
“Medicaid expansion has not reduced the ranks of the uninsured.”
More recently, from 2001 to 2003, the proportion of low-income Americans enrolled in public programs rose 6.1 percentage points. However, this increase was offset by a 4.9 percentage point decline in coverage by employer-sponsored plans. Overall, the rate of uninsured fell only about one-half point.50 [See Figure VII.] Casual empiricism suggests that it takes a 13 percentage point increase in public coverage to reduce the uninsured rate by 1 percentage point.