Medicaid Spending Growth Compared to Other Payers: A Look at the Evidence

Introduction

Medicaid provided health coverage for over 70 million individuals during the 2013 fiscal year, with a total cost (including spending by both the federal government and states) of about $460 billion.¹ In that year, total spending for Medicaid was 2.8 percent of gross domestic product (GDP). Federal spending for Medicaid was 1.6 percent of GDP, compared to 3.5 percent for mandatory Medicare outlays.² And in 2014, due to a rapid increase in spending associated mainly with the expansion of Medicaid coverage under the ACA, Medicaid expenditures totaled $494 billion, accounting for 2.9 percent of GDP, with the federal government paying about 60 percent of the total.³

A number of methodologically rigorous studies have demonstrated that Medicaid coverage helps to improve receipt of preventive health care, access to care, and out-of-pocket spending burdens and other financial outcomes.^4,5,6,7,8 Fewer studies have assessed the impacts of Medicaid on health, but studies have found positive impacts on infant mortality,⁹ child mortality,¹⁰ HIV mortality,¹¹ adult mortality,¹² disease-related mortality,¹³ and reported mental health status/rates of depression).^14,15

In addition, recent evidence suggests that Medicaid coverage has long-term positive effects. Expansions of Medicaid eligibility for pregnant women increased economic opportunity of their children when they reached adulthood through increased rates of high school and college completion and higher incomes.^16,17,18 In addition, children who gained eligibility for Medicaid paid more in cumulative taxes by age 28 compared to similar children who did not gain Medicaid coverage, such that the government is estimated to recoup 56 cents of each dollar spent on Medicaid during childhood by the time the children reach age 60.¹⁹

Despite Medicaid’s positive effects, Medicaid is a significant expenditure item and consequently, a major target for cost-cutting in federal and state budget debates. Medicaid, like other payers, has room to improve the efficiency of the services that are provided.^20,21,22 While this brief does not examine efficiency or quality of care, we recognize the potential for savings likely exists in Medicaid as well as other payers due to inefficient patterns of care (such as overreliance on emergency department care, unnecessary tests, over-prescribing of drugs), medical mistakes and other inefficient spending. At the same time, the relatively low provider payment levels in Medicaid may affect access to providers and have other negative impacts for Medicaid beneficiaries.

Given ongoing concerns about federal and state budgets, the costs of the Medicaid program are likely to be again at the forefront of state and federal policy discussions. To inform these discussions, this brief summarizes findings from studies on Medicaid program spending from both peer-reviewed journals and “grey literature” such as working papers and reports from academic and other research centers. We first examine literature that compares spending for Medicaid enrollees to spending for those with private insurance, focusing on broad, national populations and studies that were published in the last 15 years, using multiple regression methods to control for differences between the two groups in the health status and health risks of enrollees. We then examine literature that assesses the rate of growth of spending and per capita spending in Medicaid compared with other payers and other benchmarks. For this examination, due to the timely nature of the topic, we reviewed studies that were released in the last three years and included estimates of per capita Medicaid spending and growth trends, as well as selected studies of provider payment. Several other studies not included in this examination have similar findings and draw on the same data sources.

Findings from a Literature Review on Spending in the Medicaid Program

How do health care needs for Medicaid enrollees compare to the low-income privately insured population?

When comparing historical per capita spending between Medicaid and private coverage, it is critical to account for the fact that Medicaid enrollees have had far greater health care needs and a greater prevalence of disability compared to low-income privately insured population, at least prior to 2014. (These differences are likely linked to Medicaid eligibility pathways, several of which are targeted to adults with health needs. Differences could attenuate somewhat with the ACA expansion population in 2014.) A 2013 study using data from 2003 to 2009 showed that, among low-income adults (with family income below 138 percent of the federal poverty level [FPL]), Medicaid enrollees reported poorer physical and mental health, more limitations, and more comorbidities than did privately insured adults in the same income group.²³ In particular:

• Over half (53.3 percent) of adults enrolled in Medicaid reported a health limitation, compared to just one-fifth (21.0 percent) of the low-income privately insured adults.
• Medicaid enrollees had greater health care needs than low-income privately insured population on a host of other measures as well, such as a higher prevalence of multiple chronic conditions (48.3 percent versus 31.6 percent), a higher prevalence of non-chronic conditions (69.6 percent versus 59.8 percent), a higher prevalence of mental illness or substance abuse (37.8 percent versus 18.6 percent), and a higher prevalence of other conditions such as asthma, diabetes, heart disease, hypertension, back conditions, bronchitis/respiratory conditions, and digestive/gastrointestinal conditions.

How does spending per enrollee in Medicaid compare to private insurance?

Spending per enrollee is lower for Medicaid compared to private insurance after controlling for differences in socio-demographic and health characteristics between the two groups. Given the significant health and disability differences between Medicaid enrollees and those who are privately insured, the most rigorous research examining differences in per enrollee spending has focused primarily on regression-adjusted comparisons that control for these underlying differences in the need for health care.

An early study²⁴ used data from 1996 to 1999 and found that after adjusting for health and socio-demographic factors, total spending would be 18 percent higher for adults if a typical low-income adult with Medicaid were instead covered by private health insurance. The underlying driver of the difference in total expenditures was differences in provider payment rates.

• The study found no evidence that spending differences between Medicaid and private coverage for low-income enrollees were due to lower service use of Medicaid enrollees.
• If adults with Medicaid coverage were given private coverage, medical spending was predicted to increase from $3,250 to $3,848, an increase of over 18 percent, while spending for children would be virtually unchanged.

Updating and expanding on the previous research, another study²⁵ used MEPS data from 2005 and found that after adjusting for health and socio-demographic factors, total health care spending for a typical low-income Medicaid adult or child would have been far higher if covered by private health insurance. In particular, if a typical low-income Medicaid adult or child were instead covered by private health insurance, total spending would be 26 percent higher for adults and 37 percent higher for children.

• The study estimated that total health care spending would increase nearly 26 percent, from $5,671 per person per year to $7,126, if a typical low-income Medicaid adult were covered by private health insurance for a full year. In addition, total health care spending would increase 37 percent, from $909 per child per year to $1,247, if a low-income Medicaid or CHIP-enrolled child were covered instead by private health insurance for a full year.
• The study also showed that out-of-pocket spending was estimated to be six to seven fold higher under private insurance than under public insurance.

Subsequent research with the 2005 MEPS²⁶ examined use and expenditures for different types of services (e.g. inpatient hospital, outpatient hospital, emergency department, office-based physician, prescription drugs, and other)²⁷ and found that after adjusting for health status and other factors, if adults enrolled in private coverage were instead enrolled in Medicaid, physician expenditures would be 34%.

• Regression adjusted estimates showed lower spending for adults under Medicaid compared to privately insured adults for all service types, including emergency department care (8 percent), and significantly lower expenditures for inpatient hospital care (33 percent) and outpatient hospital care (40 percent).
• Regression adjusted estimates for a typical privately insured child showed lower spending under Medicaid than under private coverage for all service types except for inpatient hospital care, which the authors attributed to the possibility of either expenditure misclassification or substitution of outpatient for inpatient hospital services for children.

• The study also found that Medicaid provides access to health care services comparable to that of the privately insured sample but at significantly lower costs. In addition, the study found that the likelihood of using most health care services (e.g., primary care doctors, prescription drugs and inpatient care) would not differ significantly if Medicaid enrollees were instead covered by private insurance, with the exception of lower emergency department use and more specialist visits in private coverage.
• Additionally, given the differences in benefit design between the private plans and Medicaid, out-of-pocket spending for health care services was estimated to be over three times as high if Medicaid enrollees were instead covered by private insurance.

Lastly, after the 2012 Supreme Court Decision that rendered the Medicaid expansion optional for states under the ACA, the Congressional Budget Office (CBO) estimated per capita spending levels if individuals were covered under the new ACA exchanges as opposed to Medicaid. CBO projected that that federal spending would increase by roughly $3,000 by 2022 for each such person on average. The CBO estimated that, on average, exchange subsidies would cost the federal government about $9,000 while the Medicaid costs would have been roughly $6,000.³⁰

How does spending growth in Medicaid compare to private payers and Medicare?

In this section, we examine literature assessing the rate of growth of spending and per capita spending in Medicaid compared with other payers, as well as other benchmarks such as per capita growth in GDP and the rate of general medical cost inflation. We find that the evidence suggests that Medicaid has constrained costs as well as, or better than, Medicare and private insurance. The relevant studies are summarized in Table 2.

Medicaid spending growth between 2007 and 2013 was heavily driven by rising enrollment over and above inflation, as demonstrated in a series of annual studies.^31,32,33

• The most recent study found that total spending on medical services in Medicaid increased by an average annual rate of 5.7 percent from 2007 to 2013, with higher spending growth of 6.9 percent during the recessionary period (2007 to 2010) and slower spending growth of 4.4 percent after the recession (2010 to 2013).³⁴
• During the recessionary period (2007 to 2010), enrollment among families grew at 7.2 percent per year on average compared to 3.3 percent per year on average among the disabled and elderly. After the recession (2010 to 2013), enrollment growth among families slowed to 3.1 percent per year while remaining fairly steady at 2.9 percent per year among the disabled and elderly. In 2013, over two-thirds of Medicaid enrollees (39.8 million) were non-disabled adults and children, and of the disabled and elderly enrollees, most were dually eligible for Medicaid and Medicare.
• After adjusting for the changing composition of Medicaid enrollees (i.e. by age and disability status), annual per capita Medicaid spending growth on a per enrollee basis was just 1.7 percent from 2007 to 2013, with higher per capita spending growth of 2.2 percent during the recessionary period (2007 to 2010) and slower per capita spending growth of 1.3 percent after the recession (2010 to 2013).

This research also showed that growth in Medicaid spending per enrollee was relatively low compared to several health spending benchmarks from 2007 to 2013.³⁵

• Average annual per enrollee Medicaid spending on medical services increased by 1.7 percent per year from 2007 to 2013, compared to GDP per capita growth of 1.7 percent, average annual per capita growth of 3.1 percent for national health expenditures and average annual growth of 3.2 percent for the consumer price index (CPI) for medical care.
• At 3.1 percent, average annual spending growth per enrollee for Medicaid acute care (i.e. excluding long term care and including managed care expenses as acute care) was about 33 percent lower than the 4.6 percent average annual spending growth per enrollee for private health insurance, which provides primarily acute care benefits.

CBO analysis of historical cost growth up to 2014 showed that Medicaid spending increased as a share of GDP primarily due to rising Medicaid enrollment; it also found that excess cost growth (defined below) in Medicaid was lower than for other payers.³⁶

• CBO analysis shows that that Medicaid spending growth was mainly driven by rising enrollment and that the Medicaid program has controlled costs per enrollee more than other programs.
• The CBO’s estimate of excess cost growth, which measures the extent to which growth in adjusted health care spending exceeds growth in potential output per person,³⁷ shows that Medicaid’s excess cost growth on a per enrollee basis has been lower than other payers. Between 1990 and 2013, Medicaid spending growth was just 0.3 percentage points higher than spending growth for the overall economy, whereas overall health spending growth outpaced growth of the overall economy by 1.1 percentage points.
• In addition, the CBO analysis shows that excess cost growth in Medicaid was lower than in that Medicare and other sources since 1975.³⁸ The weighted average rate of excess cost growth in Medicaid was 1.5 percent between 1975 and 2013, compared to 1.9 percent in Medicare and 1.8 percent for other health care spending.³⁹

Consistent with the CBO findings, a paper by Iglehart and Sommers published in 2015 examined Medicaid spending over an even longer time frame—since 1966—also finding that Medicaid spending growth has been driven primarily by increased enrollment.⁴⁰ Additionally, the paper finds that inflation-adjusted per capita Medicaid spending was flat or declining between 1998 and 2014.

In a recent report, the CMS Office of the Actuary analyzed historical growth rates in per enrollee Medicaid expenditures during the 10-year period 2004 and 2013, finding that the effect of enrollment mix—or the difference between the increase in Medicaid benefit expenditures per enrollee and the increase in Medicaid benefit expenditures per enrollee if enrollment were held constant each year— on Medicaid expenditures varies widely across years.⁴¹

• After adjusting for the changing composition of Medicaid enrollees on expenditures, Medicaid benefit expenditures per enrollee grew at an average annual rate of 1.7 percent from 2004 to 2013. Over that 10-year period, adjustments for the effect of enrollee composition on spending varied widely, from −2.4 percent to 1.4 percent.

Finally, National Health Expenditure (NHE) estimates show that Medicaid spending growth in 2014 is likely faster than recent historical trends due to enrollment⁴² —and that Medicaid has generally constrained per capita spending growth more than any other payer.⁴³

• National Health Expenditure Projections from the Center for Medicare and Medicaid Services (CMS) Office of the Actuary show that in 2012 and 2013, historical data indicate that Medicaid health spending grew 5.4 percent and 6.1 percent, respectively, whereas Medicaid spending was estimated to have grown 12.0 percent in 2014, largely as a result of the expansion of Medicaid eligibility under the ACA.
• In fact, while annual spending growth rates in Medicaid were similar or higher than in Medicare or private coverage between 2007 and 2014, the NHE analysis showed that per capita Medicaid spending growth was below that of Medicare and private coverage over most of that period. For example, in 2012, the annual growth in per enrollee expenditure was 0.5 percent in Medicaid, compared to 2.7 percent in Medicare and 4.8 percent in private insurance.

How does provider payment and prescription drug payment in Medicaid compare to other payers?

The research reviewed above, which shows that Medicaid spending per capita and Medicaid spending growth have historically been relatively low despite a disproportionately sick enrollee population with more health problems, raises questions about how the cost savings has been achieved. Cost-containment efforts, such as expanded enrollee copayments and pharmacy management tools (e.g. preferred drug lists [PDLs]), as well as constrained access, have almost certainly played a role in constraining costs in the Medicaid program. However, a critical factor driving savings appears to be low payment rates.⁴⁴

A handful of studies have assessed how provider payments for particular services under Medicaid fee-for service (FFS) or Medicaid managed care compare with provider payments under Medicare or private insurance. In these studies, Medicaid is generally demonstrated to have lower payment rates. The studies are summarized in Table 3.

Between 1993 to 2014, researchers at the Urban Institute produced multiple studies that have shown that part of the reason that Medicaid is successful in constraining costs is that the program has consistently had lower fees for physician services compared with the fees paid by private payers or Medicare.^45,46

• Most recently, a 2014 Urban Institute study collected data on Medicaid physician fees for 27 procedure codes for three types of services: primary care, obstetric care, and other services.⁴⁷ The researchers computed a state-specific Medicare-to-Medicaid fee index, or the ratio of the Medicaid fee for each service in each state to the Medicare fee for the same service. The study showed that, on average, Medicaid fees in the survey were 66 percent of the Medicare fees.

A recent study of payments per inpatient hospital stay between 1996 and 2012 compared inflation-adjusted payment rates that were also standardized across patient and stay characteristics; it found that private insurance had the highest rates, followed by Medicare and then Medicaid—with Medicaid payment rates averaging approximately 90 percent of Medicare rates across the period.⁴⁸ However, the study did not include supplemental Medicaid payments to hospitals.

A 2015 analysis by the Office of Inspector General (OIG) demonstrated that Medicaid’s per unit pharmacy costs were less than half of Medicare’s per unit pharmacy costs—with much of the savings due to Medicaid’s rebate policies.⁴⁹

• The study evaluated the costs of 200 selected brand-name drugs and found that pharmacy average unit costs were similar under Medicare Part D and Medicaid. For example, the average unit reimbursement amounts in Medicare Part D and Medicaid differed by less than 2 percent for 135 of the 200 selected drugs.
• However, Medicaid’s average net unit pharmacy costs (the average unit pharmacy reimbursement amounts minus the average unit rebate) were far lower than net unit costs under Medicare’s Part D in 2012.
• For the selected brand-name drugs in the study, median Medicaid unit rebate amounts were three times higher than median Medicare Part D unit rebate amounts, and for 37 of the selected drugs, median Medicaid unit rebate amounts were over 10 times higher than those for Medicare Part D. Medicare Part D unit rebate amounts exceeded Medicaid for just two of the selected drugs in the study. (Median unit rebate amounts in dollars were not published in the study.)
• After accounting for rebates in both programs for the selected brand-name drugs in the study, Medicaid net unit costs were less than half of Medicare Part D net unit costs for 110 of the selected brand-name drugs. Medicaid net unit costs were lower than Medicare Part D net unit costs for all but five of the brand-name drugs. Overall, while Medicaid drug expenditures in 2012 were lower than Medicare Part D expenditures at $35.7 billion compared to of $66.5 billion, Medicaid drug rebates were higher than Medicare, at $16.7 billion compared to $10.3 billion. Thus, rebates totaled 46.8 percent of Medicaid drug spending, compared to just 15.5 percent of Medicare Part D spending.

A 2014 GAO study demonstrated that provider payments for selected services under Medicaid FFS and Medicaid managed care were generally substantially lower—about 30 to 65 percent lower—than private insurance.⁵⁰

• The report examined how payments for 26 evaluation and management (E/M) services (including E/M for office visits, hospital care, and emergency care) in selected states compare under Medicaid FFS and Medicaid managed care and private health insurance.
• The study found that Medicaid rates were generally lower than private insurance in 2009 and 2010, prior to the temporary payment increases mandated by the Health Care and Education Reconciliation Act of 2010 (HCERA).
• In the 40 states where data were available, Medicaid FFS payments were 27 to 65 percent lower than private insurance in 31 of the 40 states. In the 23 states where data was available to compare Medicaid managed care payments to private insurance, GAO found that Medicaid managed care payments to providers were 31 to 65 percent lower than private insurance in 18 of the 23 states. The GAO found that Medicaid payments generally were lower than private insurance for all three types of E/M assessed, and that the magnitude of the payment differences was generally largest for emergency care and smallest for office visits.

A 2015 chartpack published by the American Hospital Association (AHA) examined reimbursements for hospital-based services for community hospitals, finding that Medicaid reimbursements are far lower than those for private payers.^51,52

• This analysis of aggregate hospital payment-to-cost ratios for hospital-based services financed by Medicaid, Medicare, and private payers from 1993 through 2013 shows that Medicaid rates have historically been far lower than private payers, and similar to Medicare levels. This study included Medicaid and Medicare Disproportionate Share Hospital (DSH) payments. In 2013, while aggregate private payment-to-cost ratios were near 145 percent, Medicaid ratios were about 89 percent and Medicare ratios were about 88 percent. The gap between Medicaid and private payer ratios is larger than a decade earlier, when private payment-to-cost ratios were about 122 percent and Medicaid ratios were about 92 percent.

Conclusion

This brief reviewed literature assessing a range of evidence on Medicaid program spending. Overall, the literature shows that per capita spending in the Medicaid program is lower compared to private insurers after adjusting for the greater health needs of Medicaid enrollees; Medicaid spending growth has been primarily driven by rising Medicaid enrollment; spending growth per enrollee in Medicaid has been low compared to other payers; and lower payment levels in Medicaid have contributed to its relatively low costs. Recent federal budget proposals include provisions to reform Medicaid financing in an effort to reduce federal spending. Some savings may be found from more efficient care delivery (i.e. reductions in emergency room visits or hospital readmissions). However, given Medicaid’s already lower payment rates that contribute to lower per capita spending, such proposals could result in reductions in utilization and/or enrollment as well as additional pressure on states to lower provider payment further. These changes could have adverse effects on beneficiaries’ access to care.