Chair Grove, Chair Conklin, and members of the House State Government Committee, thank you for the honor of participating in today’s hearing. By way of introduction, my name is Aaron Yelowitz, and I have served as a professor of economics at the University of Kentucky for over 20 years. Starting with my dissertation in graduate school in the early 1990s, I’ve focused much of my research on how the Medicaid program works. My views expressed here today are informed by my research and reading of the literature on Medicaid.
Today, I would like to offer the following takeaways:
- Medicaid expansions have led to expanded coverage, including of adults with incomes substantially greater than the federal poverty line.
- Errors in enrollment in Medicaid include both enrolling ineligible adults and denying coverage to eligible Americans.
- There will be little incentive for states to audit Medicaid enrollment as long as the federal government continues to pay 90 cents of every dollar.
Like many health economists, I became intensely interested in the biggest expansion of government into the provision of health insurance in 50 years with the implementation of the Affordable Care Act. My research team published some of the first studies quantifying the gains in both public and private health insurance coverage in 2016 and 2017 using publicly available data from the Census Bureau’s American Community Survey (ACS). Our early studies found that, because the expansions in Medicaid (which were initially adopted by 26 states and the District of Columbia) provided coverage to non-elderly adults with incomes less than 138 percent of the federal poverty level (FPL), there have been large gains in Medicaid coverage for poor and near-poor adults in expansion states (and also gains in private coverage for those higher up the income distribution).
After Being Expanded, Medicaid Covered Ineligible, Higher-Income Americans
When I applied our same research methodology to higher-income individuals, I discovered that the Medicaid expansions to new adults also led to increased coverage among adults far away from the income threshold of 138 percent of the FPL. I first documented this finding in September 2016 in Kentucky. Using ACS survey data, I found approximately 38 percent of new adult Medicaid recipients, or 73,000, had incomes exceeding the Medicaid eligibility threshold, including more than 13,000 with incomes exceeding 250 percent of the FPL (an income threshold around $69,000 [in 2022 dollars] for a family of four). The findings in this study were cited as part of the motivation for Kentucky’s recently passed HB-7, which shines greater light on eligibility redetermination and presumptive eligibility practices. The bill also relies on administrative data sources such as the Kentucky Office of Unemployment Insurance to better track changes in employment and wages.
I followed up this sole-authored case study of Kentucky with a larger national study with collaborators from the University of Kentucky and Georgia State University. In a study that received prominent attention from the Wall Street Journal in August 2019, I compared 9 states that expanded Medicaid with 12 states that had not expanded by 2017. My team found that approximately 800,000 individuals appeared to gain Medicaid coverage for which they were seemingly ineligible. My study received scrutiny that I rebutted in the blog Health Affairs later that year.
Audits of Improper Enrollment Showed a Broken Enrollment Process
Toward the end of 2019, I also published a research paper for the Mercatus Center with my collaborator from the Paragon Institute that not only documents changes in improper enrollment between 2012 and 2017 for all 50 states, but also identifies local hot spots. Our reading of the audits by the Office of the Inspector General (OIG) at the US Department of Health and Human Services found many shortcomings with states’ Medicaid eligibility processes, including failing to maintain proper documentation, not properly verifying income eligibility, misclassifying individuals into the new adult category, and failing to properly verify citizenship.
Audits both by OIG and various states corroborated the concerns that our approach with survey data had raised. OIG audits in California, Colorado, Kentucky, and New York have shown large numbers of both ineligible and potentially ineligible Medicaid enrollees. State audits in Louisiana and Oregon have shown a broken eligibility process, with large numbers of ineligible or potentially ineligible enrollees. An important and intuitive reason for this consistent finding across different methodologies boils down to incentives. For new adult Medicaid enrollees, the federal government pays 90 cents of every dollar of expense, far higher than for other eligible groups. The incentive for any state to vigilantly scrutinize enrollment is far smaller than if the state were paying a higher percentage or if the Medicaid program were a block grant.
An Eligibility Process with Integrity Would Be Equitable and Reduce Waste
A Medicaid eligibility process with integrity would achieve two goals: (1) reduce the number of false positives and (2) reduce the number of false negatives. A false positive is someone who manages to enroll in Medicaid but isn’t eligible. A false negative is a person who is eligible but unsuccessful at enrolling, perhaps because of paperwork burdens. Improper enrollment (the false positives) create taxpayer waste, divert resources from intended beneficiaries, and violate the idea of horizontal equity—i.e., that those in similar circumstances should be treated the same.
Let me conclude my testimony with two thoughts. First, my analysis was done before the pandemic, and as you are aware, the Families First Coronavirus Response Act requires states to provide continuous coverage for Medicaid enrollees until the public health emergency ends as a condition of receiving enhanced federal funding. Thus, millions more adults with incomes greater than 138 percent of the FPL are now enrolled in Medicaid than before the pandemic. Second, the approach of merging administrative data sources, as in Louisiana, could fruitfully be applied in Pennsylvania. With your help, I would be eager to assist in such an endeavor.
I look forward to answering your questions.