Welcome to another installment on the misuse of the Medicaid “improper payments” metric. This conversation began in 2018, then continued in 2020, 2022, and this year. For those just joining, Brian Blase of the Paragon Institute and his colleague, Rachel Greszler of the Economic Policy Innovation Center, have launched yet another misguided assault on federal support for state Medicaid programs. They contend that CMS is understating Medicaid’s improper payment rate by more than half. They urge Congress to (among other things) charge states “the full fare” for their “excessive” improper payments, presumably by cutting federal Medicaid matching payments to states. Their paper comes as a White House statement equates targeting improper payments with “eliminating waste, fraud, and abuse in government spending.” Here’s why this is an improper use of improper payments.
Improper payments are one way the federal government measures administrative performance in Medicaid, Medicare, and other federal programs. By statute, federal agencies are required to estimate improper payments in the major programs they administer and to take action to reduce those payments. In Medicaid, the methodology for estimating improper payments is known as Payment Error Rate Measurement (PERM). The purpose of PERM is to identify improper payments made by each state, understand the root causes, and, through the use of corrective action plans, reduce the rate of improper payments in each state and nationally.
Improper payments are not fraud. They include payments made to fraudulent providers, but, as discussed below, far from all improper payments are due to fraud. CMS explains that improper payments are “payments that do not meet CMS program requirements. They can be overpayments, underpayments, or payments where insufficient information was provided to determine whether a payment was proper. Most improper payments involve a state, contractor, or provider missing an administrative step.”
Medicaid is a large public health insurance program with lots of moving parts; given this complexity, mistakes are bound to be made. Medicaid covers over 80 million Americans for needed health and long-term care services and supports. Fifty-six different state Medicaid agencies (50 states, District of Columbia, and 5 territories) administer the Medicaid program on a day-to-day basis within broad federal guidelines. The policy flexibility inherent in Medicaid’s design allows for wide variation from state to state in eligibility rules, benefit design, and provider payment arrangements. In most states, much of the program is operated by private managed care organizations (MCOs) under contract with state Medicaid agencies. In short, Medicaid is operationally complicated. Mistakes are made — by states, by MCOs, by administrative vendors, by providers, by enrollees —because there are millions of opportunities to make them, especially in collecting and maintaining documentation related to payments. PERM is the way CMS identifies the mistakes that lead to improper payments and works with states to reduce them.
CMS estimates that the national Medicaid improper payment rate in 2024 was 5.09%; in other words, nearly 95 percent of Medicaid payments were likely proper. The 5.09% rate translates into an estimated improper payment amount of $31.1 billion in federal funds. Of that amount, $23.4 billion, or 74% of the payments reviewed were classified as improper due to insufficient documentation. These payments may or may not have been incorrect; there’s simply not enough information in the claims file to know. Another $4.9 billion, or 15.6%, of payments were deemed improper because the service provided was not medically necessary or the beneficiary was ineligible. In 5% of the claims reviewed, representing $1.5 billion, the provider was not enrolled or there was monetary loss for other reasons. You can read the CMS analysis in the FY 2024 HHS Agency Financial Report (pp. 210-218).
Blase and Greszler view the 2024 CMS report as “the most meaningless, because none of the annual PERM audits that make up the average amount included eligibility verification.” PERM estimates have three components: fee-for-service claims, managed care capitation payments, and eligibility determinations. CMS, through federal contractors, measures improper payments for each component and combines the results to produce an overall rate for each state. Seventeen states are reviewed each year. The state results for each of the three annual cycles are then combined to produce a national improper payment rate, weighted by state size.
The purpose of PERM eligibility reviews is to determine whether the state properly applied federal eligibility rules and state eligibility policies and procedures. If states can’t produce records and documentation to support an eligibility determination in connection with which a payment was made, PERM treats that as an improper payment, even if the beneficiary would have been eligible if the state had the proper documentation. Examples of eligibility errors include enrolling an ineligible individual, not conducting an eligibility redetermination in a timely manner, and not documenting a required element of the eligibility process, such as income verification.
Jennifer Wagner of the Center on Budget and Policy Priorities, gave the following example of an eligibility error in her testimony earlier this week before the Government Operations Subcommittee of the House Committee on Oversight and Government Reform: “For example, an applicant may have attested to their income on their application. The eligibility worker then checked electronic income data sources and used that to verify the client statement. However, the worker may not have saved a record of the electronic data that was used in the eligibility system. So when a reviewer looks at the case, they can’t confirm that the worker’s assessment was correct. There is no reason to believe that the applicant was not in fact eligible, but since the paperwork is missing, it’s considered an improper payment. This is a technical systems and training problem, not a fraud problem.”
It’s not clear what Blase and Greszler mean by “eligibility verification,” but there was in fact an eligibility component in each of the three annual PERM cycles that made up the 2024 estimate. For the 17 states in the July 2020 to June 2021 cycle, the eligibility component estimate was 10.46%; for the July 2021 to June 2022 cycle, 3.75%; and for the July 2022 to July 2023 cycle, 1.47%. (Table 2B). During most of this 3-cycle period, the PHE continuous enrollment provision was in effect. In exchange for a 6.2 percentage point increase in their federal matching rates, states agreed to suspend disenrollments starting in March of 2020 in order to keep people insured during the COVID-19 pandemic. In April, 2023, redeterminations of eligibility resumed as part of the PHE unwinding.
The 2024 Medicaid improper payment estimate represents the third consecutive year of decline in such estimates, from 21.69% in 2021, 15.62% in 2022, and 8.58% in 2023 to 5.09% in 2024. (See Chart). CMS attributes this decline to two factors. (p. 210). First, improved state compliance with Medicaid program rules. (This is, after all, the point of the exercise). Second, “it is due to reviews that considered certain flexibilities given to states related to COVID-19, such as suspending eligibility determinations and reducing requirements for provider enrollment and revalidations, which were previously part of the improper payment reviews prior to COVID-19.”
Blase and Greszler attribute this three-year decline in Medicaid improper payments to “meaningless” measurements of eligibility errors during the PHE. But the fact is that improper eligibility rates were not the only rates to decline over this period; the fee-for-service component estimates also declined each of those years, from 13.90% in 2021, 10.42% in 2022, 6.9% in 2023, and 4.83% in 2024, contributing to the overall decline in the national improper payment rate. (The managed care component estimate was negligible, less than 0.04% in each of those years). (Table 3A).
Blase and Greszler do not address the decline in the fee-for-service improper payment rate. Nor do they entertain the possibility that the three-consecutive-year decline in the national improper payment rate reflects a welcome improvement in state administrative performance. (As my CCF colleagues have documented, there was wide variation in state performance during the PHE unwinding. When states incorrectly disenrolled children or adults who were in fact eligible, those eligibility errors were not included in improper payments because no payments were made).
Florida’s estimated Medicaid Improper Payment Rate in 2023 was 7.0%, amounting to $1.503 billion in federal improper payments. The national Medicaid improper payment rate is important, but the real point of PERM is to improve state administrative performance by estimating state-specific improper payment rates, identifying root causes, and working with states through corrective action plans to upgrade documentation and reduce mistakes. In 2022, the Biden administration CMS began posting state-specific PERM results. For example, Florida, where the Paragon Institute is headquartered, was one of the 17 states reviewed in the July 2022 to July 2023 cycle. Florida is one of five states in that group with improper payment rates of 7.0% or more (the others were Alaska, Indiana, Montana, and the District of Columbia). Breaking down Florida’s overall improper payment rate into the three components: its fee-for-service rate was estimated at 16.2% (second highest behind Arizona); its eligibility rate was 3.9% (second highest following Alaska); and its managed care rate was 0.00% (same as all 16 other states). Blase and Greszler do not indicate whether the eligibility results for Florida are “meaningless,” and if so, what they believe the “real” rate of eligibility errors in Florida to be (why not just double it?).
Blase and Greszler are making up numbers in an effort to use Medicaid improper payments to justify cutting “hundreds of billions” of federal payments to states. Context matters. The Blase and Greszler article was posted on March 3, just six days after the House of Representatives adopted a budget resolution calling for its Energy & Commerce Committee to report out at least $880 billion in cuts in federal spending over the next 10 years. Since the President has ruled out changes to Medicare, most of the cuts will have to come from Medicaid. Blase and Greszler recommend, among other things, that Congress “charge states the full fare for their excessive improper payments.” They don’t specify what that amount would be; they say only that “The President and Congress can save hundreds of billions of dollars in Medicaid spending over the next decade by reducing improper payments in the program.”
Their support for this proposition? An estimate that Medicaid made “nearly $1.1 trillion in improper payments over the past decade.” Their support for this fantastical number? Footnote 2: “Authors’ calculations; Office of Management and Budget, ‘Payment Accuracy.’” They simply take the total amount of federal Medicaid spending over the 2015-2024 period, $4.3 trillion, and multiply it by 25 percent, their evidence-free guess of the “true” improper payment rate for each of those years. Presto: $1.1 trillion—which, by happy coincidence, turns out to be twice the estimated amounts of improper payments reported by CMS for those ten years ($543 billion). It’s an attention-grabbing number, and it pairs favorably with a minimum target of $880 billion in cuts. The author’s implication: since there was such a large amount of improper payments in Medicaid over the past ten years, that will be the case going forward. Of course it will.
This is not a paragon of analysis. Here are three problems:
First, Blase and Genszler ignore the steady decline in Medicaid improper payment rates over the most recent three reporting years. Of course, that consecutive decline is inconvenient for their analysis, but it is what the data show. Simply dismissing the data as “meaningless” because they were collected during the PHE continuous enrollment period is wholly unpersuasive. The data are that some improper payments over the past three years were in fact attributable to eligibility errors: beneficiary ineligible for program or service provided accounted for 6 % of improper payments in 2022, 12% in 2023, and 16% in 2024. If the authors believe that these eligibility errors for these years are understated, they should present evidence. They fail to do so.
Second, Blase and Genszler assume that all improper payments in their wildly inflated number should in fact not have been made. The truth is, that in 2024, as noted, nearly 75 percent of improper payments were attributed to insufficient documentation. They may have been properly made (or not); we simply don’t have the information to know. (The corresponding figures for 2021, 2022, and 2023 were 89%, 87%, and 80%, respectively). The authors make no adjustment whatsoever for insufficient documentation.
Finally, Blase and Genszler don’t provide any state-by-state estimates of the impact of their proposal to charge states the “full fare” for their “excessive improper payments.” This is a telling omission. As explained above, different states have different improper payment rates, and recovering “hundreds of billions” in federal matching funds for improper payments will have very different implications for different state Medicaid programs and their treasuries.
To return to our Florida example, the state’s improper payment rate in 2023 was 7.0%. Do Blase and Gensler believe that the Florida estimate is too low because the state’s eligibility errors are greatly understated? If so, what do they think Florida’s improper payment rate is, and how much of that is “excessive?” If they believe that Florida’s reported 7.0% rates is accurate, how much of the $1.5 billion in improper payments Florida is estimated to have made would Blase and Gensler consider “full fare” and how much would they cut Florida’s current FMAP of 57.2% to recover the federal government’s losses? On all these questions: crickets.
In short, what we have here is an effort to use official agency estimates of improper payments in Medicaid to justify a massive cost shift from the federal government to the states. This effort is not just flawed. It’s improper.