Latest / The Fiscal Policy Podcast with Fexingo: Government Budgets, Stimulus, and Public Spending / How Government Inefficiency Wastes 200 Billion a Year
Transcript
- Lucas: So there's this number that's been sitting in the Government Accountability Office's annual report for years, and it's so big that most people's eyes just glaze over when they hear it. Luna: Let me guess — improper payments. The federal government's annual overpayment tab. Lucas: Exactly. The latest estimate, for fiscal year 2025, is roughly $200 billion. That's money sent to the wrong person, the wrong amount, or to someone who wasn't eligible at all. Luna: And that's just what they can measure. The actual number is probably higher. Lucas: Right. The GAO can only audit what's trackable. But even the official figure — $200 billion — is about 4 percent of the entire federal discretionary budget. To put it in perspective, that's more than what the federal government spends on housing assistance, food stamps, and unemployment benefits combined. Luna: Okay, so where does most of it happen? I'm guessing healthcare. Lucas: You'd guess right. Medicare and Medicaid account for well over half of all improper payments — about $110 billion in 2025 alone. The Centers for Medicare & Medicaid Services, or CMS, runs these programs, and they process over a billion claims a year. Luna: With that volume, some errors are inevitable. But $110 billion? That's not rounding error territory. Lucas: No, it's systemic. And the GAO has been flagging this for over a decade. In 2014, the Improper Payments Elimination and Recovery Act was supposed to crack down. Agencies were required to identify high-risk programs, report error rates, and recover overpayments. But the error rate for Medicare, for instance, has hovered around 8 to 10 percent consistently. Luna: So what's driving the errors? Is it fraud, or just bad data? Lucas: It's a mix. A big chunk is what they call 'documentation errors' — meaning the provider didn't submit the right paperwork to justify a payment. That's about 60 percent of Medicare improper payments. Another chunk is 'medical necessity' — services that were provided but weren't medically necessary under program rules. And then there's straight-up fraud, though that's harder to quantify. Luna: And these are overpayments, mostly? Not underpayments? Lucas: Overwhelmingly overpayments. The GAO says over 90 percent of improper payments are money the government shouldn't have sent. So it's not like providers are being shortchanged — it's the taxpayer who's losing. Luna: What about recovery? Don't agencies try to claw back the money? Lucas: They do, but the recovery rate is abysmal. In fiscal 2025, CMS recovered about $3.5 billion in overpayments. That sounds like a lot, but it's about 3 percent of the estimated overpayments in those programs. And the process is slow — by the time they identify an overpayment, the provider might have closed up shop or moved money around. Luna: So the incentive structure is actually backward. If you're a healthcare provider, the risk of getting caught and having to pay back is low, and even if you do, it takes years. Lucas: Exactly. And meanwhile, the payment systems themselves are ancient. A lot of Medicare claims are still processed on a system called the Multi-Carrier System, which dates back to the 1970s. It wasn't designed to catch complex billing patterns in real time. Luna: That's just embarrassing. So what's the fix? I've heard about real-time payment verification. Lucas: That's actually a proposal from the Office of Management and Budget — shift from pay and chase to verify before pay. Basically, use modern data analytics to check a claim against eligibility, medical history, and billing patterns before the check is cut. Luna: Sounds expensive to implement, though. Lucas: It is up front. But the potential savings are enormous. Even a 1 percentage point reduction in the error rate for Medicare and Medicaid alone would save about $14 billion a year. And some pilot programs have shown that real-time verification can cut improper payments by 20 to 30 percent in specific categories. Luna: But I'm guessing there's political resistance. Providers don't want slower payments, and agencies don't want to invest in IT. Lucas: That's exactly it. The American Hospital Association has lobbied against stricter prepayment review, arguing it would delay legitimate payments and increase administrative burden. And agencies have a history of underinvesting in technology — the IRS is still running on systems from the 1960s. Luna: So we're stuck in a cycle: old systems lead to high error rates, which lead to low recovery, which leads to no political will to fund new systems, which keeps the old systems in place. Lucas: That's the vicious cycle. And it's not just healthcare. The Department of Agriculture has $10 billion in improper payments in nutrition programs. The Department of Labor has billions in unemployment insurance overpayments. It's across the board. Luna: It's funny — people worry about fraud in social safety net programs, and that exists, but the bigger problem is just bureaucratic inefficiency. Lucas: Yeah. And a lot of the fraud narrative is overblown relative to the scale of simple administrative error. The GAO report actually breaks down improper payments into three categories: 'unable to determine' — which is basically missing documentation — 'program integrity' failures, and 'fraud.' The fraud piece is the smallest slice. Luna: Listen, before we move on, I want to mention something small but real. We keep this show ad-free, and the only reason we can is that some listeners chip in a couple of dollars a month on Buy Me a Coffee. If you've gotten something out of walking through the federal budget with us, that's the link — buy me a coffee dot com slash fexingo. It genuinely makes a difference. Lucas: Yeah, it's a small thing that adds up. And we really appreciate it. Now, back to the systems — there's actually a surprising source of optimism. The Pandemic Oversight Committee forced some agencies to modernize payment systems during COVID, and the results were promising. The Department of Labor, for example, cut its improper payment rate for pandemic unemployment programs from 18 percent to 4 percent in two years by implementing real-time wage verification. Luna: So the technology works. It's just a matter of applying it across the rest of government. Lucas: Right. And there's a bill in Congress right now, the Payment Integrity Improvement Act, that would mandate real-time verification for all high-dollar programs. It has bipartisan co-sponsors, but it's been stuck in committee for over a year. Luna: What's the holdup? Lucas: Cost, mostly. The Congressional Budget Office estimates it would require about $5 billion in upfront IT investment across agencies. But the GAO counter-argues that the savings would cover that in less than a year. It's a classic pay now save later dilemma. Luna: And in a tight budget environment, paying now is hard. Lucas: Especially when the savings aren't scored as deficit reduction in the same budget window. The CBO uses a ten-year horizon, but a lot of the savings from reduced improper payments accrue in years five through ten, while the costs hit in years one through three. Luna: So it's a timing mismatch. And that kills a lot of good policy. Lucas: Exactly. But here's a specific example of what's possible: the Social Security Administration has a program called 'Death Master File' matching — they use it to stop payments to deceased beneficiaries. That program alone saves about $1 billion a year, and it cost maybe $50 million to set up. Luna: So there are easy wins. Why aren't they replicated? Lucas: Part of it is that each agency operates its own payment system, and there's no central authority to force standardization. The Office of Management and Budget tries to coordinate, but they don't have direct control over agency IT budgets. Luna: So we're back to fragmentation. It's the same problem as in healthcare — multiple systems, multiple data formats, no interoperability. Lucas: Exactly. And the longer we wait, the more the legacy systems degrade. Some of these systems are running on COBOL, and the programmers who know COBOL are retiring. There's a genuine skills gap. Luna: So the $200 billion question is: will Congress ever act? Lucas: It's not a sexy issue. Nobody runs for office on 'I'm going to reduce improper payment rates.' But the money is real. $200 billion a year could fund a lot of things — infrastructure, education, deficit reduction. It's just sitting there, waiting for someone to fix the plumbing. Luna: And the plumbing is old and leaky. Lucas: Leaky is the right word. But hopefully, with enough pressure from the GAO and from watchdog groups, maybe this is the year the Payment Integrity Improvement Act moves. It's worth watching.