Plain English with Derek Thompson - The Trump-Musk Doctrine: F-ck Around and Find Out

Episode Date: February 18, 2025

For the past month, chaos and confusion have gripped Washington and the federal government. Elon Musk and his Department of Government Efficiency, or DOGE, have served as an iron fist of the Trump administration—ransacking government agencies, lighting fires in various departments, and generally firing as many people as they can get away with. Much of this work is plainly illegal. Every 12 hours, it seems, another federal judge rules that the Trump administration has exceeded its executive authority. Efficiency is a worthy goal, and some of the programs that Musk and his team cut may turn out to be wasteful. Still, the way Musk has gone about his work—destroying life-saving programs at USAID, mistakenly offering buyouts to nuclear assembly engineers and essential doctors with Veterans Affairs, slashing funds for important studies and data collection programs across government—suggests that his bureaucratic blitzkrieg isn't just illegal; it's careless and harmful. The U.S. deserves a theory of government more sophisticated than "F-ck around and find out." So, what would an effective DOGE look like? Today’s guests are Michael Geruso, an associate professor of economics at the University of Texas at Austin, and Tim Layton, a professor of health care policy at UVA. We explain why any sensible waste and fraud search-and-destroy effort should start with health care spending. Then we get very nerdy about waste and fraud in health care. Most importantly, we talk about trade-offs. It’s a myth that there is some pot of $10 billion just lying around, doing nothing, gathering dust. Every dollar of federal government spending goes to a person in a place doing a thing. And that means that every dollar we cut will have a recipient on the other end who is losing a dollar. Taking government efficiency seriously requires thinking about both sides of this equation: What do we get when we spend this dollar, and what do we lose when we take that dollar away? If you have questions, observations, or ideas for future episodes, email us at [e-mail]. Host: Derek Thompson Guests: Michael Geruso and Tim Layton Producer: Devon Baroldi Learn more about your ad choices. Visit podcastchoices.com/adchoices

What's up everybody? It's Austin Rivers here and we are back for another season of OffGuard. Me and my guy, Pasha Giggy, are hitting your podcast feeds every Monday and Thursday talking everything hoops. Austin is bringing that 11-year NBA veteran perspective and of course keeping you guys entertained throughout the season. Make sure you tap into OffGard with Austin Rivers on Spotify or wherever you get your podcasts. And don't forget to follow everything we've got going on social media. The OffGar podcast, Ringer NBA, and of course check us out on Ringer NBA's YouTube channel. We're getting better. By reputation, Elon Musk and Donald Trump are builders.

Musk has grown two of the largest hardware innovation companies in the world with Tesla and SpaceX. And as for Trump, he once told Gulf Digest magazine, quote, I own buildings. I'm a builder. I know how to build. Nobody can build like I can build. Nobody. And quote. But now these two are united in Washington, and the duumvirate of Trump and Musk, has made their mark in the first month of this administration, not by building, but rather by its opposite, which is demolition.

With the creation of the Department of Government Efficiency, aka Doge, Musk has claimed for himself an extraordinary amount of power, serving as the iron fist of the White House, rooting out what he sees as the plague of wokeism in government, halting grants, freezing payments, lighting fires in various departments, and generally firing as many people as he can get away with. For the past month, chaos and confusion have gripped Washington, D.C., and the federal bureaucracy. A slew of federal judges have already ruled that Trump and Musk have violated the law, typically by exceeding the powers of the executive branch

and attempting to defund agencies that were initially funded by Congress, the legislative branch. But what I want to focus on today is not Musk's methods or his motivations, but rather to judge him by his outcomes. Doge exists in theory

to seek efficiency, and the need for efficiency today is understandable. The federal government is deep in debt. Our interest payments now exceed what we see, spend on defense. And even if the U.S. had no issue with its debt, it would still be a mitzvah to find ways to make government work better, to take the same tax dollar further, to do one more unit of good. But like some out-of-control chemotherapy that attempts to kill a cancer and instead ravages the healthy cells, the first few weeks of Doge have showed us an out-of-control

organization wrecking blind havoc across government. Their effort to trim the fat, keeps cutting deep into bone over and over again in a way that I worry will eventually do serious damage, even if it's hard to specifically predict what the worst damage will look like. When you consider very closely what Musk and Doge have attempted to do in the last few weeks, I think the only objective conclusion one can reach at this point is that what's happening in government is not just a reign of terror. It's a reign of ineptitude. Let's start at the Department of Energy, where Doge recently laid off more than a thousand workers. Among those laid off were 300 staff at the National Nuclear Security Administration. We're talking scientists, engineers, and safety

officials responsible for safeguarding nuclear warheads. Roughly 100 people were reportedly laid off from the Pantex plant in Texas, the most important nuclear assembly and disassembly plant in the country before they were called back into the office. Daryl Kimball, the executive director of the nonpartisan Arms Control Association, said, quote, the Doge people are coming in with absolutely no knowledge of what these departments are responsible for, end quote. Next, we have Veterans Affairs, where the Trump administration offered buyouts to tens of thousands of employees before realizing that once again they'd made a mistake. Now, maybe your impression of the typical federal employee is somebody who pushes paper around all day,

but if you're familiar with the VA, you know the agency provides health and psychiatric care to millions of U.S. war vets, which means if you offer buyouts to veterans' affairs, what you're going to get is a lot of underpaid doctors, nurses, and psychologists saying, okay, see ya, and leaving offices that are already understaffed. Which is exactly what was about to happen. Until days after the buyout offer went out, thousands of doctors, nurses, and psychologists and other essential staff got a second notice. Oops, sorry, no, your buyout offer has been rescinded. At the Department of Education, which the Trump administration seems very intent on destroying,

Doge recently terminated $1 billion in contracts. but rather than end these ideological programs that Musk says he wants to eliminate, these cuts decimated the Institute for Education Science, which funds many of the most famous and long-standing studies in all of education research, including several longitudinal studies on student achievement and school effectiveness. In the big picture, it's hard to think of a better non-partisan role for government than data collection, or sort of uniquely positioned to do it. But Trump and his team have gutted now some of the best education data tools we have.

It's not just progressives who are aghast at this. Nat Malchus, a senior fellow at the American Enterprise Institute, a conservative think tank, told the Washington Post, quote, there's a lot of bloat in IES. There's a lot of problems to be solved. But these are problems you solve with a scalpel and maybe a hatchet, not a bulldozer. Doge's cuts will go much further. At the FDA, the Trump administration has fired hundreds of employees with the deepest cuts

among those involved in testing food and medical devices. At the CDC, more cuts are expected to decimate the Epidemic Intelligence Service, which pays disease detectives around the world to find and stop epidemics in other countries before they spread. At the NIH, the world's crown jewel of biotech funding, the administration is set to slash personnel and funding in a variety of ways. Now, if you're a fan of Elon Musk and Donald Trump and you're listening to the show, your hope will be that these cuts are all fat and no bone.

But I beg you to remember, this is the same organization, Doge, that in an attempt to refocus the Energy Department on Nuclear Security, initially gutted the division with the words nuclear security in the title. So far, few areas have received. more attention in this government-wide assault than USAID, the Department of Global Aid, which the Trump administration has effectively destroyed. Elon Musk seems to be on a gleeful and personal mission to dismantle USAID, placing most of its employees on leave, closing its headquarters, and moving what's left of it to the State Department.

According to one report, the administration says it plans to reduce USAID staffers from around 10,000 to 600. As Musk himself recently tweeted, quote, we spent the weekend feeding USAID into the woodchipper. There is irony here, and there is tragedy. The irony is that once upon a time,

Marco Rubio was one of the most outspoken defenders of global aid. In February 2017, he said, quote, it's critical to our national security. We don't have to give foreign aid. We do so because it furthers our national interest. In 2019, he said, quote, anybody who tells you that we can slash foreign aid and that will bring us to balance is lying to you, end quote.

Today, however, Marco Rubio is in the awkward position of being the Secretary of State overseeing the dismantling of foreign aid. The human costs are the real tragedy here. Unless we course correct again and immediately replenish our global health grants, there's no getting around the fact that a lot of poor people around the world are just going to suffer and die in order to save the typical American taxpayer about $5 a year. The U.S. pays for insecticide sprays in Uganda, for pregnancy services in Zambia, for health clinics in the poorest parts of the world.

Most notably, the president's emergency plan for AIDS relief, aka PEPFAR, has already saved an estimated 25 million lives and prevented more than 5 million babies from being born with HIV. It's not yet clear if PEPFAR will be spared, like the VA doctors, for example, or left to wither away like so much the Department of Education. This Blitzkrieg is astonished even the most famous critics of U.S. aid programs. William Easterly, an economist, has written that American aid does often prop up dictators and go to waste. But in an interview with New Yorker, he called Trump's U.S. demolition plan, quote, horrific.

It's illegal and undemocratic, he said. Elon Musk has hinted that Doge will simply reverse any measures that go too far amid rising criticism. This sounds good in theory, right? Move fast. Cut the stuff that looks bad. Add back the stuff if you miss it. But in practice, you cannot just cut 10,000 programs at once and reinstall them on a one-by-one

basis depending on whether the volume of criticism passes some imaginary threshold. Whatever you think of the failures of Biden and the failures of progressive governance, which are legion, fuck around and find out is not a suitable replacement. But fuck around and find out does appear to be our current methodology of government. Once again, an actual department of government efficiency, one that operated within the bounds of the law, one that sought to carefully understand what government does before seeking to identify waste and fraud, would be a wonderful thing to have. So what would it look like? Where would we start? What would Derek's doge begin to look for in savings? I'd start like this. In the biggest picture, the federal government has three dominant jobs.

Healthcare, Social Security, and Defense. Those three areas account for between 60 and 75% of total federal spending every year. And according to the U.S. Government Accountability Office, it's health care spending in particular that is growing the fastest and where the size of waste and fraud is likely the greatest. In particular, the GAO has scrutinized excess payments made by the government to private insurance companies under the Medicare Advantage System. So I read that GAO report, and I thought to myself, I don't know much about the Medicare Advantage system or the reasons behind excessive payments. So I want to talk to experts who do. Today's guests are Michael Jerusalem, an associate professor of economics at the University of Texas Austin,

and Tim Layton, a professor of health care policy at UVA. We talk about why it makes sense to look for savings in health care first, where excess payments come from, how so-called upcoding costs the U.S. up to $100 billion a year, why it happens, and how to fix it. But most importantly, we talk about trade-offs. It is a myth that there exists some pot of $10 billion just lying around, doing nothing, gathering dust, allocated for some dead person. Every dollar of federal government spending goes to a living person in a real place doing a thing. no bushels of cash are being shipped to Mars. And that means that every dollar we cut will have a recipient on the other end who is losing a dollar.

To take government efficiency seriously requires thinking about both sides of this equation. What do we get when we spend this dollar versus what do we lose when we take that dollar away? I'm Derek Thompson. This is plain English. Michael Jeruso, welcome with the show. Thank you so much. Happy to be here. Tim Layton, welcome to you as well. Thanks, excited to chat.

Mike, so we've got Doge out there scouring the government for savings, and I wanted to bring on some economists to talk about where a search and destroy effort for wasteful spending might actually find billions of dollars in waste. So first question, Why am I even talking to you? Why are we starting with health spending? Well, if you're looking to cut government spending, the three biggest buckets that there are,

are health care, social security, and defense. Together, those three are about 60% of the federal budget. And of those three, health care is the biggest part. About one out of every four dollars the federal government spends is on health care, either in Medicare or Medicaid, those are the two biggest components. Medicare, the program for the elderly, Medicaid, the program for lower income Americans, and the rest of what the government spends has to fit into what's left over. Medicare alone is about $900 billion a year, so almost a trillion dollars in Medicare.

And so in a program that big, even finding 1% of fat to cut is $10 billion. Most of federal agencies have budgets far below $10 billion. And so even a small slice of that very big base, you could imagine finding some real savings in. Right. Like cutting NIH by $10 billion is a 20% cut. One out of every $5 that universities are getting for medical research suddenly gone. But in Medicare, you're talking not about 20%, but 1%. So that's why it makes sense to focus in these bigger buckets. Tim, we're talking about government acting essentially as the nation's largest insurance company, Medicare, Medicaid, Veterans Affairs. And if we're going to understand how to hunt for waste and even fraud in government health spending,

I think we should understand a little bit about how government spending on health care actually works. So, Tim, let's say I go to the doctor with chronic leg pain. The doctor could prescribe me Tylenol, or he could send me off to get, you know, five blood scans, an MRI scan, a CT scan, and a prescription for new nerve pain therapy that costs $10,000 a month, right? Washington doesn't know what treatment plan is right. Is it the Tylenol that's like two cents or is it this other plan that's like 20,000 bucks a month? So how does the government acting on behalf of the taxpayer approaching this very challenging question of how to reimburse good care with an eye toward controlling costs? Yeah, I think your example is perfect here. I think first you need to understand that the government

has essentially decided that if people want to, they should be able to choose to enroll with a private insurer in most of the public health insurance programs. And today, over 50% of Medicare beneficiaries are in a private plan. That number is over 70% in Medicaid. So loads of people prefer private plans. And we want to support that decision, but we also want that private option to be budget neutral, to not cost any more than if folks had chosen the traditional public plan. And so how do we set payments to these plans to support choice and budget neutrality?

Well, we wouldn't really want the government to simply reimburse every expense that the insurers incur, plus some fee for administering the plan. That would mean that no party in that case, not the doctors, not the insurer, have any incentive to control costs. And that would lead to the second situation that you described, right? In your example, where the doctor prescribes a ton of unnecessary stuff, five blood tests and MRI scan, CT scan, et cetera. And in that type of case, like, costs can spiral out of control in this private option.

As enrollees, they want stuff, right? And the best way for insurers to get enrollees is to just give them more stuff. And insurers will just keep on giving and giving and giving because they don't have to foot the bill, because the government's just going to reimburse them. Right. And so this is essentially the incentive structure that doctors face in the traditional fee-for-service part of Medicare. And one of the reasons why we started this private option in the first place. And so instead, what you want to do is just pay the insurer a fixed amount for each month,

for each person that they enroll. We often call this capitation. And in that world, if an insurer reigns in costs by negotiating lower prices with doctors and hospitals, by directing patients to more efficient care, or even by straight up rationing and saying no for stuff, then they can profit off of part of the cost savings they generate. And hopefully they take a portion of those savings and pass them through. to the patients in the form of additional benefits like vision and dental and things they wouldn't get

in Medicare. There's some evidence that they do that. But now, more importantly, there is an incentive to restrain spending, right? To prevent costs from spiraling out of control and to maintain that budget neutrality, make it so this private option doesn't end up costing a ton more.

But a major problem with this kind of fixed payment per person setup is that plans will vastly prefer to enroll some patients versus others in that case. Say an average patient costs $5,000 a year. And so we set the payment to plans around that level,

about $5,000 per year. But say a patient with cancer costs $20,000 per year. And say, ultra-endurance bike packers, like my in-laws costs like $1,000 a year. They're super healthy. They don't need a lot of health care. And so insurers are going to do everything they can in that case to avoid the cancer patient

and to attract those cyclists, right? So basically, you'd expect private plans that always exclude oncologists and cancer clinics from their networks that exclude or put really high copays on blockbuster cancer drugs that generally makes sick patients' lives miserable with kind of layer of layer of prior authorization rules, you know, super narrow provider networks, constant denials. And their hope is that if they do this, these folks will offer a different plan or more likely for the public system instead. And you'd expect insurers to offer my in-laws subsidies for purchasing a new bike, maybe provide really good coverage for physical therapy, free gym memberships,

things like that. Or maybe they advertise in places where people are healthier, or it's widely known that like Hispanic Medicare beneficiaries are much cheaper than others. So maybe they advertise aggressively on your favorite bad bunny radio station or something like that, right? And so this is a system that we don't exactly want. We don't want a system where like the plans are trying to avoid the sick patients and attract the healthy patients. And equally importantly, if the government is paying a private insurer $5,000 per year for someone who would cost $1,000 in the public program, like my in-laws, government costs are going to spike, right? And remember, we want this private option to be budget neutral, but now it's definitely not, right? That private option is enrolling all the

cyclists, none of the cancer patients, and the cyclist costs the government $1,000 in the traditional public program, but $5,000 in the private plan. And this can lead to this type of runaway government spending, right? It seems to me like you're outlining two ways that health care spending can go awry. In a fee for service, FFS system, doctors or insurers are biased to deliver more and more and more services, even when people. patients don't need it, and that drives up costs across the system. But in the alternative scenario of capitation, now you're encouraging insurers to turn away or find ways to deny care for the sickest patients. And that means, you know, not including oncologists in your network. So there's these two

rocks that we're trying to navigate between. Mike, I want to keep the explanatory pace here slow and steady, so everyone's on the same page. But what Tim has outlined is that the government has a very serious interest in properly reimbursing doctors who care for sick patients. We want to align government payments with actual patient illness. That's the gold standard here, right? But that means that the government needs a way of actually determining, actually seeing which patients are actually very sick and need more treatment. So how are doctors telling the government how sick their patients are. Let's use, say, like, a diabetes patient as an example. Yeah, that is the hard part. The hard part is for the government to know how sick somebody is.

The doctor may know it. The patient may know their own health conditions, but for the government to know it, they need to see it in their records. And so, like Tim outlined, there's two parts of Medicare. People can choose the public option or people can choose to enroll in a private plan like United Health Care, Humana, a plan that wouldn't look very different from yours or line. In the public program, the government pays the bills. So it knows that, on average, say, a person who's coded with diabetes costs $2,000 more each year than a person who doesn't have diabetes in their records. So from that data, from the public system, the regulator can say, okay, Medicare Advantage plans. If you take on a patient who has diabetes, we'll give you an

extra $2,000 per year to take care of that patient. We'll reimburse you what we know that it costs in the public system to take care of that patient. And we'll do that for 100 different condition categories that correspond to these chronic conditions that we know persist from year to year, and we know are expensive to treat. And their goal in doing that is that they calibrate it just right so that no insurer anywhere has any incentive to avoid sick patients. If they bring in a healthy patient to their plan, they're going to make the same profit as if they bring it. bring in a sick patient to their plan. So that's the goal, but a lot hinges on that calibration.

And the way that calibration works is in these diagnoses codes that are submitted to insurers and submitted to the government on the claims that doctors send out in order to get paid. And just because that's the first time we're mentioning diagnosis codes, just define exactly what you're talking about. And you can even use a really quick example. Sure. So a diagnosis code is going to be, you know, an alpha-numeric code. it's going to say this patient has this condition. And so there might be several different codes that

correspond to a diagnosis of diabetes. Maybe the diabetes has some eye complications. Maybe it doesn't. And then what the regulator will do is sort of aggregate the information from those codes that are going to be entered on this bill that the provider sends to the government or sends to the plan that's paying him. Your doctor sends this bill out. And that information is aggregated. And is aggregated up to sort of higher levels of chronic conditions. So there might be a hundred different codes that correspond to some version of congestive heart disease, but those will all get lumped together in heart disease. So I think that if someone's listening along, they're thinking, this sounds like it makes all the sense in the world. The fee for service system is going to be biased toward

more fees for more services. We don't want that. That's totally runaway spending. The capitation system that Tim just outlined is a flat fee per patient, and that means that the sickest patients are going to be denied care that they absolutely need because they're treated or paid for the same as if they're just an average patient. So that system doesn't work. What we want is a way to make sick patients legible to government. It sounds like this is an absolutely perfect system, like it would make all this sense in the world. Now, this is what I want to do now. It's going to be a little bit nerdy and hopefully not a total failure, but I really do want an unimpeachable understanding

of how diagnostic coding works so we can understand how it goes wrong. And since there's three of us here, and we're talking about what is sort of a three-party coordination, there's a patient, there's a doctor, there's insurer, I thought we could roleplay this. So Tim, you're going to be my doctor, Mike, your Medicare, and I'm Derek.

I'm Derek, a pre-diabetic, 80-year-old, with very mild memory loss. And I've just moved to the area, Tim. So this is my very first visit to the doctor's office. It's very nice, by the way, your Sarasota office is lovely. You should know, by the way, there's a Sports Illustrator from 2007 in the waiting room

and you might need to replenish your magazine selection, but we can talk about that later. Tim, you call me into the doctor's office. What happens now? All right. So, Derek, it looks like you're doing pretty well for your age. Your blood tests. suggest you're right on the edge of having diabetes, right? And you appear to have some mild memory

loss, which could be a sign of dementia down the road. So we should keep an eye on that, but overall, not bad, right? I'll prescribe some meds to help kind of tame that diabetes and definitely come back next year so we can see how things are looking. Let's be sure to set up that appointment right now. Okay. Now, after the appointment, okay, when Derek's not with me, I'll record my notes that Derek has some evidence of diabetes and memory loss that could maybe be evidence of dementia. And then I'll send a bill to the insurer for an office visit, okay? Just a regular annual physical. I'll call in prescriptions to the pharmacy that I want Derek to be taking. And maybe I don't put the diabetes and dementia codes on the bill because they weren't full-blown yet. And this is just an annual

physical. So I don't need those diagnoses to justify this visit to justify getting paid because I'm not treating him for those things necessarily. I'm just checking him out. Okay. So I take that bill and I pass it on to Mike. Okay, so this is super important. The first thing to understand that if Tim is submitting this claim to the federal government, that's it. Nothing else happens. Everything else that I'm going to say, it would be only occurring if Tim is submitting this claim to a private insurer. So if patient Derek is insured with United Health Care's Medicare Advantage Plan or CVS's Medicare Advantage Plan, a whole different set of things is going to happen now. I'm the insurer.

I received this claim, and we take a look at it and we think, huh, there aren't any diagnosis codes here, despite the fact that last year, Derek was coded with prediabetes, and it looks like Tim's been prescribing medications for heart disease. luckily, it looks like we have an arrangement with Tim's practice. We have an arrangement with them where we can collect all of their notes and electronic health records. So we'll get those. We'll have them send them to us.

And we'll look over those. We'll send them to our coding teams and we see what they can find. All right, some time passes. We've had our teams look over the notes. Maybe AI has crawled through it. And that has turned up the word diabetes. We see that the patient is on a GLP1, maybe. the diagnosis for diabetes is substantiated enough in the notes that we can go ahead and code that

as diabetes and submit that to Medicare. And actually, now that we're looking at this patient record, we see that the patient was exhibiting some signs of memory loss. Maybe there's not enough information there to yet code dementia. But here's what we can do. We can call up the patient, and we can ask if the patient's willing to have a nurse come to you, Derek, if you're willing to have a nurse come to your home. And maybe that sounds like a hassle for you, so maybe what we'll do is we'll, we'll,

we'll, we really want to get this health risk assessment done. We'll send you a gift card for 50 bucks to, to incentivize you. So we'll send a nurse out to your home. He or she can do a test. Maybe they'll find enough information to document a diagnosis of dementia. And so now we'll get those codes onto the payment system, and we'll be reimbursed for that. We'll have that logged in the system with the Medicare administration.

And actually, as we're looking at these records, we see that, Derek, you had a diagnosis of skin cancer five years ago. Yes, yes, I know the skin cancer is in remission. No one is treating you for skin cancer right now. You're not taking any medications. You're not seeing any cancer doctors. But we think according to the rules, we can still keep the skin cancer diagnosis on your record. We mean, our goal is the insurer is to make things as complete and accurate as possible. So we're going to get that diagnosis on your record as well and get reimbursed.

for that. So Tim, Mike has just laid out a scenario where the insurance company has coded me as full-blown type-to-diabetic, full-blown dementia, and also as a person suffering from skin cancer. This is a scenario where it's the insurer with their foot on the accelerator for adding diagnostic codes. You're the doctor. Are there also some cases where it's the doctor who's so-called upcoding him or herself. Absolutely. The way that Mike described this, you know, under a lot of settings,

under like that fee-for-service type of system, if I'm just, you know, submitting these claims to Medicare, I have no reason to submit these diagnoses for any of these conditions because I get paid either way. There's no incentive for me to do this. And so I'm not going to do that. But under alternative arrangements,

sometimes the Medicare advantage, plans will set up contracts with physicians like me, right? Not really a physician, but a pretend physician like me, where I'm incentivized to diagnose these directly. And the ways that that can occur is like it could be just subtle pressure where every year when I'm renegotiating my contract with the insurer, they may say, we notice that, you know, there are a lot of cases where you weren't marking down diagnoses that we think you should be marking down. You know, if you want to stay in our network, we'd really appreciate it if, you know, you started to do that. Alternatively, they can make the incentives much more explicit, right? Like the insurers are being paid more for enrolling sicker

patients, the insurers can pay doctors more for treating sicker patients, right? With that payment based on the diagnoses. And there's nothing really wrong with that. Doctors maybe should get paid more for treating more complicated, sicker patients. But now you see how there can be a direct financial incentive for the doctor, him or herself, to record those diagnoses on the claims. And then the insurer doesn't have to go through all of this process, right? And the doctor actually wants this coding to take place. What I find so interesting about this and stories like this is I am such a sucker for cases where seemingly good intentions can go awry. Because we've established it seems to make so much sense that we would want to find a way to make

patient sickness legible to the government. And the way we discovered to do that or the way we arrived at is diagnostic codes. But Tim and Michael, as you've explained, there's all sorts of reasons for both the insurer and the doctor to pretend as if their patients are sicker than they actually are because the more severe the diagnostic code, the more money every. everybody in the system gets. So, Michael, this is a practice that is known in health care as upcoding. And both of you published a paper several years ago showing evidence that upcoding is not just a hypothetical in Medicare Advantage, but a likely fact in Medicare Advantage.

Before we talk about like the history and how to fix this, how do we know you're right? What is the best evidence that this is happening in a wide scale? Yeah, so we can talk about our paper. I'll also say that there's just many people that have researched this. The Health and Human Services Office of Inspector General, there's independent Medicare Payment Advisory Commission. Everyone sees this happening, and they get out of different ways. So, for example, this story we told about the insurer scraping through medical records,

maybe with humans, maybe with AI, that's something that the Office of Inspector General HHS has investigated and found that that's a significant portion of where the payments and codes come from. Is this scraping through medical records? Maybe the doctor doesn't even know that it's happening. Maybe the information about diagnoses never even get back to the patient. So it can be pretty disconnected from any sort of treatment plan or even knowledge of the doctor or the patient. But in our paper, what we did was we took a couple of approaches, but the easiest one, the most transparent one to explain was we looked at data from people in Massachusetts where there's tremendous database that allows you to see all the health care claims of everybody in Massachusetts.

And we would see the health care claims of these people when they're 62, 63, 64. And from those health care claims, we could construct their risk score. We knew whether they had diabetes in the claims. We knew whether they had cancer in the claims. We knew whether they were coded for dementia or depressive disorder. And then we followed two sets of people. We followed the people who turned 65 and joined a private Medicare Advantage plan, and we followed the people who turned 65 and joined the public fee-for-service system.

And as you follow the people that joined the public fee-for-service system, their conditions, their risk scores, whether they're coded with any chronic illness at all, basically is flat from 64 to when they turn 65 and 66 and so on. for the people that join MA, the moment that they become eligible for Medicare and join MA, their chronic conditions that are coded in their records immediately jump up. And by two, two and a half years after they've been in Medicare Advantage, their risk scores are 10% higher than the trend. And so people join traditional Medicare, no change in the risk score.

People charge Medicare advantage, huge bump up in their risk score. And, you know, one possibility you could say here is, oh, well, these people join traditional Medicare, people are just, the people who are joining Medicare Advantage are just getting sicker somehow. There's something systematically going on. They're getting, they truly are sicker. But there's just no way that's happening. A 10% increase in the risk score is what you'd get if within a population,

one out of 10 people became paraplegic. A 10% increase in the risk score is what would happen if 60% of the population became morbidly obese. There's just no way that the whole population is changing their health status that quickly. So this is really clear evidence that Medicare advantage enrollees are being deliberately coded as being significantly more sick than equivalent patients are in the public Medicare system. And this is costing Medicare, therefore the federal government billions of dollars a year. According to the GAO, FBI, upcoding and related misrepresentation of clinical information costs somewhere between $10 and $100 billion to the U.S. federal.

government annually. This is why we're looking here for waste or fraud. Very quickly, Michael, before I go back to Tim on the history of this practice, what's the difference between waste and fraud? And what's the right way to think about this upcoding phenomenon we're talking about today as being waste, improper payment, or fraud, someone's actually lying here? Yeah, that's a great question. And I think Tim and I, I maybe think about this a bit differently, at least I do, than the regulator. So there's just a complete continuum here between, on the one hand, getting a diagnosis coded that the patient clearly has, no one would doubt, and on the other hand, inventing something from whole cloth. And probably the

most important thing to understand is that no matter how the additional codes get on there, the taxpayer is on the hook. And so we're paying more. regardless of what is happening is legally classified as fraud or is legally permitted by the system. And the difference between those two things is just a difference of what the regulator chooses. If today the regulator says, you can do these chart reviews, you can send the patient's electronic health records to your coders at United Health Care to your AI systems. You can scrape the records. You can find the diagnoses. And we'll pay you for those. If tomorrow, they said, you can't do that. And the insurer, still did it, then that would be fraud. So what's fraud and what's legal is sort of the wrong question

to ask here. The waste is happening either way. Tim, the term that we've used today is upcoding. And contained in that term suggests the idea that there is an optimal code that we should use for people and that Medicare uses the right code and Medicare Advantage up codes, therefore uses the wrong code. Is it possible actually that Medicare is undercoding? or that historically we've been undercoding, and that in a way, what we're talking about as waste is actually a lot of doctors and insurance companies recognizing the true illness of patients

rather than exaggerating their illness. That's exactly right. I think it's important to recognize that these diagnosis codes from claims were never intended to be used in this way, right? Like the way, the reason, we have diagnosis codes is for doctors to essentially justify what they're billing, right? They weren't initially for this type of risk adjustment to plans, right? And so they're set up in such a way that a

lot of times doctors face very limited incentive to mark these diagnosis down. Sometimes they have strong incentives. If they want to get paid, they need to have them down. But in a lot of other cases, the incentive is quite weak. And so in those cases, you would expect the diagnoses on claims to be incomplete. And to be clear, like, for a long time, that was never really a concern because nobody really used these diagnoses from claims to determine how sick someone was. They were just billing kind of things that were there as part of the billing system, but they weren't what doctors used to figure out, like, how sick this person was. But now that we're using them in this way, the completeness matters, but not so much for the patient's health. Instead, it matters for

the size of the payments. So you're absolutely right that probably the bulk of what is going on here is undercoding in the kind of traditional fee for service system and a lot of what the Medicare Advantage plans are doing is being accurate and complete. It's just that that costs a lot of money because the government is paying them as though they're coding in the same way as the undercoded kind of public fee for service system. And I think an important piece of evidence about this is that, you know, estimates of how much this overpayment amounts to, range from like 50 billion to like, you know, many, you know, tens of billions of dollars.

Whereas if you look at all of like the DOJ's cases against these large insurers, the biggest one, I think, is against United. And that's for like two billion dollars total across several years, whereas like, you know, we're estimating that this, this costs the government on the order of like tens of billions of dollars every year. year, right? So the fraud piece that is being alleged is kind of small potatoes relative to the waste piece, you know, where it's legally justifiable in a lot of ways because CMS has decided to be very permissive, but is much more money than it would cost to ensure all these people in the

public program. Mike, I'm thinking about this as like blame pie. In the little play that we did, where I was the pre-diabetic 80-year-old patient, it seemed like the insurance company got the biggest slice of the blame pie. In some cases, maybe it's the doctors who have the bigger slice of the blame pie. They're the ones doing the upcoding. But I was reading some journalistic reports

of diagnostic codes. And I saw that many doctors say that they are advertised software programs that promise to make them more money, and those software programs essentially do the upcoding of the patient, themselves, right? The software developers are like, we promise that we'll make you a physician more money. How did we make you more money? We essentially take every pre-diabetic and we code them as

type 2 diabetic and voila, you end up getting a larger payment to your clinic. How big of the blame pie here should go to the software that these companies are using? Yeah, there might have been a time not that long ago when it was the insurers that were putting the software onto the doctor's tablets to make sure that they were coding everything that could be coded. But now because the insurers are just making these contracts with the physician practices so that the physicians get a bounty every time they add another code, the doctors themselves have the incentive to find these codes, to buy software products, to change their practices so that they log every eligible code. I don't usually ascribe moral blame to software. So I don't think that the

the software itself is the blame. But certainly... This is the problem with you economists. I keep wanting to cut up the blame pie with economists when I talk about health care. And they always say that we don't believe in moral systems. Well, I believe in moral systems. So pretend you're me and give me an approximation for the blame-sized pie percentage. It's going to vary a lot by practices, but in a practice where the doctors are just contracting

with the insurer and contracting on this risk basis, that could be a big fraction of pie, more than half in systems where the doctor and the insurer are the same entity like Kaiser in California, then there's actually no difference between the insurer's software and the doctor's software. But I think that the biggest chunk overall over the whole healthcare system in Medicare Advantage is coming from the insurer's chart review, not what's happening at the doctor's office. That might change over time as these contracts change, as the software becomes more nimble. but for now, or at least looking backwards, it's mostly been at the level of the insurer.

Tim, you and Michael first started looking at upcoding about 10 years ago. How much more prevalent has this practice got in the last 10 years? Yeah, that's a good question. I think a lot of people have been really interested in this question. So I actually have some new work that's under review right now with another economist Paul Jacobs, where we try to investigate this question. And so basically what we do is we follow people in Medicare who switch from the public program

to the private program. And it turns out that people who switched back in like 2013, 2014, after four years in, compared to people who didn't switch, their scores go up by like 15, 16 percent. But people who switched in 2017, 2018, which is kind of how far our data lets us go, by the time they were four years in, their scores had increased by about 23 to 24%. So I think it's getting worse and like this problem is growing. It's not clear that it will grow forever because there are also things happening in the public program that might improve coding there and such. But I think it has gotten worse since we studied this and continues to get worse from one year to the next.

Other countries have sick people. Other countries have public insurance systems. They deal with all of the problems that we've talked about today. How do we avoid incentives to overtreat? How do we avoid incentives to undertreat? How do we make the sickness of the patients that exist in each country's population legible to the government insurer? Is there something that the U.S. can learn from another country that would help us align? payments in this one? Yeah, that's a good question. I think, you know, to take a step back, there are a variety of other health insurance systems in other countries. Some of these countries, like the Netherlands, Germany, Israel use systems that look a lot like the Medicare Advantage

Program, right? Like people choose between competing private insurers, and they all have aggressive risk adjustment systems that look in a lot of ways a lot like ours. The biggest difference between us and all of those countries is that they don't have a public option, right? So they only have these private plans competing with each other. You'll note that like these overpayments come in because they pay the, we pay these private plans according to what we think they would cost in the public system. If there were no public system, then you wouldn't have this overpayment problem. You just decide how much you're willing to pay plans and you distribute the money amongst

the insurers according to how. sick each of their patient mixes are, but you don't have to worry about this kind of overpayment system. Now, there are other problems with having only private plans, but because of that, we're kind of unique in the world in operating these parallel public private systems that people have choice over, and that's kind of where these problems come from, because that is a more complicated market to regulate and to design. Mike, one thing that I really like about Tim's reply there is that you guys are economists. You study tradeoffs. And one thing that frustrates me in the conversation about government spending is when people deny the existence of tradeoffs.

Government spending tends to go somewhere. Someone is benefiting. Sometimes that money seems like it's inefficiently spent, but then you take that spending item away and you realize the good that it was doing, the people it was employing, the communities that it was propping up. Health care is no different. if you cut 10% of American health care spending, it's not as if that's like fat that's zapped from an animal's body and the animal keeps going around being like really, really ripped, right? It's not like, you know, adipose tissue that can just be burned away. That's money that's going to support jobs and companies and people and patients, right?

Economics is tradeoffs. Healthcare is tradeoffs. Government spending is tradeoffs. That's a theme that I'm picking up from what Tim's putting down. But let's talk about solutions here. Like, obviously the politics of health care reform are horrendous. People tend to dislike health care in the aggregate, but you get a lot of screaming when individual plans are changed.

And so there's this interplay between lots of fury at the system, but also a status quo bias that makes it difficult to change the system. Where do you think change can or should start here? when we're thinking about the phenomenon of upcoding and Medicare Advantage. Yeah, there are a few things that we could contemplate doing. Let me first say it's exactly right that there just is going to be somebody who loses something if we pare back $50 billion from Medicare Advantage plans. Some of that's the money that's going to the insurers,

some of the money that's passing through to doctors, hospitals, and so forth. And part of the money is passing through to the seniors themselves. if we're generously reimbursing these plans, and then we dial back that spigot and cut off some of that money, some of that is going to be borne by the seniors in these plans. Maybe there aren't as many plans as available. Maybe their co-pays go up by a little bit.

So you can see why this is politically difficult. But if you wanted to reduce these payments to plans, to try to level the playing field, so that you, me, Tim, all the taxpayers, pay the same, whether someone chooses traditional fee-for-service system or chooses Medicare advantage, there's really not a technical problem in figuring out how to do that. We know how to do that. Already, CMS deflates the payments to MA plans by a small amount because they know that these plans code more intensely. But that deflation has ever been more than 6%. And, you know, our research,

which is 10 years ago now, says that plans were upcoding at that point by at least 10%. And more recent estimates put that number at like 15 to 20 percent total overpayments to MA plans. So that's from the independent, nonpartisan Medicare Payment Advisory Commission. And so if you think that the, you know, someone might think that the overpayments are 15 percent, someone might think that they're 10 percent. Someone might think that they're 20 percent, but they're not 6 percent. And so the first thing you could do is just double the coding and deflation factor that

we deflate these plan payments by. That's not a perfect solution because there are some plans, the big plans, do a much better job, much more effective at pumping up the risk scores, and that'll hurt small plans more. But if I had to choose between doing nothing and just rationing down the payments of these plans, because everybody in the industry, all the experts, the independent watchdogs, we all know that we're overpaying these plans, I would choose to ratchet those payments down. So there are other things we can do too. There's really lots of things we can do. And you could have a more nuanced approach. We talked about this chart review earlier where it's the insurer, not the doctor, that's adding the codes. We talked about dispatching a home health visit. These are things that CMS, the regulator, could just disallow. They could say you could do these home health visits. But if you find a diagnosis code just from one of these home health visits, where we know that you're not really following up with any health care for the patient, we're not going to allow that to be included. in what constitutes the risk or what constitutes the diagnosis that you're being paid for. So you could rein back in the practices that we know are generating all these additional codes that don't really have any connection to patient care, that don't really have any connection to what the doctor is doing.

Final question, guys, this is a really interesting and unique moment of public attention on waste and fraud and government spending. You know, whether one conceives of Elon Musk and Doge as a wrecking ball to dismantle bureaucratic progressivism or a good faith effort to identify fraud or some combination of the two, there is a ton of heat right now, a ton of energy on the fact of and the question of government waste. What would be your parting wisdom to people who are listening to news about government waste? Seeking out government waste themselves, if people from the government are listening to show right now. What's the parting wisdom that you would leave with these folks? Yeah, I think, you know, when it comes to cutting government spending, cutting costs,

you want to find fat instead of bone to the extent that you can, right? And a lot of times, folks are very reticent to go into the kind of sacred categories of Medicare and Social Security and such. Within Medicare, there's likely a lot that can be done without having enormous effects on seniors' lives, right? And the primary place that I would go to do that is to these payments to the Medicare Advantage plans. As we've alluded to, there are going to be tradeoffs. If you cut spending, you cut the payments to the plans, in the past, we've found that every dollar going to the plans, about half of that, about 50 cents, goes to beneficiaries in one form or another. But at this point, the overpayments are so large that I think we're kind of in flat of the curve range here, right?

Where the things that they're giving to seniors are not things that seniors value a ton because they're just out of stuff to give them. They've run out of things that they can do to get the seniors to enroll in the plans. And so I do think that this is a place where you can cut without huge tradeoffs. As you alluded to, there's always tradeoffs, but I think here the tradeoffs may be smaller than in a lot of other places. And as I think about what, you know, Doge and these folks are doing out there, it seems like, you know, a lot of us, what we should probably be doing is helping them to identify these places where you can cut stuff without having real big impacts on people's lives. because otherwise we may worry that they're going to cut stuff that will have major impacts on people's lives. And so those of us who know kind of have ideas about what these things might be

should probably be trying to get those ideas in front of these folks. My advice is that if we're interested in cutting costs in Medicare, absolutely go after the fraud, but let's not pretend that the fraud is where the real money is. prosecute the fraud. That's flashy. It sounds like a win-win. But the real money is in the way that we legally allow these plans to inflate their payments. We've got maybe $50 billion a year in overpayments these plans that they legally extract. Most federal agencies have much smaller budgets than a billion a year. The EPA is $10 billion a year total budget. You could fit five EPAs into the overpayment that we make to MA plans. USAID was taken down recently. I'm not sure what the status of that

is as of today, but the entire budget that USAID controls is less than $50 billion. That fits in to the overage that we pay MA plans. And fixing that is not something that's going to be done overnight. It's an administrative process. You need to change the rules of the game, and that rulemaking process means that you propose rules. You let consumers, seniors, insurers, doctors weigh in on the rules. That's going to play out over time if anyone chooses to do it. But you sort of need to go through that process to have something durable that's going to stand up in court when the big health insurers decide that they don't like the new rules, for example. And so, you know, my advice is, and I don't know if anyone's listening, but my advice is do this

carefully do it in a way that's durable, and don't do it a way that's merely flashy and will immediately get overturned in court with no durable change. The one thing I want to take away from this episode is this concept of trade-offs, this idea that in the most complicated government programs, what you often have is not obvious waste or fraud, but rather good intentions gone slightly awry. and trade-off shifting ever so slightly out of balance. I really like the way that Tim described the dual-sided problems of a fee-for-service system

versus a capitation system. On the one hand, if you pay doctors to provide more services, you may accidentally or on purpose encourage health care inflation to run rampant. On the other side, if you try to control health care spending with capitation, you might accidentally, or on purpose, keep certain patients from getting the care that they need. Diagnostic codes emerge from a recognition of a problem that's real,

but they create an incentive whose costs might number in the tens of billions of dollars in terms of excess payments. I simply find that a very, very interesting illustration of the principle, that policy is hard and policy is tradeoffs. And people who pretend sometimes to have as their third job be the identification of government waste, fraud, and abuse, need to remember just how hard and complicated this system actually is. Thank you very much.

And we'll talk to you later this week. Convierre your passion in a business with Shopify and bathe records of with the form of pay with a better conversion of the world. Has heard of

the best. The incredible system of Pago of Shopify facilita the company on your site web, in the social and in in any place.

That is music for your ears. No, you'll be more your business will be a super-exit with Shopify. Empea-to-the-Provevee

your period of per year in Shopify, p.s bar records.