The amount of known fraudulent and improper payments in the U.S. is larger than the GDP of Sweden or Norway.

I keep reading articles claiming that Elon Musk and Vivek Ramaswamy’s goal of rooting out inefficiency from the federal government is dead on arrival. We’re told that the Republican majority is too slim to get anything through and that Musk and Ramaswamy “don’t have any idea what they’re doing, their private-sector successes notwithstanding.”

Yep, I am sure these guys have no clue how dysfunctional Congress and federal bureaucracies are. But I disagree that it makes their plan DOA. That’s because in the areas where Musk and Ramaswamy can make the biggest difference, reining in fraud and improper payments, Republicans and Democrats should be on board. Also, to solve this problem, we need radically different people than the ones who put us in this mess in the first place.

Think about it this way. The size of known fraudulent and improper payments in the U.S. is larger than the GDP of Sweden or Norway. According to the Government Accountability Office, “the federal government could be losing between $233 billion and $521 billion annually to fraud.” Additionally, federal agencies reported an estimated $236 billion in improper payments in fiscal year 2023. Medicare and Medicaid account for over $200 billion in fraud and improper payments annually. It’s insane.

These numbers have been growing as long as I have been following them. For instance, in 2014, I wrote in these pages, “How much money does the federal government send to the wrong recipients? In 2012, over $100 billion.” That’s a $136 billion hike in eleven years.

I don’t know, but it seems to me that if the alternative to Musk’s and Ramaswamy’s effort at reform is to continue to rely on the same people who “know” but who, for decades, have failed to rein in the explosion of improper payments, I will take my chances with the DOGE bros. Also, while it’s not the whole solution, technology will be a big part of the effort to end fraud and improper payments. And on that front, I will take the guy who built a machine that catches rockets with chopsticks (so that the rocket might perhaps be reused on our current government IT managers).

To say that the federal government’s IT is a mess would be the understatement of the year. The systems are not merely outdated; they operate with technologies and frameworks that the private sector has long abandoned. For instance, the GAO has repeatedly identified “legacy systems” as a high-risk area, with some federal agencies relying on hardware and software that are more than 50 years old. These systems, built in the 1970s and 1980s, use programming languages such as COBOL, which is so antiquated that the workforce capable of maintaining them is rapidly dwindling. Until five years ago, those responsible for the nuclear arsenal still used eight-inch floppy disks.

Maintaining obsolete technologies isn’t cheap, either. GAO reports that “of the government’s $100 billion in annual IT spending, 80% goes to operating and maintaining existing systems.”

Beyond their technical obsolescence, these legacy systems are highly fragmented. Different government agencies have independently developed and maintained their IT infrastructures over the years, leading to a patchwork of systems that cannot effectively communicate with one another. This siloed approach prevents data-sharing, reduces operational efficiency, and exacerbates fraud, waste, and duplication. It explains why the government has de facto adopted a “pay and chase” model whereby they pay out benefits first and do minimal chasing later.

This inability to prevent fraud and waste contrasts sharply with practices in the private sector, where advanced technologies are routinely used to identify and mitigate risks in real time. Private sector firms’ survival depends on it. The government doesn’t care because we are forced to pay taxes whether we like its services or not.

Modernizing government IT would require a comprehensive overhaul. To free engineers’ hands, mountains of arcane and often contradictory rules will first need to be rolled back, shifting engineers from compliance to actual coding. DOGE is already focused on deregulation, and I am sure it will quickly figure out which rules are getting in the way.

With modest rule sets, engineers can then focus on what they do best: Thinking out of the box, replacing legacy systems, consolidating data silos, and adopting cutting-edge technologies such as artificial intelligence and cloud computing to discover inefficiencies and catch improper transactions.

This is no substitute for Congress reforming all these programs, but DOGE could make a start. And the potential benefits are enormous: streamlined operations, an end to the “pay and chase” model, and savings to taxpayers that could be comparable to the GDP of Sweden.