Every government benefit has a recipient, but it also has a payer. Politics tends to encourage us to focus on the first and forget the second.

Many hardships are visible, immediate, and emotionally compelling. They provide a rationale for Congress to pass a law, for an agency to distribute money, and for elected officials to congratulate themselves on their compassion. But almost no one asks the question that the pioneering American sociologist and political economist William Graham Sumner believed mattered most: Who provided the assistance and at what cost?

Read the rest of this post →

Congress is vastly overspending, and federal debt is rising by $2 trillion a year. Without major spending reforms, the government is headed toward a financial crisis. But it is also true that even if the federal budget were balanced, many spending programs would continue undermining the economy because their costs outweigh their benefits. 

Such negative-value spending is the focus of my recent “Leaky Bucket” study with Ryan Bourne. We argue that the federal budget is riddled with wasteful “leaks.” The government scoops up trillions of dollars into its budget bucket, but many of them spill out due to inefficiencies and policy mistakes. The Trump administration has focused on reducing fraud, but that is only one type of leak from the federal budget bucket.

Read the rest of this post →

Last month, the US Department of Agriculture (USDA) released its Supplemental Nutrition Assistance Program (SNAP, formerly known as Food Stamps) payment error rate report for fiscal year 2025.

The report confirms what has long been true: SNAP is structurally prone to waste, fraud, and abuse, and the states running it have little financial incentive to fix it.

The Problems with Spending Other People’s Money

According to the report, SNAP made $10.1 billion in improper payments last year, more than 1 of every 10 dollars spent on the program, at a national error rate of 10.62 percent. More than 87 percent of this total was from overpayments — beneficiaries receiving more money than they were owed — which cost federal taxpayers $8.8 billion.

This marks the fourth consecutive year that improper payments have been in the double digits, with the FY 2025 rate nearly three times its historic low in FY 2013 (Figure 1).

This is a marginal improvement over last year’s error rate (10.93 percent). But as USDA Secretary Brooke Rollins put it, “These payment error rates are further proof that state accountability is severely lacking in SNAP.”

That problem traces back to a mismatch in how the program is financed. The states administer SNAP, but federal taxpayers foot the bill for more than 90 percent of the program’s costs. This gives states an “all gain, no pain” arrangement in which the financial consequences of lax oversight and administrative inefficiency fall almost entirely on federal taxpayers.

Giving States Skin in the Game

The One Big Beautiful Bill Act (OBBBA) aimed to address this by requiring states to pay a share of the SNAP benefits they distribute when their error rates exceed 6 percent.

States may elect to use either their FY 2025 or FY 2026 error rate to calculate that share, making this report the opening benchmark for what states will owe when the cost-sharing requirements take effect in FY 2028.

Based on FY 2025 data, 41 states and the District of Columbia face potential cost-sharing obligations (Figure 2).

Some states showed improvements in their FY 2025 error rates. New Jersey saw the largest decline, reducing its error rate from 14.33 percent in FY 2024 to 6.86 percent, potentially reflecting the state’s efforts to work more closely with county offices and implement additional quality control protocols.

Meanwhile, Hawaii’s error rate shot up from 6.68 percent to 10.92 percent, likely due to dropping net income limits for some households and absorbing caseload increases from extending eligibility to Compact of Free Association (COFA) residents.

How the Alaska Carveout Undermines OBBBA

However, six states and DC qualify for a temporary exemption under the Alaska Carveout, a policy shoehorned into OBBBA to secure Senator Lisa Murkowski’s (R‑AK) vote. It grants states with improper payment rates above 13.34 percent up to a two-year exemption from financial penalties, rewarding the worst-performing states while penalizing those working to reduce their error rates.

Case in point: Florida, Maryland, Massachusetts, New Jersey, and New York no longer qualify for an exemption and now face cost-sharing requirements because their error rates dropped below the 13.34 percent threshold.

Meanwhile, states that saw an increase in their payment errors, such as Delaware and Illinois, are granted a reprieve until at least FY 2029. To qualify for the exemption, officials in New Mexico admitted to slowing down efforts to tighten the state’s verification standards, while in Maryland, whistleblowers allege that officials conspired to leave correctable errors uncorrected to keep their improper payment rates elevated.

States have exploited the system in the other direction, too. In 2014, the USDA flagged quality control issues in 42 of 53 state SNAP agencies, with many gaming their review processes to artificially lower their reported error rates.

It is also worth noting what the FY 2025 data cannot show. The report covers October 2024 through September 2025. OBBBA did not pass until July 2025, meaning states had at most two months to respond to the new requirements before the measurement period closed. The FY 2026 report will offer a clearer picture of how OBBBA’s SNAP reforms, including the Alaska Carveout, affect state behavior — for better or for worse.

How States Are Racing to Cut Errors

The more telling signal is what states have done in anticipation of those penalties in the year since OBBBA passed. With financial penalties potentially hitting their budgets, many states have pursued a wide range of approaches to drive down their error rates:

• Virginia’s “SNAP Forward” initiative is eliminating applicants’ ability to self-attest their eligibility information, contracting with third parties to help improve the state’s payment processes, and creating a quality assurance team to identify and correct payment errors.
• Louisiana is offering cash bonuses to staff who spot payment errors, intensifying case reviews, and updating its verification systems to automate wage checks.
• Kansas is bringing in outside consultants to reduce its error rate, is participating in a mandatory federal corrective plan to address errors driven by high staff turnover and complex casework, and has proposed $4.4 million in upgrades to its verification system, including a chatbot to help staff apply eligibility rules.
• Mississippi is updating its antiquated, 40-year-old eligibility and case management software by installing new IT systems to cross-check SNAP recipient information against other programs in real time. The state is also piloting an income verification program to check SNAP recipients’ incomes against credit bureau data.
• Arkansas is investing in AI tools to improve its eligibility systems to reduce human data errors and free up caseworkers to identify mistakes in other areas. It has also allocated an additional $5 million in its FY 2027 budget to the state inspector general’s office to detect vulnerabilities in the program.
• Minnesota lawmakers allocated $90 million to replace the 35-year-old software counties use to process SNAP applications, which requires workers to manually re-enter the same eligibility data multiple times across disconnected systems, plus $15 million to upgrade fraud detection technology.

Congress Should Finish the Job It Started

What the early state responses to OBBBA make clear is that accountability changes behavior. States facing financial penalties for poor performance are investing in better verification systems and strengthening oversight.

Congress should build on this, ideally by leaving the provision of nutrition assistance up to the states. The best way to curb waste, fraud, and abuse in programs like SNAP is for states to pay for it. A program funded by state taxpayers is one that state officials have a far greater incentive to run carefully.

Short of that, Congress can help states modernize their fraud-prevention processes, including directing them to add chips to EBT cards to prevent benefit theft and to strengthen and expand their access to the Treasury’s Do Not Pay system. Congress can also eliminate error-prone policies like broad-based categorical eligibility, which is associated with disproportionately high improper payments compared to traditional eligibility, in part because states have used it to streamline program administration by weakening their verification protocols.

At a minimum, Congress should not compromise on its commitment to combating waste, fraud, and abuse in SNAP last year by delaying the start of OBBBA’s cost-sharing requirements, as some senators are pushing amid Farm Bill negotiations.

States may have made marginal improvements to their payment processes in FY 2025. But a mere 0.3 percentage-point decline shows that states still have a long way to go to rein in improper payments and comply with OBBBA’s requirements and that Congress has yet to fix SNAP’s structural integrity problems.

Congress is gearing up for another potential reconciliation bill. If it moves forward, Congress should pursue reforms that curb waste, fraud, and abuse in federal welfare programs. 

A new Cato Institute briefing paper examines the design flaws that drive misspending in five major federal welfare programs and outlines reforms to address them. The programs are the Supplemental Nutrition Assistance Program (SNAP, formerly known as Food Stamps), Medicaid and the Children’s Health Insurance Program (CHIP), child nutrition programs, Temporary Assistance for Needy Families (TANF), and the Supplemental Security Income (SSI).

Financial mismanagement in these programs takes numerous forms:

• Improper payments, including underpayments and overpayments resulting from administrative errors and outright fraud. Minnesota’s Feeding Our Future nonprofit billed a federal child nutrition program for 125 million meals never served, defrauding federal taxpayers of nearly $250 million. One Arizona-based company fraudulently billed Medicaid $650 million for substance abuse services that were never or insufficiently provided.
• Manipulation of program rules, such as exploitation of financing gimmicks, waiver authorities, and eligibility loopholes to maximize federal funding, avoid accountability, or redirect federal dollars beyond programs’ intended purposes. Forty-three states and the District of Columbia use broad-based categorical eligibility in SNAP to circumvent federal eligibility requirements, in some cases enabling millionaires and lottery winners to qualify for food stamp benefits. States have also used TANF’s loose spending rules to turn it into a slush fund to plug budget shortfalls, directing federal funds toward volleyball stadiums and college scholarships.

These problems are largely driven by two reinforcing weaknesses:

• A financing mismatch, whereby states process applications and distribute benefits while federal taxpayers finance most, or all, of the expenditures. This leaves states with weak incentives to enforce eligibility standards, invest in fraud prevention, or control waste, since the consequences of financial mismanagement fall overwhelmingly on federal taxpayers. SNAP is more than 90 percent federally funded, and Medicaid is approximately 70 percent federally funded after accounting for the financing gimmicks states use to inflate their reported spending and draw additional federal matching funds.
• Administrative complexity is made worse by outdated verification systems, reliance on self-reported eligibility data, and underutilized fraud-prevention tools, such as the Treasury’s Do Not Pay system. The Social Security Administration has failed to use readily available verification tools to check recipients’ financial accounts for SSI eligibility, with the agency’s inspector general estimating that $2 billion in SSI overpayments in fiscal year 2023 could have been prevented.

Addressing these problems requires structural reforms that align incentives, modernize verification systems, and hold program administrators accountable for poor outcomes. Congress began that work with the One Big Beautiful Bill Act (OBBBA), which strengthened program integrity in Medicaid and SNAP.

This newly released brief recommends ways for Congress to build on those reforms and bring similar accountability measures to child nutrition programs, TANF, and SSI.

Congress can save federal taxpayers nearly $6 trillion over the next decade by adopting the structural reforms proposed in this brief, namely eliminating federal funding for SNAP ($550 billion), TANF ($91 billion), and child nutrition programs ($262 billion), and converting Medicaid and CHIP into a zero-growth block grant ($5 trillion).

These reforms would help address the structural incentive problems that drive widespread financial mismanagement in these programs by giving states a direct financial stake in their performance.

Short of significantly reducing the federal government’s role in welfare provision, Congress can pursue reforms that strengthen verification systems, enhance oversight, and reduce abuse to save federal taxpayers nearly $500 billion over the next decade by eliminating provider taxes in Medicaid ($386 billion), ending broad-based categorical eligibility in SNAP ($100 billion), raising the Community Eligibility Provision threshold to 60 percent in school food programs ($3 billion), and eliminating TANF’s contingency fund ($6 billion).

If the federal government continues to operate and fund welfare programs, it has an obligation to ensure that taxpayer dollars are not lost to preventable waste, fraud, and administrative failures.

Read the full briefing paper here.

Supporters of federal spending make it sound simple: the government designs expert programs, raises money from taxes, and steers it to individuals, businesses, and nonprofit groups to solve problems in society.

Our new study, Federal Government Spending Is a Leaky Bucket, describes why this view is far too optimistic. Every step in raising and spending money is undermined by design errors, bureaucratic inefficiencies, and behavioral responses that result in most programs generating more costs than benefits. 

The study’s name borrows from economist Arthur Okun, who famously likened welfare programs to a leaky bucket. 

We argue that the government bucket is even leakier than Okun described and that his lessons apply to all federal programs—whether spending on welfare, defense, infrastructure, or health care. 

Our study provides a framework to understand why most federal spending undermines the economy and why most activities should be left to the states and the private sector. It details the many ways policymakers—carrying their program buckets—are leaking resources all over, making us worse off.

Many Americans view Brexit (Britain’s exit from the European Union) as Trumpism before Donald Trump—the first instance of an emerging nativist populism, hostile to trade and immigration. Those policy areas certainly were affected by Brexit, and the desire to restrict immigration from the EU did animate many Leave voters.

But the comparison is mostly a category error. Lord Ashcroft’s referendum day survey showed the main reason people voted for Brexit was a desire for sovereignty. Brexit wasn’t a set of policies but a constitutional instruction to return a bunch of powers from Brussels to Britain’s Parliament. How politicians would use them was always an open question.

Ten years on from the original 23 June 2016 referendum (remember, the UK only left the EU trading bloc at the end of 2020), an honest ledger infuriates people from both tribes. The Remain campaign was wrong to believe that Brexit would cause an immediate economic downturn. Classical liberal Brexiteers were too hopeful in thinking Brexit would really deliver a low-tax, lightly regulated economic power outside of Brussels’ orbit.

The truth is less dramatic. Brexit returned various economic powers at the cost of greater trade frictions. The difficulties of extrication were greater than many Leavers like me expected, but not as bad as Remainers feared.

What Brexit Didn’t Do

1. It didn’t cause an immediate recession. The Conservative government that campaigned against Brexit said a Leave vote would push Britain into recession, leaving GDP 3.6 percent lower after two years in its central “shock” case, while raising unemployment by 500,000 people. Instead, GDP grew healthily by 2.2 percent in 2016, 3 percent in 2017, and 1.6 percent in 2018, while unemployment fell right up until the end of 2019.

2. It didn’t slash immigration. Ending free movement reduced immigration from the EU. But Boris Johnson’s post-Brexit immigration reforms liberalized access for much of the rest of the world. Alongside new immigration routes for Hong Kongers and Ukrainians, the result was that non-EU immigration soared after the first year of the pandemic (see chart below). Net migration was 171,000 in the year ending December 2025, down from a March 2023 peak of 944,000. That peak happened under the post-Brexit system, not despite it. A future Reform government under Nigel Farage could cut immigration sharply. So far, post-Brexit Britain has had more immigration, just from different places.

3. It didn’t turn Britain protectionist. Brexit created new barriers with the EU, but not a general tariff turn. The EU-UK trade agreement kept most goods tariff-free, while adding paperwork and restricting services trade. Britain also unilaterally cut tariffs on thousands of goods with the rest of the world. Successive governments rolled over existing third-country trade agreements, signed new deals with Australia and New Zealand, and joined CPTPP, the Trans-Pacific Partnership. These deals are small in GDP terms relative to EU trade losses, given existing trade flows. But Britain did not follow Donald Trump into tariff protectionism.

4. It didn’t lead to authoritarian domestic politics. Britain has real civil liberties problems in the form of speech policing and protest restrictions. But Freedom House still scores the UK as “Free,” with strong political rights and civil liberties marks. For all the talk of Brexit being a “hard right” project, there’s simply been no evidence of markedly authoritarian governance since the referendum, first under the Conservatives and now Labour.

5. It didn’t deliver radical deregulation. Britain removed EU legal supremacy, but retained EU law was mostly domesticated, cataloged, and managed. The government’s own dashboard still lists 6,925 pieces of retained or assimilated law across more than 400 policy areas. The UK has diverged from EU rules in some areas, such as financial services, but most divergence reflects the EU changing its rules rather than Britain boldly rewriting its own. The result has been mostly regulatory continuity with occasional pruning, not a bonfire of Brussels red tape.

What Brexit Did 

1. It made Britain poorer. Britain has underperformed France and outperformed Germany since 2016 in GDP per capita growth. But Britain could have grown faster still. New non-tariff barriers to trade show up in microeconomic evidence as chilling substantial investment. Studies that model Britain’s pre- and post-Brexit trends find a substantial hit to GDP. It’s difficult to disentangle some of Britain’s slower growth in that counterfactual with the relative shock of the pandemic across countries, and I think studies suggesting up to an 8 percent of GDP hit from Brexit likely overstate it. But Britain lost the equivalent to at least a year or more of growth from a self-imposed supply-side tax from new trade frictions that it failed to counteract with other liberalizations.

2. It made politics more volatile. The announced resignation of Prime Minister Keir Starmer this week means Britain is now preparing for its seventh prime minister in ten years. Brexit shaped many voters’ political identities and led to a fundamental shuffling between the parties. But the resultant realignment of politics is still taking place, and the result has been deep dissatisfaction with the big parties (Labour and Conservatives), the emergence or propulsion of other parties (Nigel Farage’s Reform and the hard left Greens), and a large turnover of prime ministers.

3. It downgraded economic thinking in politics. The short-run recession warning was so overblown that many Brexiteers learned to dismiss economic analysis and economic institutions entirely, to the country’s detriment. Brexit also elevated other questions above standard economic debate, consuming time and political energy that could have gone into pro-growth reform. Economics now seems less of an aligning issue in British politics, creating space for kooky economic policy. Andy Burnham, widely discussed as the next prime minister to replace Starmer, says many of Britain’s problems owe to the privatization of state-owned industries under Margaret Thatcher. The mind boggles.

4. It increased NHS spending. The Leave campaign promised to redirect much of Britain’s net contribution to the EU to Britain’s ailing National Health Service. Theresa May’s government pledged a £20.5 billion real-terms annual NHS increase by 2023/24 and explicitly framed it as delivering more than £350m a week extra for the health service—the original referendum campaign promise. Given Brexit’s growth effects, this was not a “Brexit dividend.” It meant higher taxes. Sadly, Brexit helped sanctify more spending on an unreformed socialized health-care monopoly.

5. It let Britain avoid the EU’s AI Act. Britain has not deregulated much of what it inherited from the EU. But Brexit has allowed the UK to avoid new EU regulation of frontier technologies, as Brexit’s proponents promised. The EU chose a comprehensive, risk-tiered AI statute, with high-risk obligations on data, documentation, human oversight, logging, and robustness. Britain, for now, has chosen a more pro-innovation model, initially avoiding a new AI-specific regulator. Preserving that more liberal approach to emerging technologies is one way a real Brexit dividend could still emerge.

For Americans, the lesson is that Brexit was not Trumpism with a British accent. It was a decision to move power back from Brussels to Westminster. That choice had costs. New trade frictions made Britain poorer than it otherwise would have been, and Brexit produced substantial political volatility. But it did not produce recession, protectionism, mass deportation, or authoritarian rule. Nor did it produce much in the way of free-market reform in repatriated policy areas.

Good economic policy simply doesn’t require a pan-continental parliament, various EU presidents, harmonized labor and digital regulations, farm subsidies, and a common tariff. Yet leaving an institution with this structure is no guarantee that a country will become economically freer. The UK hasn’t capitalized much on Brexit yet. Time will tell if it does.

Was President Lyndon Johnson’s “war on poverty” the main driver of declining poverty rates after 1964?

Not according to a new study. The researchers examined a measure of poverty 

that accounts for all taxes and government benefits … [and] anchored [the] poverty measure to the official poverty rate in 1963. … [They also] adjusted income thresholds for inflation and the size of each person’s household. 

Using this measure, they found that 

poverty fell substantially prior to the War on Poverty, primarily due to increases in market income, without a substantial rise in the dependency of working-age adults and their children on government transfers for most of their income. … These trends were particularly stark for black people, who experienced a steep decline in poverty before the War on Poverty.

Cross-posted from Substack