Suchergebnisse

Republicans often say that low‐​income welfare, such as food stamps, should be only temporary, and that people should strive to stand on their own two feet. The same should be true of high‐​income welfare—that is, subsidies for farmers.

Did the debt‐​ceiling deal passed in May reflect a new congressional focus on spending restraint? Or was it a fluke win by House conservatives that will unravel as Congress addresses spending this fall?

I join Roger Pilon in expressing sadness at the passing of federal judge and U.S. senator James Buckley. I became acquainted with Jim when he was in his 90s, and I was so impressed that he was still actively considering policy issues and influencing public debate. James Buckley reached the century mark, and his mind was sharp until the end.

With Congress scheduled to reauthorize farm subsidies this fall, lawmakers should consider reform lessons from New Zealand. Facing high budget deficits in the 1980s, New Zealand cut government spending, including eliminating nearly all farm subsidies. That was an impressive reform because the country is highly dependent on agriculture. Since then, New Zealand has remained a model of market‐​based farming.

Congress is scheduled to consider a major farm bill this year, which will reauthorize many U.S. Department of Agriculture (USDA) programs, including farm subsidies and food stamps. The legislation could cost $150 billion a year and so presents an opportunity to find budget savings and reduce the flood of red ink in Washington.

Let's cut both farm subsidies and food stamps in this year's farm bill.

Chris Edwards


The next president will face a huge federal budget mess. Spending is driving up debt to unprecedented levels and is pushing federal intrusion ever deeper into state, local, and private affairs.
The most important question we should ask of 2024 presidential candidates is will they cut spending? Will they use their bully pulpit and budgeting powers to push Congress to cut?
The next president must be a dedicated spending cutter. Numerous candidates for the White House have track records on spending as governors. The table below includes governors who are in the race and some governors who are possible candidates.
The table shows general fund spending, which is the part of state budgets that governors have the most control over. The table shows spending for the first and last fiscal years each governor was in office. For governors still in office, the last year is their proposed spending for 2024. The table shows the annual average percent change in spending, and the change on a per‐​capita basis because state population growth rates differ.
By these metrics, the most frugal governors have been Asa Hutchinson, Mike Pence, Kim Reynolds, Chris Sununu, Chris Christie, and Doug Burgum, while the least frugal has been Gavin Newsom, Ron DeSantis, Gretchen Whitmer, and Nikki Haley.
Some caveats are discussed below. Also, broader fiscal assessments of these governors are available in Cato's fiscal report cards.







Notes and Caveats
The spending data is from NASBO fiscal reports. The most recent report includes 2022, 2023, and 2024. For prior years, I used the final "actual" figures from previous editions.
General fund spending growth is just one measure of a governor's fiscal performance.
Governors share fiscal powers with state legislatures, which may or may not be controlled by the same party.
The recent high inflation may have increased averages for current governors, although some current governors, such as Kim Reynolds, have nonetheless held spending down.
California's per‐​capita spending increased more than total spending because the state's population is shrinking.
I used Census actual state populations up until 2022, and then estimated 2023 and 2024 based on the growth in 2022.
Broader discussions of these governors and their fiscal policies are in Cato report cards.

Chris Edwards


The U.S. Department of Agriculture runs dozens of farm subsidy programs, which cost taxpayers more than $30 billion a year. The largest program is crop insurance, which costs about $10 billion a year. The program displaces private methods of managing risk and gives subsidies to farm businesses that do not need them.
Federal crop insurance for revenue and yield shortfalls is available for more than 100 crops, but corn, soybeans, wheat, and cotton are the main ones. The policies are provided through 14 private insurance companies.
The USDA subsidizes insurance premiums, which costs taxpayers about $8 billion a year. The subsidies cover an average 62 percent of premiums, which results in most farmers making money on this so‐​called insurance. The Congressional Budget Office found that farmers received $65 billion more in claims than they paid in premiums between 2000 and 2016.
The USDA also pays about $2 billion a year to farm insurance companies to cover their administrative costs. With these subsidies and the inflated demand for policies, the 14 insurance companies appear to make above‐​normal profits, as suggested by the Government Accountability Office and other experts.
Here are some features of the USDA's crop insurance program.
Expansion in Scope. Federal crop insurance keeps growing in size and scope. The USDA reports that the number of crops insured has risen from 112 in 2000 to 134 today, while the number of insurance plans has risen from 20 in 2012 to 36 today. Expanding the scope of farm subsidies buttresses support for the logroll in Congress.
No Income Limits. There are no income limits for the crop insurance program, with the result that subsidies are tilted upwards. An AEI study found that "farms in the top 10 percent of the crop sales distribution received approximately 68 percent of all crop insurance premium subsidies." The Government Accountability Office reported that some billionaires receive crop insurance subsidies. It is often claimed that subsidies are needed to support family farms, but most farm subsidies go to the largest producers.
No Transparency. The recipients of crop insurance subsidies are a government secret. Billions of dollars in taxpayer‐​funded aid is laundered through private insurance companies to help hide the identities of the millionaire and billionaire recipients. Congress is spending our money, but we're not allowed to know who gets it.
Crowding Out. Instead of government subsidies, the USDA itself identified market‐​based ways for farmers to manage risks, including diversifying crops, building savings, using forward contracts, using specialized equipment, and diversifying income sources. Farmers can also diversify planting locations and keep a low debt load in case emergency borrowing is needed. Federal subsidies displace or crowd out private risk management solutions, including new solutions that would arise if subsidies were repealed.
Disaster Aid. Crop insurance subsidies have been expanded over the years under the premise that they would reduce pressure for Congress to pass emergency aid packages after disasters. But farmers have enjoyed huge emergency aid packages in recent years on top of all the regular aid. Since 2017, Congress has passed "an astounding $60 billion in ad‐​hoc disaster assistance" in a series of bills.
Environmental Effects. Economists Vincent Smith and Barry Goodwin argue, "There is an extensive body of research overwhelmingly reporting that subsidized crop insurance has encouraged farmers to shift production onto more fragile lands, thereby increasing soil erosion and, by implication, agriculture's carbon footprint."
Climate Change. The Environmental Working Group says that the federal crop insurance program "doesn't encourage or require farmers to adapt to or mitigate climate change because it often pays farmers for the same type of loss year after year, like multiple years of payments due to drought." That is a fascinating point—that subsidies undermine market adaptations to changing environmental conditions. The same is true, by the way, with federal flood insurance subsidies.
The best reform would be for Congress to repeal crop insurance subsidies, which would save taxpayers about $100 billion over the next decade. A more limited reform would be to impose an income limit for subsidy recipients, as economist Eric Belasco examines here. There is no reason why farm businesses cannot plan ahead by themselves to mitigate fluctuations in their earnings from drought and other contingencies.
More on farm subsidies here, here, and here. And see farm policy analyses from AEI, EWG, TCS, and Heritage.

Chris Edwards


Congress will consider a new farm bill this year, which will likely be a logrolling extravaganza costing $1.5 trillion or more over the coming decade. Just as Lollapalooza had a diverse lineup of bands, the farm bill will include a diverse lineup of subsidies for farms, food programs, energy, rural programs, forestry, trade, environmental activities, and many other things. Hemp production used to be illegal but now gets subsidized in the farm bill.
In his book on government dysfunction, MSNBC host and former congressman Joe Scarborough described the logrolling frenzy leading to the passage of the 2002 farm bill, which he called the "largest corporate welfare scam in history." He discussed how dairy subsidies were demanded by members from Maine, Pennsylvania, and Vermont, peanut subsidies were demanded by members from Virginia, Alabama, and Georgia, and sugar subsidies were demanded by members from Florida. The logrolling continued for cotton, wheat, wool, mohair, and many other products.
Scarborough concluded, "Standing alone, not one of these corporate welfare measures could survive the bright light of public scrutiny." That is the key point about logrolling. Unfortunately, logrolling is central to the modern legislative process because the government has grown too large to consider individual provisions on their own merits.
Logrolling means that bills jammed full of special‐​interest provisions can gain majority support even if none of the provisions could gain majorities by themselves. Logrolling involves committee chairs or party leaders bundling together narrow subsidies benefiting particular states and interest groups. If democracy means majority support for specific policies, then logrolling undermines democracy.
The problem with logrolling has been observed since at least the mid‐​19th century when omnibus bills bundled dozens of Army Corps of Engineers projects across many states. At the time, people objected that these bills included low‐​value projects that did not have broad support. The federal government is much larger today, and so the logrolling problem is worse, as I discuss here and here.
Here is a June 2023 Congressional Research Service (CRS) report on the upcoming farm bill: "The omnibus nature of the farm bill can create broad coalitions of support among sometimes conflicting interests for policies that individually might have greater difficulty achieving majority support in the legislative process." That is a polite way of saying that if you bundle a bunch of loser provisions together you can end up with a legislative winner.
Farm bill logrolling is becoming more extensive says the CRS:
In recent years, more stakeholders have become involved in the debate on farm bills, including national farm groups; commodity associations; state organizations; nutrition and public health officials; and advocacy groups representing conservation, recreation, rural development, faith‐​based interests, local food systems, and organic production. These factors can contribute to increased interest in the allocation of funds provided in a farm bill.

What can we do about it? The official baseline for the farm bill this year is $1.5 trillion over 10 years. Farm bill leaders in Congress think of the baseline as the minimum pot of money they can carve up and handout to dozens of special‐​interest groups in coming months. But the federal government is hurtling toward a debt crisis, and business as usual is not acceptable. The bipartisan debt‐​ceiling deal passed in May reflected a new priority of controlling red ink. We need belt‐​tightening all around and a much lower price tag than $1.5 trillion for any farm legislation.
I look at logrolling in detail here and here and farm subsidies here.

Chris Edwards


Food stamp fraud is rising, particularly through EBT card skimming, as I discuss in this op‐​ed. EBT numbers and PINs are a great target for theft because the cards do not have chips. This is one of many failures that should prompt Congress to downsize the huge food stamp program in this year's farm bill. The food stamp program is also called the Supplemental Nutrition Assistance Program.
The fraud problem with SNAP and other benefit programs is not just that individuals fake their personal data to gain unjustified benefits, but also that criminals find systematic program weaknesses and exploit them. Federal hand‐​out programs are looted by criminal gangs in an organized fashion.
Programs paid for by the federal government and run by the states—such as food stamps—are particularly prone to looting because state administrators have little incentive to worry about costs imposed on federal taxpayers.
This July 4 piece in the San Diego Union‐​Tribune caught my eye:
The bulk purchases of Red Bull and Monster Energy drinks were huge. Within about two months late last year, the same small group had spent $305,270 buying pallets of the caffeinated drinks at grocery stores in downtown San Diego and Riverside.
The transactions were large enough that members of the group had to make special arrangements with store managers ahead of time and pick up the drinks in the Smart & Final loading docks. They presumably resold the drinks at a profit — and the profit would be high since the money didn't even come out of their own pockets.
Instead, the purchases were made with money pilfered from low‐​income public benefits recipients.
Last week, a member of the group pleaded guilty in San Diego federal court to one count related to the scheme. Beatrice Mihai, a 25‐​year‐​old Romanian citizen who lived in Anaheim, admitted in her plea agreement that she and her co‐​conspirators stole and misused the Electronic Benefit Transfer funds intended for nearly 175 victims in California, New York, Florida and Rhode Island.
… More recently, Romanian organized crime groups have plundered taxpayer‐​funded programs for low‐​income Californians, targeting those swiping state‐​issued cards loaded with EBT and unemployment benefits. Investigators believe those accounts are being targeted because the state‐​issued benefit cards are more vulnerable because they lack the chips embedded on most bank‐​issued cards. The theft has become a huge problem.
… On a near‐​weekly basis, news reports from across the U.S. — citing police records, prosecutors, court evidence and news releases — demonstrate the links between Romanian groups and ATM skimming operations.
… A few months later, investigators learned about the energy drink scheme at the San Diego and Riverside Smart & Final stores. Prosecutors alleged the group made at least 16 bulk purchases between the two locations in October and November, swiping multiple EBT cards to complete each purchase.

Let me summarize how this works. Gangs put skimmers on card machines at check‐​outs in retail stores. They gain dozens or hundreds of card numbers and PINs before the skimmers are discovered. They encode the numbers on new cards and drain food stamp accounts by buying high‐​value goods such as energy drinks. They resell the goods for cash to other retailers, typically for 50 cents on the dollar.
EBT card skimming has become a particularly severe problem in recent years. But SNAP is difficult to police in general because it includes 250,000 retailers and 42 million recipients. The program has exploded in cost from $63 billion in 2019 to about $145 billion in 2023.
The solution is to get the federal government out of food stamps and allow the states to fund their own food programs if they choose. State lawmakers must balance their budgets, and so they have strong incentives to minimize fraud and waste when funding their own programs.
More on food stamps here, here, here, here, here, here, and here.

Chris Edwards


Federal government debt is rising to dangerous and unprecedented levels. Without reforms, federal debt held by the public will grow from 98 percent of gross domestic product this year to 115 percent a decade from now. Compared to the size of the economy, federal debt and interest costs are headed toward levels never seen in our nation's history.
The recent bipartisan debt‐​ceiling deal modestly slowed the flow of red ink, but larger reforms are needed. A growing number of fiscal experts are recommending that Congress limit federal debt to 100 percent of GDP (e.g. here, here, and here). That is, Congress should restrain the budget so that debt grows no larger than the economy.
Let's look at four reasons to cut debt and then examine a plan to hit the 100 percent target with entitlement and federalism reforms.
Four Reasons to Cut Federal Debt
Funding spending with debt pushes costs onto young people in the future. But young people will have their own costs, crises, recessions, and wars to deal with, so burdening them with our costs in addition is unjust.
High and rising debt increases macroeconomic instability, and it will likely prompt a major financial crash and recession, which will cause hardship for many and undermine living standards.
Many statistical studies have found that government debt above about 90 percent of GDP slows economic growth.
Most federal spending goes toward subsidy and aid programs, which help recipients but distort the economy. As such, reducing debt with spending cuts would boost growth as resources were reallocated to higher productivity uses.
Excessive spending is causing the surge in debt. As shown in the chart, CBO expects federal revenues to remain at about 18 percent of GDP in coming years, a bit above the 50‐​year average of 17.4 percent. The problem is that spending is projected to rise to 24.8 percent of GDP by 2033, substantially above the 50‐​year average of 21.0 percent. Reducing spending to the long‐​term average would restrain debt to about 100 percent of GDP.







Entitlement and Federalism Reform Plan
Policymakers should balance the budget and cut debt in the long term, but a good near‐​term goal would be to stabilize the debt at 100 percent of GDP. The plan proposed here would achieve that in 2033 by reducing spending to 21.1 percent that year.
The table shows proposed spending reforms. The entitlement reforms would be enacted in the near‐​term and the savings would increase over time, while the federalism reforms would be phased in over 10 years. The reforms would cut program spending by $1.31 trillion in 2033 and cut overall spending including interest by $1.45 trillion, or about 15 percent of baseline spending that year.
The entitlement reforms include limiting Medicare's growth rate to the growth rate of GDP beginning in 2026. A good way to achieve that would be to restructure the program around individual vouchers, which would improve choice, encourage competition, and restrain costs. For Medicaid, the plan would convert today's open‐​ended matching grants to fixed block grants in 2024, which would control federal costs while freeing the states to innovate with their health care systems.
For Social Security, the table includes two straightforward reforms, as estimated by CBO. The first modestly reduces the annual cost of living (COLA) adjustment for benefits, and the second would modestly raise the program's full retirement age. (For both reforms, I estimated 2033 savings based on the CBO figures for 2032).
The federalism reforms would cut federal aid‐​to‐​states for programs administered by state and local governments. The federal government spends $1 trillion a year on more than 1,300 aid‐​to‐​state programs. The aid system ties the states into regulatory knots, undermines democratic control, destroys accountability, and generates waste, fraud, and abuse. State and local governments should fund their own programs for welfare, housing, transit, education, and many other things.
For programs in the table, the plan would phase in the federalism reforms over 10 years. The states could respond by funding their own programs if they chose, either by raising taxes or creating budget room by cutting other spending. Without all the top‐​down rules imposed by Washington, stand‐​alone state programs would likely be leaner and more efficient.







Final Thoughts
Policymakers may think such spending reforms are radical, but larger reforms have succeeded abroad. Facing a debt crisis in the 1990s, Canada cut its federal spending from 23 percent of GDP in 1993 to just 15 percent by 2006. The government cut entitlements, business subsidies, defense, aid to the provinces, and many other things. It privatized assets such as airports and the air traffic control system. As the government was cut, the Canadian economy boomed for 15 years.
America needs similarly large spending cuts. In addition to the above reforms, we should cut business subsidies, farm subsidies, foreign aid, and energy subsidies. We should also privatize federal assets.
I discuss further reforms here and Romina Boccia proposes reforms here and here.
____________________________________________
Data Notes: the federalism reforms generally involve zeroing out aid to the states for the specified activities. The K‑12 subsidies do not include special education subsidies. The excess highway aid is projected highway outlays that are greater than highway trust fund revenues. The values in the table were sourced from CBO projections and OMB projections.

Chris Edwards


Federal government spending is soaring and debt will soon reach record highs compared to the size of the economy. Rising spending and debt are undermining growth and may push the nation into a financial crisis.
As the federal government's size has expanded, the scope of its activities has also grown. The government subsidizes farming, health care, school lunches, broadband, rural utilities, energy, rental housing, aviation, passenger rail, public broadcasting, chip manufacturing, job training, foreign aid, education, urban transit, space exploration, and hundreds of other activities.
For decades, the government has published an official list of all its grant or subsidy programs for the states, businesses, nonprofits, and individuals. The list used to be called the Catalog of Federal Domestic Assistance but is now called Assistance Listings. The list is a rough indicator of the steadily expanding scope of federal interventions.
The chart shows that there are 2,418 federal grant or subsidy programs in 2023, more than double the number in 1990. Each new subsidy program requires higher taxes or more federal borrowing. Each subsidy generates a bureaucracy, spawns lobby groups, and encourages more special interests to demand handouts.
The rise in size and scope of federal subsidies means that Americans are losing their independence. State and local governments, businesses, nonprofits, and individuals that become hooked on subsidies become tools of the federal government. They have less incentive to work and innovate, and they shy away from criticizing government policies.
Let's all celebrate July 4, but remember that the path to freedom and prosperity is to cut the size and scope of the federal government.







____________________________
Data Notes: Counts for 2020 and 2023 are from July listings under grantsgov here. Previous years were counts based on hardcopy and electronic versions of the Catalog of Federal Domestic Assistance. The new Assistance Listings keeps the same CFDA numbers. The counts should be considered only a rough measure of federal subsidies since what constitutes one program and one grant is rather loose.
An example of a new grant program is CFDA 20.939 for safe streets enacted in 2021. You can get a sense of the bureaucracy in this one new subsidy program reading the materials here and here. There is no reason for this new federal intervention, as the states themselves are in favor of safe streets and have their own revenue sources.

Chris Edwards


The recent debt‐​ceiling deal put caps on discretionary spending for 2024 and 2025. But House conservatives want larger reforms and are seeking additional cuts in fiscal 2024 appropriations bills. House Appropriations Chair Kay Granger is right that "the debt ceiling bill set a ceiling, not a floor" for 2024 spending.
What should the GOP cut? One place to look is appropriations that are "unauthorized." Federal discretionary programs are supposed to be authorized before appropriators spend, but authorizations have lapsed for many programs and Congress spends anyway. Romina Boccia argues that this rise in "zombie spending" reduces oversight and increases waste.
Consider the Economic Development Administration (EDA), which hands out hundreds of millions of dollars a year in subsidies to state and local governments and the private sector. EDA spending has soared from $264 million in 2019 to $1.53 billion in 2023, as shown in the figure. The EDA budget received a boost from recent legislation including the American Rescue Plan and CHIPS acts.







EDA spending authorization lapsed after 2008. The last serious attempt to reauthorize it was in 2011, which failed in the Senate by a 49 to 51 vote. Supporters at the time touted such things as, "For every dollar spent in EDA, $7 of private investment is attracted," as Senator Barbara Boxer suggested. But opponent Senator Tom Coburn said that such claims were anecdotal and made up by the EDA and grantees themselves. He pointed to the lack of rigorous analysis of EDA spending and opined that the agency "has been used as a congressional slush fund to direct money to friends of Members of Congress."
The EDA was created in 1965 and says its role is to "lead the federal economic development agenda." The agency sprinkles grants widely to congressional districts across the nation in a vast range of activities. EDA leaders make lofty claims about jobs and investment that are not sufficiently challenged by members of Congress, who are eager to secure grants for their districts.
Cato has critiqued the EDA, and federal auditors have been skeptical about some of the agency's claims. The Government Accountability Office testified that the EDA's reliance on grantee estimates "may lead to inaccurate claims about program results." Analysts have long noted the "gulf between promise and performance" at the EDA.
What sorts of things does the EDA fund? The following are some projects highlighted in EDA annual reports from 2018 to 2021 (see here) that fit into two categories: 1) projects that, if useful, should have been funded by the states, and 2) projects that, if useful, should have been funded by the private sector.
Should have been funded by the states:
$8,350,057 to improve the International Parkway Bridge in Tracy, California (2021).
$7,899,000 to make levee upgrades in Hamburg, Iowa (2020).
$5 million to renovate the University of Texas Coastal Ocean Science building (2019).
$1,975,800 to Delaware Technical and Community College for a skills training facility (2018).
$3,806,761 for a new generator at the University of South Alabama (2021).
Should have been funded by the private sector:
$2,546,760 to the Institute for Advanced Learning and Research to purchase equipment for training employees (2021).
$10,214,022 for Leon County Research and Development Authority to build a business incubator in Tallahassee (2020).
$7,872,090 to Blue Lake Rancheria, California, to build a local business incubator (2019).
$2,030,000 to help Cedars‐​Sinai Biomanufacturing Center purchase equipment (2018).
$1,996,160 to purchase equipment for training at the Fairbanks Pipeline Training Center Trust (2021).
The EDA does not have special skills for boosting economic growth that the states and private sector do not have. Federal funding of local industrial parks, incubators, equipment, and other items just adds bureaucracy. Taxpayers pay for the federal EDA bureaucracy of more than 300 employees, and then those folks impose rules and regulations on local partners, which further raises costs. To receive EDA funding, for example, regions must complete central‐​planning‐​style "Comprehensive Economic Development Strategies."
The EDA often claims that it garners high returns on projects, but if true we would not need the agency because the states and private sector would eagerly fund such projects. In other cases, the EDA says that it is filling unmet needs. The head of EDA testified that "without the support and funding of EDA, many projects would struggle to attract necessary capital." But maybe those projects struggle because they have poor management or are economic losers.
Another problem with EDA projects is that they can pit jurisdictions against one another in a zero‐​sum game. In his study of the EDA, David Bier says that the agency gave $2 million to Visalia, California, to expand an industrial park, but then that subsidy induced a medical equipment manufacturer to relocate hundreds of jobs from elsewhere within the state. In such cases, EDA subsidies create winners and losers and probably no net value.
The EDA is one of many programs that spread federal subsidies around the country for local development. But local areas needing development can reform their own taxes and regulations to spur entrepreneurship and private investment. Jurisdictions that create inviting climates for businesses and skilled workers can prosper without subsidies.
Federal funding of local projects is inefficient for many reasons, and it is not affordable given ongoing federal deficits of more than $1.5 trillion a year. Both the Reagan and Trump administrations proposed eliminating the EDA. House Republicans should take another shot at reform.

Chris Edwards


The Biden–GOP debt deal adjusted work requirements in the Supplemental Nutrition Assistance Program (SNAP) but did not cut the program's spending. Cuts are needed because SNAP outlays have exploded from $63 billion in 2019 to an estimated $145 billion in 2023. Congress will have another chance at reform later this year when the program is reauthorized as part of the farm bill.
One chronic failing of SNAP is that billions of dollars are lost through fraud and abuse. Individuals, businesses, and organized gangs steal benefits. The switch from paper food stamps to electronic benefits transfer (EBT) cards two decades ago created new avenues for abuse.
My new article for National Review discusses 10 types of fraud and abuse in SNAP.
Card skimming has become a particularly severe problem in recent years. EBT numbers and PINs are a great target for theft because the cards do not have chips.
SNAP is difficult to police because it includes 250,000 retailers and 42 million recipients, who have changing income levels and other factors that affect eligibility and benefit levels. And because the federal government funds SNAP benefits, state administrators have little reason to minimize the fraud and abuse.

Chris Edwards


Senate Finance Committee Chair Ron Wyden (D‑OR) held a hearing last week to counter the House Republican plan to cut the recent IRS enforcement boost. Sen. Wyden said, "If you're looking for the big winners of the McCarthy IRS defunding plan, it's billionaires and corporations who cheat on their taxes … Repealing that funding is a $191 billion giveaway to wealthy tax cheats."
I offered a different view at the hearing. I noted that tax enforcement imposes collateral damage, that the tax gap has been stable for decades, and that the U.S. tax gap appears to be smaller than Europe's. The "tax gap" means unpaid taxes from errors and cheating.
Here's one problem with Sen. Wyden's view: IRS audits find relatively smaller errors and cheating on higher‐​income returns than lower‐​income returns. As shown in the table below, IRS audits recommend additional tax of 5 to 8 percent of income for middle‐​income households, but just 1 to 4 percent of income for high‐​income households. These are averages within income groups over 2017 to 2021.
For example, for households facing additional tax, the average is $6,100 for those earning $75,000 to $100,000, which is 7.0 percent of income, and it is $117,033 for those earning $5 million to $10 million, which is 1.7 percent of income. Note that the average audit change of $117,033 for this high‐​income group may seem large, but that is just 6 percent of the average tax paid by these households. Also note that the large relative audit change in the bottom group mainly stems from errors and cheating on the earned income tax credit.
Why does the IRS find relatively smaller errors and cheating at the top? It may because these households are more likely to hire expert accountants and lawyers who have their reputations on the line. The relatively smaller errors at the top are impressive, given that high‐​income returns often involve complex issues such as business income and capital gains.







Data Notes. The IRS publishes aggregate results of audits in its annual Data Book. My table is based on a Government Accountability Office summary of the IRS data in GAO-22–104960 (p. 34).
The income groups are based on "total positive income" not adjusted gross income (AGI). However, the IRS and GAO do not appear to provide average total positive income within the income groups, so I roughly estimated it using average AGI for 2019 within AGI income groups.
A final note is that these are the additional taxes "recommended" by IRS auditors. But many taxpayers appeal these amounts and get them reduced. Also, some taxpayers challenge IRS audit results in court and many of them win, as discussed here.