When Congress changes taxes, “tax expenditures” usually come into play. These are generally thought of as breaks, preferences, or loopholes in the tax code that distort the economy and increase complexity. Despite occasional efforts to simplify the code, the number of tax expenditures on one official list has risen from 53 in 1970 to 205 in 2023.

Policymakers should pursue tax reforms to cut tax rates and end preferences, but official tax expenditure lists are not good guides for which preferences to end. The lists are built around a tax base called Haig-Simons income, which is anti-growth and redistributionist. And the lists are biased in ways that make it appear that the tax code favors high earners.

In this policy analysis, I discuss a better way to measure and end tax preferences, which is to start from a consumption base. Such a base would be neutral with respect to saving and investment, unlike the current income tax base. I also identify tax preferences to repeal in moving toward a consumption-based tax system and discuss tax reforms for business investment, personal saving, health care, housing, municipal bonds, and the state and local tax deduction.

Congress should cut tax rates and repeal loopholes, but it needs to make sure that it is repealing actual loopholes and moving toward a more neutral tax base. The reforms proposed here would simplify the tax code, increase fairness, reduce distortions, and promote growth.

Before tax preferences were called “tax expenditures,” they were called “tax loopholes.” Loopholes were originally slits in castle walls used to fire arrows through, but at least two centuries ago the word was being used to describe discrepancies in laws.¹ Talking about a legal proceeding in 1807, Thomas Jefferson said, “What loophole they will find in it, when it comes to trial, we cannot foresee.”² By the 1930s, news and academic articles were using “tax loophole” to mean special breaks in tax laws.³ In recent decades, “tax loophole” is often used interchangeably with “tax break,” “tax preference,” “tax shelter,” and “tax subsidy.”⁴

In this study I will use “loophole” to refer to credits, deductions, exclusions, and exemptions that undermine neutral taxation. Loopholes break the smooth castle wall of the tax base. The $7,500 federal tax credit for electric vehicles (EVs) is a tax loophole. Policymakers should rid the federal tax code of such loopholes, but they first need to define a neutral tax base to measure them against.

Since the 1960s, tax scholars have used “tax expenditure” to refer to tax preferences. The term was coined by a U.S. Treasury Department official and was chosen because tax preferences can distort the economy in similar ways as spending programs. Instead of a tax credit to subsidize electric vehicles, for example, Congress could have subsidized them with a spending program.

The Joint Committee on Taxation (JCT) and the U.S. Treasury both produce annual lists of tax expenditures. I will use “tax expenditure” to refer to these flawed official lists, and “tax loophole” to refer to a subset of these provisions that are preferences against a consumption tax base. In general, these latter provisions should be repealed.

While flawed, the official tax expenditure lists illustrate the rise of tax complexity as Congress has larded up the tax code with special provisions. Figure 1 shows that the number of major tax expenditures increased from 53 in 1970 to 205 by 2023, as measured by the JCT.⁵ The jump in 2010 was partly caused by changes in JCT’s methodology, while the decline after 2010 stemmed partly from temporary provisions that expired.⁶ In addition to these provisions, the JCT lists an additional 66 tax expenditures for 2023 for which quantification is not available or had revenue effects of less than $50 million over five years (these additional provisions are not included in the counts for Figure 1).

The estimated total value of tax expenditures is $1.83 trillion for 2023, including $1.64 trillion for individuals and $187 billion for corporations. These totals are only ballpark measures of the revenue effects, but they may be roughly compared to expected federal income tax revenues of $4.8 trillion in 2023.⁷

This study argues that a subset of official tax expenditures are unjustified loopholes that should be repealed. But which ones? To find out, we need to consider the tax base—that is, what should be taxed—and distinguish between two different conceptions of that base: Haig-Simons income and consumption. Official tax expenditures are defined against an adjusted Haig-Simons base. This is an excessively broad tax base that distorts saving and investment, but it is favored by economists and policymakers on the political left because it is redistributionist.

The alternative tax base is consumption, which is simpler and neutral in its treatment of saving and investment. A consumption base is a better starting point to identify unjustified tax preferences, and a better model to guide tax reforms. The current federal “income” tax is actually a hybrid, part Haig-Simons and part consumption, and this study argues that Congress should move toward the latter.

The 2017 Tax Cuts and Jobs Act (TCJA) took steps toward simplifying the tax code and moving toward a consumption base.⁸ But then the American Rescue Plan of 2021 and the Inflation Reduction Act of 2022 moved against reform by adding and expanding tax loopholes for energy production and other activities.

After discussing the tax base and measuring tax expenditures, in the final section of this policy analysis I will identify tax loopholes to repeal in exchange for lower tax rates and discuss the taxation of business investment, personal saving, health care, housing, municipal bonds, and the deduction for state and local taxes. The proposed reforms would simplify the tax code, increase fairness, reduce distortions, and promote growth.

If you want to find loopholes in the income tax, you need to identify a consistent and neutral measure of income to judge tax provisions against. The 16th Amendment to the U.S. Constitution in 1913 allowed “taxes on incomes, from whatever source derived,” but it did not define how income should be measured. Over the past century, tax experts, economists, and policymakers have never had a fixed or unified view of the matter. As a result, the federal tax code has gyrated back and forth as policymakers moved in different directions over the decades.

Because of differing views on the proper federal tax base, there has been continuing debate over which provisions in the tax code are unjustified preferences. The official tax expenditure lists published by the JCT and the U.S. Treasury do not represent a consensus view. Indeed, far from it. The Haig-Simons income tax base underlying the official tax expenditure lists is deeply flawed and inferior to a consumption tax base.

Haig-Simons income has been the dominant conception of the ideal federal tax base. In the early 20th century, economists Robert Haig and Henry Simons proposed a tax base that includes labor income and a very broad measure of capital income. Capital income is the return to saving, which is earned as people delay consumption and channel their funds toward investment projects. To guide policymakers on the new federal income tax, Haig described what is now called the Haig-Simons tax base in a 1921 article.⁹ Simons advocated for this tax base in his 1938 book Personal Income Taxation.¹⁰

All income taxes tax the flow of returns from current production, but a Haig-Simons tax goes further and taxes all net gains in asset values, and does so on an accrual or mark-to-market basis. That is, the tax base includes the annual net change in the value of all assets owned, whether those assets are sold or not. (The Haig-Simons tax base is defined as consumption plus the net change in the value of all assets owned during the year, including all accrued gains whether realized or not.) If a person earned $50,000 in wages and their house increased in value $30,000 during a year, Haig-Simons “income” would be $80,000, even if that person did not sell the house. A tax with this base is often called a “comprehensive income tax,” but it is actually a tax on income plus net accrued capital gains.

Including capital gains as income to be taxed is a debatable idea. Capital gains are not included as income in the National Income and Product Accounts (NIPA) produced by the Department of Commerce. Within these accounts, income includes the returns to labor and capital from current production but does not include changes in asset values. Capital gains are not current earnings but rather the present value of expected increases in future earnings.

Furthermore, the idea of taxing unrealized capital gains, as under a Haig-Simons tax, is highly controversial. That would tax expected increases in future earnings that may or may not ever materialize. Unrealized capital gains are not a component of any of 11 different measures of income currently used by various federal agencies, and unrealized gains have been excluded from income since the first modern income tax law of 1913.¹¹

The proper tax treatment of capital gains has been debated since the beginning.¹² Examining debates in the early 1920s, tax law professor Marjorie Kornhauser found that “economists held widely divergent views on whether a capital increase realized from a casual, nonbusiness sale was income.”¹³ In 1921, the New York Times printed numerous articles and editorials against taxing capital gains, arguing that “capital” and “income” are distinct items.¹⁴ Indeed, capital is a stock and income is a flow. Economics professor Fred Rogers Fairchild of Yale University observed in 1921 that “the weight of economic authority supports the theory that mere growth in value of capital is not income.”¹⁵

Nonetheless, the federal government decided to tax realized capital gains, and the Supreme Court agreed in 1921.¹⁶ In other nations, it took a while for this Haig-Simons view to dominate on the issue. Other nations had income taxes, but for decades did not tax long-term capital gains. The years of adoption of taxes on long-term capital gains were: United Kingdom (1965); Canada (1972); Ireland (1975); Australia (1985); and Japan (1988). Numerous high-income countries still do not tax long-term gains.¹⁷

There is no compelling reason why Haig-Simons must be the measure of income used for taxation. Since Haig, supporters have believed that the broad base captures the “economic power” of individuals and businesses. Haig claimed that his tax base, guided by “economics and equity,” should include the “net accretion to one’s economic power between two points in time.”¹⁸ Simons thought that income should capture “the exercise of control over the use of society’s scarce resources.”¹⁹

For Simons, the goal of the proposed tax base was redistribution. His 1938 book commented favorably on the view of another expert that “taxation must be conceived as an instrumentality for altering or correcting the distribution of wealth and income.”²⁰ And his book argued that the “case for drastic progression in taxation must be rested on the case against inequality.”²¹ Looking at the history of Haig-Simons, economist Michael Schuyler noted, “Simons believed in aggressive income redistribution through taxation and thought a progressive-rate income tax, using his definition of income, was well suited to that end.”²²

Simons’ belief in “drastic progression” and his belief that individual earnings are “society’s resources” continue to undergird support of the Haig-Simons tax base today. Tax experts and policymakers on the political left are attracted to Haig-Simons because they view heavy taxation of high earners as beneficial. But that is a different goal than choosing the simplest and most neutral tax base, or the tax base that causes the least damage to the economy.

Despite the appeal of Haig-Simons to the left, Congress has always recognized that such a tax base must be modified for practical use. For one thing, taxpayers with little cash-flow cannot afford to pay an annual tax on their accrued capital gains, and thus gains have always been taxed on realization. Also, Congress has long recognized that capital gains should have lower effective tax rates than ordinary income to account for inflation and to mitigate the damage that high tax rates would impose on growth industries.

Angel investors and venture capitalists invest in risky startup businesses.²³ The value of such startups fluctuates over time, and half of them go bankrupt within five years. Haig-Simons would tax the change in value of such investments annually, even though many ventures do not survive, and those that do may not earn profits for many years. A Haig-Simons tax would severely undermine such investments and thus damage economic growth. Congress has recognized this and adopted a provision that exempts qualified investments in startups from capital gains taxes (Internal Revenue Code Section 1202). But the need for this exemption should be a strong hint to policymakers that taxing capital gains within a Haig-Simons framework is bad idea to begin with.

Another dubious component of a Haig-Simons tax base is net imputed rental income on owner-occupied homes. This is phantom rental income that one earns by simply owning one’s home.²⁴ Because Congress does not currently tax this phantom income, it is considered a tax expenditure by the U.S. Treasury. Similarly, a Haig-Simons tax base would include net imputed income on consumer durables such as refrigerators, couches, and automobiles. In their book on tax expenditures, Haig-Simons supporters Stanley Surrey and Paul McDaniel said that not taxing the imputed income from consumer durables is a “tax subsidy.”²⁵ Similarly, prominent fiscal economist Richard Goode thought that it is “discrimination” that imputed income on a home washing machine is not taxed, but income from commercial laundries is taxed.²⁶ Despite what some progressive theorists have said, imputed income from housing and consumer durables has never been taxed—both because it would be impractical and because people would view it as a bizarre imposition by the government.

Supporters of Haig-Simons view the tax base as “theoretically pure” and “ideal.”²⁷ Haig-Simons is said by many to be the “gold standard” for measuring income.²⁸ But if a tax base includes taxing phantom imputed income on home refrigerators and taxing “income” today from startups that may never generate profits, it is not ideal or pure at all.²⁹ It was also clear from the beginning that taxing broad-based income would be a complex endeavor.³⁰

Haig claimed that his definition of income was “scientific,” and he refers to it as “true income” and “economic income,” as if all economists agreed on it.³¹ But they did not, and Haig admitted that tax systems in Europe were based on different definitions of income, with numerous countries not taxing capital gains, which are central to the Haig-Simons tax. “Both the British and German statutes construct a concept [of income] much more narrow than ours,” he said.³²

Why is Haig-Simons idealized even with all the flaws? The answer is that many lawyers, economists, and policymakers are wedded to the structure for its redistribution potential. They seem fixated on attacking what they view as “economic power” with a steeply progressive tax code.

While Haig-Simons may seem like an obscure concept, it continues to have a large and often subterranean influence on tax policy. Senate Finance Committee chair Ron Wyden (D‑OR), for example, has proposed taxing capital gains on an accrual basis, meaning taxing unrealized gains.³³ But that would be so impractical that no other major nation taxes gains that way.³⁴ Similarly, the Biden administration has proposed a Billionaire Minimum Income Tax, which would “ensure that the very wealthiest Americans pay a tax rate of at least 20 percent on their full income, including unrealized appreciation.”³⁵ The ideology of Haig-Simons appears to have steered Biden, Wyden, and their advisers astray.

The same misguided ideology underlies a series of articles in ProPublica in 2021, which claimed that tax rates on high earners are exceedingly low.³⁶ The articles were based on the idea that tax rate calculations should include unrealized capital gains as income. The Biden White House has also published data claiming that high earners pay low tax rates, but this is based on faulty measures of income that include unrealized gains.³⁷

Many economists and tax experts have recognized the shortcomings of Haig-Simons and proposed an alternative federal tax base: consumption. Retail sales taxes and value-added taxes are examples of consumption-based taxes. But economists have also proposed consumption-based taxes that are collected from individuals and businesses in a manner similar to the current federal income tax. The Hall-Rabushka flat tax proposal is an example.

Consumption-based taxes have simpler and more neutral tax bases than income taxes. While Haig-Simons double-taxes saving and investment, consumption-based taxes tax all income that is consumed once. The current federal income tax base is actually a hybrid of Haig-Simons and consumption. It is a compromise resulting from decades of debate between tax experts and policymakers favoring each approach. The basic structure of the federal income tax is Haig-Simons, but it includes features that alleviate the damaging treatment of saving and investment that would occur under a pure Haig-Simons tax.

Under an income tax, when you use your earnings for immediate consumption, there is no further tax, but when you put aside earnings for the future you are taxed on the return to the saving. As a result, income taxes favor consumption today over consumption tomorrow, and because interest compounds, the bias gets worse the further in the future that “tomorrow” is. That is particularly worrisome given Americans’ low rate of saving. Income taxes favor immediate gratification over the building of financial security.

The problem with the income tax is often described as the double taxation of savings. A person earns wages, pays taxes on them, and puts aside some of those earnings in savings. Those funds are hit by both the wage tax and the tax on the savings. Saving for the longer term, such as for retirement, is hit the hardest by income taxes. This bias against saving has important implications for economic growth.

Think of taxing an apple farmer. To maximize the farmer’s harvest over the long run, it is better to tax a share of the annual harvest, not growth in the apple trees. We should tax the flow of consumption produced by the capital asset, not the capital asset itself, which is needed to produce future consumption. A consumption-based tax would tax just the apples harvested, while a Haig-Simons tax would also tax annual growth in the tree, which over time would reduce harvests and make society worse off.

The superiority of consumption taxation over income taxation has been understood for a long time. In 1884, John Stuart Mill observed, “Unless, therefore, savings are exempted from income-tax, the contributors are twice taxed on what they save, and only once on what they spend.”³⁸ So Mill decided, “the proper mode of assessing an income-tax would be to tax only the part of income devoted to expenditure, exempting that which is saved.”³⁹

When the federal income tax was imposed a century ago, an early critic of using Haig-Simons for the tax base was Yale University’s Irving Fisher, one of the leading economists of the day.⁴⁰ He said that income is best measured by the flow of the economy’s output from labor and capital that is consumed. It should not include changes in the value of the stock of capital (capital gains) nor net additions to the stock of capital (savings). Fisher argued that Haig-Simons income erroneously mixed current income with additions to capital, which would cause serious damage as a tax base. Fisher called his preferred tax base “real income” or “yield income,” but today it is called a consumption base.

The misguided taxation of capital under Haig-Simons can be seen by considering human capital. As young people build skills, they gain human capital, which allows them to earn higher incomes and contribute more to society. Fisher noted that a Haig-Simons tax “includes a tax on growing capital value or earning power. To be logical, it should include a tax on the yearly increase in capital value of the personal earning power of a young man.”⁴¹ If a college degree boosts a person’s future earning power, the Haig-Simons logic is to tax the present value of those future earnings now. Such a tax would clearly penalize progress, and so we do not tax human capital the Haig-Simons way. Yet, unfortunately, that is how we tax physical capital in the federal tax code.

Over the decades, many tax experts have come down on the side of consumption-based taxation. In a 1956 book, leading British economist Nicholas Kaldor criticized the Haig-Simons approach for double-taxing saving and being difficult to administer.⁴² In a 1974 study, Harvard law professor William Andrews concluded, “A consumption-type or cash flow personal income tax would represent an incomparably simpler tax to administer” than an income tax.⁴³

One of Washington’s top tax experts from the 1950s to the 1990s, Norman Ture, advocated reforms to remove the double taxation of saving and investment.⁴⁴ In a 1974 essay, he proposed transitioning to a roughly 23 percent flat tax on a savings-neutral base with a large personal exemption but no other preferences.⁴⁵

A landmark U.S. Treasury study in 1977 compared Haig-Simons and consumption as alternative tax bases.⁴⁶ The study found, “In some respects, a broad-based consumption tax is more equitable than a broad-based income tax. It is also easier to design and implement and has fewer harmful disincentive effects on private economic activity.”⁴⁷ That study was led by Treasury official David Bradford, who, as a Princeton University professor, wrote many tracts advocating replacing the income tax with a consumption-based tax.

In the 1980s, economists Robert Hall and Alvin Rabushka proposed their “flat tax,” which would tax income once with no preferences or loopholes. The plan would tax labor income at the individual level and capital income at the business level at a 19 percent rate.⁴⁸ Hall-Rabushka has a consumption base, and thus creates a neutral treatment for saving and investment. Bradford proposed a two-rate version of the Hall-Rabushka tax, which he called an X‑Tax.⁴⁹

To see that a Hall-Rabushka tax has a consumption base, consider a retail sales tax that covers all final goods and services. Such a sales tax is economically equivalent to a value-added tax (VAT) collected from all businesses in the production chain, but with each subtracting its purchases from other businesses. The effect is to tax each business on its value-added, which ends up collecting the same overall revenue as a tax on final retail sales.

You can transform a value-added tax to a Hall-Rabushka tax by allowing businesses to deduct wages and then taxing the wages at the individual level. So retail sales taxes, value-added taxes, and Hall-Rabushka are economically very similar. They all tax capital and labor once.⁵⁰ Hall-Rabushka would work like our current income tax with individuals and businesses filing tax returns, but the returns would be simpler and the tax structure less harmful to the economy. In the final section of this study I discuss reforms to move in the direction of Hall-Rabushka.

Next, let’s look at three advantages of consumption-based taxes over income taxes: economic growth, simplification, and fairness.

Economic growth would be maximized under a tax system that is neutral between industries and economic activities, allowing resources to flow to the highest-valued uses. An important aspect of neutrality is equal treatment between consumption and saving, and in that regard consumption-based taxes generate at least three types of benefits.

First, consumption-based taxes remove tax penalties on saving and investment. For individuals, that means increased incentives to save for retirement and other future needs. For businesses, that means increased incentives to invest in a larger capital stock, which over time generates higher productivity and worker incomes. Investments in buildings and equipment would be written off immediately, or expensed, rather than being depreciated over time. The effect would be to remove taxes on the normal returns to marginal investments, allowing investment to be optimized.

Second, it is easier to equalize marginal effective tax rates (METRs) across types of investment under a consumption-based tax than it is under an income tax. Marginal effective tax rates are the tax rates on additional units of investment, and they drive investment flows. If rates are unequal, resources get misdirected in the economy and growth suffers. Bradford argued that an income tax “makes virtually inevitable the variety of effective tax rates that are actually applied to different assets.”⁵¹ Indeed, before the TCJA, the Congressional Budget Office (CBO) estimated that “almost every combination of asset type, industry, form of organization, and source of financing yields a different [marginal] ETR.”⁵² For example, METRs for corporate investment ranged from 12 percent for railroad track to 42 percent for nuclear fuel.⁵³

The TCJA took steps toward a consumption base by cutting the corporate tax rate and implementing expensing for equipment. The law also limited business interest deductions. The changes narrowed tax differences between types of equipment investment, between debt and equity financing, and between corporate and noncorporate investments.⁵⁴ For corporations, the CBO found that METRs before the law averaged 34 percent (equity) and ‑23 percent (debt), but after the law rates averaged 22 percent and 9 percent, respectively.⁵⁵ Bradford viewed such narrowing of marginal tax rate differences as the most important economic advantage of moving toward a consumption base.⁵⁶

Third, consumption-based taxes remove tax barriers to innovation, which is crucial because innovation is the largest factor in generating rising living standards.⁵⁷ One channel of innovation is capital investment. Shifting to consumption taxation would boost business investment, as noted. That would mean not just more machines, but better machines with new technologies. Capital investment does not just build the capital stock, it also modernizes the economy.

Another channel of innovation is business research, and consumption-based taxes are superior on this front as well. Research spending is expensed under a consumption tax, but it is amortized and deducted over time under an income tax, which raises effective tax rates and undermines spending. In a misguided Haig-Simons move, the TCJA took a step backward and replaced the expensing of research with amortization over five years.

A final channel of innovation is competition from startup businesses. Many major innovations have come from startups, not existing large corporations.⁵⁸ Apple pioneered personal computers in the 1970s, not giant IBM. Like Apple, fast-growing startups today are usually financed by angel investors and venture capitalists, who take large risks in seeking long-run returns in the form of capital gains.

Such saving and investment is not deterred by consumption-based taxation, but would be seriously undermined by heavy capital gains taxes under Haig-Simons taxation.⁵⁹ This issue is crucial because fast-growing startups generate new competition for large companies, and they often become large companies themselves. More than half of U.S. stock market capitalization and more than half of all industrial research is performed by companies that were originally backed by angel and venture investment.⁶⁰

When companies innovate—through research, investment in technologies, and competing in new ways against incumbents—they can generate large spillover benefits for the economy. That is, many other firms will adopt the new approaches, thus creating broad economic benefits beyond just higher returns for the initial innovating company. When lower taxes spawn more investment, research, and startup activity, the benefits can spill across the economy and increase overall productivity.

How large are such spillovers from innovations? Economist William Nordhaus explored the question by modeling U.S. business profits and productivity over time. He concluded that “only a miniscule fraction of the social returns from technological advances over the 1948–2001 period was captured by producers, indicating that most of the benefits of technological change are passed on to consumers rather than captured by producers.”⁶¹ He found that businesses received only about 2 percent of the benefits from their innovations, with the rest accruing broadly to consumers.

As one example, mRNA vaccines were developed after a decade of research by Moderna and BioNTech, which was funded by a few billions of dollars of private angel and venture investment.⁶² Aside from vaccines, the underlying mRNA technologies may ultimately generate huge benefits in fighting cancer and other diseases.⁶³ The heavy taxation of capital gains under Haig-Simons would kill the incentives for the investors who fund such private research.

The Wall Street Journal recently profiled Illumina Inc., which has led the way in slashing the cost of sequencing a human genome from $10,000 in 2010, to $1,000 in 2014, to $600 today.⁶⁴ Those cost reductions are opening vast possibilities for medicine and health care. Sequencing has already “led to genetically targeted drugs, blood tests that can detect cancer early, and diagnoses for people with rare diseases who have long sought answers.”⁶⁵ Illumina was founded in 1998 in San Diego and was initially funded by $8.5 million in venture capital before raising equity funding in public markets.⁶⁶ Today, competitors of Illumina—funded by hundreds of millions of dollars of venture capital—are hoping to push the costs of sequencing lower, to about $100.⁶⁷ Such funding for pathbreaking innovations would dry up if risk taking investors did not have the chance to earn substantial after-tax capital gains.

Supreme Court Justice John Marshall said, “The power to tax is the power to destroy.”⁶⁸ Irving Fisher and his brother Herbert Fisher noted, “This power to destroy is many times greater when savings are taxed than when merely spendings are taxed.”⁶⁹ It is saving, they said, that fuels innovations such as railways, automobiles, and radios, and in doing so generates our rising standard of living. As such, they argued that overtaxing saving and investment is “killing the goose that lays the golden egg.”⁷⁰

In 1976, president-to-be Jimmy Carter said, “It is time for a complete overhaul of our income tax system .… It is disgrace to the human race.”⁷¹ Since his complaint, the number of pages of federal tax rules has more than tripled because Congress has continued to add provisions and preferences.⁷² Such tax complexity raises compliance costs, increases errors, and promotes tax evasion. Federal tax compliance costs the economy more than $300 billion annually.⁷³

Both income and consumption-based taxes can be complex if policymakers lard them with special preferences, such as the earned-income tax credit (EITC) and low-income housing tax credit (LIHTC). The EITC consumes inordinate IRS resources to administer and has an error and fraud rate of more than 20 percent. The LIHTC is so complicated that the IRS auditing guide is 350 pages long and a guide for businesses taking the credit is 1,790 pages long.⁷⁴

However, it is also true that income taxation is more inherently complex than consumption-based taxation. Bradford noted that “a great many of the most severe problems of measurement in the income tax fall away in a consumption tax, while the latter adds virtually no new ones.”⁷⁵ Consumption-based taxation would do away with depreciation accounting, inventory accounting, and capital gains taxation, which are some of the most complex features of the current tax code.⁷⁶ Accrual accounting under the income tax would be replaced by simpler cash accounting under a consumption-based tax.

Income taxes need complicated fixes to reduce the economic damage they cause. For example, inflation biases income taxes against longer-lived capital investments, a problem that can be fixed but would require a slew of complicated rules. By contrast, inflation is not a problem for consumption-based taxes because investments are expensed. As Bradford noted, it is “very difficult to design rules” for an income tax, and the rules need “continual patching.”⁷⁷

Income taxes need continual patching because taxpayers exploit the underlying complexity. That was the story of Enron Corporation’s infamous tax-avoidance schemes, which required a 2,700-page JCT report to unravel.⁷⁸ The report concluded that the company “excelled at making complexity an ally.”⁷⁹ Enron designed elaborate transactions to exploit depreciation, capital gains, and other features of income taxation in order to minimize its taxes.

It is true that a full-fledged consumption-based tax would share some of the complexities of income taxation, while also creating some new problems. A Hall-Rabushka tax, for example, would face challenges dealing with financial services, small businesses, and business losses.⁸⁰ Nonetheless, consumption-based taxation would appear to eliminate more complexities than any new ones it created.⁸¹ A good strategy for Congress would be to move toward a consumption base with steps that both support growth and simplify the tax code.

Many experts agree that a consumption-based tax would be simpler and less harmful to growth than an income tax, but they still favor the latter for fairness reasons. Left-leaning experts and policymakers believe that fairness requires imposing heavy taxes on high earners, and since high earners often have high saving, an income tax based on Haig-Simons would seem to be a good approach. Economist Stephen Entin concluded that “income redistribution was the main justification for the Haig-Simons definition of income.”⁸²

However, many experts do not accept that Haig-Simons taxation is fairer than consumption-based taxation. Indeed, Bradford argued that a “principal argument in favor of a consumption approach is, rather, one of equity.”⁸³ There are at least three reasons why Bradford and other economists believe consumption-based taxation is favorable regarding fairness or equity.

First, if fairness means that a tax should have a progressive rate structure, then a consumption-based tax can be designed in that manner. That was the idea behind Bradford’s X‑Tax, which has a consumption base and multiple rates. Even the single rate Hall-Rabushka tax has a large standard deduction so that low earners would pay little, if any, tax.

Second, because consumption-based taxes are simpler than income taxes, it is more likely that similar individuals and businesses will pay similar taxes, which is called horizontal equity. Income taxes require many jury-rigged features, which can cause a greater dispersion of tax liabilities.

Third, consumption-based taxes fall equally on the spendthrift and the frugal, but income taxes fall harder on the latter. Consider two brothers with equal earnings. An income tax favors the spendthrift brother who blows his money shopping and discriminates against the frugal brother who saves for retirement. Yet by saving, the frugal brother is the one who funds investment, innovation, and growth in the economy, which benefits all of us. Also, because he has savings, the frugal brother is less likely to become dependent on government programs, and thus unfairly impose costs on the rest of us.

Consider a family that goes from rags to riches to rags over a number of generations. A Haig-Simons tax would fall heavily on the earlier harder-working generations, but lighter on the later spendthrift generations. A consumption-based tax would do the opposite, thus supporting the socially beneficial behavior of early generations while taxing the “idle rich” of later generations.⁸⁴

Sometimes it is said that consumption-based taxes do not tax capital income, which would seem unfair, but that is not the case. Economist Glenn Hubbard noted that “consumption and income taxes actually treat similarly much of what is called capital income.”⁸⁵ Capital income can be divided into four components: the risk-free return to waiting (the time value of money); the return to risk; an inflation premium; and above-normal or inframarginal returns, also called rent.⁸⁶ The treatment of the latter three items is thought to be the same under income and consumption-based taxes, although income taxes usually don’t fully index for inflation.

The main difference is that the first item, the return to waiting, is taxed by income taxes and not by consumption-based taxes. That is a small slice of capital income. Bradford noted that “the difference between income and consumption taxes is the treatment of the risk-free reward to waiting, certainly below 2 percent per year.”⁸⁷ But that small slice makes a big difference in terms of structuring a tax system that does not distort saving and investment.

Policymakers concerned about fairness should note that consumption-based taxes fully tax above-normal returns, which means high profits from market power, windfalls, and other unique profits that are available only to certain businesses and investors.⁸⁸ Monopolies and particularly successful technology firms, for example, have their exceptional profits taxed under both income and consumption-based taxes.⁸⁹ Hubbard concludes that the claim “that consumption tax reform is a sop to the rich is almost certainly unfair.”⁹⁰

U.S. Treasury Assistant Secretary Stanley Surrey and his staff published the first list of tax expenditures in 1968. Surrey and McDaniel noted that tax preferences resembled spending programs, even claiming that “a tax expenditure is a spending program.”⁹¹ Surrey was an advocate of Haig-Simons income as the federal tax base, and so that base was the starting point for the Treasury’s tax expenditure lists.

The Congressional Budget Act of 1974 mandated that the administration prepare a list of tax expenditures, a list that is prepared by the U.S. Treasury and included in the annual federal budget. Meanwhile, the congressional Joint Committee on Taxation began publishing its own, somewhat different, list of tax expenditures in 1972.

The Treasury and JCT lists use versions of Haig-Simons income for measuring tax expenditures. Both agencies recognize that taxing a pure Haig-Simons base is unrealistic, so they use a modified version called a “normal” tax base. Because of the Haig-Simons starting point and the many ad hoc features of these normal tax bases, both official lists are highly flawed. As such, they do not provide good guides for steering tax reforms.

The 1974 Budget Act defined tax expenditures as “revenue losses attributable to provisions of the federal tax laws which allow a special exclusion, exemption, or deduction from gross income or which provide a special credit, a preferential rate of tax, or a deferral of tax liability.”⁹² But “special” and “preferential” compared to what? Compared to the “normal” tax bases that Treasury and JCT have somewhat arbitrarily defined. As economists Rosanne Altshuler and Robert Dietz noted, defining what is a “normal” tax base is “inherently a subjective exercise.”⁹³ Indeed, the JCT admitted that its normal baseline results “from a series of ultimately idiosyncratic or pragmatic choices.”⁹⁴

Criticisms of the inconsistencies of official tax expenditures began soon after the lists were first published.⁹⁵ One problem is that the lists include provisions that cause undertaxation, but they do not include major provisions that cause overtaxation.⁹⁶ These items have been called tax penalties, tax surcharges, or negative tax expenditures. For example, the lists consider the current double taxation of corporate equity to be normal, even though the treatment clearly overtaxes and distorts. Another example of overtaxation ignored by the tax expenditure lists is the tax code’s lack of inflation indexing for capital gains.

The Treasury and JCT lists are biased in the selection of provisions included and excluded. For the corporate income tax, the lists treated reduced tax rates for small companies prior to the TCJA as tax expenditures. But for the individual income tax, the treatment is different. The lists do not consider reduced tax rates at the bottom end as tax expenditures. Rather, the current highly progressive tax-rate structure is considered “normal,” and not a special preference for lower-income households.

Similarly, the standard deduction is not considered a tax expenditure, even though it creates preferential treatment, particularly for lower-income households. The lists also assume that the double taxation of saving and investment is normal, as discussed below. These features of the official lists reveal a left-of-center ideological bias that accepts that the income tax should be highly redistributive.

The JCT examined its tax expenditure methods in a 2008 study and admitted that the

Former Treasury official J. D. Foster agreed, saying that the official tax expenditure reports “are simply too fundamentally flawed to serve as guides” for tax reform.⁹⁸ Nonetheless, the JCT and Treasury continue publishing their flawed, but influential, lists.

The main flaw in the official tax expenditure lists is that the modified Haig-Simons baselines accept the double taxation of saving and investment as normal. Provisions that reduce double taxation—such as 401(k) plans—are deemed tax expenditures. That approach signals that provisions such as 401(k)s are misguided loopholes, but the opposite is true since they reduce the anti-saving bias of the income tax.

The anti-saving and anti-investment approach of the tax expenditure lists is a flaw that has been recognized a long time. Economist Norman Ture argued in 1991:

The George W. Bush administration tried to reform the tax expenditure lists. Its fiscal 2002 budget noted, regarding the assumed normal tax base, “Because of the breadth of this arbitrary tax base, the administration believes that the concept of ‘tax expenditure’ is of questionable analytic value.”¹⁰⁰ To begin correcting the problem, the administration presented a separate tax expenditure list for a consumption base.¹⁰¹ From this perspective, the analysis found that provisions that reduced the double taxation of saving, such as 401(k)s, were not tax expenditures.

The JCT says that its tax expenditure list can help determine the relative merits of provisions, but that “no judgment is made, nor any implication intended, about the desirability of any special tax provision as a matter of public policy.”¹⁰² Despite the disclaimer, the official tax expenditure lists have a powerful effect on narratives about tax policy. Economist Bruce Bartlett noted that the official approach “reinforces the supposed superiority of an income base and is a barrier to adoption of a consumption-based system.”¹⁰³

Table 1 lists the largest items on the current JCT tax expenditure list with dollar values for 2023. The dollar values are the estimated reductions in revenues from the provisions.¹⁰⁴ The top part of the table shows 10 provisions that are tax expenditures under both the current JCT method and a consumption base; many of these items should be repealed.

The bottom part of Table 1 shows 10 provisions that are on the JCT list but are not loopholes under a consumption-based tax, as they provide relief from the double taxation of saving and investment. These items should be generally retained, although many of them should be simplified. There is ambiguity in how some of the tax expenditure items in the table should be classified.¹⁰⁵

Figure 2 presents a dollar breakdown of all 205 provisions on the JCT list for 2023.¹⁰⁶ About 48 percent of tax expenditure dollars are for saving and investment relief provisions, which are not loopholes under a consumption-based tax. About 45 percent are true loopholes for individuals and 7 percent are true loopholes for corporations.¹⁰⁷

People often say that Congress should cut tax rates and “broaden the base.” But that is not correct because the base can be broadened in ways that increase distortions. If policymakers were guided by the JCT list and repealed 401(k)s, for example, that would increase economic distortions. Tax reform should be about cutting rates and moving toward a more neutral tax base.

The false notion of “broadening the base” often steers reformers astray. In 2010, the Simpson-Bowles fiscal commission proposed tax reforms to reduce deficits. But the commission mistakenly assumed that the official tax expenditure lists were all unjustified loopholes. The commission’s report declared, “The current tax code is riddled with $1.1 trillion of tax expenditures: backdoor spending hidden in the tax code.”¹⁰⁸ Its basic reform option was to “eliminate all income tax expenditures,” although another option identified some provisions to retain.

The same false notion of broadening the base played a large role in the landmark Tax Reform Act of 1986 (TRA86). Reporters Jeffrey Birnbaum and Alan Murray wrote about the law in their book Showdown at Gucci Gulch.¹⁰⁹ Their story, and the story of TRA86, is of repealing loopholes and cutting tax rates. But Birnbaum and Murray appeared to assume that all official tax expenditures are unjustified breaks. To them, accelerated depreciation, reduced capital gains taxes, and Individual Retirement Accounts are special-interest preferences, schemes, and shelters. But these widely available provisions reduce the code’s anti-saving bias and reduce distortions, not increase them.

Based on a misguided understanding of tax loopholes, TRA86 moved the tax code away from a consumption base. The tax rate cuts under TRA86 were impressive, and the act eliminated some actual loopholes. But TRA86 also made changes that raised taxes on saving and investment, and it made the tax code more complicated by moving toward Haig-Simons.¹¹⁰

In the years after TRA86, Congress reversed course on the law’s anti-saving and anti-investment features. This change in direction after TRA86—such as the capital gains tax cut of 1997—suggests the lack of durability of a Haig-Simons base. Most recently, the Tax Cuts and Jobs Act of 2017 moved toward a consumption base with its embrace of capital expensing for business equipment. Nonetheless, Haig-Simons retains support on the political left because of its redistribution potential.

Analysts on the left use the official tax expenditure lists to complain that high earners enjoy most federal tax breaks or loopholes. Many provisions used by high earners—such as 401(k)s—are on the lists but are not real loopholes, while other breaks important to low earners are not on the lists, such as the standard deduction and the progressive rate structure. These factors bake into the cake the notion that high earners have an unfair advantage in the tax code, and analysts on the left relentlessly push that false message.

The originator of tax expenditures, Surry, along with McDaniel, complained that tax expenditures are “upside down” because the top 1.4 percent received 31 percent of the “subsidies.”¹¹¹ More recently, the Center on Budget and Policy Priorities claimed that “spending through the tax code skews towards the top” and found that 59 percent of the “subsidies” go to the top one-fifth of households.¹¹² Tax expenditures are “upside down,” the group complained.

Others come to the same conclusion. Economist Bill Gale at the Brookings Institution said, “High income households are more likely to use tax expenditures, creating ‘upside-down’ subsidies that disproportionately benefit the well-off.”¹¹³ The Committee for a Responsible Federal Budget wants to cut tax expenditures, which are “costly and regressive” and “skew toward high-income households.”¹¹⁴

However, it is a false narrative to imply that high earners benefit more than others from the tax code. It is based on the arbitrary choice of items in the official tax expenditure lists. If the Treasury and JCT assumed that a proportional tax structure was “normal,” rather than a highly progressive structure, then lower-income households would be shown to receive a massive tax expenditure.

Economists Altshuler and Dietz studied this issue.¹¹⁵ They performed calculations for 2005 assuming that income tax rates below the top rate were tax expenditures. They found that the dollar value of the lower rates on lower- and middle-income households was huge, almost exceeding the value of all other tax expenditures combined. Thus, treating today’s progressive tax structure as “normal” is a critical bias that makes assertions about the overall fairness of tax expenditures meaningless.

Fairness can be better judged by the distribution of overall federal tax payments. Figures 3 and 4 show CBO data for 2019 indicating that high earners are penalized by the tax code, not favored. Looking at individual income taxes, the top fifth of households paid an average effective tax rate of 15.4 percent, compared to 2.4 percent for the middle fifth and −11.1 percent for the bottom fifth, as shown in Figure 3.¹¹⁶ The rates for the bottom quintiles are negative because a number of federal tax credits are refundable, meaning they provide payments to households that pay no income tax. Looking at all federal taxes, the top fifth paid an average tax rate of 24.4 percent, compared to 13.0 percent for the middle fifth and 0.5 percent for the bottom fifth, as shown in Figure 4.

A final note on fairness regards the interplay of tax rates and tax expenditures. As tax rates rise, the dollar value of many tax expenditures increases. So if Congress raised tax rates on high earners, ironically, analysts on the left could then complain that tax expenditures were even more unfair because provisions taken by top households would have even higher dollar values.¹¹⁷

In sum, here are five things to remember about tax expenditures:

• Official tax expenditure lists do not provide good guides for policymakers for pursuing tax reform.
• Almost half the dollar value of JCT tax expenditures are provisions that relieve the double taxation of saving and investment. These provisions are not loopholes under a consumption tax base.
• The actual federal tax base is a hybrid of Haig-Simons and consumption, so the Treasury and JCT should present two tax expenditure lists, one based on each model.¹¹⁸
• Tax expenditure analyses should include a full discussion of tax surcharges or penalties, such as the double taxation of corporate equity.
• Because of the ideological bias in the official lists, claims that tax expenditures favor high-earners are off-base. The more important fact is that average federal tax rates rise sharply as income rises.

To better guide lawmakers on tax reforms, the Treasury and JCT should fix their tax expenditure presentations. Based on fuller and less-biased information, Congress should pursue major tax reforms.

Congress should transition the federal tax code to a structure with lower rates and a simpler base. The base should not double-tax saving and investment, nor have preferences for particular industries or groups of taxpayers. The tax code should interfere as little as possible with household and business choices.

Table 2 shows reforms that would move in that direction.¹¹⁹ In this section I discuss repealing loopholes and cutting tax rates; the tax treatment of business investment, personal saving, health care, housing, municipal bonds; and the state and local tax deduction.

Congressman Richard Gephardt (D‑MO), a tax policy leader in the 1980s, said that he favored closing special-interest breaks to improve efficiency. In a 1985 Cato Journal article, he wrote:

Gephardt’s support was important to the passage of the Tax Reform Act of 1986. The general thrust of the act—closing loopholes and cutting tax rates—was generally in the right direction, but lawmakers were led astray by the Haig-Simons income definition. The Tax Cuts and Jobs Act of 2017 was better aimed at cutting tax rates and moving toward a more neutral consumption tax base.

Congress should build on the 2017 act with further reforms. As Gephardt noted, lawmakers are poor central planners. They often try to fix economic problems with narrow tax breaks, but it is unlikely that they can manipulate the tax code to allocate resources better than markets.¹²¹ Narrow breaks increase compliance costs and are often plagued by fraud and abuse. Congress should let markets allocate resources and allow entrepreneurs to fix problems in the economy.

Tax breaks are often aimed at fixing problems that the government itself created. The low-income housing tax credit, for example, aims to reduce high housing costs, but that problem is created by excessive zoning, land use, and building regulations. The LIHTC is an ineffective solution because it is plagued by abuse and delivers most of the benefits to developers and banks.¹²² Table 2 proposes LIHTC repeal. A better solution for housing affordability is to repeal regulations that strangle the market supply of housing.

Another tax-code attempt to solve a government-created problem is the earned income tax credit. The credit was created to increase work incentives and offset high Social Security payroll taxes. But the EITC is a poor solution, as it has a high error and fraud rate, and for many recipients it creates a disincentive to increase work effort.¹²³ The refundable part of the EITC imposes costs on other taxpayers, and the tax credit itself may suppress market wages if it increases labor supply. Economists Michael Keen and Joel Slemrod note, “To the extent that wages fall, the benefit of the EITC to the intended beneficiaries—low-income, often low‐​skill, workers—is reduced, and some of the intended transfer redounds to employers.”¹²⁴ By one estimate, employers of low-skilled workers receive roughly three-quarters of the EITC’s benefits.¹²⁵ Table 2 includes EITC repeal.

A better way to improve work incentives would be to convert Social Security payroll taxes to private retirement account contributions.¹²⁶ The 12.4 percent Social Security payroll tax puts a wedge between what employers pay and what workers earn after tax. But with private accounts, retirement contributions would create direct value to workers and thus not undermine work incentives.¹²⁷ In advocating private accounts, economist Edward Prescott argued, “If people are in control of their own savings, and if their retirement is funded by savings rather than transfers, they will work more because they will have more to gain. And everyone will be better off.”¹²⁸

Table 2 includes the repeal of 39 energy tax expenditures on the JCT list, many of which were added or expanded in the Inflation Reduction Act of 2022.¹²⁹ Each item raises administrative and compliance costs, which benefits tax lawyers but is wasteful for the economy. Consider the complexity of one new energy break that was passed in 2022:

The initial Treasury guidance for this one tax credit runs 34 pages.¹³¹ The IRS will need to train a team to understand the details of aviation fuel and parameters of the credit to administer it and police likely abuse. Apparently, experts do not even agree whether the aviation fuel that is promoted by this credit will actually be a net positive for the environment, which underscores Gephardt’s point about central planning.¹³²

Repealing loopholes would raise revenues that could be used to cut tax rates. That is important because high tax rates magnify the damage of taxation, which is called deadweight losses or excess burdens. As tax rates rise, individuals and businesses reduce productive activities such as working and investing, and they increase unproductive activities such as avoidance and evasion. Keen and Slemrod note, “The excess burden suffered by the taxpayer is the same whether the response is in terms of real economy activity … or takes the form of evasion and avoidance.”¹³³

The deadweight losses of the federal tax system are large. The CBO found that “typical estimates of the economic cost of a dollar of tax revenue range from 20 cents to 60 cents over and above the revenue raised.”¹³⁴ That means for every $1 billion in higher taxes, the harm to the private economy is between $1.2 billion and $1.6 billion. According to Keen and Slemrod, when the British government used to impose taxes on fireplaces and windows, it led to houses with fewer fireplaces and windows.¹³⁵ This collateral damage is deadweight loss.

As tax rates rise, deadweight losses rise more than proportionally. Harvard University’s Greg Mankiw explains: “It is a standard proposition in economics that the deadweight loss of a tax rises approximately with the square of the tax rate.… If we double the size of a tax, the deadweight loss increases four-fold.”¹³⁶ Thus a 40 percent tax rate is four times more damaging than a 20 percent rate. That is why a flatter tax structure with lower rates would be more efficient than today’s progressive tax structure.

As tax rates were cut, the harms caused by discontinuities in the tax base would fall. For example, the income tax puts corporate equity at a disadvantage to debt, which induces corporations to overleverage and risk bankruptcy. When the corporate tax rate is cut, the advantage of debt over equity is reduced. The TCJA cut the corporate tax rate, which in turn reduced the dollar value of corporate tax expenditures.¹³⁷

Tax rate cuts would reduce lobbyist pressure to carve new loopholes. Writing in 1974, Norman Ture said that the high tax rates at the time “exerted enormous pressures for changes in the law to afford exceptions from the full application of the high, graduated rates of tax, with respect to particular groups of taxpayers, particular types of income and expense, and particular uses of income.”¹³⁸

What tax rates should Congress cut? Some goals should include cutting the corporate tax rate from 21 percent to 15 percent, cutting the top dividend and capital gains rates from 23.8 percent to 15 percent, and cutting individual income tax rates from a seven-rate structure ranging from 10 to 37 percent to a two-rate structure of 10 and 25 percent. Congress should also consider a reform option recently proposed by the Tax Foundation. It would eliminate many of the tax loopholes discussed here, end the double taxation of corporate equity, and establish a 20 percent flat tax rate for businesses and individuals.¹³⁹

In sum, repealing the loopholes listed in Table 2 would reduce central planning and allow resources to flow to the best uses. It would raise revenues to use for cutting tax rates, which would encourage working and investment and discourage avoidance and evasion. Cutting tax rates would also reduce distortions from remaining loopholes and reduce incentives for lobbying.

In moving toward a consumption-based tax from an income tax, business depreciation deductions would be replaced with capital expensing. Businesses would immediately deduct the costs of equipment and structures rather than deducting them over time. With depreciation, the time value of money erodes future-year deductions, which denies recovery of the full original capital cost. The result is higher effective tax rates on investments.

There are at least four advantages of expensing. First, expensing removes the tax on the normal returns to marginal investments, which in theory eliminates the tax code’s anti-investment bias.¹⁴⁰ More capital would be added to production, which would raise productivity and increase incomes. Above-normal returns on investment would continue to be taxed under a system with expensing.

Second, expensing eliminates the damaging effects of inflation on investment.¹⁴¹ Inflation further erodes the value of future deductions under a depreciation system, which increases effective tax rates. Economist Martin Feldstein noted, “When inflation was at double-digit levels in the late 1970s, the taxation of nominal interest and nominal capital gains and the use of historic cost depreciation raised the effective tax rate substantially, to more than 100 percent in some years and for some types of investment.”¹⁴²

Third, expensing eliminates distortions caused by varying effective tax rates on different types of assets, as previously discussed. Income taxes tend to distort effective tax rates in favor of shorter-term over longer-term investments. The reduction of such distortions is one of the major advantages of consumption-based taxes over income taxes.¹⁴³

Fourth, expensing is simpler than depreciation. Depreciation is one of the most complicated features of the income tax. To create greater neutrality between investments under an income tax requires complex rules, but with expensing neutrality it is achieved with simple rules.

The TCJA implemented capital expensing for equipment, but the provision is phasing down and expires after 2026. Expensing is not available for structures, which are generally depreciated over 27.5 years for residential buildings and 39 years for commercial buildings. The eventual goal should be permanent enactment of expensing for both equipment and structures.¹⁴⁴

In the TCJA, capital expensing was enacted alongside limitations on business interest deductions. Interest used to be deductible when paid, but net interest deductions are now limited to 30 percent of adjusted taxable income, which is earnings before interest and taxes.¹⁴⁵ This section 163(j) limitation applies to taxpayers with gross receipts of $25 million or more, as indexed for inflation. Many economists think that allowing expensing combined with full interest deductions allows a sort of double deduction. But other economists dispute that view, arguing that if lenders are taxed on interest income, then denying an interest deduction double-taxes interest.¹⁴⁶ Either way, the interest deduction limitation in the TCJA serves to reduce the general advantage of debt over equity in the tax code.

While generally embracing capital expensing, the TCJA took one step backward. Businesses have long been allowed to expense research spending, which is the proper consumption tax treatment. But the provision was considered an official tax expenditure because Haig-Simons requires amortization. Unfortunately, the TCJA moved toward Haig-Simons by requiring amortization over five years, which raised effective tax rates on research.¹⁴⁷ That was a major policy mistake, and there are bipartisan efforts to reinstate full expensing for research.

An irony with the Haig-Simons approach to tax policy is that it punishes the same activities that the federal government often subsidizes with spending. The government spends vastly on research and technology, as it did with the CHIPS and Science Act of 2022, which subsidizes the semiconductor industry. But then the government punishes private research and technology investment by not enacting permanent expensing in the tax code. The best policy would be expensing in the tax code and no spending subsidies.

Under the Hall-Rabushka tax, all investments in research, equipment, and structures would be expensed. Debt and equity would receive equal treatment. The system would tax above-normal returns in a consistent and equal manner across the economy. That is the direction that Congress should move.

The tax code includes numerous vehicles to provide relief from the double taxation of saving, including 401(k) plans, traditional IRAs, Roth IRAs, and multiple types of education saving accounts. Each of these vehicles includes rules for eligibility, contributions, earnings, withdrawals, penalties, and rollovers. To simplify personal saving and expand saving opportunities, Congress should replace some of these vehicles with Universal Savings Accounts (USAs).

USAs would be similar to expanded and more flexible Roth IRAs. Contributions to the accounts would come from after-tax income, but all account earnings would be tax-free. Unlike with Roth IRAs, individuals could withdraw USA funds tax-free at any time for any reason, which would maximize liquidity and encourage greater saving.

There is growing interest in USAs. Republicans included the accounts in their Tax Reform 2.0 legislation, which passed the House in 2018.¹⁴⁸ The legislation included USAs with annual contribution limits of $2,500, but a better limit would be $20,000 or more. Both Canada and the United Kingdom have enacted USA-style accounts that have been very popular with individuals at all income levels.¹⁴⁹

If Congress enacted USAs with high contribution limits, it could simplify the tax code by phasing out IRAs, Roth IRAs, and education savings accounts. Those accounts would not be needed because USAs are superior due to their flexibility. Individuals and families could use USAs to save for retirement, as well as for health care, college expenses, buying a home, covering unemployment, starting a business, and many other purposes.

All personal saving should be encouraged, not just the types of saving favored by some members of Congress. All saving increases personal financial security and provides fuel for economic growth. As such, Congress should follow the lead of Canada and the United Kingdom and enact Universal Savings Accounts.

One of the largest tax expenditures is the exclusion for employer-provided health insurance. While wages are subject to income and payroll taxes, worker compensation in the form of employer health coverage is not. That is a lot of compensation—the average employer-sponsored family premium in 2021 was $21,222.¹⁵⁰ Economists generally agree that the exclusion distorts labor markets, favors larger employers over smaller ones, reduces choice, and encourages excessive insurance coverage and the overconsumption of health care.

However, it also true that much of America’s private health care system has been built around the exclusion. About three-quarters of U.S. workers are eligible for employer-provided coverage, and about three-quarters of those individuals enroll in the coverage.¹⁵¹ Thus, many people rely on the coverage structured around the exclusion, and so fully repealing it would be disruptive and not likely to pass Congress. Instead, policymakers should consider reforms to reduce the distortions caused by the preferential tax treatment of health care.

One reform option proposed by the Cato Institute’s Michael Cannon is to replace the current exclusion with a different mechanism that would also be tax-advantaged but would reduce current distortions in health care.¹⁵² The plan would transform the tax exclusion while expanding Health Savings Accounts (HSAs) into what Cannon calls Large HSAs.

In 2023, individuals who are in qualified high-deductible health insurance plans can deposit up to $3,850 each, or $7,750 per family, into tax-free HSAs.¹⁵³ HSA funds can accumulate over time and be moved from job to job with workers. These accounts reduce the tax distortion in favor of employer-provided insurance. They were enacted in 2003 and have been very successful: by 2022, Americans owned 32 million HSAs with assets topping $100 billion.¹⁵⁴

Cannon’s proposal would convert the current employer exclusion into an exclusion solely for HSA contributions, while increasing HSA contribution limits substantially. Health insurance would be added to the list of items people can purchase with their HSAs, and current insurance requirements would be loosened so that people could purchase health coverage that suits them. Finally, the exclusion would be capped to limit the overall amount of tax preferences provided to health spending.

The Cannon plan would be revenue neutral, as current health-related tax expenditures would be replaced by a tax expenditure for Large HSAs. The accounts would allow individuals to control their health care funds, and the funds would be portable as workers moved between jobs. Large HSAs would be a step toward greater competition, flexibility, and choice in the nation’s health care system.

Many tax experts complain that owner-occupied housing receives large subsidies in the tax code. The Treasury and JCT tax expenditure lists include numerous housing-related provisions. Both lists include the mortgage interest deduction and the partial exclusion of capital gains on home sales. The Treasury list also includes net imputed rental income on owner-occupied homes, while the JCT list excludes this item for administrative reasons. From a Haig-Simons perspective, these three provisions are tax subsidies, but from a consumption tax perspective the situation is different.

Net imputed rental income is one of the largest items on the Treasury list. Under the Haig-Simons income definition, homeowners gain a benefit because the government does not tax them as landlords renting their own homes to themselves. That may seem bizarre, but net imputed rental income is estimated as income in the National Income and Product Accounts to create equal treatment between rental and owner-occupied housing.

This item is “net” income because it is the gross rental value of one’s home less depreciation, interest, and other expenses. After deductions, the leftover net would mainly include the time value of money related to owning an asset. Homeowners earn this capital income whether or not their home has a mortgage. Under Haig-Simons, this income should be taxed annually.

Consumer durables are viewed the same as housing in the Haig-Simons definition.¹⁵⁵ Cars, couches, dishwashers, clothes, and every other long-lived consumer asset is thought to generate capital income, as if one were renting the items to oneself. To be consistent with Haig-Simons, official tax expenditure lists should include net imputed income from these household items as well, but they do not for practical reasons.

What about a consumption-based tax? Whether homebuyers pay cash up front or contract a mortgage, they pay off the principal from after-tax earnings, which means that housing is taxed when the principal is paid. Buying a house is like buying a Roth IRA: a single layer of tax is paid up front, and then future income or consumption is not taxed again. The same is true for consumer durables. Thus, the current treatment of housing and consumer durables in the tax code is basically the correct one for a consumption-based tax.¹⁵⁶ Net imputed rental income is the normal return to saving, which is not taxed under a consumption-based tax and thus does not represent a true loophole.

The best treatment for the mortgage interest deduction (MID) is more ambiguous. Currently, the MID is limited to the first $750,000 of principal value, down from $1 million before the TCJA was passed. If homeowners were required to pay tax on their home’s imputed income, then a mortgage deduction would be appropriate. But since such imputed income is not taxed, the general view is that the mortgage deduction is a tax expenditure under an income tax.¹⁵⁷

However, mortgage providers generally pay tax on interest income, and the MID offsets that interest inclusion to create roughly neutral tax treatment from a consumption tax perspective. In that case, the MID would not be a tax expenditure. However, another consideration is that under a full consumption-based tax, such as Hall-Rabushka, interest would be removed from the tax base, so there would be no MID in that case. Given that the current tax code is a hybrid of an income and consumption base, there is some ambiguity regarding the best reform option for the MID.¹⁵⁸

That said, economists generally agree that the tax code currently favors owner-occupied housing over fully taxable assets in the economy. Prior to the TCJA, the CBO estimated that the marginal effective tax rate on owner-occupied housing was –2 percent, which compares to the average METR on investment through C corporations of 31 percent.¹⁵⁹ As a result of the tax differential, too many resources flow to owner-occupied housing relative to other assets, which undermines growth. The MID also complicates the tax code. For these reasons, the MID should be repealed and the added revenues used to cut overall tax rates.

Finally, the tax code allows homeowners to exclude a portion of capital gains on the sale of their homes, which is both a Treasury and JCT tax expenditure. The exclusion is $250,000 for single filers and $500,000 for joint filers. A consumption-based tax would not tax capital gains, and thus would not treat this provision as an unjustified loophole.

State and local governments issue debt to finance infrastructure investments. The interest on such “muni bonds” is generally tax-free under the federal income tax. Tax-free muni bonds are either “public purpose” for projects such as schools, or “private activity” for projects such as housing, energy, and broadband.¹⁶⁰ Table 2 shows that the tax expenditure for tax-free muni bonds is $39.8 billion a year.

The muni bond tax exemption is considered a tax expenditure under the income tax but not under a consumption-based tax because the latter exempts the normal return to saving. Nonetheless, the tax exemption is a problematic provision because it adds substantial complexity to the tax code and it favors government borrowing over private borrowing. So for the following four reasons, the tax exemption should be repealed.

First, by reducing the cost of borrowing, the muni bond tax exemption encourages governments to borrow in excess. Governments generally do not need to borrow to fund infrastructure because they can finance projects on a pay-as-you-go basis, meaning paying with current taxes and user charges. Pay-as-you-go funding is better than borrowing because it increases transparency and political responsibility. It also avoids interest costs and Wall Street fees to issue the debt, which for state and local governments amount to about $4 billion a year.¹⁶¹

Second, tax-free muni bonds stack the deck against the private provision of facilities such as airports, seaports, and transit systems. Numerous other nations have more privately provided infrastructure than does the United States, partly because they do not favor government projects in the tax code. For example, while all major U.S. airports are government owned, almost half of all major airports in Europe are privately owned.¹⁶² The muni bond exemption stands in the way of state and local privatization.

Third, the tax exemption generates ongoing lobbying as special interests try to secure tax benefits for their projects. Congress imposes limits on the issuance of tax-free private activity bonds through state volume caps and by prescribing allowable projects. But since 1968, “the number of eligible private activities has been gradually increased from 12 activities to 30” as lobbyists have pried the loophole wider.¹⁶³ The federal rules specifying the details of bonds and allowable purposes are enormously complex, which enriches an industry of high-paid lawyers. One guide to the use of tax-free bonds for housing is 700 pages long.¹⁶⁴ Such bureaucracy is wasteful for the overall economy.

Fourth, the tax exemption tilts the economy toward government-favored infrastructure over unsubsidized private infrastructure. It gives a financing advantage to government jails and sports stadiums over private manufacturing plants and cell phone systems. A 2020 study examined 57 major league stadiums built since 2000 and found that 43 were partly funded by tax-free muni bonds.¹⁶⁵ It makes no sense to favor sports stadiums over manufacturing plants or over other entertainment facilities that may be fully taxable, such as bowling alleys and movie theaters. The CBO estimates that government investment is less productive than private investment, so the tax advantage for the former undermines growth.¹⁶⁶

Prior to passage of the TCJA in 2017, there was no direct limit on the amount of state and local taxes (SALT) that individuals could deduct on their federal returns.¹⁶⁷ The deduction is only available to taxpayers who itemized deductions, which in 2017 was 31 percent of tax-filing households.¹⁶⁸ Nearly all households that itemized took the SALT deduction.

The TCJA capped the annual SALT deduction at $10,000 for single and married tax filers, and the law also nearly doubled standard deductions. Those changes reduced the number of households taking the SALT deduction from 46 million in 2017 to 17 million in 2018.¹⁶⁹ The $10,000 cap is not adjusted for inflation.

Is the SALT deduction an unjustified loophole? There is some theoretical ambiguity, and that is true whether the goal is either income tax reform or consumption-based tax reform.¹⁷⁰ State and local taxes pay for government services, which residents generally do not pay taxes on. If taxes are deductible, it advantages government services over privately provided services in the economy. Thus, eliminating the SALT deduction is a way to impose tax on government services to put them on a roughly equal footing with privately provided services. However, some economists argue that a share of state and local spending is investment spending, which should be tax-exempt under a consumption-based tax. In that case, a SALT deduction would be a way to roughly exempt the government investment spending from tax.¹⁷¹

Nonetheless, the SALT deduction biases taxpayers and policymakers toward favoring a bigger government and this distortion outweighs other concerns. The deduction softens the blow to taxpayers of state and local income, property, and sales taxes, which encourages policymakers to increase those taxes, particularly on higher earners. Before the law change, 91 percent of SALT deduction benefits went to households with incomes above $100,000.¹⁷² The deduction favors higher-income and higher-tax states over other states.

Taxes should signal to residents the full cost of government services, but the SALT deduction undermines that economic signal. As a result of the SALT subsidy, “too many of those services may be supplied, and state and local governments may be bigger as a result,” noted the CBO.¹⁷³ By reducing the perceived cost of government, the SALT deduction induces residents to demand too much of it.

Leading up to the Tax Reform Act of 1986, the Reagan administration proposed getting rid of the deduction, with President Ronald Reagan arguing, “Perhaps if the high-tax states didn’t have this federal crutch to prop up their big spending, they might have to cut taxes to stay competitive.”¹⁷⁴ The 1986 law succeeded in eliminating the deductibility of sales taxes, but that reform was later reversed.

The SALT deduction cap under the TCJA was a long-needed reform, but it is scheduled to expire after 2025 along with most individual provisions of the law. Congress should not just extend the SALT cap but proceed to repeal the individual SALT deduction altogether. That would simplify the tax code, end the incentive for state and local governments to expand, and generate revenues to cut federal tax rates for all taxpayers.

The federal tax code will continue changing. The next administration may propose tax cuts or tax increases, and Congress will need to decide whether to extend the TCJA reforms. Policymakers should repeal loopholes and cut tax rates. The goal should be a simpler tax code that minimizes distortions and boosts economic growth.

To guide reforms, policymakers need an accurate tabulation of loopholes in the tax code. The current Treasury and JCT tax expenditure lists fall far short. They depend on the faulty Haig-Simons definition of income, which endorses the double taxation of saving and investment, and they depend on ad hoc rules reflecting a redistributionist bias.

Claims that high earners gain the most from tax loopholes are off base. The claims stem from arbitrary choices made in producing the official tax expenditure lists. If the Treasury and JCT had not assumed away the progressive income tax rate structure, that tax-code feature would be a massive tax expenditure for lower-income households, which would greatly alter the usual narratives about loopholes.

There are some ambiguities in tax expenditure analysis, and no list is perfect. But the Treasury and JCT should remove obvious inconsistencies and the redistributionist bias, and they should include a fuller discussion of penalties such as the double taxation of corporate equity. Also, the agencies should present two separate tax expenditure lists reflecting the two basic views of federal taxation: Haig-Simons and consumption. The actual federal tax code is a hybrid of these views, so the two alternatives should be presented on an equal basis.

Congress and the next administration should pursue major tax reforms. Tax loopholes to repeal include the earned income tax credit, the mortgage interest deduction, the exemption for municipal bond interest, the state and local tax deduction, and dozens of energy tax breaks. Policymakers should enact Universal Savings Accounts and reform the tax treatment of health care and business investment. Congress should move toward a consumption-base tax with a lower and flatter tax rate structure.

The author appreciates detailed comments from Stephen Entin, J. D. Foster, Adam Michel, William McBride, and Chris Hesse. Author responsible for any errors.

Edwards, Chris. “Tax Expenditures and Tax Reform,” Policy Analysis no. 954, Cato Institute, Washington, DC, July 25, 2023.