In U.S. federal income tax, the standard deduction, along with the personal exemptions, provides taxpayers with a minimum amount of untaxed income, effectively creating a “zero bracket amount.” For historical and political reasons, however, the standard deduction also operates as a simplified substitute for the itemized deductions, such as the deductions for extraordinary medical expenses, charitable contributions, and home mortgage interest. This seemingly reasonable compromise in fact leads to substantial, and surprising, conceptual complexity. In particular, close analysis of each of the two roles shows that their effects, and related criticisms, are often contradictory, which in turn makes it difficult, if not impossible, to have coherent policy debates regarding the proper roles of the standard deduction and the personal deductions. This flawed compromise between progressivity and simplification is not necessary. We can replace the standard deduction with a true, independent zero bracket amount and a floor under the itemized deductions while staying revenue- and distribution-neutral. This would effectively divorce the two roles of the standard deduction—zero bracket amount and simplification of the itemized deductions—leading to more coherence in individual income taxation and giving more flexibility to policymakers. This article proposes further to disaggregate the single floor under the itemized deductions into multiple, independent floors under each itemized deduction. This also would lead to greater coherence and flexibility in tax system design. While creating multiple floors would marginally increase complexity for some taxpayers, the costs of such complexity are justified in light of the benefits of more accuracy and coherence.
- I. INTRODUCTION
- II. CURRENT LAW AND HISTORY OF THE STANDARD DEDUCTION
- A. Current Law
- B. The Original Standard Deduction: 1944-1964
- C. The Introduction of Progressivity and the Breakdown of Deduction Coherence: 1964-1977
- D. The Zero Bracket Amount—But Only in Name: 1977-1986
- E. The Current Era: 1986 to Today
- III. A CRITIQUE OF THE STANDARD DEDUCTION
- A. The Equivalence of a Standard Deduction and an Exemption (or ZBA) Plus a Floor
- B. Mixing Income Measurement with Progressivity
- 1. Simplification: No ZBA
- 2. Progressivity: ZBA Equal to the Standard Deduction Amount
- 3. Hybrid: ZBA Less Than the Standard Deduction Amount
- 4. Other Views
- C. Policy Debate Distortion
- 1. Unresolved Political Debates
- 2. Distorted Political Incentives
- 3. Inefficient Subsidies
- 4. Undermining Respect
- D. Summary
- IV. A PROPOSAL: THE ZERO BRACKET AMOUNT
- A. Replace the Standard Deduction with a Zero Bracket and Floors
- B. Examples and Complications
- C. Flat or Variable ZBA
- D. The Decreasing Benefit of Simplification
- 1. Simplification as a Goal of Tax Policy
- 2. The Cost of Itemizing
- 3. Summary
- E. Policy Benefits of Tax Floors
- V. CONCLUSION
The standard deduction is a provision of the U.S. individual income tax system that serves two purposes: First, it provides for a minimum amount of untaxed income, thus acting as an element of progressivity. Second, it provides a simplified alternative to the itemized deductions for taxpayers with relatively low itemizable expenses, thus relieving them of the record-keeping and computational burden of itemizing deductions. Having one provision serve each of these purposes is a compromise, but it is a compromise that is unreasonable and unnecessary.
The essence of the problem is that these two purposes—progressivity and simplification—are in conflict. The progressivity purpose, by which I mean the goal of having higher-income taxpayers pay a higher share of income tax via a progressive rate structure and other provisions, is served by having a relatively large amount of otherwise-taxable income go untaxed, through what is in essence a zero-percent tax bracket made of the standard deduction and personal exemptions. I refer to this as the “zero bracket amount” or “ZBA.” Thus, the standard deduction and the personal exemptions allow for an unmarried taxpayer with no dependents in 2010 to pay no tax on the first $9350 of adjusted gross income, of which $5700 is due to the standard deduction. As a result of this and other provisions, 43% of taxpayers pay no income tax at all.
The simplification purpose, by which I mean the goal of limiting complexity and allowing for more straightforward tax computation (by the taxpayer) and easier enforcement of compliance (by the government), is served by allowing some taxpayers to substitute a single standard deduction for the many itemized deductions when those deductions would otherwise be small and thus too burdensome to calculate (and examine) relative to the benefit from doing so. As noted above, in 2010 the standard deduction for an unmarried taxpayer is $5700—meaning that the taxpayer would not itemize deductions if his itemized deductions would be less than $5700. As a result of this, nearly two-thirds of taxpayers do not itemize deductions and instead take the standard deduction.
The problem with having one provision play both a progressivity and simplification purpose is that those two purposes are in conflict. In particular, it is just not possible for a single tax provision to consistently play both a simplification and progressivity role at the same time. If tax is a product of a tax base and a tax rate, one provision cannot define both the base and the rate, at least not well. If the standard deduction reduces taxable income—by approximating the itemized deductions—then it is defining the base; but if it is stating an amount of taxable income not subject to tax—a ZBA—then it is part of the rate structure. It cannot do both simultaneously without causing taxpayers to face disparate tax bases and effective rate structures.
Analysts and commentators tend to see the standard deduction and personal exemptions as primarily a tool of progressivity—as part of the rate structure. But in grafting this tool of progressivity onto what was once a pure simplification provision, Congress inadvertently created of a provision of surprising complexity. Indeed, it is actually quite difficult to conceptualize a system where a single standard deduction provision plays both a progressivity and a simplification role. One way to view it is that the standard deduction is essentially a ZBA that phases out with the amount of itemized deductions—the greater the expenses, the lower the “net” excess of the standard deduction over itemized deductions. Another view is that the standard deduction is really a flat ZBA plus a floor under itemized deductions—all taxpayers benefit from a flat ZBA, but no itemized deductions are allowed below a floor equal to the same amount as the ZBA.
Although the same computational result follows in each case, these views are largely contradictory, incorporate different assumptions about the relative priority of progressivity versus simplicity, lead to different criticisms of the tax system, and, importantly, imply different policy responses. For example, the first view implies that all taxpayers receive the tax benefit of the itemized deductions on the first dollar of itemizable expense, but that the system violates horizontal equity because taxpayers with the same AGI might face different, and likely regressive, effective ZBAs. The second view implies the opposite: that many taxpayers are not receiving a tax benefit from their itemizable expenses, but that there is no harm to horizontal equity from the distribution of the ZBA. Thus, one can essentially choose whichever view of the standard deduction, and corresponding criticisms, one wants.
This lack of a shared conception of the role of the standard deduction makes it difficult to resolve important tax policy questions, such as those related to low-income tax relief or to the itemized deductions themselves. How to discuss extending the charitable contributions deduction to non-itemizers when it is not entirely clear that they do not already get the benefit of that deduction? How to discuss the proper size of a ZBA when we cannot say with precision what its size actually is?
To put this in concrete terms, suppose a non-itemizer says, “I think it’s unfair that I don’t get a tax benefit from my charitable contributions.” One might respond, “But you do—you get the standard deduction instead, which is an even better deal.” The non-itemizer might respond, “Then I think it’s unfair that I get less benefit from the standard deduction than someone who gives less to charity.” One might respond to this, “No, you get the same benefit—everyone gets the value of the standard deduction, which is really a zero bracket amount—it’s just that neither of your expenses are high enough to qualify for a tax deduction on top of that.” An absurd circular conversation, to be sure, but in a sense this is where our public discourse regarding the standard deduction is since we have no way of deciding which of these views is proper.
In addition to the conceptual incoherence, the current structure of the standard deduction likely results in a suboptimal choice of the actual standard deduction amount. For example, under the current standard deduction the ZBA is the same size as the simplified substitute for the itemized deductions. But why must these amounts be the same? It is almost certain that the optimal amounts would not be equal to each other, since they are driven by independent policy considerations. The optimal substitute for the itemized deductions might only be a few thousand dollars—much more than that and it no longer seems too costly for a taxpayer to manage—yet our desire for progressivity pushes the standard deduction as high as $5700 for an individual. Conversely, the optimal ZBA is likely greater than $5700, or even $9350, yet the attempt to balance simplicity and accuracy forces the amount down. By forcing ourselves to pick one number, we likely miss both our progressivity target and our simplification target by a lot. This is just one example of the ways in which policymakers lack the flexibility to make targeted policies due to having only a single crude tool.
Despite the relative significance of the standard deduction, surprisingly little has been written analyzing its effects in depth. There are vastly more articles in the legal literature discussing the merits and problems of the itemized deductions—despite the fact that the standard deduction affects nearly twice as many taxpayers as the itemized deductions. Perhaps this is because the itemized deductions raise questions about tax expenditures, targeted policy goals, and economic incentives that do not at first appear to be present with the standard deduction. Perhaps it is also because there is some general understanding that simplification and progressivity can be at odds, in the standard deduction and elsewhere. What is lacking, and what this article provides, is a full articulation of the problems and their implications, and a clear proposal for reform. In particular, existing commentary tends not to fully address the fact that the standard deduction embodies two contradictory purposes, that each purpose implies a particular set of criticisms, and that any particular criticism of the standard deduction necessarily implies a particular view of the standard deduction’s primary role. It is this conceptual problem that is at the core of this article’s criticism of the standard deduction.
This article further proposes to solve that problem by separating the progressivity and simplification purposes of the standard deduction. First, Congress should implement a true, independent ZBA (whether in the form of an actual zero-percent tax bracket or a larger personal exemption) and replace the standard deduction with an explicit floor under the itemized deductions. This would effectively divorce the progressivity and simplification roles now combined in a single provision, without affecting tax revenue or overall distribution of tax burdens. It would positively affect horizontal equity by making the determination of the proper amount of tax-free income independent of the amount of itemizable expenses. Furthermore, it would allow policymakers greater flexibility in targeting low-income tax relief, since they could adjust the ZBA without affecting tax deductions.
In addition, rather than a single floor under the itemized deductions, the floor itself should be disaggregated into multiple, independent floors under each itemized deduction. This would separate the determination of the proper deduction amount for a given itemizable expense from the amounts of any other itemizable expenses. It would thus positively affect horizontal equity by treating all taxpayers the same with respect to a given category of itemizable expense. This article further argues that deduction floors have particular, and underused, tax policy benefits, since they can allow policymakers to more effectively manage deductions by adjusting floors to optimize trade-offs between simplicity, revenue, incentives, and equity for each given deduction. While disaggregating the floors would introduce some new complexity in return-preparation, that complexity is more than balanced by the benefits, given that the floors would also continue to serve simplification by limiting the situations where expenses are deductible. While others, most notably Louis Kaplow, have discussed the general computational equivalence between a standard deduction and a floor, and factors to consider in choosing between them, this article goes further by making the affirmative normative case for the superiority of disaggregated floors over the current standard deduction.
This article proceeds as follows. Part II reviews the current operation of the standard deduction and its history, especially its shift from a pure simplification measure to one that one that was commandeered to play a progressivity role. Part III provides a comprehensive critique of the standard deduction, focusing on the conceptual contradictions between simplification and progressivity. Part IV introduces the reform proposal and discusses the additional advantages of disaggregated deduction floors as a policy tool. It also addresses the primary argument in favor of the standard deduction—administrative simplicity. Part V concludes the article.
Before turning to the detailed criticisms of the standard deduction, this article first sets out the changes in the standard deduction over its 60-plus years in the Tax Code. The criticisms that follow in Part III largely result from the fact that, for historical and political reasons, a zero bracket was grafted on to what began as a pure substitute for the itemized deductions. While the standard deduction has operated in essentially its current form since 1977, the earlier versions of the standard deduction were quite different. In particular, the original standard deduction in 1944 was calculated as a percentage of AGI, rather than a flat amount. As discussed below, this important feature made the original standard deduction almost entirely a simplification measure. However, in the 1960s and 70s, policymakers became more concerned about low-income tax relief, particularly in the face of high inflation, and the standard deduction appeared to be the most straightforward way to target that relief. Thus began a period of gradually grafting more progressivity-centric features onto the standard deduction and chipping away at the earlier simplification-centric features. While policymakers have been generally explicit about their intent to turn the standard deduction into a vehicle for low-income tax relief, they do not appear to have considered the downsides, discussed in Part III, nor the alternatives, discussed in Part IV.
Under current law, the amount of an individual’s income subject to tax (“taxable income”) is calculated, by default, by subtracting the standard deduction and the deductions for personal exemptions from adjusted gross income (“AGI”). Thus, the standard deduction provides for an amount of AGI that is untaxed. In 2010, the basic standard deduction amounts were $5700 for a single taxpayer or for a married taxpayer filing separately, $11,400 for a married couple filing a joint return, and $8300 for a taxpayer filing as a head of household.
Instead of using the standard deduction, however, a taxpayer can elect to “itemize,” that is, to take certain other deductions in lieu of using the standard deduction. Deductions are amounts subtracted from gross (or adjusted gross) income in order to calculate taxable income. They generally represent expenses that are either believed to be properly subtracted from earnings before determining what is “income”—such as expenses in the production of income—or believed to be worthy of subsidizing via the Tax Code—such as charitable contributions. Deductions for individuals are generally grouped into two categories: “above-the-line” deductions from gross income to calculate AGI, and “below-the line,” or itemized, deductions from AGI to calculate taxable income. It is these below-the-line, or itemized, deductions for which the standard deduction is a substitute. The above-the-line deductions are available to all taxpayers, while the below-the-line deductions are taken only by those whose itemizable expenses exceed the relevant standard deduction amount. This structure results in a taxpayer deducting from AGI an amount no less than the standard deduction in computing tax, and often more, if the taxpayer’s itemized deductions exceed his standard deduction amount. The most important itemized deductions for individuals are for state and local taxes, mortgage interest, medical expenses, charitable contributions, certain expenses in the production of income (typically relating to investments, income-producing property, and other non-business activity), unreimbursed employee expenses, and casualty and theft losses.
This article uses the terms “below-the-line deductions” and “itemized deductions” interchangeably. It uses the term “itemizable expense” to refer to the expense itself, as distinguished from the respective itemized deduction, which may differ in amount from the expense itself, or may not be available at all.
During World War II, the United States faced an enormous need to raise revenue, which in turn led to an expansion of the income tax beyond the small percentage of high-income taxpayers that it had previously reached. While high-income taxpayers generally had the sophistication to navigate the complexities of the tax system, Congress was concerned that this was not necessarily the case for the new middle- and low-income taxpayers. In 1944, Congress made some of the war-time tax increases permanent, but alongside those it also expanded a simplified tax system and instituted the “optional standard deduction.” The optional standard deduction was set at 10% of AGI, up to a maximum of $500 for single taxpayers, $1000 for married taxpayers filing jointly (or roughly $6250/$12,500 in 2011 dollars). Taxpayers with AGI less than $5000 (~$62,500 in 2011)—and thus those below the cap on the percentage standard deduction—computed tax according to simplified tax tables, while those with AGI greater than $5000, and all those with itemizable expenses greater than 10% of their AGI, computed tax themselves. Congress estimated that no more than 20% of taxpayers would compute their own tax, with everyone else using the tax tables.
Along with the introduction of the standard deduction, Congress also created the concept of “adjusted gross income,” or AGI. Congress believed that for purposes of determining the correct percentage standard deduction, it was necessary to allow first for some deductions to all taxpayers. In particular, Congress distinguished between business-related deductions, which were deductible from gross income in computing AGI—they were placed “above the line”—and all others, for which the optional standard deduction would be a substitute, which were deductible from AGI in computing taxable income—they were placed “below the line.” The new above-the-line deductions that all taxpayers would use in calculating AGI included those for expenses attributable to a trade or business or to rents and royalties; to deductible employee expenses (such as meals and lodging while traveling on business); and to losses from the sale or exchange of property. The intent was to come up with an income measure that put wage-earners on similar footing as business owners for purposes of then calculating the percentage standard deduction. The Senate Report accompanying the bill stated that the deductions thus permitted to be made from gross income in arriving at adjusted gross income are those which are necessary to make as nearly equivalent as practicable the concept of adjusted gross income, when that concept is applied to different types of taxpayers deriving their income from varying sources. . . . For example, in the case of an individual merchant or store proprietor, gross income under the law is gross receipts less the cost of goods sold; it is necessary to reduce this amount by the amount of business expenses before it becomes comparable, for the purposes of . . . the standard deduction, to the salary or wages of an employee in the usual case.
The remaining deductions—those now “below the line”—would thus be those that could be expected to be relatively uniformly distributed across all taxpayers, regardless of source of income. The below-the-line deductions for which the standard deduction would substitute included those for interest, state and local taxes, charitable contributions, and medical expenses. Through the 1954 Code, the structure of the standard deduction remained relatively unchanged.
In the reports accompanying the 1944 Act, Congress nowhere identified progressivity or reducing the tax burden on low-income taxpayers as a goal of the legislation; instead, Congress focused entirely on simplification. This is consistent with the design of the provision: itemizable expenses tend to increase with income, so it follows that a simplified substitute for the expenses would also increase with income (while an element of progressivity certainly would not).
In 1964—twenty years after the standard deduction was introduced—progressivity finally appeared in the design of the standard deduction. With the Revenue Act of 1964, Congress created the “minimum standard deduction,” which was intended “to remove from the tax rolls those persons with minimum incomes and also to provide those with incomes just slightly above these levels a somewhat larger tax reduction than is made available generally through the rate cuts.” This is the first time a progressivity goal was articulated in the legislative history of the standard deduction, and the first time that the standard deduction provision was used to ensure some baseline amount of untaxed income. The new minimum standard deduction was calculated as $200 plus $100 for each dependency deduction (or ~$1420/$710 in 2011). A taxpayer could take the larger of the minimum standard deduction, or the percentage standard deduction (officially renamed the “10-percent standard deduction”). The result was that all taxpayers could get at least $900 (~$6400 in 2011) of income tax-free—a $300 minimum standard deduction plus the $600 personal exemption—even if that amount was greater than 10% of their AGI.
In 1969, the connection to progressivity was made even more explicit. In the Tax Reform Act of 1969, Congress renamed the “minimum standard deduction” the “low income allowance” and redesigned it to be precisely the poverty level of income. Thus, taxpayers could be assured of paying no tax on incomes at or below the poverty level. The low income allowance was $1100 (~$6600 in 2011), so that a taxpayer faced no tax for income less than $1700 (~$10,200 in 2011) plus each additional $600 dependency deduction, which, according to Congress, was exactly what the then-Department of Health, Education and Welfare estimated to be poverty level income. This was an $800 (~$4800 in 2011) increase over the minimum standard deduction put in place only five years earlier.
Simplification purposes continued to drive other changes, however. By 1969, higher medical costs, interest rates, and states taxes; increased homeownership; and more expensive homes had eroded the reach of the percentage standard deduction. Where 82% of returns took the standard deduction in 1944, only 59% did in 1965. To counter this, the 1969 Act gradually increased the percentage standard deduction from 10% of AGI, capped at $1000 (as it had been since 1948), to 13% of AGI, capped at $1500, for the 1971 tax year, 14%/$2000 in 1972, and 15%/$2000 thereafter. The Joint Committee on Taxation estimated that this would increase the percentage of taxpayers taking the standard deduction from 58% to 70%, thus simplifying tax return preparation for a significant number of taxpayers.
As the standard deduction became more entrenched, the 1944 distinction between business-related and personal deductions for purposes of calculating AGI began to break down. Significantly, almost none of the important deductions for individuals added after 1954 were added below the line, despite the fact that many, if not most, related to personal expenditures rather than costs associated with a trade or business.
One of the earliest personal deductions added after the creation of the standard deduction was the deduction for job-related moving expenses. The Revenue Act of 1964 added the deduction above the line, thus making it deductible to all taxpayers, itemizers and non-itemizers alike. The House Committee Report states that it added the deduction above the line in part because, where the expenses were large, “it occurred to your committee that it would be undesirable to, in effect, make taxpayers choose between taking this deduction and the standard deduction in lieu of itemized personal deductions.” The reasoning here is faulty, for some of the reasons discussed infra in Part III.B, but regardless it could also apply to any personal deduction—any non-itemizer must “choose between” itemizing and taking the standard deduction. A potential response is that moving expenses occur irregularly, unlike charitable deductions, mortgage interest, and state taxes, and that placing them below the line would force some non-itemizers to learn the full set of itemization rules for one year simply because of one large, non-recurring expense. Thus they would be forced to “chose between” continuing simply to be non-itemizers and getting the benefit of the deduction. But the same argument could be applied to the deduction for extraordinary medical expenses or casualty losses, which are also likely to be non-recurring.
More importantly, even if placing moving expenses above the line could be justified on simplification grounds, it is not consistent with the original 1944 distinction between above- and below-the-line expenses. Recall that above-the-line deductions were originally those that would be appropriate in order for AGI to be a relatively uniform base from which to calculate the percentage standard deduction fairly across different types of taxpayers. Job-related moving expenses are more like the expenses of a wage-earner than a business-owner, and thus should not have been used in calculating AGI, under the 1944 definition. Placing moving expenses above the line thus effectively lowered the percentage standard deduction amount available to that taxpayer compared to a similarly situated taxpayer who did not move.
Another example is the deduction for alimony paid. Prior to 1976, the deduction was below the line and thus subject to the standard deduction. In the Tax Reform Act of 1976, Congress moved the deduction above the line, stating only:
The committee believes that the splitting of income or assignment of income through the payment of alimony is not properly treated under current law which permits only an itemized deduction for alimony. Instead, the committee believes it is more appropriate to take the payment of alimony into account as a deduction in arriving at adjusted gross income, rather than as itemized deductions which are generally limited to personal expenses. As a deduction from gross income, the alimony deduction would be available to taxpayers who elect the standard deduction as well as to those taxpayers who elect to itemize their deductions. Because prior to 1976 alimony was below the line, the pre-1976 standard deduction effectively included an estimate of alimony payments, at least to the degree it includes any of the itemized deductions. Thus moving the deduction above the line without any change in the standard deduction effectively gave alimony-payers an additional subsidy on top of the already existing effective tax-free treatment. But the existence of the standard deduction obscured this effect. It could be argued, as the Senate Finance Committee appears to in the above passage, that it is important for the alimony deduction to match the inclusion of alimony as gross income under § 71, but that same argument could apply to mortgage interest, medical expenses, or other personal expenditures that can be matched with income of other taxpayers.
Because alimony payments are recurring, we would not have the same concern about one-time itemizers that we would in the moving expense case. But only a minority of taxpayers pay alimony, so how can it be fair that an alimony-payer and a non-alimony-payer face the same standard deduction, if alimony is below the line? But, again, the same can be said for homeowners, residents of high-tax states, large charitable donors, etc. As with moving expenses, the key is that AGI began as a relatively coherent concept in the context of the percentage standard deduction, but over time began to lose that coherence as more and more personal deductions were placed above the line. And as it lost that coherence, the argument for grouping disparate items under the standard deduction weakened as well.
In 1977, Congress simplified the standard deduction, but in so doing it also pushed the standard deduction farther away from its original simplification purposes and more toward being a zero bracket. Just prior to the 1977 changes, the standard deduction consisted of a minimum standard deduction (the low income allowance), a percentage standard deduction (at that time 16% of AGI), and a cap on the total standard deduction. This was a relatively confusing patchwork, particularly given that the minimum and maximum standard deductions were only $700 apart.
Perhaps recognizing that the existing framework had become unwieldy and that the standard deduction had been effectively playing a strong zero-bracket role since 1969, Congress replaced the framework with one flat standard deduction, much as we have today under current law. However, instead of treating the standard deduction as a deduction from AGI, as was in effect prior to 1977 (and after 1986), Congress created the “zero bracket amount” and built the deduction into the tax tables, the rate schedules, and the definition of “taxable income.” The changes allowed for the first $2200 ($3200 for joint returns) (~$8000/$11,700 in 2011) to be tax-free. Making the change explicit, Congress redesignated the standard deduction the “zero bracket amount.”
To be clear, the 1977 standard deduction/ZBA was not a true, independent ZBA in the sense used in this article. Itemized deductions were still subject to this standard deduction/ZBA. In particular, the only itemized deductions allowable in computing taxable income were those that were in excess of the standard deduction/ZBA. For those taxpayers that, for one reason or another, were not entitled to the full standard deduction/ZBA, such as a dependent of another, the concept of the “unused zero bracket amount” allowed for the amount by which the standard deduction/ZBA exceeded itemized deductions to be added back into taxable income, effectively denying such taxpayers any benefit of the standard deduction/ZBA, much as under current law. In its report on the bill, the Senate Finance Committee switched back and forth between calling this provision by its new name, the zero bracket amount, and simply calling it a standard deduction, highlighting the fact that the concepts were effectively interchangeable.
Nonetheless, by replacing the percentage standard deduction with a flat standard deduction, the 1977 Act removed one of the few remaining pure simplification features of the standard deduction. Furthermore, by adding the standard deduction to the rate schedules, Congress made clear the effect of a standard deduction, namely, that it functions as a floor on itemized deductions. Instead of calculating taxable income by subtracting itemized deductions, the Code after 1977 provided for subtracting only the excess of itemized deductions over the standard deduction/ZBA. The Senate Finance Committee titled the section of its report on this provision of the act, “Conversion of standard deduction into zero bracket amount and floor under itemized deductions.”
It is thus ironic that Congress named primarily simplification purposes in making the changes. While these changes allowed for a greater amount of tax-free income to low-income people, the shift from a standard deduction to a zero bracket seems to have been motivated out of belief that the simplifying provision—the standard deduction—had itself become too complex. Because the percentage standard deduction occupied a relatively narrow slice of itemizable expenses, the belief was that simply making it flat would remove a large degree of complexity, with relatively little negative effect. As discussed in the next section, this was not the case.
From 1977 to 1986 there were few changes to the structure of the standard deduction/ZBA itself. In 1978, the amounts were increased to $2300 for single taxpayers, and $3400 for joint taxpayers. The change was made to offset the effects of inflation, but because of continuing inflationary erosion, the amounts were ultimately indexed to inflation by the Economic Recovery Tax Act of 1981.
The 1981 Act also contained a particularly glaring example of the breakdown of the distinctions between above- and below-the-line deductions. The Act provided that, for a four-year period, charitable contributions would be moved partly above the line, thus making them partially deductible in calculating AGI. The 1944 distinction between business-related deductions—placed above the line—and personal deductions—placed below—is, admittedly, fuzzy at the margins. Nonetheless, charitable contributions are perhaps the purest example of personal deductions, having almost no business or income-producing purpose. Placing the charitable contribution above the line while keeping, e.g., the miscellaneous itemized deductions below the line is difficult to justify. Furthermore, as in the case of alimony, since the standard deduction was already intended to include an estimate of charitable contributions—as it had since 1944—moving the deduction above the line essentially allowed non-itemizers to get double the tax benefit.
Congress’s intent in creating the zero bracket amount in 1977 and moving the calculation of taxable income to the tax tables and rate schedules was ostensibly simplification. However, commentators realized fairly early that the change likely introduced too much additional complexity in exchange for relatively little simplification benefit. The value of the change to non-itemizers was relatively small—just the removal of one straightforward subtraction step, since they no longer had to separately subtract the standard deduction. Indeed, that benefit could have been achieved without altering the rate schedule and the definition of taxable income at all, since the tables themselves could have given tax net of the standard deduction (as they did prior to 1977). The change to the rate schedule and the definition of taxable income, therefore, was really only a simplification for the estimated 4% of taxpayers who were both non-itemizers (since those who were itemizers would perform that subtraction step anyway) and whose incomes would have exceeded the income ceilings on the tax tables, since they were the only taxpayers who would have necessarily had to do the additional subtraction step.
In exchange for that minor benefit, the change to the rate schedules and the definition of taxable income required a number of technical and conforming changes to the then-effective Code provisions dealing with net operating losses, lump-sum distributions, short tax-year computations, percentage depletion limitations, trust distributions, income sourcing, capital losses, income averaging, and others. Commentators quickly determined that they preferred that the standard deduction be treated as a deduction from income, rather than a zero bracket. In 1986, Congress relented and created the basic structure for the standard deduction as it still exists today.
In this section, this article discusses the principal problems with the standard deduction. First, it notes the theoretical problems with having a single provision serve both a simplification purpose and a progressivity purpose, and especially that such a dual role leads to conflicting policy arguments for and against the current standard deduction. Second, it notes that partly as a result of these conflicts, the standard deduction exerts a significant destabilizing force on policy debates regarding definitions of income and the proper role of deductions.
The discussion that follows relies on the equivalence between (1) a standard deduction and (2) an exemption (or ZBA) plus a floor under itemized deductions, each of the same amount as the standard deduction. To see this equivalence, consider the following example. Suppose there is a tax system, T, much like ours, with a standard deduction of $10,000, and where taxpayers deduct the larger of the standard deduction or their total itemized deductions. And suppose there is another tax system, T’, which is the same as T, except that there is no standard deduction, a $10,000 exemption (or ZBA) for all taxpayers, and a $10,000 floor under the itemized deductions—expenses are only deductible to the extent they exceed $10,000. There is a single tax rate of 50% in both systems.
Example 1a: TP has AGI of $100,000 and deductible expenses of $20,000. Under T, TP itemizes, since his expenses are greater than the standard deduction. His taxable income is thus $100,000 – $20,000 = $80,000, and his total tax is .5 * $80,000 = $40,000. Under T’, TP can only deduct $20,000 – $10,000 of expenses, but gets a $10,000 exemption. His taxable income is thus $100,000 – $10,000 – $10,000 = $80,000, and his total tax is again .5 * $80,000 = $40,000.
Example 1b: If on the other hand TP has no deductible expenses, then, under T, he deducts only the standard deduction amount. His taxable income is thus $100,000 – $10,000 = $90,000, and his total tax is .5 * $90,000 = $45,000. Under T’, he receives only the exemption. His taxable income is thus $100,000 – $10,000 = $90,000, and his total tax is again .5 * $90,000 = $45,000.
The result is thus the same for the taxpayer under either T or T’. This is true generally for any choice of AGI, floor and exemption amounts, and standard deduction amount, provided that the floor and exemption amounts are both equal to the standard deduction amount. Thus, if we simply replaced the standard deduction with a ZBA of the same amount and instituted a floor on total itemized deductions of the same amount, we would have a system precisely equivalent to our current system.
The standard deduction performs two functions. First, it is a substitute for the itemized deductions for those with relatively few itemizable expenses, in order to lower the administrative costs—for the taxpayer and the government—of trying to measure the expenses precisely. Second, it is an element of progressivity, providing, along with the personal exemptions, for a portion of income that is not taxed. The problems with the standard deduction arise because neither of the above statements is entirely accurate—indeed, that would be impossible. The standard deduction cannot both lower taxable income and provide for some amount of taxable income to be untaxed (or at least not well, as we will see).
In other words, a single provision is being used both to define the tax base and as part of the rate structure applied to that base, and that is simply contradictory. If the standard deduction is a substitute for itemized deductions in calculating taxable income, then it cannot also act as a ZBA with respect to an amount of taxable income equal to the deductions—in essence, part of the ZBA has already been “used up.” Even though the result might be computationally equivalent—in either case, the tax on the relevant amount is zero—the policy implications and criticisms are quite different depending on whether one sees the standard deduction primarily as a substitute for the itemized deductions—that is, whether it is part of the tax base calculation—or as an element of progressivity—that is, whether it is part of the rate structure. What commentary exists on the standard deduction understands that the dual role creates problems, and it discusses some, but not all, of the associated problems discussed here. But the commentary largely fails to explain that particular criticisms are associated with particular assumptions about the primary role of the standard deduction, and it frequently lumps together criticisms that derive from contradictory assumptions. For example, a leading treatise states that as a result of the standard deduction a non-itemizer does not get a tax benefit from his charitable contributions, medical expenses, or casualty losses—a criticism that, as discussed below, only applies if we view the standard deduction primarily as an element of progressivity—but that a fixed standard deduction ought to be replaced with one calculated as a percentage of AGI—an appropriate solution only if we see the standard deduction as primarily an element of simplification. The problem and the solution are unrelated.
To show the contradiction, I imagine two extreme, but computationally equivalent, cases: one, in which Congress has effectively decided not to have a zero bracket at all—to tax the first dollar of income—and to have the standard deduction act entirely as a substitute for the itemized deductions; and a second, in which Congress has effectively decided to have a zero bracket amount—a ZBA—exactly equal to the standard deduction, but to have a floor on combined itemized expenses also exactly equal to that standard deduction. I also examine a third case, where the ZBA is implicitly fixed at some amount less than the full standard deduction. As I will show, the policy criticisms vary significantly depending on this choice of frame.
It is important to note at the outset that these cases are revenue- and distribution-neutral; a given taxpayer’s tax return would not be affected by this choice of frame. The purpose of this exercise is not to argue that one or the other view changes anything computationally—as shown in Part III.A, supra, the two are equivalent. Rather, the purpose is to show the contradictory criticisms that result from the two frames, and to ask what conclusions, if any, we can draw about the current system’s trade-offs between progressivity and simplification.
First, consider the case where the standard deduction is not intended to provide for some amount of untaxed income—no zero bracket—but where the standard deduction is entirely a simplified substitute for the itemized deductions. In theory, this allows for lower administrative costs to both the taxpayer and the government. The taxpayer does not have to worry about record-keeping or difficult calculations, and the government does not face the risk of miscalculations or the higher cost of ensuring compliance. But taxpayers and the government are paying a high price for this benefit.
The standard deduction, in this simplification-centric view, acts as a payment by the government to taxpayers in exchange for taxpayers waiving their right to itemize. Viewed in this way, the payment is grossly over-generous and distributed exactly backwards. A non-itemizer with low itemizable expenses and a non-itemizer with high itemizable expenses each get exactly the same gross amount from the government in exchange for not itemizing. But net of their itemizable expenses, the payments are distributed upside down: the taxpayer with the least to benefit from the itemized deductions receives the largest net payment, while the taxpayer with the most to gain receives the least.
Example 2: TP1 and TP2 each have AGI of $50,000. TP1 has $1000 in itemizable expenses; TP2 has $4000. The standard deduction is $5000. The tax rate is 50%. Because the taxpayers could deduct their itemizable expenses to calculate taxable income, TP1’s net gain from instead taking the standard deduction is $2000, while TP2’s is only $500, even though TP2 has given up a greater amount of itemized deductions.
If we believe that the itemized deductions are proper in measuring Haig-Simons income, then the standard deduction is functionally an instrument of regressivity in the simplification-centric case, since taxpayers with lower taxable income receive lower tax benefits.
Example 3: The same facts as in Example 2. Using the itemized deductions, TP1 has taxable income of $49,000, and TP2 has taxable income of $46,000. If they use the standard deduction, each taxpayer instead has taxable income of $45,000. Using the standard deduction, TP1 lowers her taxable income by $3000 more than TP2 does, and thus receives a net benefit that is $1500 more than TP2’s, despite having a higher “true” taxable income.
Note, however, that the conclusion that the net payments are distributed backwards is not dependent on our view of whether the personal deductions are proper in measuring income. Even if the deductions are viewed entirely as subsidies, the government is simply paying the wrong people—it pays the most to those with the lowest cost of itemizing and the least to give up. While not necessarily regressive, the payments still appear inequitable and wasteful, if only because of their arbitrary distribution.
Some have called the denial of itemization a taxpayer equity issue by arguing that it privileges, say, charitable spending by one kind of taxpayer over another. But that is not the case in the simplification-centric scenario, since non-itemizers are getting a better deal—the full value of what would have been their itemized deductions plus an additional exemption amount. Indeed, by granting that additional exemption, the government is still somewhat subsidizing charitable giving, even if it not at the margin. Because this is the no-ZBA case—where we take as fixed the intent to tax the first dollar of taxable income—the baseline comparison should be to a tax system with no standard deduction and no zero bracket. It follows that the grant of an additional deduction on top of that would in most cases increase charitable giving among those who receive that benefit, since the taxpayer has more after-tax dollars to spread around. In economics terminology, the taxpayer’s budget constraint has shifted such that consumption of most goods—including charitable giving—will increase, despite the fact that there is no change in the after-tax marginal cost of giving.
Therefore, the actual equity issue in the no-ZBA case is the distribution of these tax windfalls. If the windfall is the payment for the denial of itemization, then the windfalls are being distributed exactly in reverse—Congress effectively decided to provide non-itemizers with a generous tax benefit in order to have a more simplified tax system, but that benefit is distributed inequitably among non-itemizers; those with high itemizable expenses get relatively little benefit, while those with low itemizable expenses get a large benefit.
Now consider the case where the intent of the standard deduction is to be a ZBA—to provide for an amount of taxable income that is subject to a 0% tax rate. Because the standard deduction is linked to the itemized deductions, a ZBA equal to the standard deduction amount implies a floor on itemized deductions also equal to that standard deduction amount—taxpayers can only deduct itemizable expenses in excess of the standard deduction amount. This progressivity-centric case is somewhat easier to understand, since floors already apply to a number of the itemized deductions and because of the clear Congressional intent for the standard deduction to play a progressivity role. But there are several problems with this current view. First, there is little reason for the floor for itemized deductions to be equal to the ZBA; the proper amount of itemized deductions should not relate to the proper amount of untaxed income. Second, there is no good reason why the ability to deduct one kind of expense depends on other, unrelated expenses. Why should the size of one’s mortgage determine whether or not one can deduct charitable contributions? Here, the faulty equity argument mentioned in the prior section regarding the different treatment of itemizable expenses becomes more relevant, since the Tax Code is distributing the benefits of itemization arbitrarily. It is important to be precise here, though: this equity problem is not with the standard deduction per se. Combining unrelated deductions is not objectionable when the standard deduction is substituting for those deductions and is by definition equal or greater than the total of those deductions. What is objectionable is applying a single floor to a group of personal deductions in order to determine the availability of any deduction.
Third, there are already existing floors that serve particular purposes, but by lumping all the personal deductions into one category, the purposes of those floors become blurred and muted. For example, the floor for the medical expense deduction could be said to exist primarily because our concern is with people who are abnormally unhealthy. As William Andrews argued, the distribution of medical expenses can be seen as the distribution of differences in health, and treating those with bad health as “consuming” health care would be equivalent to taxing them based on that health. But there is still some baseline level of health care that everyone uses, even in good health. Therefore, the 7.5% of AGI floor for medical expenses could be said to crudely limit the deduction to those in bad health. But any careful calibration of the floor to measure bad health is wasted for non-itemizers, because the amount above the floor just gets tossed back into the mix to be subject to the standard deduction floor.
Example 4: TP1 has AGI of $50,000 and $5000 in medical expenses; TP2 is the same but also has $5000 in mortgage interest. The standard deduction is $5000. Since 7.5% of TP1’s AGI is $3750, only $1250 of TP1’s medical expenses will be considered the extraordinary costs of bad health and thus be deductible—but because that is her only deductible expense, it will not end up being deducted because she has not surpassed the overall standard deduction floor. TP2, however, has surpassed the standard deduction floor due to her mortgage interest, and thus she may take the additional $1250 deduction, despite having equally bad health as TP1.
So if any of the 63% of taxpayers who take the standard deduction get sick, then Congress’s determination of what expenses constitute bad health simply does not apply to them. This argument is related to the argument that the standard deduction distributes the benefits of itemized deductions arbitrarily, but the issue here is system design rather than equity—the work that the itemized deduction floors are supposed to be doing is wasted.
At this point, a reader would be justified in questioning whether either of the extreme cases described above reflects how the standard deduction was intended—likely a mix of the two was intended, where the government does desire some non-zero ZBA, but at an amount less than the full standard deduction amount. While this may seem reasonable, it requires introducing substantial additional complexity, bordering on absurdity, to keep the dual roles fully bifurcated and still get the same result as under the current system—so much so, that it is impossible to believe that this is how Congress intended the standard deduction to be viewed. First, it requires separating the deduction amount from the itemization threshold such that the deduction amount is less than the threshold amount above which one is allowed to itemize, with that deduction equal to the difference between the threshold and the ZBA. Second, it would require that itemizers only be allowed to deduct their expenses above a floor equal to this implied ZBA.
Example 5: Assume that instead of a standard deduction of $5000, there is a ZBA of $4000 and a deduction of $1000 given to those whose itemizable expenses do not exceed a threshold of $5000. Those whose expenses do exceed $5000 are permitted to deduct all expenses above a floor of $4000. Thus a non-itemizer (one whose expenses do not exceed $5000) is not taxed on a) $4000 plus b) an additional $1000 (total of $5000), while an itemizer (one whose expenses exceed $5000) is not taxed on a) $4000 plus b) the amount of his itemizable expenses above $4000 (total of at least $5000). This is equivalent to having no ZBA while allowing non-itemizers a standard deduction of $5000 and allowing itemizers to fully deduct their itemizable expenses.
While the mental gymnastics required to think coherently about this mid-ZBA option ought to alarm us, this scenario has some theoretical support, even if it would be administratively impossible. Louis Kaplow, in discussing the effects of separating a standard deduction from an itemization threshold, notes that having a deduction less than the threshold is likely to be more accurate, and thus have lower equity costs, than having the deduction equal to a threshold. A standard deduction that does not play a ZBA role is in essence an estimate of the non-itemizers’ itemizable expenses. To minimize the error cost of that estimation, we should use the mean of the distribution of itemizable expenses as the optimal standard deduction amount, and if itemizable expenses cluster toward $0 for non-itemizers, as is plausible, then the optimal standard deduction may actually be less than half of the threshold. On the other hand, this interpretation reintroduces the asymmetry between itemizers and non-itemizers that was effectively removed in the simplification-centric and progressivity-centric cases. In this hybrid case, we deny a deduction to itemizers for expenses below $4000, but also provide an additional $1000 of deductions to non-itemizers with expenses below $5000 (most of whom would have expenses below $4000). In other words, non-itemizers could deduct an estimate of their expenses below $4000 (the $1000 standard deduction), while itemizers could not deduct any amount below $4000.
Furthermore, in this hybrid case we have each of the costs discussed in the simplification- and progressivity-centric cases, though each is somewhat lower in magnitude. The net standard deduction payments continue to be distributed inequitably, with those with the lowest itemizable expenses receiving the greatest benefit. Moreover, the single floor and its arbitrary connection to the ZBA remain.
There are yet more ways to conceptualize the dual roles of the standard deduction. I briefly note two others here. As with the frames presented above, each of these frames also implies its own set of criticisms.
One could describe the difference between the standard deduction amount and actual itemizable expenses as a variable ZBA that shrinks as itemizable expenses grow. One could attempt to justify this structure by noting that itemizable expenses tend to rise with income, and so the ZBA would tend to shrink as income grew. I focus on the above conceptions because they allow the clearest bifurcation of the simplification and progressivity roles. Moreover, conceiving of the standard deduction as a ZBA that phases-out with itemizable expenses is flawed. First, if the intent is to have a ZBA phase out as income rose, then it should be indexed to AGI or another income measure, not itemizable expenses. Second, while itemizable expenses may tend to rise with income generally, the distribution of expenses at a given income level is more likely to reflect the opposite—the higher the expenses, the more limited the ability to pay. Thus we end up with the same criticisms as in Part III.B.1.
Another possible interpretation of the standard deduction is that it effectively creates two different tax bases and rate structures. Itemizers—generally higher income taxpayers—are taxed based on taxable income and using a graduated rate structure that does not have a zero bracket, while non-itemizers—generally lower income taxpayers—are taxed on the basis of AGI, but with a rate structure that includes the zero bracket. In other words, instead of calling the standard deduction a trade-off between itemizing and simplifying, we can call it trade-off between a lower tax base and a zero bracket. In essence, the two distinct frames presented in Parts III.B.1 (no ZBA) and 2 (high ZBA) are both true at the same time, but for different taxpayers.
The prior section addressed theoretical and conceptual incoherence of the current standard deduction. Now, I turn to the practical effects of that incoherence. First, the fact that it is doing “double duty” as an income measurement tool and a ZBA means that we cannot resolve issues relating to the equity of particular personal deductions. Second, the fact that the personal deductions hit a minority of taxpayers means that most voters will not have the proper incentives to care about these provisions, which in turns enables interest groups to capture the political process. Third, it minimizes and distorts the ability of itemized deductions to motivate certain behavior, such as contributing to charity. Finally, by introducing additional arbitrariness and by adding to the erosion of a clearly identifiable tax base, the standard deduction may aid in undermining respect for the tax system. This increases enforcement costs and makes it more difficult for the government to raise revenue in the future.
Consider the potential expansion of the charitable deduction to non-itemizers, proposed most recently in 2005. A large part of the public argument for doing so was that it is “unfair” to allow the deduction for some people and not for others. But that criticism implicitly assumes that the standard deduction is really a ZBA, not a simplified tool for measuring income. If the standard deduction is intended to be a proxy for personal deductions like the charitable deduction, then non-itemizers are already getting the full benefit of their charitable deductions—and then some. But to many taxpayers, it probably does not feel like they are getting the benefit, and in part they are right, since they do not feel the incentive effects of the tax deduction at the margin. This disparate treatment lends support to oversimplified fairness arguments about extending the charitable deduction, even though whether the current deduction is in fact unfair is indeterminate. Indeed, as discussed above, under the simplification-centric view, the lump-sum benefit of a large standard deduction likely does partially subsidize charitable giving and other itemizable expenses.
Moreover, if Congress ended up extending the charitable contribution to non-itemizers, as they have in the past, they may feel increased pressure to do the same for the other itemized deductions. After all, it would be just as “unfair” to deny the benefits of the deduction for state and local taxes to a majority of taxpayers. The advantage of such bills is that they may end up de facto answering the question of what, exactly, the standard deduction is intended to do. The more below-the-line deductions move above the line, the more the standard deduction is doing the work of a ZBA. But it would be preferable if those questions were answered through a reasoned policy analysis, rather than just as a byproduct of political forces that the standard deduction distortions helped create in the first place. And the ad hoc way in which deductions are placed below or above the line has already removed much of the original coherence between those two categories of deduction.
It is not just the equity arguments that are distorted. The political debate itself is skewed because only some taxpayers will receive any direct benefit as a result of incremental change in the itemized deductions. One can see this in, for example, the debates around the home mortgage interest deduction. Although the issue is complicated, there is some general acceptance that the home mortgage interest deduction is more of a tax expenditure than a tool in properly measuring income. As such, it is an “upside down” subsidy, since those with higher marginal tax rates receive larger subsidies as a percentage of their interest expense than those with lower marginal rates. Repealing the mortgage interest deduction or replacing it with a tax credit would thus seem to be consistent with general equitable principles, but the conventional wisdom is that the mortgage interest deduction cannot be touched. One possible reason for this is that those who currently take the mortgage interest deduction would immediately see their tax bill increased as a direct result, but non-itemizers might not see any direct change, at least at first. In other words, the change initially would harm the current beneficiaries, but not directly benefit non-itemizers (or non-homeowners in general). Even if a repeal were accompanied by, say, a lowering of tax rates in order to be revenue neutral, the benefits would likely be spread out—including going to those who had seen their interest deduction lowered—and thus not felt as strongly by non-itemizers, even if they knew that those benefits were a result of the repeal. And a distributionally neutral adjustment would heavily favor rich renters. Repeal is politically unlikely in that situation. Furthermore, as Samansky has noted, the effect can also cut in the other direction. While the standard deduction may make it more difficult to repeal some personal deductions that ought to be repealed, it can also give the government cover when trying to repeal personal deductions that perhaps should stay. During the debates over the Tax Reform Act of 1986, the Reagan Administration argued that the state and local tax deduction should be repealed in part because its benefit fell on so few people. Samansky writes:
[A]ny policy justification for the itemized deductions became irrelevant after it was recognized that most individuals could not deduct them. The issue became one of political power, including consideration of the effect of the tax burden on various groups and states. It is unlikely that attempts to rationalize the itemized deductions can obtain much political support under these circumstances.
In short, the standard deduction has the effect of minimizing policy arguments related to the itemized deductions and elevating the effects of political and interest group pressure.
The standard deduction also has the effect of minimizing any well-intentioned incentives written into the Code. Whether the Tax Code should be used to penalize or subsidize particular behavior unrelated to measurement of income has long been a debated issue, but there is no doubt that the Code is so used. But if those incentives are in the form of itemized deductions, then at best they hit only around 35% of taxpayers. The existence of the standard deduction thus undermines an incentive’s efficacy. And the problem is compounded by the existing deduction floors, such as the 7.5% of AGI floor for medical expenses. The combination of the standard deduction and the existing floors may create many complex or even perverse incentives.
Increasingly, Congress is structuring tax incentives as tax credits or placing them above the line. This avoids the standard deduction problem, but creates another by making the Tax Code more complicated (when the point of the standard deduction was simplification) and less coherent (since it is no longer clear why some provisions merit above-the-line treatment while others do not). For example, the deduction for contributions to health savings accounts is above the line, while the deduction for medical expenses is below; the deduction for work-related expenses incurred by teachers is above the line, while that for most other work is below.
Finally, there is the issue of overall respect for the tax system. As discussed above, the current system could be seen as two different tax systems: non-itemizers are taxed on AGI, but with a rate structure that includes zero bracket, while itemizers are taxed on taxable income, but with a rate structure that does not include a zero bracket. While the computational results are the same in either frame, the change in perspective could be profound. If middle-income taxpayers see themselves as being taxed on a different tax base than high-income taxpayers, it could undermine belief in the tax system as fundamentally fair. Even if the zero bracket is a fair trade-off for a higher tax base in terms of overall tax benefit, there still may be a belief that vertical equity has been compromised. It is beyond the scope of this paper—and perhaps impossible—to tie the tax gap or current anti-tax movements directly to any disrespect bred by the standard deduction. But it certainly does not help.
In Part II, this article showed the historical progression of the standard deduction from a purely simplifying provision to one that plays an increasingly important progressivity role. As that part shows, there has been a clear Congressional intent to have some relatively large amount of untaxed income. In this part presented two extreme versions of the standard deduction—one in which it functioned entirely as a simplification measure with no ZBA role and one where it was entirely a ZBA (plus a floor under the itemized deductions). It also presented several other ways of framing the standard deduction: as a combination of ZBA and simplified substitute for the itemized deductions (involving a separation of the standard deduction amount and the threshold above which such deduction can be taken); as a variable ZBA that phases out with itemizable expenses; and as two distinct tax bases and rate structures depending on the amount of itemizable expenses. Of these different views, the most coherent is the progressivity-centric, high-ZBA view—that the standard deduction is simply a ZBA coupled with a single floor under the itemized deductions. The flaws in the other views are too great to withstand scrutiny, and the current structure of the standard deduction is too far divorced from those view’s purposes. The progressivity-centric view is also most consistent with the legislative history and Congress’s intent to allow for a large amount of untaxed income, especially to low-income persons. As shown in Part III.B.2, this high-ZBA view is not without flaws, but as shown in the following section, it is also the easiest to fix.
In what follows, the article proposes and discusses a reform to the standard deduction and the itemized deductions: replace the standard deduction with a true zero bracket amount, independent of any implied floor on itemized deductions, and disaggregate that single implied floor into explicit floors on individual deductions. This would address most of the criticisms in Part III with only a marginal increase in complexity. This part also discusses why floors such as those proposed here have particular tax policy benefits.
The two largest drivers of the problems I have discussed in Part III are, first, the combination of income-measurement and ZBA functions—combining a definition of the tax base with a tax rate schedule—and, second, the grouping of unrelated deductions under one standard deduction. To fix these problems requires two steps.
First, the single standard deduction should be separated into a true, independent ZBA and a separate floor under the itemized deductions. This change would make explicit what is already the case in the minds of most analysts and commentators: that the standard deduction operates primarily as an element of progressivity—a ZBA. It would be consistent with Congress’s explicit intent for the standard deduction to play a ZBA role. It would separate the determination of the proper amount of untaxed income—the ZBA—from the proper amount of itemized deductions, while still retaining the overall progressivity of the current system. It would thus positively affect horizontal equity by making the determination of the proper amount of tax-free income independent from the amount of itemizable expenses a taxpayer has. Finally, it would allow policymakers to more effectively target tax relief, since they would have the ability to do so simply with adjustments to the rate structure without affecting tax deductions and credits.
Second, rather than a single floor under the itemized deductions, the floor itself should be disaggregated into multiple, independent floors under each itemized deduction. This would separate the determination of the proper deduction amount for a given itemizable expense from the amounts of any other itemizable expenses. It would thus positively affect horizontal equity by treating all taxpayers the same with respect to a given category of itemizable expense. It would allow policymakers to more effectively optimize deductions for itemizable expenses, since they could adjust the floors to optimize trade-offs between simplicity, revenue, incentives, and equity for each deduction. And it would largely remove the no-longer-coherent distinction between above-the-line and below-the-line deductions. This distinction was meaningful when the standard deduction was intended primarily to approximate personal deductions, but since that has not been the case since at least 1977, the distinction now only leads to more arbitrariness in the Tax Code.
The strongest arguments against such a reform are as follows: First, it would entail some additional complexity; this criticism is addressed in Part IV.D. Second, that it would have some negative distributional changes as a result of the interaction of multiple floors, as shown in Part IV.B; this criticism is addressed in part by the affirmative argument for the benefits of tax floors, presented in Part IV.E.
To see how this proposal would work in practice, consider the following simplified example. Suppose a standard deduction SD covered deductions for expenses a, b, c, and d. If a taxpayer had expenses such that a + b + c + d < SD, then he would take SD as the deduction rather than itemize a, b, c, and d. As shown above, this is equivalent to having an itemized deductions floor equal to SD and a zero bracket (or exemption) equal to SD. However, suppose instead of the single floor we create separate floors A, B, C, and D for each of the deduction categories. That is, if a taxpayer had expenses such that a < A, b < B, etc., then he would not itemize those expenses. A taxpayer would only itemize the amount by which a exceeded A, b exceeded B, etc. For taxpayers with low itemizable expenses, little complexity would be added. But some previous non-itemizers with expenses approaching the standard deduction may find themselves itemizing in some cases.
Suppose that we initially set the individual floors such that A + B + C + D = SD (and create a new ZBA equal to SD). Suppose further that there is a taxpayer with expenses a + b + c + d < SD, such that a < A, b < B, and c < C, but d > D. Such a taxpayer would be under the floors for a, b, and c but would itemize her d expenses under the new regime. Her total itemized deductions would thus be d – D, which would be equivalent to deducting an amount SD + (d – D) > SD under the old regime.
Example 6: Suppose that we replace a $10,000 standard deduction with a $10,000 ZBA and with floors of $2500 on each of deductions a, b, c, and d. TP has expenses such that a = $0, b = $100, c = $1000, and d = $7900. TP’s total expenses are less than $10,000, and thus TP would have taken the standard deduction under the old regime. However, now TP’s expenses on d exceed that floor of $2500. TP continues to get the benefit of the $10,000 reduction in taxable income (now in the form of a ZBA), but can reduce taxable income further by the amount by which TP’s d expenses exceed $2500, i.e., $5400. Therefore TP will face tax on $5400 less of income under the new regime than under the old regime (assuming no other changes).
In addition to becoming a partial itemizer, which would introduce some additional complexity, the taxpayer in the example would also have less taxable income under the new regime than the old, even though the total of the new floors equals the old single standard deduction. Thus, the introduction of the floors effects both simplification and this taxpayer’s tax burden. As Louis Kaplow notes, the choice of thresholds and floors in this context will always affect the degree of simplification and the distribution of tax burdens within the same income class, even if revenue and overall progressivity are held constant.
To see the effect on relative tax burdens, suppose there are two taxpayers with the same income and the same total of itemizable expenses, but one of whom has an even distribution of expenses across all the categories, while the other has expenses concentrated all in one category. Suppose further that both taxpayers were itemizers under the old regime, but only by a narrow margin. If we set the total of the floors equal to the old single standard deduction, the second taxpayer would see her taxes decrease while the first would see little or no change, despite their having the same taxable income if floors were not considered.
Example 7: Assume the same standard deduction and floors as the prior example. TP1 has itemizable expenses a = b = c = d = $2600. TP2 has itemizable expenses a = b = c = $0, and d = $10,400. Under the prior regime with a standard deduction of $10,000, each is an itemizer and reduces his taxable income by $10,400. Under the new regime, TP1 can deduct $100 of each of his expenses, the amount by which each exceeds the $2500 floor. Thus, TP1 continues to reduce his taxable income by a total of $10,400 ($10,000 ZBA plus $400 total itemized deductions). TP2, however, can take itemized deductions of $7900, the amount by which his d expenses exceed the $2500 floor. Thus TP2 can now lower his taxable income by a total of $17,900, or $7500 more than TP1 and more than under the prior regime.
But if we raised the total of the floors so that the second taxpayer would have no change in her taxes, then the first taxpayer who had been itemizing with flat expenses narrowly above the old standard deduction would no longer be able to itemize and would lose the net of that amount over the new ZBA of SD.
Example 8: The same facts as the prior example, except that the new floors are set at $7500 instead of $2500. Under this regime, TP2 would have itemized deductions of only $400, the amount by which his d expenses exceed the $7500 floor. Thus, TP2 can reduce his taxable income by $10,400, the same amount as under the prior regime. But TP1 can no longer itemize, and thus can lower his taxable income by only $10,000 (the ZBA) rather than $10,400.
Increasing exemptions or the ZBA would not change the relative tax burdens, since such progressivity changes would apply to both taxpayers equally. There will thus necessarily be some changes in distribution within income classes, even if overall distribution of the tax burden between income classes can be held constant. Some calibration would therefore be required to optimize simplicity, horizontal equity, and revenue. These optimal values for the new floors would probably be such that their sum is greater than SD to avoid creating too many new partial itemizers, and any effects that this would have on tax burdens for those who would otherwise have itemized should be made up for in the rate structure or personal exemptions, such as by having an initial ZBA greater than SD.
An additional feature of this reform proposal is that the system could be adjusted to provide additional progressivity without explicitly raising rates. This is explicitly the case with the increase in personal exemptions with family size. But in addition, the zero bracket could itself phase-out with income, which would cause a flattening of the marginal rates that apply to higher-income taxpayers, such that the marginal and average tax rates converge somewhat, particularly when considering the itemized deduction phase-out. To be clear, however, such changes are equivalent to raising marginal rates during the phase-out period. Thus, the advantage to doing so through the ZBA (or variable deduction floors, as discussed in Part IV.E, infra) is really one of political economy and optics, rather than substance.
In the 1960s, when policymakers effectively changed the standard deduction from being primarily a simplification of the itemized deductions to being the “low income allowance” equal to poverty-level income, a major reason for doing so was a belief that the standard deduction would allow for more targeted tax relief. Instead of simply creating a zero bracket (or increasing exemptions), which would apply to everyone, the government decided, in effect, to have a zero bracket that phased out with itemized deductions, as a rough proxy for income. This was said to be “the most equitable and efficient method available of directing tax relief to persons in the lowest income ranges.” In other words, the concern was that merely adding a zero bracket (or increasing exemptions) would be less progressive than increasing the standard deduction. Yet, as discussed here, that is simply not the case when the zero bracket is accompanied by a floor (or floors) on the itemized deductions. As noted above, one can actually achieve more progressivity (that is, more focused effective rate increases) though this article’s proposal via the combination of a variable ZBA and variable floors.
The primary argument in favor of keeping the standard deduction is simplification. As of 2007, 63.3% of taxpayers take the standard deduction, which partially accounts for the fact that 32.6% of individual tax returns are nontaxable. Forty-two percent of tax returns are the simplified 1040EZ or 1040A forms. All of this allows for relatively low compliance costs, both on the part of taxpayers and the government. Given the complexity of the tax system as a whole, we should be wary of anything that would upset these achievements.
However, the simplification value of the standard deduction is overstated. First, tax simplification itself has costs, which are better understood today than when the standard deduction was first enacted. Second, the costs of itemizing likely have dropped as technology and third-party reporting has eased the record-keeping and calculation requirements. Third, the change proposed above—adding a ZBA and floors to the itemized deductions—results in only marginally more complexity than the current system.
Simplification has been a goal of tax policy for almost as long as we have had an income tax, yet scholars have in recent years pushed back on this somewhat. This article does not intend to step fully into this debate; it intends only to note that the ground in the debate has shifted somewhat, and that current thinking about the value of simplification points towards relatively limited simplification benefit from the standard deduction.
Commentators typically point to three types of tax complexity: compliance complexity, transactional complexity, and rule complexity. The standard deduction, and the reform proposal in this article, primarily implicate compliance complexity. Indeed, the costs largely boil down to increased computation and record-keeping. Thus, when people speak of the admittedly daunting complexity of the income tax system generally, such invocations do not do a lot of work in supporting the standard deduction, which addresses only a small source of relatively limited complexity. In addition, when commentators discuss complexity for individual taxpayers, they tend to cite elements such as the alternative minimum tax; the earned income tax credit; or the complexities related to domestic relations, dependents, and child care. Rarely do itemized deductions make the list.
In addition, this article’s proposal arguably decreases rule complexity, at least for those taxpayers who are not certain whether their itemized deductions will exceed the standard deduction. While the current system has a single rule for determining when deductions are available, combining unrelated deductions for purposes of satisfying that rule means that it is relatively difficult for some taxpayers to determine whether a given marginal expense will be deductible. If a taxpayer on the cusp of itemizing wished to know whether her next charitable contribution, say, would push her over the standard deduction (and thus be subsidized, allowing her to increase the amount donated), she would have to estimate her likely state taxes for the year, out of pocket medical expenses, and the like. It is much simpler to only consider the total charitable contributions for the year, and whether the next donation will cause her to exceed the charitable contributions deduction floor. Admittedly many itemizers know well ahead of time that they will be itemizing, usually based on their mortgage interest payments, but for some taxpayers disaggregating the floors likely makes the system more transparent, and allows the incentive effects of the deductions to be more salient.
Finally, complexity in the service of accuracy has particular equity benefits, namely the reduction of the risk of over- (or under-) taxation. Louis Kaplow has described this in terms of the risk premium: the value of accuracy to an individual is equivalent to what the individual would be willing to pay to insure against the risk of paying the wrong tax. Assuming decreasing absolute risk aversion, “the value of accuracy will be greater for low-income individuals.” In other words, accuracy in the service of greater equity yields real benefits to taxpayers, particularly the lower-income taxpayers more likely to use the standard deduction.
Of course, the standard deduction could be said to result in under-taxation of individuals, since by definition those who take the standard deduction have itemized deductions less than the standard deduction (or at least less than the standard deduction plus the cost of itemizing). Thus it is not quite as simple as just pointing to a quantification of the benefit of accuracy. Leaving aside the question of whether taking the under-taxation view repeats the conceptual errors discussed in Part III.B, however, recall that under this article’s proposed reform, the additional complexity is largely in the form of marginally more computation and record-keeping for some subset of current non-itemizers. Putting forth that effort will, in many cases, result in reduced taxation for that individual compared with doing nothing under this article’s proposed reform (and likely compared with the current status quo, if there is no adjustment to the ZBA size or the rate structure)—thus resulting in a very real benefit. Again, the purpose of this discussion is not to argue whether the standard deduction’s simplification benefits are net positive or negative; the argument is that the small increase in complexity for a subset of current non-itemizers under this proposal is likely offset in part, and perhaps in whole, by the benefits to those taxpayers.
To put this in concrete terms, consider the likely costs of additional computation and record-keeping. First, in an age of computer-aided tax return preparation, computational complexity is less and less of a burden on taxpayers. In 2006, the most recent tax year for which data is available, 89% of tax returns were computer-prepared, and the percentage has been rising steadily. Computer preparation is not without cost, of course. All but the most basic Turbo Tax software for 2010, for example, costs at least $20. The average minimum fee for basic professional preparation is $79, but can go up quickly. However, since 2002 the IRS has had a “Free File” program, under which private companies have agreed to provide free computer-aided preparation and filing for taxpayers with AGI less than $58,000. This is in addition to the long-standing Volunteer Income Tax Assistance (VITA) and Tax Counseling for the Elderly programs, under which volunteers provide tax help, preparation, and filing for low- and middle-income taxpayers, and those 60 or over. In all, access to computer-aided preparation is likely available to nearly all taxpayers at little or no cost.
The cost of record-keeping and entering the information is somewhat harder to pinpoint. One study put the average cost of itemizing in 1982 at $43, or roughly $100 in 2011 dollars, though for taxpayers with income less than $25,000 in 1982 (~$57,000 in 2011), costs were $25 (~$57 in 2011) or less. However, these cost estimates include the computational and return-preparation costs now largely being offset by software.
Furthermore, the growth of third-party reporting and the like has eased the record-keeping burden. For home mortgage interest, for example, taxpayers that pay over $600 in interest in a given tax year toward their bank-financed mortgage will receive a statement from their bank showing how much interest they paid during the year. The cost of itemizing is relatively small—holding onto the statement, knowing where to enter it, etc. Homeowners might have two or three statements to deal with, considering home equity lines or second homes, but that is hardly a daunting number.
State and local taxes often are similarly well-documented, though there might be more than two or three numbers to deal with. Employees receiving W-2s have the information handed to them, while independent contractors and the self-employed can look at last year’s tax returns (and refunds) to see how big a check they wrote during that tax year. For real estate and other property taxes, homeowners generally receive bills from the city or town showing the taxes due. If a taxpayer pays the taxes directly, he needs to keep track of when he did so (since taxes might have been billed at the end of the year, but not due until the next); if, as is common with home mortgages, the taxpayer’s bank pays the taxes out of escrow, they typically provide that information to the taxpayer. On the margins there may be people with more complicated state-tax situations, but for most people, state taxes are relatively straightforward and well documented.
In the case of medical expenses and casualty and theft losses, the largest expenses, such as an elective surgery or a car theft, are not likely to be recurring and would involve relatively little record-keeping. On the other hand, unreimbursed employee business expenses are more likely to be more recurring and greater in number. So itemization of these expenses may be more of a burden. But that burden is also likely to be borne by more sophisticated taxpayers, who may already be keeping records of their income-producing activities in order to measure profit. Furthermore, because these expenses are generally agreed to be properly deductible in measuring income, this is likely a case where the accuracy arguments carry more weight than the simplification arguments. The inequity of having one’s, say, charitable deduction determined by the size of one’s mortgage may be easier to swallow if we believe that both deductions are essentially handouts from the government. But where someone clearly has expenses of earning income that should lower her tax, tying the ability to do so to the size of her mortgage or state taxes is less acceptable.
Simplification may be the most justified for the charitable contributions deduction, since many people give small, un- or poorly documented amounts to charity, such as cash in the church collection plate or used clothing to thrift shops. But for many gifts, charities already have in place a system of receipts and documentation in order to serve current itemizers. Perhaps some giving that had been ad hoc would move into these more organized settings following a change—writing an annual check to one’s church, for example, instead of dropping cash in the collection plate. Whether this would affect the overall character of giving is beyond the scope of this article, but any change probably would not be dramatic. However, given that the charitable deduction faces perhaps the most political pressure to be moved above the line—and, indeed, was at least partially above the line from 1982 to 1986 and was nearly made so again in 2005—it seems that public opinion does not consider the additional complexity to be large.
The previous two subsections argue that the costs of itemizing are not large and may be justified much of the time. In addition, under this article’s proposal, the costs themselves are limited. First, replacing the standard deduction with floors will not cause most current non-itemizers to begin itemizing. Many non-itemizers will continue to have itemizable expenses below the floor for a given deduction. Second, those that do begin itemizing are likely to do so for only a fraction of the total categories of deduction, perhaps only one. This limits the costs further, since they would not have to bear the full record-keeping costs that current itemizers do. Ultimately, however much additional complexity is created is an empirical question that depends on the size of the ZBA and the floors, and thus requires additional research.
This section address an additional argument in favor of this proposal, separate from the benefits of separating the progressivity and simplification purposes of the standard deduction, namely the particular policy benefits of using floors. As noted above, we can achieve the same effect as a floor with a combination of a standard deduction and a change in exemptions, and the same is true at the level of the individual itemized deductions—rather than having disaggregated floors, we could create disaggregated independent standard deductions for each of the itemized deductions and end up in the same place in terms of distribution and revenue. Yet floors have a number of advantages over individual standard deductions from a policy standpoint.
In particular, floors can be designed to support the underlying policy goals of a particular deduction. The size of the floor, how it is calculated, its distributional effects, etc., can all be tailored to dovetail with the nature and purpose of the deduction itself. Part IV.B presented a simplified example using fixed floors totaling close to the current standard deduction amount. But actually choosing the type and magnitude of floor would likely be the most difficult part of reform.
A floor for a deduction can exist for a number of different reasons. It can be purely to simplify calculation and recordkeeping, by denying small deduction amounts. But it could also be designed to limit tax avoidance behavior, or to optimize the trade-off between incentives and revenue (because the last dollar of spending may need more encouragement than the first, the government may want to focus its money there), or to pinpoint the true effect on income (as with the floor on the medical expense deduction and the costs of bad health), or perhaps simply to limit the revenue effects of deductions we would like to remove but cannot for political reasons.
Consider further the use of floors to try to measure income more accurately. In the examples in Part IV.B, the disparities between the two taxpayers are less under high floors than low floors. But this does not necessarily mean that high floors are the more equitable solution. If a floor exits purely for simplification reasons, then there is a clear trade-off with equity. But if we believe that only expenses above the floor should be deductible in the first place, then it is not actually the case that our two taxpayers have the same taxable income in the view of the government, and our concerns about horizontal equity are less pressing.
Furthermore, any adverse effect on horizontal equity that this article’s proposal would have is less than under the current structure, where taxpayers with the same level of, say, charitable contributions are treated differently. Each personal deduction is independent of the others, and they each accomplish different things. Some are clearly measuring income, some are clearly tax expenditures, some are a little of both. Some are more valuable from a policy standpoint, some are less. The fact that a taxpayer with high expenditures on a particular itemizable expense gets some additional benefit is not unreasonable if that benefit is a result solely of the provisions of that particular deduction and the policies that guide it. Although there are likely to be some tax windfalls handed out, at least they would not be arbitrary.
For the most part, the floors currently in place on the itemized deductions are variable and calculated as a percentage of AGI, rather than being fixed. Variable floors have the appeal of being more progressive, but they may not be appropriate for the given type of expense. For example, the floor of 7.5% of AGI on medical expenses seems inappropriate if the purpose of the deduction is to treat extraordinary medical expenses as reflecting an endowment of bad health rather than consumption. Having the floor be a percentage of AGI implies that rich taxpayers with, say, cancer are healthier than poor taxpayers with cancer. A fixed floor thus seems more appropriate for deductions that are more about income measurement than subsidization.
Put another way, one could view the variable floor on extraordinary medical expenses as a crude approximation of a tax credit, rather than a deduction, for expenses above some fixed floor.
Example 9: Suppose annual medical expenses above $3000 are universally those attributable to bad health, and that there are two taxpayers each with $9000 in medical costs. TP1 has AGI of $40,000 (such that 7.5% of AGI is $3000) and is in a 20% bracket. TP2 has AGI of $80,000 (such that 7.5% is $6000) and is in a 40% tax bracket. TP1 will be able to deduct $6000 of medical expenses, for a net tax benefit (at his 20% rate) of $1200. TP2 will be able to deduct $3000 of medical expenses, for a net tax benefit (at his 40% rate) also of $1200. This is equivalent to both taxpayers receiving a 20% tax credit on the $6000 that, by assumption, is the cost of their bad health.
But tax credits tend to be more favored for tax expenditure subsidies than income-measuring deductions. Thus, again, arguably a variable floor is inappropriate for an income-measuring deduction like that for extraordinary medical expenses. The same logic applies to the 2% floor on miscellaneous itemized deductions and the 10% floor on casualty and theft losses.
Indeed, under a pure Haig-Simons definition of income, there is an argument that the miscellaneous itemized deductions and the deduction for casualty and theft losses should not face a floor at all. For income-measuring deductions, floors should represent the thresholds above which certain expenses should no longer be treated as voluntary consumption but instead should be treated more as differences in endowments. Thus, the floor on medical expenses can be said to reflect a decision that only extraordinary medical expenses should be deductible. The case for a floor on the miscellaneous itemized deduction and the casualty and theft loss deduction is less clear. The miscellaneous itemized deductions are generally for expenses related to the production of income, and casualty and theft losses are clearly subtractions from wealth, which should be deductible under a pure Haig-Simons income tax., 
A variable floor would, however, be more appropriate for the mortgage interest and charitable deductions. The mortgage interest deduction especially is the classic example of Surrey’s “upside-down subsidy,” whereby the subsidies increase with marginal rates and thus with income. Providing a variable floor mutes that effect, by roughly approximating a tax credit with a fixed floor. Assuming mortgage interest tracks income, a variable floor also provides a more treasury-efficient way to target the marginal dollar of spending, since fewer of the dollars below the floor are subsidized.
Under the current system, floors do not appear to be used in this way, and rather seem to be used for a combination of revenue and tax-avoidance reasons. This is likely due, again, to the distorting effect of the standard deduction. Consider the floor for medical expenses. Under the Patient Protection and Affordable Care Act, the 7.5% floor for medical expenses was raised to 10%, likely with the primary purpose of raising revenue to pay for health care reform—the Joint Committee estimates that raising the floor will generate about $15.2 billion from 2010-2019. While there may be policy reasons (other than revenue) for wanting a higher floor, the 7.5% floor has been in place since 1986, so it is only in the face of a pressing need for health care–related revenue that the floor was adjusted. Arguably a more broadly based floor would be a less tempting source of cash if it were explicitly intended to capture the point at which medical expenses should be considered in measuring a taxpayer’s ability to pay.
The use of floors to further specific policy goals need not be limited only to the current itemized deductions. This article’s proposal would remove much of the distinction between above- and below-the-line deductions. Were that to happen, there would be little reason to confine the use of floors to the below-the-line deductions. As noted in Parts II and III.C, the distinction between above- and below-the-line deductions has gradually eroded over the years, such that many above-the-line deductions are just as deserving of floors as the current below-the-line deductions.
Furthermore, even if the optimal floor is fixed rather than variable (and thus the same effect could be achieved with an independent standard deduction), floors continue to have additional policy benefits. First, their perceived effects line up with their actual effects. To put this another way, recall that one of the primary criticisms this article directs at the standard deduction is that it appears to be both a zero bracket and a substitute for the itemized deductions at the same time, and that this leads to confusion in policy debates and in the public mind. Most analysts may think of the standard deduction (and personal exemptions) as a ZBA combined with a floor under the itemized deductions, but it requires an analytical step or two to see that, something most taxpayers are not prepared to do. On the other hand, using a floor makes this effect explicit in the minds of taxpayers and less-informed policymakers. For example, a high floor makes clear to a taxpayer that there is a policy judgment to subsidize only extraordinary expenses. If, instead of a floor, we used a high standard deduction (and a correspondingly lower ZBA) that policy judgment would not be as apparent.
Second, floors provide for more straightforward calibration over time, both of the floors themselves and also of the ZBA. Using floors, rather than standard deductions, requires a substantial independent ZBA. This would give policymakers some slack with which to adjust both the proper amount of untaxed income and the proper amount of tax deduction. If, instead, we used standard deductions, there would be a much smaller independent ZBA, which would allow only more limited policy adjustments lest overall progressivity be affected.
Despite being a major source of distortion in the Tax Code, the standard deduction has received relatively little close analysis and criticism. Perhaps the lack of attention is because tax policymakers and commentators believe they understand its weaknesses, but its ubiquity and importance still merit a closer review than it has been given to date. Upon such a review, there is little to support its continued existence. The compromise between simplification and progressivity at the core of the standard deduction leads to conceptual incoherence and adversely affects equity, progressivity, tax incentive policy, tax reform, and overall clarity of the Tax Code, all in the name of simplicity. And the benefit of that simplicity is overstated, and declining.
Instead, Congress ought to repeal the standard deduction and replace it with an independent zero bracket amount and with independent floors for each itemized deduction. Treating each itemized deduction separately would mean taxpayers in the same income class would be treated the same with respect to each deduction; deductions of one type would no longer be dependent on other, unrelated deductions; the Tax Code would move closer to using true taxable income as a base for all taxpayers; tax incentive policy would be more effective; policy debates over the personal deductions would be more substantive; and the overall sense of the Tax Code’s fairness would improve. In addition, whatever progressivity role the standard deduction played could be easily played by either a true zero-percent tax bracket or expanded personal exemptions.
 See I.R.C. § 63(c) (2010).
 To compute taxable income a taxpayer subtracts the greater of the standard deduction or the itemized deductions. I.R.C. § 63(b), (e) (2010). The itemized deductions are those deductions allowable under Chapter 1 of the Internal Revenue Code other than those allowable in calculating adjusted gross income under § 62 and those for personal exemptions under § 151. I.R.C. § 63(b) (2010). See infra Part II.A.
 For simplicity, I generally disregard the existing personal exemptions under § 151 throughout this article when speaking of a ZBA. Nonetheless, the personal exemptions provide an important complication of the issues discussed herein, namely because (a) they already function as a “pure” ZBA, that is, without any simultaneous simplification purpose, and (b) they provide that the effective ZBA varies with family size.
 The standard deduction for 2010 for an unmarried taxpayer is $5700. Rev. Proc. 2009-50 § 3.11(1), 2009-45 I.R.B. 617. The personal exemption in 2010 is $3650. Id. at § 3.19. This assumes adjusted gross income below $166,800 for a single taxpayer; above that amount, the personal exemption amount begins to phase out, Rev. Proc. 2008-66 § 3.19(2), 2008-45 I.R.B. 1107, though the phase-out did not apply for 2010, Rev. Proc. 2009-50 § 2.08, 2009-45 I.R.B. 617. A married couple with two children does not pay tax on $28,000 of AGI, $11,400 of which is due to the standard deduction. Rev. Proc. 2009-50 §§ 3.11(1), 3.19. The exemption phase-out for a married couple is $250,200. Rev. Proc. 2008-66 § 3.19(2). The $28,000 figure includes $2000 in child tax credits. See I.R.C. § 24(a) (2010).
 Tax Policy Center, Tax Units with Zero or Negative Tax Liability, 2009-2019, available at http://www.taxpolicycenter.org/numbers/Content/PDF/T09-0202.pdf.
 In 2007, the standard deduction was claimed on 63.3% of tax returns. Justin Bryan, Internal Revenue Serv., Individual Income Tax Returns, 2007, Statistics of Income Bulletin, 2009, at 7, available at http://www.irs.gov/pub/irs-soi/09fallbulindincomeret.pdf [hereinafter 2009 SOI Bulletin].
 See infra Part III.
 See, e.g., Staff of joint Comm. on Tax’n, 110th Cong., A Reconsideration of Tax Expenditure Analysis 22 (Joint Comm. Print 2008) [hereinafter A Reconsideration of Tax Expenditure Analysis] (“The JCT Staff views the personal exemptions and the standard deduction as defining a zero-rate bracket . . . .”).
 See infra Part III.B.4.
 Horizontal equity captures the notion that similar taxpayers should face similar tax burdens. See infra note 183.
 The 2010 poverty-level income was $10,830 for a single person and $22,050 for a family of four. Delayed Update of the HHS Poverty Guidelines for the Remainder of 2010, 75 Fed. Reg. 45628-02 (Aug. 3, 2010).
 I know of only two articles in the legal literature that primarily discuss the standard deduction. Louis Kaplow, The Standard Deduction and Floors in the Income Tax, 50 Tax L. Rev. 1 (1994) [hereinafter Kaplow, Standard Deduction]; Allan J. Samansky, Nonstandard Thoughts About the Standard Deduction, 1991 Utah L. Rev. 531. Samansky’s article is the only one I know that directly challenges the structure of the standard deduction. Others include some of the criticisms of the standard deduction that I discuss here. See, e.g., Glenn E. Coven, The Decline and Fall of Taxable Income, 79 Mich. L. Rev. 1525, 1556-64 (1981); Alan L. Feld, Fairness in Rate Cuts in the Individual Income Tax, 69 Cornell L. Rev. 429, 441 (1983); Theodore P. Seto & Sande L. Buhai, Tax and Disability: Ability to Pay and the Taxation of Difference, 154 U. Pa. L. Rev. 1053, 1088-93 (2006); see also Kaplow, Standard Deduction, supra, at 3 & n.10 (noting a similar lack of literature).
 See 2009 SOI Bulletin, supra note 6, at 7.
 See infra Part IV.D.1.
 See, e.g., infra notes 93, 100, 101, & 126 and accompanying text. An important exception to this is Seto, supra note 12, at 1088-93, which describes the tension in the dual purposes of the standard deduction in the course of making an argument for a tax system that more effectively measures ability to pay. See also Feld, supra note 12, at 441 (briefly noting the standard deduction’s “double duty” problem).
 See supra note 3. An important difference between the two is that the total exemption amount increases with number of dependents. Thus the balance between the two (and other provisions, such as the Child Tax Credit, I.R.C. § 24 (2010)) ought to reflect policy decisions about the proper amount of untaxed income relative to family size. Such considerations are beyond the scope of this article.
 See infra Part III.A.
 Readers knowledgeable in the history of the standard deduction will recall that for a period from 1977 to 1986 the standard deduction was in fact labeled the “zero bracket amount” and existed in the rate structure. To be clear, that version of the ZBA was not an independent ZBA in the sense that this article uses the term; it still operated in almost exactly the same way as the current standard deduction that is the subject of this article. See infra Part II.D.
 See Kaplow, Standard Deduction, supra note 12.
 I.R.C. § 63(b) (2010). AGI, in turn, is calculated by deducting certain amounts from gross income. I.R.C. § 62 (2010). The Tax Code also provides for the exclusion of certain amounts from the definition of “gross income.” I.R.C. §§ 61(b), 101-140 (2010).
 Rev. Proc. 2009-50 § 3.11(1), 2009-45 I.R.B. 617. The standard deduction amounts are indexed to inflation and thus adjusted annually. I.R.C. § 63(a)(4) (2010). The standard deduction provisions also provide for other adjustments. For the aged and blind the standard deduction is increased by $1100. I.R.C. § 63(c)(1)(B), (c)(3), (c)(4), (f) (2010); Rev. Proc. 2009-50 § 3.11(3). The amount increases to $1400 for an unmarried taxpayer who is not a surviving spouse. Id. For a person claimed as a dependent on another’s tax return, the basic standard deduction is, for 2010, the greater of $950 or $300 plus the individual’s earned income, up to the typical basic standard deduction amount described above. I.R.C. § 63(c)(5) (2010), Rev. Proc. 2009-50 § 3.11(2). The principal purpose of this reduction is to reduce opportunities for tax avoidance by parents shifting investment income to their children. See 2 Boris I. Bittker & Lawrence Lokken, Federal Taxation of Income, Estates, and Gifts ¶ 30.5.2 (3d ed. 2000).
 I.R.C. § 63(b), (e) (2010) (on election to itemize).
 I.R.C. § 164 (2010).
 I.R.C. § 163(a), (h) (2010).
 I.R.C. § 213 (2010). The deduction for medical expenses is limited to amounts above 7.5% of AGI. I.R.C. § 213(a) (2010).
 I.R.C. § 170(a)(1) (2010). The deduction for charitable contributions can be no more than 50% of a taxpayer’s “contribution base” (and possibly less). I.R.C. § 170(b)(1) (2010). “Contribution base” is AGI computed without regard for any net operating loss carryback to the taxable year under § 172 (2010). I.R.C. § 170(b)(1)(G) (2010).
 I.R.C. § 212 (2010).
 I.R.C. § 162(a) (2010). Section 212 and 162 deductions (among others) are subject to the 2% of AGI floor on miscellaneous itemized deductions. I.R.C. § 67 (2010).
 I.R.C. § 165 (2010). The losses are limited to amounts above $100 per casualty and only to the extent the total losses exceed 10% of AGI. I.R.C. § 165(h) (2010).
 The percentage of households subject to the income tax grew from 5% to 74% during the period. See Lawrence H. Seltzer, The Personal Exemptions in the Income Tax 62 tbl. 9 (1968). The percentage dropped to 56% in 1946.
 Individual Income Tax Act of 1944, Pub. L. No. 78-315, § 9(a), 58 Stat. 231, 236. The cap was later increased to $1000 for single taxpayers. Revenue Act of 1948, Pub. L. No. 80-471, § 302(a), 62 Stat. 110, 114.
 Inflation calculations performed herein using the Bureau of Labor Statistics Inflation Calculator. Bureau of Labor Statistics, CPI Inflation Calculator, http://www.bls.gov/data/inflation_calculator.htm (last visited Apr. 16, 2011).
 For simplicity, this article uses the term “percentage standard deduction” to refer to a standard deduction calculated as a percentage of AGI (as opposed to a flat minimum or maximum standard deduction).
 S. Rep. No. 78-885, at 4 (1944). In 1944, 82% of returns claimed the standard deduction. See Samansky, supra note 12, at 532-33 (citing IRS data). By 1950, the percentage still remained 80%. Statistics of Income Historical Table 7: Standard, Itemized, and Total Deductions Reported on Individual Income Tax Returns, Tax Years 1950-2009, I.R.S., http://www.irs.gov/pub/irs-soi/histab7.xls (last visited Apr. 16, 2011).
 Individual Income Tax Act of 1944 § 8(a), 58 Stat. at 235.
 See H.R. Rep. No. 78-1365, at 3 (1944) (noting that AGI is “in general . . . gross income less business deductions”).
 Employee travel expenses were moved below the line by the Tax Reform Act of 1986, Pub. L. No. 99-514, § 132(b)(1), 100 Stat. 2085, 2115.
 Individual Income Tax Act of 1944 § 8(a), 58 Stat. at 235.
 S. Rep. No. 78-885, at 24-25 (1944).
 The original standard deduction was also a substitute for certain tax credits: foreign tax credits, credits with respect to taxes withheld at the source under then-section 143(a) (relating to interest on tax-free covenant bounds), and all credits against net income with respect to interest on certain government obligations. See Individual Income Tax Act of 1944 § 9(a), 58 Stat. at 236.
 Internal Revenue Code of 1954, Pub. L. No. 83-591, 68A Stat. 3.
 The stated goals of the legislation were “1. To relieve the great majority of taxpayers from the necessity of computing their income tax. 2. To reduce the number of tax computations. 3. To simplify the return form. 4. To decrease the number of persons required to file declarations of estimated tax. 5. To eliminate some of the difficulties and uncertainties in the making of estimates required for declarations.” H.R. Rep. No. 78-1365, at 1 (1944); see also S. Rep. No. 78-885, at 1 (1944) (same).
 Non-itemizers with AGI over $5000 faced a fixed standard deduction of $500. This could be seen as small element of progressivity, since these taxpayers would be treated to a somewhat steeper effective rate schedule than those with AGI under $5000. However, these taxpayers were earning more than twice the median income and thus in a relatively high income cohort. See Bureau of the Census, Current Population Reports: Consumer Income: Family and Individual Model Income in the United States: 1945, at 1 (1948), available at http://www2.census.gov/prod2/popscan/p60-002.pdf (reporting 1945 median family and individual income as $2379). Furthermore, many, if not most, of these taxpayers could be expected to be itemizers. Thus, the percentage of taxpayers subject to a fixed standard deduction was likely small enough not to affect overall progressivity more than marginally.
 H.R. Rep. No. 88-749, at 1333 (1964).
 Revenue Act of 1964, Pub. L. No. 88-272, § 112(a), 78 Stat. 19, 23. The $1000 cap remained in place. The $100 per dependent was in addition to the existing deductions for personal exemptions under I.R.C. § 151 (1964).
 The initial low income allowance was actually quite a bit more complicated. For the 1970 tax year, part of the low income allowance phased out as incomes increased above the poverty level. Tax Reform Act of 1969, Pub. L. No. 91-172, § 802(a), 83 Stat. 487, 676-77. The phase-out could potentially have allowed for an additional degree of progressivity by shrinking the effective zero bracket amount as income rose. However, the phase-out was to be in effect only for the 1970 and 1971 tax years. § 802(e), 83 Stat. at 678. For 1971 onward, the bill provided that the low income allowance would simply be a flat $1050. Furthermore, the phase-out provision was eliminated for the 1971 tax year by the Revenue Act of 1971. Pub. L. No. 92-178, § 203(a), 85 Stat. 497, 511. This was done in order to accelerate tax relief because of the “depressed economic conditions.” H.R. Rep. No. 92-533, at 7 (1971). The end result is that for all but one year the low income allowance was a fixed amount of tax-free income given to all taxpayers, regardless of income.
 A single taxpayer received his own $600 dependency deduction plus the low income allowance of $1100.
 H.R. Rep. No. 91-413, at 205-06 (1969), reprinted in 1969-3 C.B. 200. Prior to 1969, the minimum amount of tax-free income (the standard deduction plus dependency deductions) was $900 for a single taxpayer, an amount which jumped to $1700 in 1969. There were thus a large number of taxable returns below the poverty level before the 1969 change. See Staff of the Joint Comm. on Tax’n, 91st Cong., General Explanation of the Tax Reform Act of 1969, at 218 (Joint Comm. Print 1970) [hereinafter 1969 Bluebook] (5.2 million taxable returns below poverty level in 1969).
 See, e.g., H.R. Rep. No. 94-658, at 3-6 (1976); text at infra notes 51-55.
 1969 Bluebook, supra note 49, at 216.
 Id. The introduction of the minimum standard deduction in 1964 accounted for a slight increase in use of the standard deduction, from 56% in 1963 to 58% in 1965.
 See supra note 31.
 Tax Reform Act of 1969, Pub. L. No. 91-172, § 802(a), 83 Stat. 487, 676-77.
 1969 Bluebook, supra note 49, at 218.
 In 1971, the increases for 1973 to 15% of AGI and $2000 were accelerated to 1972, along with other tax-relief provisions, in order to provide tax relief during an economic recession. See H.R. Rep. No. 92-533, at 7-8 (1971). The 1971 Act also increased the low income allowance to $1300 (~$7100 in 2011), though the Committee Reports attributed this increase to inflation, not further tax relief; the allowance was still pegged at the poverty level of income. Id. at 8. In 1975, the percentage standard deduction was nudged up again to 16% of AGI, up to $2400 ($2800 for joint filers) (~$9800/$11,500 in 2011), while the low income allowance was increased to $1700 ($2100 for joint filers) (~$7000/$8600 in 2011). Revenue Adjustment Act of 1975, Pub. L. No. 94-164, § 2(a)(1), (b), 89 Stat. 970, 970-71; Tax Reform Act of 1976, Pub. L. No. 94-455, § 401(b), 90 Stat. 1520, 1556 (making temporary changes permanent).
 Revenue Act of 1964, Pub. L. No. 88-272, § 213(a) (creating new § 217), (b) (adding § 217 deduction to AGI calculation in § 62), 78 Stat. 19, 50-52.
 H.R. Rep. No. 88-749, at 48 (1963).
 In particular, it confuses the simplification and progressivity effects described infra in Part III.B. By stating that a taxpayer must “choose between” the two, Congress is implicitly rejecting the idea that making the moving expenses deduction subject to the standard deduction would be de facto making it subject to a floor. Were the expenses below the line, there would be no “choice” because the only expenses properly deductible would be those that exceeded the standard deduction floor. The statement is thus not consistent with a zero-bracket view of the standard deduction. On the other hand, the simplification view ought to cut off any fairness arguments of this type, since a non-itemizer is, by definition, always getting a better deal than he would if he were instead to itemize.
Furthermore, adding below-the-line deductions would narrow the gap between the standard deduction and actual below-the-line expenses, which, as discussed infra in Part IV.D.1, has positive equity effects. These effects may be offset in part by the increase in complexity as more taxpayers are pushed to being itemizers, so the overall equity effects are unclear. Perhaps these issues are why the Committee chose to describe making the moving expenses deduction subject to the standard deduction “undesirable” rather than “unfair.”
 See infra notes 197-198 and accompanying text. The casualty loss deduction at the time did not have an overall floor (only the $100 per-loss floor), I.R.C. § 165(c)(3) (1964), and thus likely was used more often than today’s deduction subject to a floor of 10% of AGI, I.R.C. § 165(h) (2010). However, the medical expense deduction in effect in 1944 was subject to a 5% floor. Individual Income Tax Act of 1944, Pub. L. No. 78-315, § 8(c), 58 Stat. 231, 235.
 See supra text accompanying notes 35-40.
 Suppose a taxpayer in 1965 had $500 in moving expenses. This would lower his AGI by $500, which would in turn lower his percentage standard deduction by $50. Thus the net deduction is actually only $450. If this taxpayer were a non-itemizer with total itemizable expenses close to the itemization threshold, he would have been better off had moving expenses been placed below the line, since he may have been able to deduct more than $450 of his expenses.
 The statutory predecessor to § 215 was enacted in 1942. Revenue Act of 1942, Pub. L. No. 77-753, § 120(b), 56 Stat. 798, 816-17.
 S. Rep. No. 94-938(I), at 124-25 (1976). The Finance Committee Report seems to be implying that it is important for the § 215 deductions to match the § 71 inclusions, but that same argument would apply, e.g., to mortgage interest, medical expenses, or other personal expenditures that can be matched with income of other taxpayers.
 Indeed, rather than decrease the standard deduction, the 1976 Act made permanent the increases in 1971 and 1972. See Tax Reform Act of 1976, Pub. L. No. 94-455, § 401(b)(1), 90 Stat. 1520, 1556.
 An alternative way to view this benefit is as an increase in the net amount by which the standard deduction exceeds deductible expenditures. See infra Part III.B.1.
 See supra note 56 and accompanying text (noting a low income allowance of $1700 and a percentage standard deduction capped at $2400 in 1975).
 Tax Reduction and Simplification Act of 1977, Pub. L. No. 95-30, § 102(a), 91 Stat. 126, 135-36.
 While a decrease from the 1976 maximum standard deduction of $2400, this was also an increase in the amount of tax-free income for low-income taxpayers, since the prior low income allowance was $1700. See Tax Reform Act of 1976 § 401(b)(1), 90 Stat. at 1556. Congress estimated that 7.3 million more taxpayers would become non-itemizers as a result of the change, 3.7 million more returns would become entirely nontaxable, and that the total number of itemizers would be reduced from 31% to 23%. S. Rep. No. 95-66, at 50 (1977).
 See Tax Reduction and Simplification Act of 1977 § 102(a), 91 Stat. at 135; S. Rep. No. 95-66, at 51 (“Conversion of standard deduction into zero bracket amount and floor under itemized deductions”).
 The 1977 Act redefined taxable income for individuals as “adjusted gross income (1) reduced by the sum of (a) the excess itemized deductions, and (b) the deductions for personal exemptions provided by § 151, and (2) increased (in the case of an individual for whom an unused zero bracket amount computation is provided by subsection (e)) by the unused zero bracket amount (if any).” Tax Reduction and Simplification Act of 1977 § 102(a), 91 Stat. at 135.
 See id.
 See supra note 21.
 “The House bill eliminates the present minimum, percentage and maximum standard deductions and replaces them with what is, in effect, a flat standard deduction. . . . By incorporating the flat standard deduction in a zero rate bracket in the tax tables and rate schedules, the House bill eliminates the need for the separate concept of the standard deduction in the Code and the subtraction of the standard deduction in computing tax liability.” S. Rep. No 95-66, at 50-51 (1977).
 Id. at 51.
 See id. at 4-5, 15; H.R. Rep. No. 95-27, at 38-39 (1977).
 Revenue Act of 1978, Pub. L. No. 95-600, § 101(b), 92 Stat. 2763, 2769-70 (1978).
 See H.R. Rep. No. 95-1445, at 30 (1978).
 Economic Recovery Tax Act of 1981, Pub. L. No. 97-34, § 104(a), 95 Stat. 172, 188-89.
 In 1982 and 1983 the non-itemizer deduction in calculating AGI was limited to 25% of the amount donated, and the deduction was capped at $100. In 1984, it was 25% and $300. There was no cap in 1985 and 1986, though only 50% of donations were allowed in 1985. The full amount would be deductible in 1986, but there would be no deduction in years following. Economic Recovery Tax Act of 1981 § 121(a), 95 Stat. at 196.
 This is not to argue that they are not properly deductible in calculating personal income, however. That discussion is beyond the scope of this article, but has been discussed in great detail elsewhere. See supra note 99.
 See text at supra notes 63-65.
 See, e.g., Gene O. Sjostrand, Why Not Go Back to the Standard Deduction, 56 Taxes 265, 265 (1978); Donald C. Marshall & James E. Parker, Zero Bracket Approach Adversely Affects Income Averaging, 56 Taxes 179, 188 (1978).
 See S. Rep. No. 95-66, at 48 (1977) (“To prevent itemizers whose tax table income is above the table ceiling levels from being required to subtract the floor on itemized deductions from their income, the bill builds the floor into the rate schedules, as well as the tax tables, as a zero rate bracket. Taxpayers not using the tax tables will subtract their personal exemptions from tax table income in order to determine taxable income, against which the rate schedules are to be applied. As a result of this change, the present law concept of taxable income is redefined to reflect the floor under itemized deductions.”); Sjostrand, supra note 83, at 265; Marshall & Parker, supra note 83, at 188.
 S. Rep. No. 95-66, at 55-59; see Sjostrand, supra note 84, at 265-66.
 See, e.g., ABA Tax Section Recommendation No. 1979-9, 32 Tax Law. 1482, 1482-83 (1979); see also Bittker & Lokken, supra note 21, ¶ 30.5.1.
 Tax Reform Act of 1986, Pub. L. No. 99-514, § 102(a), 100 Stat. 2085, 2099.
 Note that an exemption and a zero bracket amount are equivalent. Whether you subtract an amount from income or multiply that amount by 0%, the result is the same—that amount goes untaxed. I use an exemption here to simplify the calculations.
 To see this more generally suppose taxes are determined by the equation
T = t(I – D)
where t is the tax rate, I is adjusted gross income and D is the greater of itemized deductions ID or the standard deduction SD. Suppose that the calculation of tax is changed to introduce an exemption amount E, and a floor F under itemized deductions, such that deductions D’ are equal to ID – F if ID > F, and 0 otherwise. Taxes would then be determined by the equation
T’ = t(I – E – D’)
If we set T = T’ then
D = E + D’ = E + ID – F
Thus, when ID ≥ SD
ID = E + ID – F
E = F (1)
Furthermore, when ID < SD, D’ must equal 0, because otherwise we would have the equation
SD = E + ID – F
which cannot be true if E = F but ID < SD. Thus
SD = E + D’ = E + 0 = E (2)
Therefore, combining results (1) and (2), the two tax systems are equivalent if and only if SD = E = F, or where the standard deduction amount in the first scenario is equal to the floor and to the exemption (or ZBA) in the second scenario. For a similar derivation see Kaplow, Standard Deduction, supra note 12, at 5-6.
 Note that the contradiction is not because calculating the tax base involves subtraction while applying a tax rate involves multiplication; a ZBA can be implemented either through the rate structure or as a larger personal exemption. Rather, the contradiction is in having the ZBA also substitute for items that are independently deducted from the tax base.
 It should be noted that the standard deduction is a compromise between simplification and progressivity, and thus both sets of criticisms can, in a sense, be partially true at the same time. But this only underscores the conceptual complexity of the standard deduction and its effect on debates regarding the personal deductions.
 While it is the case that a non-itemizer’s charitable giving is not subsidized at the margin, it is not the case that the encouragement to give is “wholly absent” for a non-itemizer if we consider the standard deduction to be largely or entirely a substitute for the itemized deductions, and not a ZBA. See text at infra note 100.
 Bittker & Lokken, supra note 21, ¶ 30.5.1. The treatise notes that this fix would “not have the progressivity effects of the present deduction,” id., which is partly an acknowledgement of the conflict. But the treatise identifies the progressivity effects as the phase-out of a variable ZBA as one’s itemizable expenses increase. Id. Under this view of the standard deduction only the net amount of the standard deduction above itemizable expenses is the ZBA. Because itemizable expenses tend to track income, the treatise argues, the ZBA tends to be largest for those with the least income, and disappears for those with higher income. Id. Thus, the loss of progressivity under the treatise’s proposal comes from losing a variable ZBA that phases out with a proxy for income. Id. However, this interpretation of the current standard deduction implies that the standard deduction does act, in part, as an approximation of the itemized deductions (since the ZBA portion of the standard deduction is only that amount in excess of the itemized deduction), and this is in conflict with the treatise’s primary criticism of the standard deduction. In other words, seeing the standard deduction as a variable ZBA that phases out with the level of itemizable expenses implies that a non-itemizer is getting the benefit of a deduction for her itemizable expenses (though at the expense of a larger ZBA). See also Samansky, supra note 12, at 555 (making a similar error in saying that “[t]he most troubling consequence of the standard deduction is that it systematically deprives most moderate income persons of any tax benefit from charitable contributions”).
 Although the legislative history does not show that Congress engaged with these arguments when they were expanding the standard deduction, one person did raise many of the same arguments discussed herein in Congressional testimony during consideration of the Tax Reform Act of 1969. A summary of testimony by Norman Ture (an “uncompromising advocate of supply-side economics and an architect of the 1981 tax cut,” Irving Molotsky, Norman Ture, Architect of the 1981 Tax Cut, Dies at 74, N.Y. Times, Aug. 13, 1997, at A21) stated that increasing the standard deduction would “move directly contrary to improving horizontal equity, so long as provisions for itemizing deductions remain in the law”; that “[i]n effect, the standard deduction permits two taxpayers in the same family circumstances and with the same [AGI] to pay the same tax even though incurring substantially different amounts of itemizable expenses”; that “the argument regarding simplification is substantially overstated from the taxpayers’ point of view”; that the argument that the standard deduction “would reduce tax liabilities . . . really is an argument for reducing effective tax rates for [low-income] taxpayers. . . [which] would be better accomplished by reducing the statutory tax rates in the relevant income range”; and that “the presumption that itemized deductions contribute to fairness . . . is open to challenge,” and if they do, then the deductions should “not . . . be constrained by such devices as [variable floors] or the  proposal to raise the price for itemizing.” Staff of the Joint Comm. on Tax’n, 91st Cong., Summary of Testimony on the Standard Deduction 4-5 (Joint Comm. Print 1969) (testimony of Norman B. Ture), reprinted in 21 Tax Reform – 1969: A Legislative History of the Tax Reform Act of 1969, Doc. No. 67 (Bernard D. Reams, Jr., ed., 1991).
 It is possible that the standard deduction actually increases complexity for taxpayers on the margin of being itemizers, since they may bear the costs of record-keeping and computation but still fail to exceed the threshold (or, alternatively, fail to itemize when they should have). In practice, it is likely that many taxpayers do not itemize unless they have expenses exceeding the standard deduction by an amount approximating the costs associated with itemizing. See Mark M. Pitt & Joel Slemrod, The Compliance Cost of Itemizing Deductions: Evidence from Individual Tax Returns, 79 Amer. Econ. Rev. 1224, 1224 (1989) (estimating the cost of itemizing via forgone reductions in tax liability).
 Technically, a taxpayer must elect to itemize deductions, though that election is not limited to those with expenses greater than the standard deduction. See I.R.C. § 63(e) (2010).
 Bittker & Lokken characterize this net payment differently. Instead of seeing it as a windfall payment for waiving the right to itemize, they describe it as a variable ZBA that shrinks as income goes up. Bittker & Lokken, supra note 21, ¶ 30.5.1; see infra Part III.B.4. While there may be some vertical equity benefits of having such a variable ZBA, it still would violate horizontal equity in the way that I discuss.
 “Personal income may be defined as the algebraic sum of (1) the market value of rights exercised in consumption and (2) the change in value of the store of property rights between the beginning and end of the period in question.” Henry C. Simons, Personal Income Taxation: The Definition of Income As a Problem of Fiscal Policy 50 (1938). Under this definition, a deduction is proper if the expense should not be considered consumption. See William D. Andrews, Personal Deductions in an Ideal Income Tax, 86 Harv. L. Rev. 309, 313-14 (1972). Thus, a taxpayer with more deductions has a lower “true” income than a taxpayer with the same gross income but fewer deductions.
 The literature over whether a particular deduction is proper in measuring Haig-Simons income or otherwise measuring ability to pay, or is instead a subsidy—or “tax expenditure”—is rich, though ultimately unresolved for many of the deductions. See, e.g., Stanley S. Surrey, Pathways to Tax Reform: The Concept of Tax Expenditures (1973); Andrews, supra note 98; Boris I. Bittker, Income Tax Deductions, Credits, and Subsidies for Personal Expenditures, 16 J.L. & Econ. 193 (1973); Thomas D. Griffith, Theories of Personal Deductions in the Income Tax, 40 Hastings L.J. 343 (1989); Jeffrey H. Kahn, Personal Deductions—A Tax “Ideal” or Just Another “Deal”?, 2002 L. Rev. M.S.U.-D.C.L. 1; Daniel N. Shaviro, Rethinking Tax Expenditures and Fiscal Language, 57 Tax L. Rev. 187 (2004); Victor Thuronyi, Tax Expenditures: A Reassessment, 1988 Duke L.J. 1155 (1998); David A. Weisbach & Jacob Nussim, The Integration of Tax and Spending Programs, 113 Yale L.J. 955 (2004). This article does not intend to wade into this debate. As noted in the accompanying text, the arguments herein are valid regardless of which view one takes. Even if all the personal deductions are tax expenditures, it remains the case that those deductions are used to define the tax base, however that concept might vary from ideal Haig-Simons income, and variations from that base should not be as arbitrary as the standard deduction makes them. That being said, some discussions in this article assume that certain deductions, such as those for extraordinary medical expenses and casualty and theft losses, are more likely to be proper under a Haig-Simons definition, while others, such as that for mortgage interest, are more likely to be tax expenditure subsidies.
 See Coven, supra note 12, at 1562; Samansky, supra note 12, at 544; see also Bittker & Lokken, supra note 21, ¶ 30.5.1.
 Although Samansky is critical of this inequity, supra note 12 at 544, he suggests that if all itemized deductions were proper in measuring income—if all affected ability to pay and none were “tax expenditures”—then this system might in fact be justified for simplification reasons, since the sum of the itemizable expenses is a coherent measure of a taxpayer’s ability to pay. Id. at n.85. He therefore ties at least part of his criticism of the standard deduction to the fact that many itemized deductions do not actually address ability to pay. This is an example of why it is necessary to bifurcate the effects of the standard deduction, since this conclusion does not follow in the simplification-centric case. Indeed, if itemizable expenses did not alter a taxpayer’s ability to pay, then we ought to feel more comfortable that the net payments are distributed arbitrarily, since the difference in ability to pay between a non-itemizer with high itemizable expenses and one with low itemizable expenses is smaller than the difference in their actual itemizable expenses. The greater the difference in ability to pay between these two taxpayers, the less comfortable we should be with the inequitable distribution of the net standard deduction tax benefit. In short, it is preferable that the net payments be distributed arbitrarily rather than regressively. See also Kaplow, supra note 12, at 11 n.28. But see Jeffrey H. Kahn, Beyond the Little Dutch Boy: An Argument for Structural Change in Tax Deduction Classification, 80 Wash. L. Rev. 1, 36 (2005) (arguing that the simplification benefits exceed the cost to equity).
 See infra Part IV.A.
 The existence of the implied floor still adds some simplicity to this frame, of course—removing the floors entirely would generate substantially more complexity. However, there may be many other policy reasons for having a floor beyond just simplicity. See infra Part IV.E
 See, e.g., I.R.C. §§ 165(h)(2) (2010) (floor of 10% of AGI on casualty losses), 213(a) (2005) (floor of 7.5% of AGI on medical expenses, going to 10% in 2013).
 See supra Part II.C.
 See Kaplow, Standard Deduction, supra note 12, at 27-29; Samansky, supra note 12, at 548.
 See supra note 100 and accompanying text.
 Andrews, supra note 98, at 335. But see Louis Kaplow, The Income Tax as Insurance: The Casualty Loss and Medical Expenses Deductions and the Exclusions of Medical Insurance Premiums, 79 Cal. L. Rev. 1485, 1486, 1505-06 (criticizing analysis that focuses on definitions of what is “income” or “consumption”).
 Supra note 104.
 This is a gross simplification, of course, not least because having the floor be a percentage of AGI implies that rich taxpayers with, say, cancer are healthier than poor taxpayers with cancer. This issue is discussed more fully infra in Part IV.E
 2009 SOI Bulletin, supra note 6, at 7.
 See Bittker & Lokken, supra note 21, ¶ 30.5.1 (noting this problem).
 See supra Part II.C.
 For a full discussion of separating a standard deduction from its threshold, see Kaplow, Standard Deduction, supra note 12, at 12-19.
 See id.
 See id. at 14 & n.38.
 See id. at 14-15 & n.39.
 If non-itemizers have itemizable expenses evenly distributed between zero and some threshold T, then the mean amount of expenses would be T/2. However, if the expenses cluster toward zero, the mean would be somewhat less than T/2. See id. at 14-15.
Kaplow’s analysis becomes more complex when he introduces the likely taxpayer responses to having a deduction less than a threshold. First, because there is a greater advantage to exceeding the threshold than under the current system, more taxpayers will do the necessary calculations, thus decreasing the simplification benefit. Second, having a deduction less than a threshold may introduce excessive incentives for taxpayers to boost their itemizable expenses to get over the threshold or to cluster their expenses in alternate years. These costs would likely push the optimal deduction to be higher than the mean of non-itemizers’ itemizable expenses. See id. at 17-19.
 See Bittker & Lokken, supra note 21, ¶ 30.5.1.
 Note that even though this view appears to allow for both simplification and a ZBA, it still has the effect of denying the taxpayer an effective marginal incentive to increase spending, since instead of providing a subsidy by lowering taxable income, the increase in spending shrinks the effective ZBA. Indeed, that would result in a marginal penalty, rather than a marginal subsidy.
 See Coven, supra note 12, at 1558-59. At the time Coven was writing, the percentage standard deduction was still in place, so he characterized the trade-off as itemizers being taxed on taxable income while non-itemizers were taxed on 84% of AGI, but with the same rate structure.
 Ignoring personal exemptions and similar provisions.
 Note that there are still more ways of framing an analysis of the standard deduction, for example as a presumptive tax. See, e.g., Joel Slemrod & Shlomo Yitzhaki, Analyzing the Standard Deduction as a Presumptive Tax, 1 Int’l Tax & Pub. Fin. 25 (1994); see also supra note 91.
 Feld, supra note 12, at 441.
 Tax Relief Act of 2005, S. 2020, 109th Cong. (2005).
 See, e.g., Independent Sector, Fact Sheet on the Charitable Deduction Changes for Itemizers and Non-Itemizers in S.2020, the Tax Relief Act of 2005, at 2, 4 (on file with author). Some have argued that this would motivate increased overall charitable giving, but the evidence is mixed. See, e.g., Jane G. Gravelle, Cong. Research Serv., RL 31108, Economic Analysis of the Charitable Contribution Deduction for Non-itemizers 5-6 (2005).
 See supra text accompanying note 100.
 Moreover, they may in fact feel a tax penalty rather than a tax subsidy. See supra note 120.
 See supra note 120.
 See supra note 80 and accompanying text.
 The “fairness” argument is somewhat different in the case of state taxes, since the argument would not be focused on the “fairness” of whether the government provided an incentive to the taxpayer’s marginal state-tax dollar, since taxes arguably are not voluntary expenditures in the way that charitable deductions are. But see Brian Galle, Federal Fairness to State Taxpayers: Irrationality, Unfunded Mandates, and the “SALT” Deduction, 106 Mich. L. Rev. 805, 812 n.29 (2008); Louis Kaplow, Fiscal Federalism and the Deductibility of State and Local Taxes Under the Federal Income Tax, 82 Va. L. Rev. 413, 417 (1996). A non-itemizer is unlikely to argue that it is not fair for an itemizer to be given an incentive to pay more state taxes than she is. Rather, the more politically salient argument is simply the disparate treatment of different taxpayers.
 The fewer below-the-line deductions there are, the more the standard deduction becomes a de facto ZBA.
 See supra Part II.C.
 See Samansky, supra note 12, at 547-48.
 See, e.g., Office of Mgmt & Budget, Analytical Perspectives, Budget of the United States, Fiscal Year 2007, at 288, available at http://www.gpoaccess.gov/usbudget/fy07/browse.html (including the mortgage deduction on official list of tax expenditures).
 See Surrey, supra note 99, at 37. But see Simons, supra note 98, at 112 (criticizing the failure to tax imputed rent); Kahn, supra note 99, at 48-49 (noting that the mortgage interest deduction can be justified as neutralizing the bias between purchasing a home with one’s own funds versus borrowed money, as a result of the failure to include imputed rent in income).
 See, e.g., Mortgage Tax Deduction is Safe, Bush Says, L.A. Times, Feb. 18, 2006, at C2; see also Jeffrey H. Birnbaum & Alan S. Murray, Showdown at Gucci Gulch: Lawmakers, Lobbyists, and the Unlikely Triumph of Tax Reform 57 (1987) (discussing political pressures that resulted in the mortgage interest deduction remaining in the Tax Code after the Tax Reform Act of 1986, despite the repeal of the deduction for other personal interest).
 Since rich renters would share the benefit of a rate reduction in order to make up for the loss of the mortgage interest deduction by taxpayers in the same income cohort, which could be quite large because of the upside-down nature of the subsidy.
 Samansky, supra note 12, at 547-48.
 Tax Reform Act of 1986, Pub. L. No. 99-514, 100 Stat. 2085.
 See Samansky, supra note 12, at 548; see also Birnbaum & Murray, supra note 137, at 128-29. This is not to argue that the deduction for state and local taxes is or is not a “tax expenditure,” only that, if it were proper in measuring income, that argument would be difficult to put forth when only a minority of taxpayers receive the deduction. See generally Galle, supra note 131; Kaplow, supra note 131.
 Samansky, supra note 12, at 548.
 See supra note 99.
 See supra note 6.
 Whereas the standard deduction alone forces different, unrelated deductions to be considered together, adding a floor calculated as a percentage of AGI adds a third dimension—income—to the mix. It is difficult to predict the effects of this mish-mash given the diversity of taxpayers.
 E.g., I.R.C. §§ 21 (expenses for dependent care necessary for employment), 25A (Hope and Lifetime Learning credits) (2010).
 E.g., I.R.C. §§ 62(a)(17), 221 (interest on education loans) (2010).
 I.R.C. §§ 62(a)(19), 223 (2010).
 I.R.C. § 213 (2010).
 I.R.C. § 62(a)(2)(D) (2010).
 I.R.C. § 162 (2010).
 See supra Part III.B.4.
 That is, a simplification-centric standard deduction ought to be designed as a percentage of AGI. See supra Parts II.B, III.B.1. A variable ZBA ought to be designed to phase out with income, not itemized deductions. See supra Part III.B.4. The proposal that follows depends on adopting the progressivity-centric view explicitly, but that does not mean that it is the only solution. If, e.g., one preferred a simplification-centric view, then one could propose a percentage standard deduction, as Bittker and Lokken do. See supra note 21, ¶ 30.5.1.
 See supra Part II.C.
 See supra note 8.
 See supra Part II.C.
 Note that the same effect could be achieved simply by creating multiple standard deductions and leaving the current rate structure untouched. See supra Part III.A. The problems associated with combining income-measuring and progressivity functions would remain, however.
 Note that this does not necessarily mean removing the concept of AGI from the Tax Code entirely. The Tax Code currently uses AGI for a number of purposes, such as limitations, floors, and phase-outs on existing deductions, credits, and other benefits. See, e.g., I.R.C. §§ 21(a)(2) (phase out of dependent care credit), 24(b) (phase-out of child tax credit), 36C(b)(2) (phase out of adoption expenses credit), 67(a) (2% of AGI floor on miscellaneous itemized deductions), 68(a) (overall limitation on itemized deductions), 213(a) (7.5% of AGI floor on medical expense deduction), 408A(c)(3) (limitation on contributions to Roth IRAs) (2010). For these provisions AGI plays a role similar to its role in the early years of the standard deduction, namely as a relatively uniform base from which to calculate a rule of general applicability (here, the relevant phase-outs; in 1944, the percentage standard deduction). Neither gross income nor taxable income provides a satisfying replacement in the case of deduction phase-outs—gross income, because its meaning varies so much across different types of income, as Congress noted in 1944, see supra text accompanying notes 35-40; taxable income because the calculation becomes circular when a deduction used in calculating taxable income itself phases out with taxable income. It would add substantial complexity to have to calculate taxable income without regard for a certain deduction in order to determine the applicable deduction amount, particularly across many different deductions. Thus some sort of AGI concept could still play a role. However, its importance and salience to the average taxpayer would be substantially reduced, and policymakers would no longer feel such intense pressure to move deductions above the line. Thus under this proposal AGI could return to being a more coherent and uniform measure across all taxpayers.
On the other hand, with the important exception of § 68, the AGI-based phase-outs apply largely to credits, rather than deductions, and thus we might ask whether the deductions that we limit or phase out for progressivity reasons might not work better as tax credits, which would avoid the circularity problem.
 See supra Part III.A; see also Kaplow, Standard Deduction, supra note 12, at 6.
 See Kaplow, Standard Deduction, supra note 12, at 6-12 (“These effects [a change in tax burdens and simplification] should dictate policy on thresholds [or floors] because such effects are inherent: They arise regardless of whether or how one adjusts [the ZBA]. By contrast, effects on revenue and the overall income distribution are, in principle, wholly independent of how thresholds [or floors] are set: Any such effects of changing a threshold can be nullified if that is desired or can be achieved by adjusting [the ZBA] without changing the threshold [or floor].”).
 See Kaplow, Standard Deduction, supra note 12, at 9-10.
 Currently, the standard deduction could be viewed as a zero bracket that phases out with itemized deductions, which in turn are partly correlated with AGI. However, as discussed throughout, that is an imperfect proxy and has negative horizontal equity effects. If the intent is to have the ZBA phase out with income, then we should just do that.
 I.R.C. § 68 (2010).
 A phase-out based on AGI is essentially an increase in marginal tax rates on AGI. If an additional dollar of income results in a loss of $x in tax benefits (in the form, e.g., of an increase in the floor on a deduction or a decrease in the ZBA), that is equivalent to a tax of $x on that marginal dollar.
 See supra Part II.C.
 U.S. Treasury Department, 91st Cong., Tax Reform Studies and Proposals, pt. 2, at 127 (Joint Comm. Print 1969), reprinted in 22 Tax Reform – 1969: A legislative History of the Tax Reform Act of 1969, supra note 94, Doc. No. 102; see also President’s 1963 Tax Message: Hearings Before the H. Comm. on Ways and Means, 88th Cong. 15 (1963) (“One way to provide relief to low-income taxpayers . . . would be to raise the personal exemption above its present level of $600. This is an extremely costly approach, however, and one which would not fulfill our objective of giving relief where it is needed most.”), 42-43 (Treasury Secretary C. Douglas Dillon noting that merely increasing exemptions would “offer greater tax savings the higher the income of the taxpayer, thus wasting much of the revenue that it would cost if [low-income tax relief] were to be achieved through this route”), 195 (Treasury staff noting that increasing the standard deduction would allow “[t]he benefits of the proposal [to] be distributed in favor of persons with low incomes. Persons with high and middle incomes will find the provision of no benefit. Equivalent relief to low-income taxpayers granted in the form of an increase in the per capita exemption would involve a much greater loss of revenue and would provide the greatest tax savings where they are least needed.”), reprinted in 26 Internal Revenue Acts of the United States: Revenue Acts of 1953-1972 With Legislative Histories, Laws, and Congressional Documents, pt. I (Bernard D. Reams, Jr., ed., 1985).
 It is odd that Treasury in 1963 did not offer this as an alternative, since the 1963 proposals contain, in addition to the increase in the standard deduction, a separate floor on the itemized deductions, for many of the same reasons discussed in this article. See President’s 1963 Tax Message, supra note 166, at 19 (“I, therefore, recommend that itemized deductions, which now average about 20 percent of adjusted gross incomes, be limited to those in excess of 5 percent of the taxpayer’s adjusted gross income. This 5-percent floor will make $2.3 billion of revenue available for reduction in individual tax rates. At the same time incentives to homeownership or charitable contributions will remain. In fact, this tax program as a whole, providing as it does substantial reductions in Federal tax liabilities for virtually all families and individuals, will make it easier for people to meet their personal and civic obligations. This broadening of the tax base which permits a greater reduction in individual income tax rates has an accompanying advantage of real simplification.”).
 Furthermore, while having a zero bracket that phases out with itemized deductions may have some progressivity from a vertical equity standpoint, to the degree that incomes correlate with itemizable expenses, it has negative horizontal equity effects for taxpayers within the same income class. See supra Part III.B.1.
 2009 SOI Bulletin, supra note 6, at 7.
 Id. at 22, Table 1.
 Michael Parisi, Individual Income Tax Returns, Preliminary Data, 2008, Statistics of Income Bulletin, Winter 2010, at 5, 12, available at http://www.irs.gov/pub/irs-soi/10winbulindincretpre.pdf.
 See, e.g., Edward J. McCaffery, The Holy Grail of Tax Simplification, 1990 Wis. L. Rev. 1267 (1990); Martin J. McMahon, Jr., Individual Tax Reform For Fairness and Simplicity: Let Economic Growth Fend For Itself, 50 Wash. & Lee L. Rev. 459 (1993); Deborah H. Schenk, Simplification for Individual Taxpayers: Problems and Proposals, 45 Tax L. Rev. 121 (1989); Stanley S. Surrey & Gerard M. Brannon, Simplification and Equity as Goals of Tax Policy, 9 Wm. & Mary L. Rev. 915 (1968).
 See, e.g., Steven A. Dean, Attractive Complexity: Tax Deregulation, the Check-the-Box Election, and the Future of Tax Simplification, 34 Hofstra L. Rev. 405 (2005); Samuel A. Donaldson, The Easy Case Against Tax Simplification, 22 Va. Tax Rev. 645 (2003); Louis Kaplow, Accuracy, Complexity, and the Income Tax, 14 J. L. Econ. & Org. 61 (1998) [hereinafter Kaplow, Accuracy].
 See, e.g., Dean, supra note 173, at 412; Schenk, supra note 172, at 127; David F. Bradford, Untangling the income tax 266-67 (1986).
 Transactional complexity generally describes the lengths that taxpayers sometimes go to in order to structure transactions in tax-efficient ways. Rule complexity refers to the difficulty in interpreting some aspects of the law. Neither is likely to affect low-income taxpayers most of the time, and hardly ever in relation to their use of the standard deduction or itemized deductions. An exception to this might be in trying to determine whether a particular expense qualifies as deductible.
 By contrast, recall that the first standard deduction approximated not only the itemized deductions but a number of complex tax credits, including the foreign tax credit. See supra note 40.
 See, e.g., Joseph DiSciullo, AICPA Tax Simplification Report Includes AMT, Estate Tax Reform, 125 Tax Notes 409 (2009); Sam Goldfarb, Complexity of AMT Leads to Administrative Errors, TIGTA Says, 120 Tax Notes 641 (2008).
 See, e.g., Taxpayer Advocate Service, National Taxpayer Advocate 2007 Annual Report to Congress, Vol. II, 94-118 (2007).
 See generally, e.g., Schenk, supra note 172.
 This may be in part a result of the forces discussed in Part III.C, supra. The itemized deductions have seen relatively little political meddling, in part due to the limited effect that changes to them would have on the majority of individual taxpayers. And where there has been substantial added complexity related to itemized deductions—such as in the rules for evaluating particular charitable gifts, see I.R.C. § 170 (2010); Treas. Regs. §§ 1.170A-1 et seq.—the rules are unlikely to create complexity for middle- and low-income taxpayers.
 Kaplow, Accuracy, supra note 173, at 62-63. Kaplow also discusses the efficiency benefits from accuracy, namely that a more accurate income tax keeps the labor-leisure distortion to a minimum. Id. at 63.
 Id. If risk aversion is highest at lower incomes—because of their relatively lower ability to afford over-taxation errors—then they would pay a greater premium to insure against that risk. Kaplow writes that for a standard logarithmic utility function, a given fixed error would have a ten times greater utility cost on a person with an income of $10,000 than on a person with an income of $100,000. This error cost could be even greater due to the sometimes large marginal tax rates on low-income individuals due to welfare program phase-outs. Id. at 66.
 Kaplow’s equation of the accuracy benefit and the risk premium quantifies the benefits of vertical equity—that is, the benefit from not being taxed as if one were richer. This formulation likely also encompasses the horizontal equity concern that is said to motivate at least some of the itemized deductions, and that this article argues are some of the major harms of the standard deduction. The horizontal equity concern is that the standard deduction causes taxpayers with the same “true” income to be taxed differently—which is another way of saying that a lower-income person has been taxed too much.
Kaplow is critical of horizontal equity as an independent norm, since the basic proposition of equal treatment of equals follows necessarily from compliance with vertical equity. Louis Kaplow, Horizontal Equity: Measures in Search of a Principle, 42 Nat’l Tax J. 139 (1989); Louis Kaplow, A Note on Horizontal Equity, 1 Fla. Tax Rev. 191 (1992). But see Richard A. Musgrave, Horizontal Equity, Once More, 43 Nat’l Tax J. 113 (1990) (arguing that horizontal equity has independent meaning in a second-best world). The more rigorous statements of horizontal equity as an independent norm tend to relate to the effects of tax changes on similarly situated individuals. That is, where two individuals have the same utility given a certain tax (or in the absence of a tax), horizontal equity requires that they be equally well off following a change in the tax. See, e.g., Martin Feldstein, On the Theory of Tax Reform, 6 J. Pub. Econ. 77, 83 (1976). Cast in these terms, the horizontal equity problem with the standard deduction is that taxpayers who might have had the same utility levels in the absence of the standard deduction, but different itemized deductions, no longer have the same utility level when the standard deduction is introduced. See supra Part III.B.1. But the difficulty with this conception is treating the existence of the itemized deductions as a given. Since some itemized deductions reflect income measurement or ability to pay, while others reflect pure subsidies (and yet others a mixture of the two), it cannot be said with any rigor what the baseline utility level ought to be.
 Taking the view that non-itemizers are under-taxed is implicitly adopting a no-ZBA, simplification-centric view of the standard deduction.
 I include in this the complexity effects of simply having the longer form necessary to calculate the various floors. However, a key benefit of computer-aided tax preparation is that form length alone is increasingly irrelevant.
 See generally Lawrence Zelenak, Complex Tax Legislation in the Turbotax Era, 1 Colum. J. Tax L. 91 (2010).
 Internal Revenue Service, Tax Year 2006 Taxpayer Usage Study Report 16 (2007), available at http://www.irs.gov/taxstats/article/0,,id=184856,00.html [hereinafter 2006 Taxpayer Usage Study Report]. See Zelenak, supra note 186, at 95. This percentage includes electronically filed returns, but even looking only at paper-filed returns, 83% of Forms 1040 were prepared by computer. Id.
 88% of tax year 2005 returns were prepared by computer. 2006 Taxpayer Usage Study Report, supra note 187. To compare, just over 67% of tax year 1997 returns were prepared by computer. Internal Revenue Service, Taxpayer Usage Study 1997 Data Release 1 (1998), available at http://www.irs.gov/pub/irs-soi/97tpus.pdf.
 TurboTax, http://turbotax.intuit.com (last visited Apr. 10, 2011). For 2010 the Turbo Tax free version applied only to 1040-EZ returns, with state returns costing extra. The more deduction-focused “Basic” and “Deluxe” versions cost $20 and $30, respectively (not including state returns).
 National Association of Tax Professionals, 2010 Tax Professional Study 6 (2010), available at http://www.natptax.com/Public Documents/Join NATP/2010 Fee Study Sample.pdf.
 See Free File: Do Your Federal Taxes for Free, I.R.S., http://www.irs.gov/efile/article/0,,id=118986,00.html (last visited Apr. 11, 2011).
 See Free Tax Return Preparation for You by Volunteers, I.R.S., http://www.irs.gov/individuals/article/0,,id=107626,00.html (last visited Apr. 16, 2011).
 Pitt & Slemrod, supra note 95, at 1229.
 See I.R.C. § 6050H(d) (2010) (requiring mortgagee to furnish mortgagor with statement showing interest paid); Treas. Reg. §§ 1.6050H-1, -2 (as amended in 2000).
 In fairness, the cost of itemizing mortgage interest is rarely raised as a concern because most people with mortgage interest itemize. See President’s Advisory Panel on Tax Reform, Simple, Fair, and Pro-Growth: Proposals to Fix America’s Tax System 74 (2005) [hereinafter Tax Reform Panel] (stating that 54% of homeowners who pay mortgage interest receive a tax benefit from the deduction). Fifty-four percent is only a narrow majority, however, so a large number of taxpayers could still face some incremental complexity.
 As an alternative, we could require states and municipalities to issue a statement showing the total of all taxes paid in a given year. This would impose some additional cost, but it likely would not be unreasonable given that the information ought to be easily compiled by the relevant tax authorities. States already provide taxpayers with Forms 1099-G to document any state tax refunds paid during the tax year.
 Someone with a chronic medical condition might have costs that would routinely break the 7.5% AGI floor—but such a person is also likely to be insured (realistically, an uninsured chronically ill person is not likely to be able to bear those costs ongoing) and thus not bear the brunt of the costs himself.
 Elderly insured taxpayers may be an exception to this. Whereas most insured taxpayers are unlikely to spend enough in prescription and doctor-visit copayments to exceed the 7.5% of AGI threshold, the combination of a high number of copayments plus relatively low income mean that elderly insured taxpayers may exceed it more often. (Though the elderly also have a higher standard deduction, and thus a higher floor on the total of their itemized deductions. See supra note 21.) The Staff of the Joint Committee on Taxation does not include copayments by the insured elderly in its discussion of common § 213 deductions. See Staff of the Joint Comm. on Tax’n, 110th, Tax Expenditures for Health Care 23 (Joint Comm. Print 2008) [hereinafter Tax Expenditures for Health Care].
 See supra notes 181-183 and accompanying text.
 See supra note 101.
 See Treas. Reg. § 1.170A-13(a)(1)(ii) (as amended in 1996) (naming tax receipt from donee charitable organization as one form of required proof of donation for purposes of claiming the § 170 deduction).
 See supra note 80 and accompanying text.
 See supra note 125.
 A current itemizer must keep records across all categories of itemizable expense in order to get the full benefit. For example, a taxpayer who buys a home must begin keeping records not only of mortgage interest, but also of charitable contributions and other itemizable expenses, even though those expenses may not have changed with the home purchase.
 See supra Part III.A.
 See generally Kaplow, Standard Deduction, supra note 12.
 Kaplow criticizes the idea that a floor can effectively subsidize only “extraordinary” expenses, since, as he shows, a floor is equivalent to a standard deduction, provided that there are offsetting adjustments to exemptions or the rate structure. Kaplow, Standard Deduction, supra note 12, at 3 n.12. Thus a floor can have the same effect as a standard deduction, namely to provide an additional deduction to those with low itemizable expenses, which is contrary to the purpose of deducting only extraordinary expenses. But the source of the equivalence is the offsetting adjustment to exemptions or the rate structure. Here, I am looking instead at what results from taking as fixed a high ZBA equal to the standard deduction. The particular level of any floor on deductions should be a separate determination based on Congress’s view about the role of the particular deduction. If the ZBA reflects a fixed amount of untaxed income, then adjustments to the deduction floors, given that fixed ZBA, would indeed reflect a policy to subsidize only extraordinary expenses.
 See Andrews, supra note 108 and accompanying text.
 See infra Part IV.D.1.
 Mismeasurement errors as a result of floors do tend to partially cancel each other out, allowing for some equity benefit to having a single floor (in terms of lowering the total degree of mismeasurement). See Kaplow, Standard Deduction, supra note 12, at 27-28 & n.80 (noting, however, that the “main virtue of grouping items” is simplification). Nonetheless, there are other policy considerations beyond accuracy in the choice of floor, as discussed herein.
 It is a simplification to say that extraordinary medical expenses do not correlate with income at all. It is very likely that a wealthy taxpayer with an illness consumes more medical services than a poorer taxpayer with the same illness (and thus may actually be healthier). But in most cases, the expenses of the wealthy taxpayer will be borne by insurance. On the margins, there may be some taxpayers who purchase their own insurance and choose to under-insure so as to get the benefit of partial deducibility of medical expenses. See Kaplow, supra note 108, at 1487; Tax Expenditures for Health Care, supra note 198. But see Tax Expenditures for Health Care, supra note 198, at 23 (“In practice, however, it is unlikely that those individuals who do not purchase health insurance will have the liquidity to pay medical expenses that are a large portion of their annual income.”). According to the Staff of the Joint Committee on Taxation, the deduction is taken largely for un- or under-covered medical expenses, such as metal health care, dental care, and nursing home care. Id. In 2007, 91% of the returns claiming the deduction reported AGI less than $100,000 (accounting for 68% of the total tax expenditure). Id. at 22.
 The numbers in this example were obviously chosen to get an equivalent result. This is unlikely to be the result in most actual cases. The larger point is that having a variable, and increasing, floor offsets to some degree the “upside-down” subsidy inherent in a deduction, and thus has some credit-like features.
 See, e.g., Surrey, supra note 99, at 98-100.
 See, e.g., Bittker, supra note 99, at 197. But see Daniel I. Halperin, Valuing Personal Consumption: Cost Versus Value and the Impact of Insurance, 1 Fla. Tax Rev. 1 (1992) (challenging the deductibility of losses, particularly where there is asymmetric tax treatment of consumption gains and losses); Kaplow, Insurance, supra note 108, at 1500-04.
 A variable floor is somewhat more defensible with respect to these income-measuring deductions for simplification reasons, however. If we provide a floor purely to relieve some taxpayers and the government from the burden of itemization, then it is reasonable for the floor to increase with income. Higher income taxpayers are more likely to incur itemizable expenses, and likely place a higher cost on the time that would be spent to itemize.
 This is not to say, of course, that adding a variable floor to the mortgage interest deduction is preferable to replacing it with a tax credit or repealing it altogether; it is only to say that, were the deduction retained, a variable floor would alleviate some of its flaws. The President’s Advisory Panel on Tax Reform proposed replacing the mortgage interest deduction with a 15% tax credit. Supra note 195, at 70-75.
 Patient Protection and Affordable Care Act, Pub. L. No. 111-148, § 9013(a).
 See Martin Vaughan, Senate Looks to Trim Tax Break for Personal Medical Costs, Wall St. J., June 22, 2009 (discussing proposal as a revenue-raiser).
 Staff of the Joint Comm. on Tax’n, 111th Cong., Estimated Effects of the Revenue Provisions Contained in the “Patient Protection and Affordable Care Act” 1 (Joint Comm. Print 2009).
 For example, the fact that health costs as a percentage of income are rising implies that any floor should rise as well in order to continue to target the same group of taxpayers. See Vaughan, supra note 218 (quoting Henry Aaron on this rationale).
 Tax Reform Act of 1986, Pub. L. No. 99-514, § 133(a).
 But see supra note 158.
 See supra Part III.C.
 See A Reconsideration of Tax Expenditure Analysis, supra note 8.