Ironing out the flat tax.
The business tax is computed much like a VAT.(2) Businesses are taxed on the difference between gross receipts from sales of property or services and the cost of business inputs, wages, and retirement contributions. The tax provides current expensing of all business purchases, businesses may not deduct interest or dividends or other financial payments, and financial income is not included when received.
Flat Tax proposals offer few additional details. We do not know, for example, the rules for such everyday transactions as the formation or liquidation of a business or the sale of property on credit. Given the immense size and complexity of our economy, the simple outline of the Flat Tax offered by its proponents does not come close to legislation that could actually be enacted.
Without additional details, evaluation of the Flat Tax is difficult. Much of the appeal of the Flat Tax lies in its alleged simplicity--the vaunted postcard returns resonate with many. If the claims of simplicity are not correct, the Flat Tax may be significantly less attractive. Moreover, the efficiency claims for the Flat Tax rely in part on simple, clearly stated economic incentives and relative unavoidability. If, once implemented, the Flat Tax is complex and avoidable, the efficiency claims may falter.(3) And the fairness claims for the Flat Tax, regardless of their merit, become weaker if the tax is avoidable. In particular, an avoidable tax will have different distributive consequences (most likely more regressive) than the simple, unavoidable tax championed by proponents of the Flat Tax. The devil for the Flat Tax might truly be in the details. To date, however, no details of the design of the Flat Tax have been offered.
This article attempts to fill this gap by beginning the study of the design of the Flat Tax.(4) The article is organized around a central feature of the Flat Tax that has not previously drawn attention--something I call "openness." By openness I mean that deductions claimed by one taxpayer are not necessarily offset by inclusions of another (and vice versa). After explaining how the Flat Tax is open, I will explore five central design elements that are affected by openness. I will also briefly consider additional issues that must be resolved before implementing the tax.
This article will address only the design issues relating to the Flat Tax. It will not discuss the merits of an income tax versus a consumption tax.(5) Nor will it discuss important questions of economic efficiency or the equitable distribution of the tax burden under the Flat Tax, except as they arise from design problems. The transition to the Flat Tax will be discussed briefly, but mostly with respect to design issues, not efficiency or fairness concerns. In addition, the article will assume that the Flat Tax is enacted in relatively pure form, so that political compromises that introduce additional complexity are not discussed. These issues are all important but are well covered in prior literature.(6) The major hole remaining is the design of the system.
Although this article is written at the level of implementation rather than theory, there are two important underlying theoretical problems. First, suppose we identify an anomaly in the treatment of a transaction under the Flat Tax. Someone might be over- or under-taxed, or there may be an unintended incentive--maybe a loophole--that causes taxpayers to structure transactions inefficiently to avoid tax. The question is whether the tax should be amended to fix the anomaly or whether it should be left as is. This requires a trade-off between the administrative costs of fixing the problem and the inefficiency, unfairness, and revenue effects of leaving the anomaly. How these trade-offs should be made is not fully resolved.(7) The original designers of the Flat Tax, Robert Hall and Alvin Rabushka had a strong preference for simplicity as witnessed by the proposed elimination of all personal deductions. But they failed to identify a large number of issues, and decisions must be made on these issues. The recommendations here are based on my judgment about the costs of complexity compared to the costs of a given anomaly. These judgments may be completely wrong (although I don't think so), but the point of the article is not to recommend final resolution of the issues but rather to identify the issues that must be dealt with in designing the Flat Tax and to evaluate the costs of solutions.
Second, in evaluating claims of administrative costs and complexity, there is no easily defendable baseline for comparison. Theoretically, one would compare either absolute or marginal administrative costs of each system. But in practice it is convenient to compare one system to another. Most of the comparisons in this article are to the current income tax. The reason is that most people have some sense of the administrative costs and complexities of the current system, so comparing the Flat Tax to the current system provides information. In considering tax reform, other important comparisons would be to other potential reforms, such as a simplified income tax or a VAT.
The Flat Tax proposed by Hall and Rabushka is a tax on consumption, not income. Part I of this article, therefore, gives background on consumption taxes and the basic mechanic's of the Flat Tax. Part II introduces the concept of openness and shows how the Flat Tax is open domestically and internationally. Part III considers five major design issues that stem from the openness of the Flat Tax. Part IV gives a very brief discussion of other design issues that will have to be resolved to implement the Flat Tax. Part V evaluates the design and provides a conclusion.
A. Consumption Tax Basics
This Part provides background on consumption taxes in general, not limited to the Flat Tax. This background is necessary for understanding the Flat Tax. All of the material provided in this Part can be found in prior literature.(8)
The goal of a consumption tax is to capture all consumption in the economy. Instead of measuring consumption directly, say through a tax on consumption purchases, consumption can be derived from income. Income in a given period is equal to the sum of a taxpayer's consumption and his change in savings during that period. By simple algebra, consumption is equal to income less the change in savings (i.e., minus savings plus dissavings).
The change in savings in a given period is equal to the difference between amounts saved and amounts withdrawn from savings to be used for consumption. We can measure tiffs difference by measuring difference in receipts from the sale of investments and the outlays for the purchase of investments. Net receipts in a period means that the taxpayer withdrew savings to consume and net payments means the taxpayer saved. As a result, we can tax consumption by measuring cash flows, including receipts, and deducting outlays (other than consumption outlays). A tax following this pattern is called a personal cash-flow consumption tax.
One important consequence of this logic is that the major difference between an income tax and a consumption tax is the timing of basis recovery. In an income tax, there is no deduction for savings. Instead, investments are given tax basis which is recovered when the investment is recovered (e.g., through depreciation or on sale). In a cash-flow consumption tax, basis is recovered immediately through a deduction for outlays (i.e., investments are expensed).
A second consequence of this logic is that, under certain assumptions, a consumption tax does not tax (exempts) the yield on investments. The intuition is that the immediate deduction in the cash-flow consumption tax creates tax savings, which can be invested. When the investment is sold, the taxpayer must pay tax on the full amount realized, but the tax is exactly equal to the invested value of the original tax savings.
Example Suppose a taxpayer earns $100 and wants to invest it. Assume the taxpayer has two choices: an investment that is immediately deductible but which is fully taxed on sale, and an investment that is not deductible but whose yield is exempt. Assume the tax rate is forty percent and that the pre-tax return on investments during the relevant time period is fifty percent. If the taxpayer invests in the asset with the exempt yield, the taxpayer must first pay tax on the $100 earnings. Thus, the taxpayer must pay a tax of $40 and has only $60 to invest. The $60 will earn a fifty percent return, or $30, which is not taxed. Withdrawing the initial $60 invested is tax free, leaving the taxpayer with $90. If the taxpayer invests in the deductible investment, the taxpayer can invest the full $100. The $100 will earn a fifty percent return giving the taxpayer $150. When the cash is withdrawn from the investment, the taxpayer must pay a tax of forty percent of $150, or $60, leaving $90. Thus, the immediately deductible investment and the yield-exempt investment leave the taxpayer in the identical place, with $90.(9)
A cash-flow consumption tax allows immediate deductions for all investments and, therefore, under the assumptions, exempts the yield on investments. Thus, we can replicate a cash-flow consumption tax by simply not taxing the yield on assets but also not allowing a deduction for purchases. This method of taxation is called yield-exemption. The Flat Tax uses yield exemption for nonbusiness assets such as housing.(10)
If this equivalence holds, the cash-flow consumption tax is just a tax on wages. The reason is that there are only two types of resources, labor and capital. If the return to capital is exempt, only labor is taxed, effectively creating a wage tax (treating all returns to labor as wages). There are, however, two major exceptions to the cash-flow, yield-exemption equivalence.
First, a cash-flow tax taxes certain returns in excess of the market rate of return, known as inframarginal returns. An investment has an inframarginal return if one cannot invest additional cash at the same rate. The inframarginal return is the return on the investment above the normal rate of return (i.e. the rate of return on any additional cash invested). For example, monopoly profits are inframarginal returns.
The reason that a cash-flow tax taxes inframarginal returns is that taxpayers cannot fully "gross-up" their investments by the tax savings because the savings must be invested at a lower rate than the original investment. The return on their investment of the tax savings from a deduction earns only the marginal rate of return, not the inframarginal rate. These returns will be insufficient to pay the tax on the original investment (which will have grown more quickly) when the investment is sold.
Example The facts are the same as above in which the taxpayer earns $100 and wants to invest it. Suppose, however, that the marginal rate for investments is ten percent, but the taxpayer has an opportunity to invest $60 at a fifty percent return. The remaining $40 (the tax savings from not paying immediate tax on the earnings) must be invested at ten percent. Under a cash-flow system, the taxpayer invests the full $100. After one year, the taxpayer has $90 from $60 of the investment and $44 from the investment of the remaining $40, for a total of $134. When the taxpayer withdraws the money, the $134 is fully taxable. The tax liability is $53.60, leaving the taxpayer with $80.40. Under a yield exempt tax, taxpayer invests only $60, which grows at a fifty percent rate to $90. The difference between yield-exemption and cash-flow taxation is $9.60. The $60 investment produced a return forty percentage points greater than the market return, or $24 (the $90 super-return is $24 greater than the $66 regular return on the same $60 investment). A tax on this excess would be $9.60, which is the difference between total exemption of the yield and cash-flow taxation. Thus, cash-flow taxation taxes the return in excess of the marginal return.(11)
It is not clear what portion of investments produce inframarginal returns.(12) Inframarginal investments are not just investments that produce extraordinary returns. Instead, they are investments in which one cannot invest additional money at the same rate. So a risky start-up business may not produce inframarginal returns as additional cash would be welcome. Instead, it may produce large returns to human capital (i.e., talent or skill) and to risk.(13)
Second, depending on the particular transition rules that are adopted, a cash-flow consumption tax would tax all existing capital (old wealth) on a one-time (present value) basis. Consider a taxpayer who makes a $100 investment today under the income tax. She gets basis equal to $100. Suppose tomorrow we impose a cash-flow consumption tax, and the next day she sells the investment for $100. When she sells the investment, she receives $100, fully taxed under the cash-flow consumption tax. The effect is that wealth existing on the date the consumption tax is imposed gets taxed, regardless of whether it had previously been taxed (had tax basis) under the income tax.
Whether this happens depends on the transition rule for changing to a new system. For example, if we eliminated the income tax and imposed a broad-based retail sales tax (a type of consumption tax), all spending for consumption would be subject to the tax. In particular, there would be no recovery of existing income tax basis because all purchases would be subject to the retail sales tax, even if made out of cash from the sale of assets with a high basis under the no-longer existent income tax. If we instead imposed a yield-exempt regime, there would be no tax on existing capital because all returns, such as our taxpayer's receipt of $100 on the sale of her investment, would be explicitly exempt.
While many economists treat consumption taxes as necessarily taxing existing wealth (and taxes that do not are different types of taxes),(14) there is nothing inherent in the concept of consumption that prevents us from choosing any form of transition relief desired. For example, we can choose to allow recovery of existing income tax basis when we switch to a consumption tax, which would reduce the tax on existing capital to the extent it has been taxed under the income tax (i.e., to the extent of income tax basis less liabilities). The Flat Tax has no explicit transition rules and, therefore, (subject to exceptions discussed below) attempts to tax existing wealth.(15)
Transition effects raise important economic, fairness, and political issues which may cause us to allow or deny transition relief. In particular, the efficiency of a consumption tax depends substantially on transition. The assumption in most efficiency analyses is that the transition tax is unavoidable. In economists' jargon, it is lump sum. Lump sum taxes are efficient precisely because they are unavoidable--there is no effect on behavior so there is no efficiency cost. The transition tax, which falls on all existing wealth, is a substantial portion of the tax base in a consumption tax, and elimination of this huge lump sum tax significantly reduces the efficiency of a consumption tax. In some models, the effect may be large enough to change the ranking between different types of taxes: A consumption tax with transition relief may be less efficient than an income tax, while a consumption tax without transition relief may be more efficient than an income tax.(16)
Similarly, the distributional effects of the transition tax are substantial. The tax falls on existing wealth, which tends to be held by high-income taxpayers. A consumption tax without transition relief, therefore, may be substantially more progressive than a similar tax with transition relief. The transition tax also redistributes between the old and the young. Because older people hold most of the existing wealth, the transition tax falls primarily on older individuals. A consumption tax that imposes a transition tax, therefore, redistributes from the old to the young as compared to a system that grants transition relief.
The Flat Tax offers no transition relief which simplifies much of the design as transition rules need not be considered.(17) As will be discussed in Part II, however, the transition tax in the Flat Tax is easily avoided. This will significantly affect the conclusions about the efficiency and distributional effects of the Flat Tax. Avoidance of the transition tax will generally make the Flat Tax less efficient and less progressive. To reduce avoidance of the transition tax, the Flat Tax can impose a host of complex rules, which may improve the efficiency and incidence of the tax, but also will make it more costly to administer.
Although the taxation of inframarginal returns and transition are the only theoretical differences between yield-exempt and cash-flow taxes, there is an important difference in implementation. In the real world, we cannot adequately police the border between wages and capital. Entrepreneurs are likely to take some of their returns to human capital in the form of a return to physical (including intangible) capital rather than as wages. Thus, some of Bill Gates's stock in Microsoft is best characterized as return to human capital or wages. Many patents may primarily represent human capital. The yield-exempt mechanism requires the tax system to make the distinction between the two types of returns because returns to physical capital are taxed under a completely different system (yield exemption) than returns to human capital (taxed as wages). The cash-flow mechanism does not have to distinguish between these two types of returns because all cash flows are treated identically: Returns to human capital disguised as returns to other capital are taxed under the cash-flow mechanism exactly the same way as wages.
Example Suppose a business owned by an individual creates a patent worth $100. Suppose also that the business pays the individual wages of $30 and then the individual sells the stock of the business for $70. Assume there were no inputs other than the individual's labor used to create the patent so that the entire $100 is a return to labor. Under a yield-exempt tax, the individual would pay taxes on the $30 of wages and not pay taxes on the sale of the stock. If the wages are below the appropriate amount, as they are here, the individual avoids taxes. Under a cash-flow tax, the individual would pay tax on the $30 of wages and the $70 of sales receipts. Regardless of the allocation of receipts between wages and capital, the individual would pay the same tax.
To the extent this border is difficult to police, there may be large differences between yield-exempt and cash-flow taxes.(18)
To summarize, a cash-flow tax taxes wages, inframarginal returns to capital, and, on a one-time basis, all existing capital. The two major differences between a cash-flow tax and a yield-exempt tax are the tax on inframarginal returns and the tax on existing capital. A yield-exempt tax would not tax inframarginal returns as all returns to capital are explicitly exempt. Nor would a yield-exempt tax impose a tax on existing capital without a special rule doing so because the yield on all capital is explicitly exempt.(19) Therefore, a yield-exempt tax taxes only wages. In addition, a yield-exempt tax requires a determination of whether a given return is a return to human capital or some other type of return while a cash-flow tax does not.
B. The Flat Tax--General Background
This Part will consider the general functioning of the Flat Tax as set forth by Hall and Rabushka. Three points will be made. First, the Flat Tax is a progressive consumption tax. Second, the business-level tax in the Flat Tax exists only to tax existing wealth and inframarginal returns. Third, the Flat Tax imposes multiple methods of taxing capital.
Begin by considering a retail-sales tax. A retail-sales tax is a direct tax on consumption purchases. If applied to all consumption purchases in the economy, it would be a consumption tax, equivalent in overall effect to the cash-flow tax discussed above.
The explanations of why the retail-sales tax and the cash-flow tax are consumption taxes are quite different, but it is easy to see the connection. In a cash-flow consumption tax, the taxable mount for a given taxpayer is equal to his consumption purchases because investments are deducted. Thus, the taxpayer effectively pays a tax if he purchases a good or service. The collection mechanism is for individual taxpayers to pay a tax on all of their consumption purchases during the year (or other accounting period) once, in a single tax filing. No special listing or tracking of each consumption purchase is required. In a retail sales tax, a tax is paid when a good or service is purchased for consumption. Businesses remit the tax, and the tax is computed for each purchase rather than on an annual basis. The overall effect, however, is the same.(20)
Next consider a valued added tax, or VAT. A VAT is simply a complicated method of collecting a retail-sales tax. A VAT collects the tax on each good at each stage of production rather than only at the retail level. The reason most nations impose a VAT rather than a retail-sales tax is that avoidance of a VAT is more difficult.(21)
Example Suppose Bigco creates a good from scratch and sells it in the market. To impose a VAT or a retail-sales tax, we tax the value of the good by requiring Bigco to pay tax equal to the value of the good times the tax rate. In this simple case, the VAT is identical to the retail-sales tax. Suppose Bigco purchases inputs from another company. In a VAT, the company selling the inputs would be taxed on the value of the goods sold. Bigco would then sell the finished good to the public and be taxed. If we taxed Bigco on the full sales price of the good, we would double-tax the good simply because Bigco purchased its inputs from another company rather than making the product from scratch. To prevent this, we allow Bigco to deduct its purchases, or claim a credit for the taxes paid on the purchase, and the tax savings from the deduction or credit offsets the tax on the seller of the input.
A deduction for purchases and an inclusion for sales means businesses are taxed on a cash-flow basis. A VAT that allows a deduction for purchases and an inclusion for sales is called a "subtraction-method" VAT. Note that there is no deduction for wages and no direct tax to the workers on wages. Deductions are allowed to prevent duplication of the tax and, therefore, should only be allowed for goods or services purchased from a business that has already paid the tax. Workers are not taxed, so the purchase of their labor is not deductible. Like a retail-sales tax, individuals pay the tax only on purchases.
European countries do not impose this method of VAT. Instead, they impose what is called a "credit-invoice" VAT. In a credit-invoice VAT, businesses get a credit against taxes for any taxes paid by the sellers of their inputs instead of a deduction. Conceptually, the credit and the deduction are the same--they both provide the same dollar offset against taxes for taxes paid by sellers of inputs. Both types of VAT use the same method to measure consumption. As will be discussed below, the real difference is the use of invoices.
The Flat Tax operates like the subtraction-method VAT described above, except that it allows businesses a deduction for wages and taxes individuals on wages. Individuals are not taxed on investment income, just like in a VAT, or sales tax. The genius of the Flat Tax is that, by taxing wages at the individual level, they can be taxed at a progressive rate. In a VAT, wages cannot be taxed at a progressive rate because there is no tax at the individual level, making the VAT more regressive than the Flat Tax. The Flat Tax, then, is a progressive consumption tax.
Once again, the difference between a consumption tax and an income tax is the recovery of basis. If an income tax were collected at the business level, businesses would not get an immediate deduction for the costs of inputs. Instead they would recover the cost over time. The only difference between the discussion here and the discussion of individual consumption taxes above is that here we are imposing the tax at the business level rather than the individual level, but the principles for measuring (and taxing) consumption in the economy are the same.
Thus, the crucial feature of the Flat Tax that makes it a tax on consumption is the use of immediate write-offs for expenditures. If the Flat Tax required businesses to depreciate assets based on their economic life, the Flat Tax would measure income. This type of tax has been labeled variously as an "income VAT" or the Comprehensive Business Income Tax (CBIT).(22) Note also that one might be cautious in designing the Flat Tax around expensing as it would be an easy change in the future for Congress to require depreciation. The design should, if possible, be sufficiently robust to cover this possibility.
Given that the Flat Tax has an explicit wage tax, and that a consumption tax taxes wages, inframarginal returns and existing capital, the only reason for the business-level tax must be to collect the tax on existing capital and inframarginal returns.(23) This is significantly different from the European systems in which the business-level tax taxes all consumption.
The business-level tax does not apply to all capital. While businesses are not defined by Flat Tax proposals, not all capital will be treated as held by a business in any likely definition. For example, personal residences, consumer durables, collectibles owned by individuals, and assets such as land or commodities held by individuals purely as investments are unlikely to be treated as part of a business. These assets represent a significant portion of capital in the United States. Personal residences alone are twenty-two percent of the capital stock.(24) The return on these assets is explicitly exempt instead of being subject to the business cash-flow tax. Because these assets are exempt from the business tax, they are not subject to the transition tax or the tax on any inframarginal returns.
Having multiple methods of taxing capital creates distributional and efficiency issues. Not imposing the transition tax and the tax on inframarginal returns on capital held outside of a trade or business means the overall tax on capital is lower, and more of the tax burden will necessarily fall on labor. It also reduces the size of the presumably efficient transition tax, which will reduce the efficiency benefits. In addition, if the reason the transition tax is efficient is because it is unavoidable, not imposing the tax on all assets will reduce its efficiency by making it avoidable--it will be based on individuals' decisions to purchase assets subject to the business tax.(25)
To summarize, the Flat Tax is a progressive consumption tax, collected at the individual level for wages and at the business level for existing capital and inframarginal returns. The key to the Flat Tax taxing consumption is the cash-flow mechanism at the business level, which allows expensing rather than capitalization and depreciation. Progressivity is created through the progressive tax on wages. Nonbusiness capital is taxed under the yield-exempt method so the Flat Tax imposes multiple methods of taxing capital.
C. Additional Details of the Flat Tax
Hall and Rabushka provide a number of additional details for the Flat Tax. To give a more complete sense of the tax, the most important of these details are briefly described below, but further discussion will be taken up in the rest of this article.
* Interest is not deductible by borrowers and not taxed to lenders. More generally, financial transactions are ignored, generating neither deductions nor inclusions. Only real transactions are subject to the cash-flow tax. The resulting system, used in all VATs, is known as an R-based system.
* Pensions are deducted by businesses when earned by an employee but not included by employees until paid (i.e., at retirement).
* Fringe benefits are not included by employees (as roughly true under current law) but, unlike current law, businesses may not deduct the cost of fringe benefits.
* State and local governments and, roughly, charities as defined under current law, are exempt from tax.
* Businesses withhold taxes on wages.
* Losses are not refundable, but can be carried forward indefinitely with interest at a rate equal to the average risk-flee short term yield during the year the loss is incurred (i.e., unused losses increase each year by that interest rate).
II. THE OPENNESS OF THE FLAT TAX
The most unusual feature of the Flat Tax is that it is open. As noted in the introduction, by openness I mean that the Flat Tax allows businesses to claim deductions that are not necessarily offset by corresponding inclusions elsewhere. For example, a business may deduct the cost of land purchased from an individual, but the individual is not taxed on the sale.
Credit-invoice VATs, by contrast, are generally "closed" in the sense that credits are only allowed for purchases from taxpayers under the invoice mechanism. In a closed system, transactions generally have no net tax effect unless they are consumption purchases because a business purchasing an asset may claim a credit (deduction) only if there is an offsetting tax on the seller. This is consistent with the structure of a VAT which is designed so that credits or deductions offset tax liability at earlier stages in production. Only when the good or service is ultimately purchased to be consumed is there a net tax on a transaction.
The openness of the Flat Tax stems from two sources: the lack of invoices for domestic transactions and the treatment of international transactions, each of which is discussed below.
A. Lack of Invoices
Recall that the Flat Tax has multiple regimes for taxing capital. The cash-flow regime applies to all assets used in a trade or business. Assets not used in a trade or business--for example, land or durable assets held by individuals--are taxed under a yield-exempt method.
The openness of the Flat Tax comes from the interaction of these two regimes. In particular, the Flat Tax does not have an invoice system governing the treatment of assets that switch between the regimes. This allows businesses to deduct purchases from individuals and other nontaxpayers. A business that purchases land from an individual may deduct the cost of the land even though the individual does not pay tax on the receipts.
The economic effect of an open system is that it does not impose a transition tax on assets held by nontaxpayers on the transition date that are subsequently used in a business.
Example Suppose an individual owns land worth $100 on the transition date. At some later date, a business buys the land for $100 and then resells it to a third party for $100. If the system is closed, the business does not get a deduction for the purchase of the land from the individual because a nontaxpayer was the seller. The business has gross receipts of $100 on the sale of the land and no deduction on the purchase. Its net receipts are $100, and it pays tax on $100. If the system is open, the business deducts the cost of the land, which means the business has no tax liability because the $100 deduction for the purchase offsets the $100 receipt on the sale.
Thus, the closed system taxes the full value of the land held by the individual while the open system exempts the value of the land held by the individual. The difference between the open and closed systems is the transition tax. Closed systems impose a transition tax on assets that switch from nonbusiness to business use, while open systems do not impose a tax on these assets.
Realistically, the value of the land will vary between the transition date and the time the individual sells it to the business and will further vary between the time the business buys it and sells it. These fluctuations in value do not affect the basic conclusion. For example, if the land changes value while held by the individual, the expected present value of the land on the transition date will still be $100 and the expected present value of the tax in the closed system will be the tax on $100. If the land appreciates in value while held by the business, the appreciation is subject to the cash-flow mechanism which will exempt normal returns and tax inframarginal returns. Thus, the effect of the closed system as compared to the open system is to impose a transition tax on assets held by nontaxpayers on the transition date that are ultimately used in a trade or business.
There are some exceptions to this conclusion. First, closed and open systems are different with respect to the resale of used goods. Suppose a consumer purchases a car subject to a subtraction-method VAT and then resells the car to a used car dealer. Tax was paid once by the individual when the car was originally purchased. If the car dealer cannot deduct the purchase price of the used car (because the system is closed), a double tax is imposed when the dealer resells the car. If the dealer can deduct the purchase price of the used car (because the system is open), only a single tax is imposed on the use (by two different individuals) of the car. To avoid the double tax created by a closed system, European VATs have special rules for used goods. An open system would require no special rules.
Second, an individual may create an asset through his own efforts and sell the asset to a business. An open system will not tax the value of the asset (unless the individual creating the asset is treated as engaged in a trade or business). A closed system would tax the value of the asset. Third, in the example above, if the value of the land increased unexpectedly, it might be difficult to argue that this windfall is merely part of the expected future value on the transition date. If windfalls are not included in the present value calculus, openness exempts post-transition windfalls as well as transition capital (or, phrased alternatively, a closed system taxes post-transition windfalls for assets that are eventually used in a trade or business).
The discussion above used an individual as the seller. But nontaxpayers potentially include foreigners (explored below), charities, governments, special classes of businesses such as small businesses, as well as individuals. In addition, most systems are not completely open or completely closed. European VATs are generally closed but olden allow deductions for purchases from special classes of nontaxpaying businesses.(26) The Japanese consumption tax is open domestically but closed internationally. The Hall and Rabushka system is completely open but could be partially open instead. Note also that other literature, particularly that by Charles McLure, calls a closed VAT "sophisticated," and an open VAT "naive."(27) I prefer the less weighted terms open and closed.
The decision to have an open tax will have distributional and efficiency effects. For example, if the tax is closed, assets that switch to business use after the transition date will be subject to the transition tax while, if the tax is open, assets that switch regimes will not be affected. The additional transition tax imposed by a closed system may be inefficient because the tax is avoidable (it is based on a decision to shift assets from personal to business use). But an open system creates an incentive to shift assets to nontaxpayers immediately prior to transition and shift them back after transition, which makes the transition tax less efficient. It is not clear without more information which system is more efficient. Similarly, an open system has distributive effects: The transition tax falls more on current wealthholders. Reducing the size of the transition tax means the tax burden necessarily falls more on labor.
The major advantage of an open system is that, at least at the surface level, it is cheaper to administer. In a closed system, each purchase has to be traced to a particular seller, and the purchaser must verify that the seller is a taxpayer. Effectively, invoices and tax registration are required in a closed system, exactly as in credit-invoice VATs. Conceivably, the Flat Tax could be closed, in which case many of the administrative and implementation issues would be similar to those of a European system. The Hall and Rabushka plan, however, is open. To highlight the administrative issues in the Hall and Rabushka plan, I will assume for the remainder of this article that the system is open with respect to payments to domestic nontaxpayers.
Many of the implementation issues for the Flat Tax arise from its openness. While detailed examination of specific issues is left for Part III below, some examples of how the open system affects design are useful here to show the pervasiveness of the issue.
First, consider the treatment of financial transactions. As discussed below, the Flat Tax does not allow deductions for payments on financial instruments and, correspondingly, does not require an inclusion for receipts from financial instruments. For example, interest is neither deductible by the payor nor includible by the recipient. Consider the treatment of a simple financial instrument: a contract to purchase fungible property in the future.
Example: Forward Contracts A business and an individual enter into a contract in which the business promises to purchase property for $100 in six months (a long forward contract). The forward can be settled in cash or property, at the election of the business. If the value of the property has gone up, the contract will be settled in cash, creating no income to the business, as the cash is from a financial transaction. If the value of the property has gone down, the contract will be settled by delivery of the property; its subsequent resale into the market produces a deduction for the business.
There are two possible outcomes. When the value of the property has gone down, the business will have deductible losses; when the value has gone up, the business will have exempt gains. Thus, a contract that has no expected economic value will generate expected tax losses. Moreover, the parties can take offsetting short and long positions in the same commodity (a straddle). The parties would then have no risk, and the business would always end up with a deduction (on the long position if the property value goes down and on the short position if the property value goes up). Given that no risk is involved, businesses can use this type of transaction to eliminate business taxes at any time by simply doing it in greater size.
This transaction works because the Flat Tax is open, and ignores financial transactions. If the tax were closed, the purchase of the asset from the individual or other nontaxpayer would not be deductible and the problem would be solved. European VATs, for example, do not have problems with this transaction. Many other problems with the taxation of financial transactions are the result of the openness of the Flat Tax.
Second, consider valuation problems created by an open system. An individual who owns a profitable business could sell an asset to the business for a wildly inflated price (using funds loaned to the business by the individual to make the payment). The business could claim a deduction and the individual would have no corresponding inclusion. This means the government has to police the price of sales between businesses and owners in an open system. Effectively, transfer pricing problems familiar in the international income tax context become prevalent domestically in the Flat Tax.
Third, consider the treatment of losses. European VATs allow full refundability of losses. That is, the government makes a payment to businesses that have credits in excess of their tax liability. This is necessary to ensure that the net result of transactions between businesses is zero (with net tax receipts attributable to purchases by consumers from businesses). If there are no invoices, the possibility of losses being claimed improperly all but precludes their full refundability. The individual described above who mispriced an asset to offset taxes paid by the business could, if losses were refundable, require the government to write checks. This is generally viewed as intolerable, and Hall and Rabushka therefore proposed that losses not be refundable. Instead, losses are carried forward and increased each year by a very low interest rate.(28)
Nonrefundability of losses, however, creates enormous administrative problems because transactions between businesses can generate tax liability. For example, suppose a business forms a subsidiary that will engage in a speculative research venture that will not produce profits until far into the future. If the business contributes assets to the subsidiary and the contribution is treated as a taxable sale or exchange, the business would have an inclusion equal to the value of the assets but the subsidiary would have nonrefundable losses. This means that special rules, such as the current nonrecognition rules, will be needed to prevent mistaxation of these types of transactions. Effectively, large portions of the current corporate or partnership tax rules might have to be incorporated into the Flat Tax.
Finally, consider the treatment of small businesses. Suppose that, for administrative reasons, small businesses are exempt from tax. Consider a small business that provides services to larger businesses, e.g., a small law firm or computer support firm. That is, the small business operates at the wholesale rather than retail level. Suppose the price of the inputs (including wages) of the small firm is $100 and it charges $110 for its services. The small firm cannot deduct the cost of its inputs, although the seller of the inputs will pay taxes. Effectively, exemption from tax means the small business is treated as a consumer of its inputs.
If the tax system is open, the purchasing business will deduct the $110 cost of the services. This deduction covers not only the $10 of value added by the small business but also $100 for inputs of the small business, which the tax-exempt small business could not deduct. The net benefit of exemption in an open system is the tax on the value added by the exempt business.
If the tax system is closed, the large business gets no deduction for the $110 purchase. If it resells the product for $110, all $110 in gross receipts are taxed. That is, both the $10 value of added by and the $100 of inputs used by the small business are taxed. But the $100 of inputs to the small business were already taxed because the exempt small business could not deduct its costs on purchase. This means the inputs of small businesses operating at the wholesale level are double taxed.(29) This is why small business exemptions in the European system are used almost entirely by retail businesses, and other small businesses often elect to be taxed.(30) For wholesale businesses, being taxed results in lower net tax payments than exemption!
These examples show that the openness of the system will affect many design decisions, making some, such as the treatment of financial instruments and losses, more difficult, and some, such as the treatment of small businesses, easier.
B. International Operation of the Flat Tax
The Flat Tax is unique in its treatment of international transactions because it taxes exports and exempts imports. As explained below, all other VATs exempt exports and tax imports. Although one might initially think that the difference is economically important, it turns out to have no major effects on trade. Instead, the most important implication is that this treatment makes the Flat Tax open internationally. This Part describes the Flat Tax's international system and the economics behind it. Design issues are sketched in this Part and are discussed in more detail in Part III below.(31)
The Flat Tax is a territorial, origin-based consumption tax. A territorial tax does not tax foreign earnings, dividends, or interest of U.S. taxpayers. By contrast, the current income tax taxes worldwide income of U.S. taxpayers (although the taxation of the income is often deferred until repatriation). Territorial jurisdiction is typical of consumption taxes imposed throughout the world.
Under an origin-based consumption tax, a taxpayer gets a deduction for imports and pays tax on exports. This treatment applies whether there is a cross-border purchase or sale, or file business simply ships the good across the border to a branch (in which case the deduction or inclusion is based on a hypothetical sale at fair market value at the border). Effectively, the tax is imposed on domestic production, regardless of where the product is consumed.
Example Export. An automobile is produced domestically and sold to a retailer in Germany for resale and use in Germany. When the car is exported, a tax is imposed on the sale of the car, just like any other sale by the business. The value of the production in the U.S. is taxed, but any value added in Germany, for example, through retailing, is not taxed by the U.S. Import. A German automobile is purchased by a U.S. retailer for resale in the U.S. The retailer deducts the cost of the purchase, like any other business purchase. Only the value added in the U.S. by the retailer would be taxed (i.e., the difference between the price paid by the retailer and the sales price to the consumer).
The Flat Tax is open internationally because it is origin-based. The German car maker is not subject to the Flat Tax when it sells the car to the U.S. retailer, so the retailer gets a deduction even though there is no offsetting inclusion.
No existing VAT is origin-based.(32) Instead, VATs are uniformly imposed on a destination basis. Under a destination-based tax, imports are not deductible and exports are not taxable. The tax base in a destination-based tax is domestic consumption.
Example Export. Under a destination-based tax, the same car sold in Germany would bear no U.S. tax. Under a subtraction-method tax, like the Flat Tax, no tax would be imposed on the sale, but the selling business would still get a deduction for its inputs, effectively offsetting any tax imposed at prior levels of production. Import. If a car is imported into the U.S., no deduction is allowed to the importer. When the importer sells the car, the receipts are taxed, so that the full value of the consumption in the U.S. is taxed here.
In the above example, Germany will probably impose a destination-based consumption tax. Thus, if the U.S. imposes an origin-based tax and Germany imposes a destination-based tax, a car exported to Germany from the U.S. will bear U.S. tax and German tax. If the car were produced in Germany and sold in the U.S., then Germany would not impose a tax under its destination-based system, nor would the U.S. impose a tax under its origin-based system. U.S. exports, therefore, bear a double tax and imports bear no tax. This is obviously not the strongest political selling point of the Flat Tax. (Note that this leaves aside questions about foreign income taxes.)
Economists argue, however, that exchange rates will adjust to eliminate adverse effects as long as the tax is imposed uniformly among on all goods in the economy. This argument is summarized below. The major design issues created by an origin-based system stems from its openness, and the problems created by an international open system are largely the same as the domestic problems highlighted above. The following discusses the basics of the economic analysis.
Destination-based consumption taxes do not alter international cash flows relative to a world with no taxes. Consider the initial imposition of a uniform destination-based VAT. All prices in the economy would increase by the tax. There would, however, be no effect on trade flows. Because all taxes would be removed on exports, exports would not be affected. Imports would bear a tax, but so would all competing goods; therefore, relative prices would remain unchanged, and imports would not be affected either.
Origin-based taxes are equivalent on the margin to destination-based taxes. Suppose the destination-based consumption tax was replaced with an origin-based consumption tax. Now a tax is imposed on exports, making U.S. goods more expensive than other goods, and no tax is imposed on imports, making foreign goods cheaper than U.S. goods. There will be less demand for U.S. dollars by foreigners and more demand for foreign currency by U.S. nationals. Relative to other currencies, the dollar will weaken and, given consumers' preferences for U.S. and foreign goods, the change in the value of the dollar will eliminate the effect of the tax.(33)
These currency adjustments are not intuitive to most people. The double taxation of exports and the zero taxation of imports will be immediately obvious to any observer, but there will be no easy way to prove that currencies have adjusted. We will have to rely on the economists' logic and, potentially, complex empirical studies. Thus, despite assurances by economists, origin-based taxes like the Flat Tax are likely to face serious political problems.
Also note that the preceding logic only applies to uniform taxes. Inevitably, some goods will not be taxed (e.g., goods sold by small businesses or any good given a special preference). Currency prices, however, will only adjust in the aggregate to the overall level of taxation. Thus, a sector of the economy that is taxed more heavily than the economy as a whole would be disadvantaged under the origin-based tax, and a sector that is taxed more lightly would be at an advantage.
Origin-based and destination-based taxes differ with respect to inframarginal returns and the transition tax. An origin-based tax taxes inframarginal returns on inbound investments and imposes a transition tax on foreign investors in U.S. assets. A destination-based tax taxes inframarginal returns on outbound investments and imposes a transition tax on U.S. taxpayers investing abroad.
Consider inframarginal returns first. Suppose that a U.S. corporation makes an investment abroad. Say it builds a plant abroad to produce goods sold here. Under a destination-based tax, any U.S. tax is rebated at the border and any import is taxed at the border. Effectively, the export of the plant (i.e., building it abroad) is deducted and the import of the goods produced by the plant is taxed. Destination-based taxes give cash-flow treatment to foreign investments. An origin-based tax does not allow a deduction for exports and does not tax imports and, therefore, gives yield-exempt treatment. Following the usual difference between cash-flow and yield-exempt taxes, if there are inframarginal returns, the corporation bears a tax under the destination principle but not under the origin principle.
The same analysis, flipped, applies for inbound investments. Foreign businesses bear a tax on inframarginal returns on investments in the U.S. under an origin-based tax but not under a destination-based tax.
The transition effects are more complex because they depend on relative price levels. The analysis breaks down into four pieces: the effects on a foreign owner of a U.S. asset under a destination-based tax; the effects on a U.S. owner of a foreign asset under a destination-based tax; and the effects on each of these under an origin-based tax. Begin by considering the case of a U.S. owner of a foreign asset under a destination-based tax. The cash flow from the foreign asset is not affected by the tax, but an increase in the relative U.S. price level reduces the value of the cash flow, which means the owner of the asset bears a transition tax. Under an origin-based tax, there is no change in relative price levels, so the U.S. owner of the foreign asset escapes tax. The opposite analysis holds for a foreign owner of a U.S. asset--the change in price levels under a destination-based tax ensures that the real value of the cash flows from the U.S. asset are unchanged. Under an origin-based tax, the lack of change in price levels means the foreign owner of a U.S. asset bears a tax.
One can say much more about the international economic effects of the Flat Tax, but the preceding analysis is sufficient for the consideration of design issues. One important question remaining is why the Flat Tax is origin-based if all other VATs are destination-based. The answer technically appears to be the GATT, although one suspects cosmetic issues of design also play a role.
2. Why is the flat tax origin-based?
The 1994 GATT imposes conditions on tax rebates for the export of goods to prevent nations from using tax rebates to create export subsidies.(34) Generally, the GATT allows rebates for indirect taxes but prohibits rebates for direct taxes.(35) The Flat Tax would probably be considered a direct tax because it taxes wages at the individual level.(36) Thus, even though the Flat Tax is economically identical to a VAT (with a cash payment to individuals based on family size), which is an indirect tax, the Flat Tax is treated as a direct tax and may not grant border rebates under the GATT (i.e., it must be origin-based).
If the U.S. were truly serious about the Flat Tax and wanted it to be destination-based, one can imagine some accommodation at the international level. It would be difficult for other nations to insist on a distinction that makes no economic sense in the face of U.S. pressure. A second reason for the origin basis of the Flat Tax, however, is cosmetics. Hall and Rabushka have styled the Flat Tax as an income tax and frequently refer to it as such, under the apparent assumption that this helps politically. Adopting an origin-based system makes it look more like an income tax. Moreover, it means there are fewer changes from current law. The question is whether this cosmetic difference is worth the disadvantages: being inconsistent with other consumption taxes, relying on currency adjustments to avoid adverse effects on U.S. exports, and being open internationally.
3. Effects on design.
The effects of the Flat Tax's international tax system on the design of the tax are discussed in Part III below. It is worth pointing out here that the most important implication of an origin-based tax on the design of the tax system is that it is open. The effects are similar to those of having an open system domestically: For example, an open system has transfer pricing issues, problems with losses, and difficulty taxing financial transactions. Allowing the system to be open internationally, however, compounds the effect because foreign businesses and their large, concentrated pools of capital are exempt. Problems with financial products and transfer pricing compound significantly merely because of the potential to increase their size and sophistication.
III. FIVE MAJOR DESIGN ISSUES
The number of individual design issues presented by the Flat Tax is too vast to cover in a single paper of reasonable length. To get a sense of the overall system, this Part considers five issues: (i) financial transactions; (ii) losses and the structure of the business tax; (iii) accounting methods; (iv) international transactions; and (v) small businesses. The first four are chosen because they are the core of modern business tax planning and are responsible for much of the complexity of the current tax law. Because of the complexity of the issues raised by these transactions, the discussion here is necessarily an overview. Small businesses raise extremely difficult and somewhat unique issues in the Flat Tax and are also worth close examination. These five issues, therefore, should give a good sense of the overall administrative and compliance issues in the Flat Tax. As discussed below, many of the design considerations in these areas are driven by the openness of the Flat Tax. Other issues, which may be equally important, are discussed much more briefly in the next Part.
A. Financial Instruments
As noted above, the Flat Tax, like a VAT, is R-based, which means it does not allow deductions for payments on financial instruments and, correspondingly, does not require an inclusion for receipts from financial instruments. Thus, interest and dividends are not deductible to payors and not includible to recipients. The nominal tax on a financial investment, therefore, falls on the operating business rather than on the investor. For example, if an individual lends a business $100 at a ten percent interest rate, the individual investor has no nominal tax liability. The business, however, will not be able to deduct the interest and, therefore, will have nominal tax liability on the amounts it earns with the $100.
Because the Flat Tax is R-based, it must distinguish between payments on financial instruments and payments on other investments. If the Flat Tax does not properly distinguish between these two, taxpayers may be able to shift the tax on investments. For example, interest can be disguised as another type of cash flow that is deductible. This shifts the nominal tax on the interest to the investor contrary to the premise of the R-based system.
The ability to shift the tax on the return to an investment is a particular problem in an open system because a deduction to the business (for disguised interest) will not necessarily be offset by an inclusion to the lender. Thus, to the extent the Flat Tax imposes a tax on an investment (because the return is inframarginal or a hidden return to human capital, or on transition), shifting the nominal liability to a nontaxpayer avoids the tax. For example, in the case of disguised interest, the business will be able to deduct the disguised interest but the lender will have no inclusion.(37)
The Flat Tax, therefore, will need a set of rules designed to distinguish the return on financial instruments from the return on goods and services. This distinction, however, is not coherent. Investments, whether financial or real, produce expected cash flows. There is little difference in theory between the two types of cash flows. Financial investments are ultimately (often through a series of intermediaries) investments in productive assets and the overall expected cash flows have to be the same. This means any set of rules in the Flat Tax that distinguish between financial and real returns will be manipulable.
The current income tax has some rules that distinguish financial returns from others. For example, many rules apply only to positions in actively traded property, which, on a rough basis, describes many financial instruments.(38) Moreover, current law attempts to identify disguised interest in some situations.(39) But these rules are not complete and do not need to be because the current tax does not have the same dichotomy between the two types of payments. The Flat Tax, effectively, will need many of the current rules for identifying interest and other financial flows plus a host of additional rules to complete the regime. These rules would be complex given the wide variety of ways financial and real flows can be intertwined. Because of the lack of conceptual coherence, the Flat Tax is also likely to need anti-arbitrage rules. Similar types of cash flows can be created through different combinations of financial and real assets. These cash flows will have different tax treatments which can be used in an arbitrage.
On the other hand, the timing of payments is less important in the Flat Tax, so many of the current rules on timing of income from financial instruments can be eliminated. (Part III.C below discusses accounting methods and when timing can be important in the Flat Tax.) Moreover, the capital gain and loss rules for financial instruments can be eliminated in the Flat Tax. There will clearly be some simplifications that can be made. The net effect is probably close to a wash--the Flat Tax rules will have to be about as complex as those of current law.
To make this discussion more concrete, consider the following four examples which illustrate the problem of distinguishing financial from real flows. The first example is the forward contracts example given above. Recall that, in this example, the taxpayer enters into a forward contract and settles it in either cash or property to generate deductible loss or exempt gain. If the transaction is done as a straddle, so that the business is both long and short the same commodity, the parties would have no risk and the business would always ends up with a deduction (on the long if the property value goes down and on the short if the property value goes up). Given that no risk is involved, businesses could use this transaction to eliminate business taxes at any time by simply doing it in greater size.
On its surface, the transaction seems to work because of electivity. If the taxpayer could not choose the treatment after the fact, the taxpayer could not be guaranteed to win. But taking away the explicit electivity would not solve the problem. The electivity arises because identical transactions can be structured as either real or financial. As long as the Flat Tax treats identical transactions differently, taxpayers will have effective electivity and will be able to take advantage of it. For example, suppose the explicit electivity were eliminated by always treating forward contracts as real. Virtually identical financial instruments would continue to be treated as financial. Taxpayers could buy and sell combinations of forwards and other instruments to produce zero net cash flow with net tax deductions or inclusions. If forward contracts were always financial, real business operations could be structured with long-term contracts and treated as financial.
The underlying problem is that the Flat Tax is R-based and open. This combination, R-basis and openness, means identical transactions will be treated differently in ways that can easily be arbitraged. If, for example, the tax were not R-based, the forward sale would be treated symmetrically regardless of whether the business settled with cash or with property. Businesses could no longer generate exempt gain. (This is why the transaction is not a serious problem under current law.)(40) Similarly, if the tax system were closed, the purchase of the asset from the individual or other non-taxpayer would not be deductible, again solving the problem. European VATs, for example, do not have problems with this transaction because they are closed.
There are, of course, a variety of approaches to stopping the particular transaction given in the example. One approach is to identify the financial and real elements of transactions and tax them appropriately. For example, the Flat Tax could treat assets that are purchased through forward contracts as purchased for their fair market value on the purchase date, regardless of the price paid. Any difference between the fair market value of the asset and the price paid would be a financial transaction, nondeductible to the buyer and exempt to the seller. In the above transaction, if the business took delivery of the property, it would only get a deduction for the fair market value of the property, which would make the forward purchase the same as any spot purchase. The forward contract would offer no advantage.
This approach, however, requires valuation. In addition, it would be complex for everyday business transactions such as long-term fixed price contracts. Perhaps for certain contracts closely related to purchases of inventory, businesses could agree to treat all payments on the contract as payments for a physical good. This would eliminate the asymmetry present in the example. The details of this regime could be worked out, but they would be complex. One can imagine many other regimes, but they all come at a complexity cost.
Example: Installment sale from individual to business A business purchases property from an individual for use in a trade or business for immediate delivery. The property is worth $100 today and the business promises to pay the seller $110 next year. The transaction is documented as a purchase for $110 without any payment of interest.
This transaction involves an implicit loan. Effectively, the business has borrowed $100 from the seller at a ten percent interest rate. If we separated the interest element from the purchase of the physical property, we would give the business an immediate deduction for the $100 purchase and no deduction for the $10 of interest.
If we tax the business on a cash-flow basis we produce the same present value result. The business would get no deduction for the purchase until it paid cash. But paying $110 next year is the same as paying $100 today, so the business is indifferent between the two methods and, therefore, documentation does not generally matter in this case.
The reason policymakers might care about this transaction is that it allows businesses effectively to elect to treat loans on a cash-flow basis. This allows businesses to shift the return on the financial investment to individuals, a result contrary to the basic decision to have an R-based tax. Given the equivalence between yield-exempt and cash-flow taxes, policymakers should only care where the equivalence does not hold, namely in the presence of inframarginal returns or hidden returns to human capital. For example, if the interest rate was not just ten percent interest, but instead was contingent on a return, cash-flow treatment would allow shifting of inframarginal returns to yield-exempt individuals. The problems presented here would be eliminated if the tax system were closed.
If we reverse the transaction, there are still problems with identifying interest, but the problems with misidentification do not stem from openness.
Example: Installment purchase by an individual from a business An individual purchases a $100 good from a business by promising to pay $110 in one year. The transaction is documented as the purchase of a $90 good with $20 of interest.
By overstating interest on the sale, the business reduces its taxable receipts without changing its cash flows. In the example, the business reduces taxable receipts by $10 as compared to a true statement of the interest in the transaction. The individual on the other side of the transaction is indifferent because no payment on the transaction is deductible. The parties, therefore, do not have adverse interests and there is a strong incentive to mischaracterize the transactions. There is no deduction for a payment to a nontaxpayer, so openness has nothing to do with this transaction.
Correctly identifying the interest element in the transaction will not be easy because the parties often will have no incentive to negotiate a true interest rate. For example, in a typical retail installment sale (say, the purchase of furniture or a car on credit), there is not likely to be a comparable, nondistorted price to use to determine the true price. Determining the true price would involve some measure of profitability of sellers or the appropriate interest rate. One can imagine solutions that involve publishing maximum interest rates for various types of loans, but any such solutions would be complex.
This transaction creates problems in a closed system. One would expect, therefore, that European VATs would have solutions. Nevertheless, as far as I know, VATs generally have no rules attempting to limit interest on consumer lending. The problem may simply be too difficult, although one would expect that the amount of revenue at issue would be large and, therefore, worth a few complex rules.
The next example shows how taxpayers can arbitrage the distinction between real and financial assets. It requires a preliminary example.
Preliminary Example: Sale of property and repurchase A business sells property to an individual for $100 and promises to repurchase the property next year for $110.(41)
Effectively, the business has borrowed $100 at ten percent on a nonrecourse, collateralized basis. Because the loan is documented as two transactions in property, the business is taxed on a cash-flow basis.
This transaction is similar to the first example of the forward contract --it involves a similar forward purchase of property for a fixed price. But the tax-planning goals are different. In this transaction, the goal is to separate the interest element in the return to a physical investment from the risk element. By separating the interest element in a physical asset, this transaction allows interest to be taxed on a cash-flow basis. This example more closely resembles the example of an installment sale from an individual to a business because it allows a loan to be treated on a cash-flow basis. The same concern is present: the shifting of inframarginal returns to nontaxpayers. And here again, this transaction is not a problem in a closed system.
Example: Arbitrage A U.S. business enters into an arrangement with a shell corporation located in an offshore tax haven. In the arrangement, the shell borrows money from the business and promises to repay a contingent amount in the future. The shell then uses the money to purchase an asset from the business. In addition, the business agrees to repurchase the asset around the same time the loan is due for the same contingent purchase price. The contingency will be set so that it is very likely to go up at a very high rate.
The cash flows on the transaction wash out. The business can neither gain nor lose and at all times is in an identical position to doing nothing. The tax result for the business, however, is an immediate inclusion from the sale of the asset and a very large deduction when the asset is repurchased. The loan has no results under the Flat Tax. If the repurchase is set at a high enough price, the business can simply manufacture deductions.
This example uses the ability to structure investment returns as either financial or real to create an arbitrage. The arbitrage puts the tax liability in a nontaxpaying entity and creates an offsetting deduction to a taxable business. Anti-arbitrage rules might prevent transactions as bald as this one, but either the rules will be overbroad or there will be some gaps.
The underlying problem behind all these examples is that there is no principled distinction between financial and real flows. For example, asset pricing models apply equally to financial and real returns. Economically, the two are simply not different and can, therefore, be intertwined in an indefinite number of ways. The transactions considered here are extremely simple, and I have no substantial economic incentive to find the really big holes. Taxpayers will have strong incentives to find others, and one can imagine that many more exist.
Switching to a closed system (i.e., using invoices and a destination basis) much like the European systems would substantially reduce the problems. All of the problems illustrated above, except the problems in the retail installment sale example, would be eliminated under such a system. The openness of the Flat Tax is central to the design of financial products rules.
If the Flat Tax must remain open, one place to start would be the various time value of money rules in current law.(42) For example, the installment sale rules would reduce the problem with understatement of interest in the sale from an individual to a business.(43) New rules imposing maximum interest rates would reduce the problem of installment sales from businesses to individuals. Special rules for forward sales that allow deduction or require inclusion only of the fair market value of the underlying asset would help (treating the rest as a payment on a financial transaction). Integration rules or other antiarbitrage rules would be needed to reduce the problems associated with creating loans through the sale and repurchase of property. The scope of these rules is likely to be broad and uncertain and their application complex.
The Flat Tax, however, does offer some simplification in the financial products world. In particular, the character of gains and losses as ordinary or capital would no longer matter. In addition, timing would matter much less in the Flat Tax, so many of the timing rules could be eliminated. Overall, the set of rules needed would be on the same order of magnitude in terms of complexity as the financial products rules under current law.
B. Losses and the Operation of the Business Tax in a Flat Tax World
This Part considers two related subjects: (i) the treatment of net losses in a given year; and (ii) the rules governing business transactions such as formations, liquidations and mergers. These two topics are covered together because the treatment of losses is central to the design of the business-level tax.
Hall and Rabushka propose not to allow refunds for losses (i.e., a cash payment from the government equal to losses in the accounting period multiplied by the tax rate), contrary to the uniform rule for VATs.(44) While they do not clearly state their reasons, the apparent reason is the lack of invoices, meaning that the system is open.(45) In an open system there is no necessary tension between a deduction (producing refundable losses) to one party and an inclusion at the same tax rate to another party. Without this tension, pricing problems and outright fraud become a significant concern.(46) Instead of allowing refunds, Hall and Rabushka propose to allow losses to be carded forward (but not back) indefinitely, increased each year by the average daily yield on three-month Treasury Bills during the first year.(47)
The interest rate is set lower than the market rate. Although Hall and Rabushka do not give the reason why, presumably, if the interest rate were the market rate, a carryforward regime would create no fewer problems than a refund regime as the two would have the same present value (assuming losses of liquidating bankrupt companies could somehow be used, say through a merger with a profitable company). Thus, the low interest rate might be intended as an intermediate solution between full refundability or its equivalent and current law, which allows carryforwards without interest.(48)
A low rate of interest on loss carryforwards will change some of the conclusions about the effects of the Flat Tax given in Part II above. To the extent the interest rate is lower than the cost of capital, the Flat Tax overtaxes some consumption. The value of the initial deduction for an investment that creates a loss will not be sufficient to offset the gain on the future sale because the deduction grows at a rate that is lower than the market rate. The greater the difference between the cost of capital and the carryforward rate, the higher the tax. The cost of capital will be higher for riskier projects but the refund rate will stay the same. Therefore, the analysis showing that a consumption tax does not tax the returns to risk would not apply: high-risk investments would be taxed at a higher rate than low-risk investments. In essence, the carryforward regime creates a sliding-scale tax rate: Goods and services produced by entities with higher costs of capital are subject to a higher tax than goods and services produced by entities with lower costs of capital.
While the nonrefundability of losses has economic consequences, the immediate implications for implementation are straightforward. Businesses would have to keep records of loss carryforwards, adjusting them by the interest rate and the government would have to publish the appropriate interest rates. Some complications could arise in the Hall and Rabushka proposal because carryforwards arising in different years would use different interest rates, requiring complex calculations and stacking rules to determine which carryforwards are used first. Presumably, a single interest rate could be used instead. The main impact of nonrefundability, however, is in the design of the business tax.
To illustrate the connection between the treatment of losses and the business tax, consider the tax system if losses in a given year were refundable. Transactions between businesses would create no net tax as taxable receipts to one would produce deductible payments to the other, and the refund from the deduction would exactly equal the tax on the receipts. The treatment of transactions between businesses, therefore, would be relatively unimportant so long as businesses treat transactions the same. All transactions between businesses, therefore, could be treated as taxable. Avoidance of taxable treatment would also have no net effect (as long as nothing left the business tax base, say as a distribution to shareholders, without tax). Few business tax rules would be needed. This is demonstrated by the European VATs which refund losses and have almost no special rules governing transactions between businesses.(49)
If losses are not refundable, transactions between businesses can have tax effects. Suppose losses were not refundable and could not be carried forward or back. That is, losses not used against receipts in the year incurred would be forfeited. In this case, a contribution of property by a business to a newly formed joint venture, if treated as a taxable transaction, could produce net tax--there would be an immediate tax to the contributor on the exchange and, if the joint venture did not immediately produce revenues, no offsetting deduction to the new venture. We would need nonrecognition rules for such formations. Similarly, a sale from a profit-making company to a company with no net receipts (say a bankrupt company) would generate net tax revenues. The profit-making company would have taxable receipts but the bankrupt company would get no deduction. There would be a strong incentive for profitable businesses to acquire money-losing business and rules might be needed to prevent distortions in the market for corporate control. The various nonrecognition regimes and the anti-loss trafficking rules that are central elements of the current corporate tax regimes will probably need to be duplicated in the Flat Tax because of the treatment of losses. Losses are the pivotal element in the design of the business tax.
The easiest place to begin exploring the business tax is with the formation of a business. There are two polar cases: the formation of a new company by an individual (or other nontaxpayer) and the formation of a subsidiary by a business. Suppose that in both cases the contributor transfers property to the business in exchange for equity in the business.
The sensible rule for the formation of a business by an individual is to treat it as a taxable sale. Under current law, a formation is generally tax-free to both the business and the individual.(50) That is, neither the individual nor the business recognizes gain or loss on the contribution. Because individuals would otherwise be taxed, they like this treatment and generally do not seek to avoid it. In the Flat Tax, however, nonrecognition would deny a deduction to the business with no offsetting benefit to the individual, because individuals are not taxed on the sale of property anyway. That is, if the same transaction were structured as a sale, the business would get a deduction without an offsetting gain to the individual. Nonrecognition, therefore, would be taxpayer adverse rather than taxpayer friendly and would give individuals a strong incentive to structure contributions as sales. As such, it would require enforcement, such as a prophylactic rule treating all sales between any substantial owner of a business and the business as tax-free contributions and preventing third parties from facilitating such sales. A good case can be made that the Flat Tax should instead simply treat all formations by individuals as sales.
Now consider the formation of a subsidiary (sub) by a parent corporation (parent). If the formation is treated as a sale, parent would have taxable receipts and sub would have deductible payments. But, as noted above, if sub is a new business that will not produce net receipts for a number of years, parent's tax would not be offset by sub's deduction. Each year the loss is carried forward it loses value because the carryforward interest rate is below the market rate, making the rule particularly punishing for long-term speculative or research ventures. A tax-free contribution regime is necessary to facilitate the formation of businesses by other businesses.(51)
Coordinating the two regimes, taxable treatment for contributions by individuals and nonrecognition treatment for contributions by businesses might be complex. Rules will be needed much like current law to determine which contributions are tax-free and which are taxable.
The same problem might occur for distributions of property from a business. Again, consider individuals and businesses separately. If a business distributes property to an individual, the distribution must be taxable to the business. The business deducted the cost of the property when it was purchased. The distribution must be taxed to recapture the initial deduction. If not, business could purchase consumption goods, deducting the cost, and distribute them tax-free to the owners of the business. Liquidating distributions to individuals would have to get the same treatment as operating distributions.
Suppose that a subsidiary distributes property to a parent or liquidates into a parent. If losses were refundable, the distribution would never produce overall gain or loss because any gain to the subsidiary would be offset by deductions to the parent. But if losses are not refundable, we must count on the parent being able to use the deduction (for the acquisition of sub's property) or the distribution will have a net tax effect. It would, therefore, be desirable to have a nonrecognition rule somewhat like the rule under current law for distributions to and liquidations into parent businesses.(52)
The rules governing corporate acquisitions raise similar issues. Because the Flat Tax is R-based, the purchase of the stock of one business by another is tax-free. But, without a special rule, the purchase of assets of a business would be taxable. If losses were refundable, the two treatments would be the same, except for administrative costs. European countries, for this reason, generally "zero rate" (tax but at a tax rate of zero) or exempt the purchase of an ongoing line of business, effectively allowing an asset purchase of a line of business to get nonrecognition treatment.(53) If losses are not refundable, the treatment of a transaction as a tax-free stock purchase or a taxable asset purchase can make a difference. Buyers with losses would get unusable deductions in an asset purchase and sellers with losses would have exempt gains on a stock sale.
To eliminate these differences, the Flat Tax could have rules allowing taxpayers to treat stock purchases as asset purchases, or vice-versa. Current law has rules of this sort. For example, current law has an election to treat a stock purchase as an asset purchase and has nonrecognition rules that allow asset acquisitions to get treatment similar to that given to stock acquisitions.(54) The Flat Tax could adopt similar rules.
If the Flat Tax attempts to harmonize stock and asset purchases, it could have relatively loose rules as compared to current law for two reasons. First, the Flat Tax does not have a two-level corporate tax. Many of the rules of current law are complex because of the need to defend the two-level system. Second, nonrecognition treatment under an income tax produces different results than taxable treatment. Under a consumption tax, nonrecognition treatment for transactions between businesses is usually the same as taxable treatment (with the exception being if one of the businesses is a loss company). Therefore, the Flat Tax could mimic the European rules and simply have a line-of-business type rule, rather than have all the bells and whistles of current law.(55)
The general pattern from these examples is that transactions between businesses can and often should be treated as nonrecognition transactions, primarily to prevent the nonrefundability of losses from distorting the results. At the same time, transactions between businesses and individuals (and other nontaxpayers) should be taxable.
The limitation on this logic is the concern about loss trafficking. Nonrecognition rules in the Flat Tax would allow businesses to transfer assets to loss companies to use up losses. For example, without limitations on nonrecognition treatment, a profit-making company could purchase a readily marketable commodity, claim a deduction for the purchase, and transfer it to a loss company in exchange for (preferred) stock in a nonrecognition transaction. The loss company would sell the commodity and use its losses to offset the tax on the receipts. The net result is to transfer the loss to the profit-making company.
Transferability of losses may be a good thing, but it is not the regime that Hall and Rabushka proposed. There is a significant tension between the need for nonrecognition rules and the rule that losses are not freely transferable. The question is, assuming transferability will not be literally allowed, whether or to what extent rules should be adopted to limit loss transferring transactions.
One possibility is to impose limitations on nonrecognition similar to those found under current law. For example, current law does not allow transfers to corporations to be tax free unless the transferors have at least eighty percent control of the corporation. Similarly, only liquidations into eighty percent corporate parents are tax free. Similar limitations on loss transferability might need to be written into the acquisition rules, eliminating the hope that many of the complexities of current law can be eliminated.
Current law also has rules that explicitly limit the ability of businesses to acquire other businesses to take advantage of unused losses.(56) The goal of these rules is to prevent the treatment of losses from causing distortions in the market for corporate control caused by the loss regime. These rules are very complex under current law and similar rules would be equally complex under the Flat Tax. There is no clear way to differentiate the bad transactions from the good so the rules are inaccurate, over-taxing some transactions yet missing others that should be covered. Given that increasing losses each year by the interest rate will mean the need to transfer losses will be less than under current law, the scope of explicit anti-loss trafficking regimes should be limited.
Stepping back, it becomes apparent that the treatment of losses means the Flat Tax generally will need nonrecognition rules to prevent everyday transactions between businesses from inappropriately generating tax liability. But these nonrecognition regimes will need all sorts of bells and whistles both to ensure that transactions between businesses and individuals are taxable and to prevent free transferability of losses. In addition, the Flat Tax will very likely need explicit loss transfer limitations, such as those found in current law. Effectively, much of current law ends up being recreated in the Flat Tax.
The one area where the business tax would be simpler than current law is in the treatment of distributions. The current rules for both partnership and corporate distributions are complex--the partnership rules are designed to prevent disguised sales; the corporate rules enforce the double-level tax. In the Flat Tax, distributions would simply be tax-free and few rules would be needed.
The overall assessment of the business tax rules in the Flat Tax is that they would be somewhat simpler than current law, although many of the basic elements of current law, such as nonrecognition rules and distribution rules, would remain.(57)
C. Accounting Methods and Periods
This Part considers two separate but related questions: (i) what is the appropriate period for cumulating transactions (the accounting period); and (ii) what is the appropriate method of determining when transactions are to be accounted for (the accounting method).
An initial intuition might be that the accounting method and period do not matter in the Flat Tax because, as demonstrated in Part I above, timing does not matter in a consumption tax. The savings from the initial deduction for an investment exactly pays for the tax on the sale of the investment. If the sale is accelerated or deferred, the value of the tax savings from the initial deduction decreases or increases accordingly because both the investment and the tax savings grow at the same rate. Accelerating or deferring a transaction, therefore, would have no benefit. Accounting methods and periods (by definition) are used solely to determine the timing of transactions, and if timing does not matter, accounting methods and periods should not matter. In fact, one scholar has argued that the fact that timing does not matter in a consumption tax (so that the timing of realization is not an issue) is one of the principal administrative arguments in favor of such a tax over an income tax.(58)
The examples demonstrating that timing does not matter were correct within their assumptions. The examples, however, assumed that the accounting period was the same as, or shorter than, the duration of a transaction. If a business purchased an asset in one year and sold it ten years later, the equivalence between yield-exempt and cash-flow taxation could be demonstrated with ten-year or shorter accounting periods. Suppose, instead, that the accounting period is longer than the length of the transaction.
Example A business purchases an asset from an individual or foreigner for $100 on December 31, 1999 and resells the asset to an individual or foreigner for $100.01 on January 1, 2000.
The business deducts $100 in 1.999 and includes $100.01 in 2000. If returns are filed on an annual basis, the deduction would come a full year before the inclusion. For the cost of the tax on one penny, the business gets the use of the value of the $100 deduction for one year.(59) If, however, the property were sold by the business at the end of the year 2000, so that the accounting period matched the length of the transaction, there would be no net benefit or tax on the transaction. The timing of the purchase and sale affect the net tax on the transaction. There is an incentive to accelerate the purchase of property from, and defer the sale of property to, nontaxpayers. Moving a purchase up by a few days to an earlier year accelerates the deduction at a very low cost. Deferring a sale by a few days can delay tax on the receipts for a year. Timing can matter in the Flat Tax and, therefore, we must determine the appropriate accounting period and method.
The openness of the Flat Tax makes the problem with timing worse, but the timing problem would exist even in a closed system. In a closed system, timing would matter only for final sales to consumers, as the tax from any other sale is offset exactly by a corresponding deduction to the other party. In an open system, timing can matter anytime one of the parties to the transaction is not a taxpayer. Openness, therefore, greatly expands the scope of the problem.
The obvious solution to this problem is to have short accounting periods. If the accounting period were a single day (or a single minute), there would be no distortions. One day's (or one minute's) delay or acceleration of a transaction would produce one day's (or one minute's) change in the timing of the tax deduction or inclusion. Timing would no longer matter.
The problem with daily accounting periods is complexity. While a very short accounting period would not necessarily require a filing for each period, it would require taxpayers to track exactly when each transaction took place and to make adjustments, such as an interest charge, to reflect the timing of the transaction. This would be impractical.
Most VATs require quarterly filings, and some allow monthly filings (mostly to allow businesses to get refunds more quickly).(60) Quarterly filing reduces the effect of timing disparities without imposing enormous administrative costs. Some similar approach might be necessary under the Flat Tax. Quarterly filing need only be required of businesses as the accounting period problem is a problem with the cash-flow mechanism used by businesses, not the yield-exempt mechanism used by individuals.(61) Current law already requires businesses to make quarterly payments, so quarterly filing would not be a big increase in filing costs.(62) If quarterly filing is a problem, businesses could be required only to collect information on a quarterly basis and to file returns on a yearly basis (based on the information collected each quarter).
The shorter the accounting period, the lower the benefits of manipulating the timing of transactions (and the lower the penalties for those inadvertently on the wrong side). The question is whether quarterly accounting periods are sufficiently short that we need not worry about the accounting method. The answer depends on the cost of imposing accounting method rules that reduce problems with the timing of transactions. (Note that eliminating the problems through an accounting method [as opposed to a very short accounting period] is not possible because taxpayers will have incentives to actually change the timing of transactions, not just to manipulate the accounting rules.)
Hall and Rabushka propose to put all businesses on the cash method of accounting.(63) The reason is that the cash method measures actual inflows and outflows, so it is uniquely appropriate for a cash-flow tax such as the Flat Tax. The cash method, however, would be extremely easy to manipulate. For example, if a receipt is delayed from the end of one quarter until the beginning of the next, the loss of the use of money for a short time may be less than the benefit of paying taxes a quarter later. Similarly, accelerating a payment from one quarter to an earlier quarter may cause the taxpayer to lose the use of money for a short time but may be offset by the acceleration of the deduction.
Current law requires the accrual method of accounting in part because it is thought to be less manipulable than the cash method.(64) Many elements of the accrual method, however, would have to be rethought for the Flat Tax. For example, the accrual method of accounting requires inclusion and deduction when the fact of the income or liability is fixed and the amount can be reasonably determined.(65) This may occur at a completely different time than when cash is received and no adjustment is made for the time value of money. In addition, the accrual method is complex and questions of the appropriate timing of transactions arise frequently.(66)
An alternative would be a set of rules similar to those used by European VATs. VATs have detailed timing rules, often with different rules for the supply of goods and for services.(67) Generally, however, they treat the supply of the good or service as the taxable event. One reason for looking to the supply of goods or services might be that supply is less manipulable than the payment of cash. The problem with adopting these rules is that they would be largely unfamiliar to domestic businesses.(68) Moreover, the theoretical basis for these rules is uncertain as the time of supply might be different from the time cash is received, creating implicit loans which would have to be accounted for properly.
Given that accounting rules will matter and that there is likely to be more than one acceptable method of accounting, the Flat Tax will probably need a counterpart to the rules in the current law that prevent double counting when taxpayers change methods.(69) These are among the more complex sets of accounting rules under current law--the new Flat Tax counterparts are likely to be equally complex.
While the Flat Tax will need some accounting rules, the problems with accounting methods should not be overstated. The single biggest area of simplification in the Flat Tax is probably accounting issues. The Flat Tax will eliminate many of the most troublesome aspects of accounting under current law. The capitalization requirement as embodied in common law principles(70) and in complex statutory rules(71) would be eliminated, substantially simplifying accounting problems. Inventory rules(72) would no longer be necessary as inventory would be deducted when purchased. In addition, timing problems are limited by the length of the accounting period in the Flat Tax, but are not under the income tax. For example, special rules for long-term contracts,(73) which are extremely complex and important under the income tax, would have far less importance under the Flat Tax because taxpayers could only achieve deferral for one accounting period.
The net result for accounting methods and periods is that the rules would be substantially simpler than current law. But accounting methods and periods would still matter. Moreover, many of the accounting method rules would probably have to change.
D. International Transactions
This Part considers design issues relating to international transactions. Recall that the Flat Tax is origin-based, so that businesses deduct the cost of imports and are taxed on exports. The basic rules for an origin-based system are straightforward. They are the same as for domestic transactions: Businesses get a deduction for purchases and an inclusion for sales. If goods or services are transported across borders without a sale, (e.g., a transfer to a foreign branch of a domestic business), they are treated as sold at the border for their fair market value. Similarly goods and services imported into the U.S. without a sale are treated as purchased at the border for their fair market value.
The most important problem with origin basis is that the system is open. The problems created by having an open system, however, are not generally treated here as separate international tax problems, although they could be because the relevant transactions are across international borders. Thus, a problem with a cross-border financial transaction is treated here as a problem with the taxation of financial products. This Part explores other problems with international transactions and also briefly mentions transfer pricing which is a problem with openness not explicitly discussed elsewhere. The general conclusion is that most of the implementation issues for international transactions (other than the consequences of openness) are not that serious, which is important given that some have claimed that these issues are significant.(74)
1. Transfer pricing.
Because goods must be valued when they cross the border, valuation and, in particular, transfer pricing (i.e., pricing of goods transferred between controlled entities), is a problem. For example, consider a taxpayer who begins the manufacture of a car in the United States, ships it to Mexico for some stage of production and then ships it back to the United States to be finished and sold. If the transfer price when the car is shipped to Mexico is artificially low and the price when the car returns is artificially high, the taxpayer can shift income to Mexico from the United States. Similarly, a service provider might sell services from an offshore tax haven while providing all the work domestically. Only to the extent value is added in the tax haven are the receipts properly allocable to the tax haven, but enforcing appropriate transfer pricing will be difficult.
It is likely that the transfer pricing regime in the Flat Tax would be similar to current law in both scope and complexity. Because the Flat Tax is territorial while current law has worldwide jurisdiction, locating earnings offshore may create more of an advantage under the Flat Tax than under current law, but the extent of the difference is not clear, given the ability to defer the taxation of foreign earnings under current law. At most, more enforcement or a slightly stronger set of regulations might be needed. Examining the details of transfer pricing enforcement, however, is not as important as noting that the need for transfer pricing enforcement is a very significant fact in evaluating the simplicity of the Flat Tax.
2. Creditability of the flat tax.
If a country has an income tax that taxes worldwide income, generally only other nations' income taxes are creditable against the country's income tax. This is the case, for example, for the U.S. worldwide income tax with respect to foreign taxes. The issue is whether the Flat Tax would be creditable against foreign income tax. Reuven Avi-Yonah suggests that this might be a problem.(75)
I do not believe this issue is very serious. On the margin, the Flat Tax imposes no tax on capital income, so, on a present value basis, credits do not matter. Effectively, if the Flat Tax is not creditable against foreign income taxes, the initial reduction in credits when the investment is expensed exactly offsets the value of the credits from the U.S. tax on the future returns from the investment.
As noted above, an origin-based tax taxes inframarginal returns on inbound investments, so whether the tax is creditable matters if a foreigner is receiving inframarginal returns for an investment in the United States. But if the inframarginal investment is specific to the United States, whether the foreign country grants a credit will not affect the decision where to invest. Whether to grant the credit will be up to the foreign country and, while it may matter to individual businesses, it need not concern us as a serious issue in implementing the Flat Tax. The only case remaining, where the inframarginal return is not specific to the United States, is unlikely to be a large category.
Avi-Yonah has also suggested that the Flat Tax would not qualify under existing treaties, thus requiring renegotiation of our entire treaty network.(76) Moreover, because the Flat Tax eliminates the withholding tax on dividends, one of the main negotiating tools in the treaty process, renegotiation may not be possible. The Flat Tax would, then, effectively toss out the entire network of tax treaties of the United States. And treaties are important to U.S. businesses as they provide reductions or exemptions from foreign withholding taxes, scale back the tax reach of host countries, and prevent discriminatory treatment of foreign investment by host countries. Renegotiation of treaties would be a major implementation cost of the Flat Tax.
Tax treaties apply to income taxes. Avi-Yonah notes that the Flat Tax is not an income tax and concludes that, therefore, it is not covered by existing treaties. Nevertheless, as a matter of treaty interpretation, existing treaties should apply to the Flat Tax. Treaties typically apply to "federal income taxes" and any identical or substantially similar taxes.(77) Generally, income taxes are not defined in treaties. While the Flat Tax would tax consumption, not income, it is not labeled a consumption tax, which seems to be the key factor. For example, the United States has previously had periods under its so-called income tax where, because of accelerated depreciation and investment tax credits, the tax on capital was zero or negative, effectively creating a consumption tax. Abrogation of treaties because we adopted accelerated depreciation and an investment tax credit was not an issue. There is no reason to believe that treaties would be less applicable to the Flat Tax than to the income tax in periods when the tax on capital was zero.
If the Flat Tax qualifies as an income tax under the treaties, the only issue is whether other countries would abrogate their treaties in response to some perceived or real threat from the Flat Tax. It is difficult to determine the response of other nations. There are arguments on both sides and commentators have differed in their speculations.(78) All I can offer here is that if other nations abrogate their treaties, renegotiation would be difficult as there would have to have been some reason for abrogation. The Flat Tax eliminates withholding taxes on foreign investments in the United States. Withholding is the most important leverage we have to induce other nations to sign treaties. To help with renegotiation, the U.S. should consider retaining the withholding tax.
4. U.S. as tax haven.
The decision by other nations to abrogate treaties may depend on whether the U.S. would become a tax haven under the Flat Tax because of its low tax on capital. If the U.S. were a significant tax haven, given the size of the economy and richness of investment opportunities, capital would flow to the U.S. and away from more productive uses, particularly in developing countries, which could create serious consequences. Other countries would be forced to abandon their taxes on capital, forcing the world into less progressive taxation and limiting other countries' choices for financing their governments.(79) Avi-Yonah, consistent with his prediction on treaties, predicts that the U.S. would become a tax haven.(80)
I think this prediction is simply incorrect. The United States previously has had a very low, even negative, tax on capital income, and problems with foreign investors sheltering income in the United States were not sufficient to cause serious international concerns. And, under the prior low capital tax regimes, interest was deductible to the payer while, under the Flat Tax, interest would not be deductible, making it much more difficult for a foreigner to repatriate gains without tax. Moreover, to the extent capital is located in the U.S., there will be an additional demand for dollars, so currency adjustments should eliminate any benefits of investing in the U.S.
The simplification potential of the Flat Tax with respect to international tax rules is significant. For example, the foreign tax credit rules(81) and the anti-deferral rules,(82) both of which are significant elements of foreign tax planning and complexity, could be eliminated. Moreover, the source rules(83) could be substantially simplified as we would not need separate baskets or expense allocation rules. The rules for cross-border nonrecognition exchanges(84) could be eliminated as any movement of an asset across the border would be taxable and movements of stock across the border would be irrelevant to the tax base.
6. Summary of international issues.
Despite the complexity of the economic issues, the design considerations for international taxation under the Flat Tax are mostly good news. The Flat Tax would allow substantial simplification of the international tax rules. The major caveat to this conclusion is that the effects of openness are treated as independent problems. The large concentrated pools of international capital will allow taxpayers to take advantage of the problems created by openness more easily than a system that is closed internationally (but open domestically). These problems, therefore, could be classified as international tax problems instead of problems with openness.
E. Small Businesses
This Part considers the treatment of small businesses. Hall and Rabushka have no definition of a business (other than the useless statement that each sole proprietorship, partnership, and corporation constitutes a business),(85) and no explicit exception for small businesses. Discussion in their book indicates that the definition of a business is intended to be broad.(86) There is no small business exception under current law--small businesses, while subject to many simplifying rules, must file returns and pay taxes like any other business. Nevertheless, despite the lack of small business exceptions in the Hall and Rabushka outline and in current law, there are good reasons for having a small business exemption in the Flat Tax. And like the other issues discussed, the openness of the Flat Tax is central to the taxation of small businesses.
To motivate the problem, consider an example given by Alan Feld.(87) A taxpayer owns a home and uses it as a personal residence. This is a durable, nonproductive asset currently used jointly for consumption and investment. It would, under the Flat Tax, be taxed under the yield-exemption method and avoid the transition tax and the tax on any inframarginal return. Suppose the taxpayer decides to rent out a room for a six month period while continuing to live in the rest of the house.
If the rental were treated as a business, we would, as noted above, treat the formation of the rental business as the purchase of an asset by a new business, the rental business, which would get a deduction for the fair market value of the room. The business would be taxed on the rents, and, at the end of the six month period, be treated as selling the room back to the original owner. This system is complex and, most importantly, requires valuation. The homeowner would have to value the room on the date of the formation of the business and on the date of the liquidation of the business.
The valuation requirements probably make this treatment infeasible for many and subject to significant abuse by the aggressive. An initial reaction, therefore, is that the example illustrates an intolerable situation. The task is to consider the costs and benefits more explicitly. Note that much of the problem is caused by the openness of the Flat Tax. If the Flat Tax were closed, no deduction would be allowed on the formation of a business by an individual as the business would be acquiring the assets from a non-taxpayer. For this reason, most European VATs have successful small business exceptions.(88)
There are four reasons for having a small business exception and several countervailing factors. First, as indicated in the rental example, taxing all small businesses would impose significant valuation problems. Assets would be overvalued on contribution, creating large deductions, and undervalued on distribution, creating small inclusions. (This is simply another version of the transfer pricing problem mentioned above.) The government would find it difficult to challenge these valuations given both the large number of transactions at issue and the difficulty of challenging any individual transaction.
Second, allocation of costs between personal consumption and the business would be difficult if small businesses are taxed. In the Feld example,(89) cutting the lawn will be partially a business activity and partially consumption, and the costs would have to be allocated between the two, as would furnace repairs, shoveling snow, and paying property taxes. There are strong incentives to allocate consumption costs to businesses as doing so generates a tax deduction.
This problem exists under the current income tax but it might be worse under the Flat Tax. The reason is that there is a sharper distinction between businesses and individuals under the Flat Tax than under current law. The Flat Tax eliminates all business-related deductions by individuals, such as deductions for unreimbursed employee expenses. But the same expenses incurred by an independent contractor taxed as a business would be deductible. There will, therefore, be a strong incentive under the Flat Tax to be classified as an independent contractor. Under the current income tax, the two percent floor on miscellaneous itemized deductions creates this same dichotomy,(90) but for larger expenses, independent contractors and employees are treated similarly. A small business exception would reduce the problem as there would be no advantage to allocating costs to an exempt business.
Third, the administrative costs of taxing small businesses are substantially higher than for other businesses. For example, a New Zealand study reported that, on average, firms with under approximately $16,000 (U.S.) gross receipts spent 500 times as much (as a percentage of sales) to comply with the New Zealand VAT as businesses with over $26 million in receipts.(91) The United States General Accounting Office reports that the government's administrative costs can be substantially reduced by exempting small businesses.(92)
Finally, a small business exception will not cost very much because the vast majority of sales are by larger businesses. For example, the GAO reports in the same study that in the United States, when sole proprietors and farmers are included as businesses (which they would be under the Flat Tax), 0.4 percent of the business income tax returns account for seventy percent of the income tax revenue. The Japanese consumption tax exempts sixty percent of businesses from tax, but sales by these businesses account for only two to three percent of total domestic taxable sales.(93) Moreover, these figures may overestimate the revenue loss because a small business exception will reduce taxpayers' ability to treat consumption expenditures as business expenses. A small business exception may significantly reduce the complexity of the Flat Tax while not losing very much revenue.
These four considerations create a strong argument for a small business exception. Note, however, that only the valuation problem is unique to the Flat Tax--under the current income tax or under a credit-invoice VAT, formation of a business is nontaxable so no valuation is necessary. The administrative benefits, low revenue cost, and the benefit of simplifying the allocation of personal and business items are all part of current law (although there are some additional advantages to being classified as independent contractors under the Flat Tax). Nevertheless, these problems are sufficiently difficult that, combined with the valuation problem, some type of small business exception may be warranted.
There are some countervailing factors, however. If small businesses were exempt, business owners would have incentives not to pay wages above the personal allowance as any wages above the personal allowance would be taxable. If, instead, the business paid wages to the owner up to the personal allowance and paid all other earnings as dividends, the business and owner together would pay no taxes. While the Flat Tax could impose reasonable compensation rules, a sizable small business exemption would realistically exempt a good portion of the wages of small business owners. Exempt small businesses could also provide employees with tax-free fringe benefits which would be difficult to challenge through a reasonable compensation claim.
To the extent a reasonable compensation rule cannot be adequately enforced, a generous small business exception may make the independent contractor-employee distinction worse rather than better, as suggested above. Many employees would seek to become independent contractors exempt from tax as small businesses and pay themselves below-market wages. This could be policed through the definition of independent contractor, but like policing a reasonable compensation rule, doing so would be difficult.(94)
In addition, a small business exception might make valuation problems worse. Without any small business exception (or a very small one), taxpayers will be able to misvalue assets like the rented room in the above example to reduce taxes. But an expansive exception for small businesses may exacerbate the valuation problem rather than reduce it because valuation will be required when the business crosses the now higher threshold. That is, for very small businesses, the range of possible valuations will be relatively small, As the business gets larger, it may more legitimately claim very large valuations of its assets, particularly as the business develops intangible assets--think of a start-up biotechnology company that can claim to have the cure for the common cold, or even a one-in-a-thousand chance of a cure. It will claim a huge deduction up front and corresponding receipts will never be taxed. Moreover, the business could eventually be sold to a profit-making business that can use the losses generated by the huge up-front deduction.(95)
The only obvious conclusion from the discussion is that there is no simple way out. Having no small business exception would impose large administrative costs, particularly when one considers examples like Feld's rental of a room. A large small business exception creates other problems, particularly the need to value assets when the threshold is crossed. A moderate-sized exception with complex limitations to prevent abuse may be the best we can do, but further searching for a solution is needed.
IV. ADDITIONAL DESIGN ISSUES
This Part will briefly consider ten additional design issues to get a sense of the overall complexity and administrative costs of the Flat Tax. These ten issues (and the five considered above) are only a sampling of the issues that would have to be covered in a complete version of the Flat lax.
A. The Distinction Between Investment and Consumption
The difference between investment and consumption is basic to both the income tax and the consumption tax. In a cash-flow consumption tax, investments are deducted while consumption is not. In an income tax, investments receive basis or an immediate deduction while consumption does not. But both taxes must distinguish between investment and consumption, and there are few reasons to believe it would be easier under the Flat Tax than under current law.
One advantage of the Flat Tax is that it taxes individuals on a yield-exempt basis. This means there is no difference between investment and consumption for individuals acting purely in their individual capacity. For example, whether a work of art is investment or consumption would not matter. Whether the purchase of a house is investment or consumption (or some of both) would not matter. While this seems to offer some simplification, most of the issues under an income tax would be the same under the Flat Tax because most issues involve a business. Consider three typical problems involving mixed consumption and business expenditures: (i) fringe benefits; (ii) home offices; and (iii) personal expenses related to the production of income, such as the costs of commuting or child care.
1. Fringe benefits.
Fringe benefits raise difficult issues under current law because they combine elements of business expense and personal consumption. These problems will be the same under the Flat Tax.
Consider an employer that purchases a car and allows an employee to use it without restriction. Under an income tax, if the car is not taxed as a fringe benefit but is deducted by the employer, the income (and the consumption created by the income) escapes taxation. For example, if the company earns a $100 profit selling widgets and uses the profit to compensate the employee with a car, some of the profits go untaxed. The business would have to pay an immediate tax on the $100 but could depreciate the car, so the difference between $100 and the present value of the depreciation deductions gets taxed, but the remainder of the $100 of earnings goes untaxed. Under a cash-flow tax, the business would get an immediate deduction for the car, meaning the entire $100 would go untaxed. This distinction makes the Flat Tax worse than the current tax, though not significantly so.
Hall and Rabushka would deny businesses deductions for fringe benefits.(96) This approach to fringe benefits is not limited to the Flat Tax and similar approaches have been proposed for the income tax.(97) That is, there is no reason to believe that the solution to fringe benefit problems is any easier under the Flat Tax than under current law.
Moreover, Hall and Rabushka's simple statement of their proposal for fringe benefits conceals the complex nature of the problem. It would be very complicated to define fringe benefits. We would have to determine which meals eaten on the job are fringe benefits and which are not. When is travel a fringe benefit? Would a fancy room at the Royal Hawaiian Hotel, provided free to the manager, or meal vouchers for policemen be fringe benefits?(98) What about parking spaces, company-provided gyms, and discounts at the company store?(99) What about a corner office with expensive art, fancy furniture, and a secretary to do your bidding? Current law has struggled with all of these problems. Many non-cash benefits involve mixed consumption and business motivations, and a rule denying a deduction for fringe benefits does not reduce or change in any significant way the problems with distinguishing the two elements. While some fringe benefits, such as health care, are easy to identify, the complexity of current law stems from difficult classification problems, all of which would be present in the Flat Tax. Record-keeping and classification rules such as those found in multiple, lengthy, and complex parts of current law would be necessary.(100)
Denying the deduction would also, in many cases, overtax the benefits because many fringe benefits have some business element. The appropriate treatment of a fringe benefit that has both compensatory and productive elements is to tax the compensatory element but not the productive element.(101) For example, under current law, we only deny fifty percent of the expense for meals and entertainment on the theory that some element of these expenses is for business purposes. To the extent that the Flat Tax would overtax fringe benefits, it would introduce inefficiencies in the opposite direction from those of current law.
In any event, there is no reason to believe that fringe benefit taxation would be any easier under the Flat Tax than under current law. The incentives would remain about the same as would the complexity of the issues and transactions.
2. Home offices--claiming business deductions for personal expenses.
Claiming business deductions for personal expenses is a common method of deducting consumption expenses under current law. Absent a small business exception, the incentive would be the same under the Flat Tax (and maybe even stronger as businesses can fully expense all purchases).
One of the most common examples is home offices. Under current law, draconian rules are needed to prevent abuse of home office deductions. Under the Flat Tax, no deduction is allowed for any business costs of employees so only those claiming to run their own business from their home would be able to claim the home office deduction. But solving the problem by flatly denying deductions for employees does not help that much. Independent contractors are the guts of the problem and the problem would be identical under the Flat Tax and the current income tax.
The only reason the Flat Tax might be simpler is if, as suggested above, a small business exception is adopted. Then individuals would be unable to claim a home office deduction until the business exceeded the threshold for taxation. If the threshold is reasonably large, many cases would disappear. This would still leave a category of sole proprietors with reasonably sized businesses who could claim home office deductions, but the category would likely be smaller than under current law. (Note that adopting a closed version of the Flat Tax would solve this problem as well.) Of course, the income tax could very well solve the problem the same way by exempting small businesses.
3. Mixed expenses.
Individuals incur a variety of expenses that have both business and consumption elements, such as the costs of commuting, work clothing, and child care. These costs are associated with labor income which makes accurately measuring labor income difficult.
The Flat Tax proposal would deny all of these deductions presumably because these costs reflect consumption choices. This approach is similar to that of current law, although current law does offer some exceptions, such as the nontaxation of employer-provided subsidies for public transportation or parking(102) and the child care credit.(103) The problems with properly measuring labor income under the Flat Tax should be the same as under the current income tax and there is no reason to expect that the implementation would be or should be any different.
The Flat Tax will need rules similar to current law to distinguish consumption from business costs. A small business exception would eliminate some problems, and some of the Hall and Rabushka proposals would be an improvement to current law. To the extent Flat Tax proposals simplify these rules, however, the simplifications would work for current law. Design choices here are readily accessible by looking to current law or the large number of articles discussing changes. There is nothing special about the Flat Tax that makes these problems easier or more difficult.
B. Independent Contractors
As mentioned several times above, the Flat Tax will have to distinguish between independent contractors and employees. The distinction is extremely problematic under current law. The IRS uses a twenty factor test(104) that is frequently at odds with taxpayers' asserted classifications and there are frequent significant disputes. These disputes and something like the twenty factor test will continue under the Flat Tax.
The effect of the distinction between independent contractors and employees depends on whether there is a small business exception. Suppose there is no small business exception. In this case, independent contractors will have to file business tax returns calculated on a cash-flow basis, subject to the various business tax, financial products, accounting, and other rules discussed elsewhere in this article. The separate filing requirement, and the resulting complexity, will be a surprise to many who believed they could use the postcard return. On the other hand, many will attempt to structure relationships as independent contractor relationships, happy to file the extra return for the tax benefits. For example, costs that are deductible to independent contractors would not be deductible to an employee. If, on the other hand, there is a small business exception, many employees will seek to structure their relationships as independent contractor relationships. The most important reason is that independent contractors can pay themselves below-market wages, taking the remainder of their true wages as return on equity.
Regardless of whether there is a small business exception, therefore, we can fully expect significant controversy, complexity, and litigation over this distinction. The level of controversy and complexity should be as great or greater than that of current law.
C. Pensions and Deferred Compensation
The treatment of pensions under the Flat Tax has already been the subject of preliminary analysis.(105) This Part will briefly discuss some of the conclusions.
Under the Flat Tax, investment income is generally exempt from tax, either under the cash-flow mechanism or the yield-exempt mechanism. Pension income in a qualified plan under current law is subject to the cash-flow mechanism. Employers deduct the contribution to the pension like the payment of any other wages. Employees have no immediate inclusion, which can be understood as an immediate inclusion followed by a deduction for the pension investment. When the employee withdraws the money, the employee is taxed. Effectively, pensions are taxed under the cash-flow method. Therefore, the general treatment of investments under the Flat Tax is the same as the treatment of pensions under current law.(106)
Current law, however, imposes a number of restrictions on cash-flow treatment. For example, current law imposes nondiscrimination requirements, which prevent employers from offering pensions only to highly compensated employees. In addition, current law has withdrawal restrictions and funding requirements. All investments under the Flat Tax receive cash-flow treatment, so as Dan Halperin and Michael Graetz point out, there will be an incentive to avoid pensions under the Flat Tax if they are subject to these requirements.(107) Effectively, under the Flat Tax, the law could not impose any significant requirements on pensions except perhaps to the extent the pensions provide some market benefit above and beyond private savings. Requirements more costly than such benefits would simply drive savings out of pension plans.
The Flat Tax removes many but not all requirements for pensions. Qualified plans would no longer need to satisfy the nondiscrimination requirements, benefit limits, and the restrictions on the timing of distribution. The employee protection requirements of ERISA are retained, however, including standards as to eligibility, vesting, funding, and fiduciary rules. The empirical question is the effect of retaining these rules in a world where most of the benefits of a qualified plan can be achieved outside the plan and the general conclusion is that the reduction in pension coverage would be nontrivial.(108)
While the economic and policy issues associated with the change in pension rules are significant, there are few surprising implementation issues. To the extent the Flat Tax imposes requirements in qualified plans, the rules will have to be implemented, and one imagines that they would look like those of current law. Given the reduction in requirements from current law, it is clear that the Flat Tax will be simpler than current law. Nevertheless, the remaining pension rules will create significant complexity.
D. Death and the Estate Tax
Current law allows taxpayers to step up the basis of any assets held at death to their fair market value.(109) The Flat Tax has no particular rules about death, but it does not need to. The yield on investments is exempt, so a rule exempting gain on death would have no effect.
Hall and Rabushka would eliminate the estate tax.(110) This raises many economic and fairness issues. From an implementation point of view, however, elimination of the estate tax would be an enormous benefit. The estate tax, of course, could be eliminated under the current tax structure, so that the benefits from eliminating the estate tax should be viewed as independent of the decision to adopt the Flat Tax.
E. Tax-Exempt Entities
Hall and Rabushka retain tax exemption for a category of entities that roughly corresponds to charities under current law.(111) Other tax-exempt entities under current law, such as labor unions and trade associations, would be subject to the business tax. Employees of tax-exempt entities under the Flat Tax would be subject to the wage tax like any other employee. The entity would pay a special tax on fringe benefits to mimic the nondeductibility of fringe benefits for taxable entities. Contributions to tax-exempt entities would not be deductible, unlike under current law.
The benefit of tax exemption under the Flat Tax will often be lower than under current law. The marginal return to capital is not taxed under the Flat Tax so the business tax only taxes inframarginal returns and transition capital. Exemption under the Flat Tax, therefore, means exemption from these taxes. Some tax-exempt entities, such as hospitals and educational institutions, may have substantial operating assets and exemption from the transition tax would be valuable for these entities. Other entities (as well as educational institutions and hospitals) may have large endowments. These endowments, however, are generally invested in taxable businesses and the transition tax will be paid on these investments at the business level.
Because the benefit of exemption is less under the Flat Tax, many of the details of the existing tax-exempt regime will be less important. Nevertheless, entities will care about exemption and many of the current rules will be needed.
For example, the need to classify entities as exempt or not means detailed rules for classification will be needed. Hall and Rabushka define exempt entities as "[e]ducational, religious, charitable, philanthropic, cultural, and community service organizations that do not return income to individual and corporate owners."(112) Each of these terms will need a definition and the notion of returning income to owners will need substantial clarification. Although the concept of an owner of a nonprofit is not clear, I assume Hall and Rabushka mean to impose some sort of private inurement rules like those of current law.(113)
The taxation of non-exempt but nonprofit entities such as labor unions and trade associations is not clear. The most likely treatment would be for a labor union or a trade association to be treated as selling services to its members in return for dues. Dues would then be taxable receipts offset by the cost of services provided. Any net receipts retained by the union or association would be taxable. If in any year dues exceed expenses, the union or association could face tax liability.
Suppose a tax-exempt entity loses its tax-exempt status, either because it does not meet one of the required purposes or there is private inurement. Its purchases would have occurred in earlier years so no deduction would, without a special rule, be available for the purchases, but any receipts would be fully taxable. Effectively, there would be a one-time tax on all its capital. This is a severe penalty to pay. Some sort of lesser sanction would be necessary. One possibility is to treat the cessation of tax exemption as the formation of a business, which would mean the business could deduct the fair market value of its assets. This, however, would require valuation. Another alternative is some sort of intermediate sanctions regime under which entities could be penalized without losing their exemption.
Suppose the entity basically retains its tax exempt purpose but runs a candy store on the side. Under current law, the profits of the candy store are subject to the unrelated business income tax (hereinafter "UBIT").(114) Would UBIT be necessary under the Flat Tax? There would seem to be no reason to exempt assets of unrelated businesses held by tax-exempt entities from the transition tax or the tax on inframarginal returns. Therefore, as Hall and Rabushka acknowledge, the UBIT rules would be needed, including the rules classifying activities as related or not.(115)
The private foundation rules would probably not be necessary under the Flat Tax (unless the estate tax were retained). There would be no tax advantage to forming a tax-exempt entity to control funds because direct control by an individual would be tax-exempt and there would be no deduction on contribution. Even on transition there would be little or no benefit because a private foundation is likely to have its assets invested in taxable businesses which would be subject to the transition tax. If, however, a deduction is allowed for charitable contributions, the private foundation rules might be needed.
F. Financial Intermediaries
The treatment of financial intermediaries under a consumption tax, including the Flat Tax, has been the subject of extensive analysis.(116) The general problem is that services such as free checking or intermediation are not priced separately from the lending of money--the prices of these services are built into the interest rate. Interest, however, is exempt under the Flat Tax, making it difficult to capture the consumption element of these services. Observers have noted that the rules for financial intermediaries are among the most complex rules in a typical VAT.(117)
The Flat Tax does not create problems for financial services different from those created generally by a VAT. In fact, the Flat Tax may have some advantages over a VAT in this regard. The reason is that VATs commonly exempt the services provided by financial intermediaries. If the Flat Tax had such an exemption, it would not cover the value added by employees of the intermediary as they are explicitly taxed on wages. Thus, failure to capture the value of financial services provided by intermediaries is of less consequence in the Flat Tax.
Hall and Rabushka provide a regime that attempts to capture the value of financial services. They would require banks and insurance companies to report the price of the services they provide to depositors, measured as the difference between the market interest rate and the lower rate that the bank pays on accounts that have bundled services.(118) Similarly, the service element in mortgage interest, in the form of an interest rate higher than the market rate, would be added to the tax base of a bank. This regime will be very complex to implement.
David Bradford points out that bundled financial services are not taxed under current law.(119) It is not clear, therefore, that it is worth the complexity to impose the valuation regime proposed by Hall and Rabushka. In addition, there is at least some argument that financial services are not generally consumption goods and, therefore, should not be subject to tax.(120) Without a special regime, however, financial institutions will consistently generate losses--they will have no taxable income but will be able to deduct their expenses, such as wages. There could be strong incentives to transfer these losses which could create large inefficiencies.
In any event, the most that can be said here is that the VAT rules applicable to financial services are among the most complex in the entire VAT system. There is no reason to believe that the Flat Tax rules would be substantially simpler.
G. Low-Income Taxpayers and the Earned Income Credit
Hall and Rabushka do not include the earned income credit (hereinafter "EIC") in their outline of the Flat Tax. Elimination of the EIC means elimination of a significant poverty assistance program. One in five American families now collects the EIC.(121) In addition, without the EIC, the Flat Tax is likely to be significantly less progressive than current law, but the Flat Tax with the EIC may be a reasonable facsimile of current progressivity (for all but for the highest income taxpayer).(122) Pressure to maintain some version of the EIC may be strong.
The current EIC is generally based on wages but is phased out in part based on overall income. The Flat Tax, however, does not require taxpayers to retain records or determine income other than wage income. Implementation of the EIC, therefore, faces two choices.
First, the EIC could be based solely on wage income. This would significantly simplify the EIC. Adding only one or two lines to the postcard return would allow such an EIC to be included in the Flat Tax. The problem with this approach is that Congress has never thought it appropriate to base the EIC solely on wages because those living off of investments with low wages could claim the credit notwithstanding ample resources. Since its inception, the EIC has included a provision to prevent those with ample resources from claiming the credit and these rules have recently been strengthened.(123) It is not clear that Congress would be any more willing to base the EIC solely on wages under the Flat Tax that it is under current law.
The second option for the EIC under the Flat Tax is to require some level of income computations for all those claiming the EIC. This computation could be relatively simple, such as a net worth test or a realized income test. The problem is that doing so would effectively put those who claim the EIC, at least in part, back in an income tax system. Their net payments to or from the government would depend in part on income. To the extent there are benefits to a consumption tax for low wage individuals, this second option would reduce these benefits by partly putting them on an income tax system. In addition, this option would increase record-keeping requirements considerably for many individuals.
Moving the EIC out of the tax system to the welfare system does not change the analysis at all. A different agency and different bureaucrats would administer the system, but the implementation and economic issues would remain the same. Combining the EIC with existing welfare programs might reduce costs, but this decision can be made notwithstanding the adoption of the Flat Tax.
H. State and Local Governments
The most significant issue facing state and local governments would be the elimination of their ability to base their tax systems on the federal income tax. Currently, many states "piggyback" their systems on the federal system, greatly simplifying administrative and compliance costs.
Unless states switched to a base similar to the Flat Tax, few of the implementation benefits of the Flat Tax would be achieved. Taxpayers would have to compute their income for state tax purposes and their Flat Tax liability for federal purposes. In fact, subjecting taxpayers to both systems would likely increase implementation costs. Therefore reduction in implementation costs requires a change in state tax laws.
All interest is exempt under the Flat Tax. Therefore, the rules for tax-exempt bonds under current law could be eliminated. This would be a great simplification but it would have the effect of eliminating the preference for state and local bonds. Retention .of the preference would likely require rules similar to those of current law.
State and local governments would presumably have to be subject to the special fringe benefits tax on fringes provided to their employees. This raises comity issues well beyond the scope of this article.
I. Filing Unit
Problems with the filing unit under the Flat Tax should be similar to those under current law. Current law compromises between three goals: progressivity, taxing married couples with the same total income the same, and not changing individuals' tax situations upon marriage. These three principles are mathematically incompatible. Under the joint filing system of current law, the third principle is compromised because tax liability may go up (in the case of equal earners) or go down (in the case of unequal earners) upon marriage.
The Flat Tax should have the identical problem with the only difference being that it is generally less progressive than current law. Thus, the Flat Tax will either have marriage penalties or bonuses or it will tax equal earning couples unequally.
The treatment of children is less problematic under the Flat Tax than under current law. Current law includes the so-called "kiddie tax" which taxes children at their parents' rates. The most important reason for the kiddie tax is to prevent parents from nominally giving capital income to their children to take advantage of lower tax rates. In the Flat Tax, capital income is not generally taxed, so the kiddie tax will not be needed (so long as assignment of wage income is adequately policed).
Hall and Rabushka propose no transition relief on the change to the Flat Tax. Businesses would be subject to the cash-flow tax without regard to existing tax basis. Having no transition relief, while raising political and economic issues, would seem to greatly simplify implementation.
Even with no transition relief, however, there will be opportunities for taxpayers to avoid the transition tax and antiavoidance rules will be complex. The transition tax is effectively the business tax. Any avoidance of the business tax avoids the transition tax. Therefore, many of the rules concerning financial products, accounting methods, and the business tax can be thought of as transition tax rules. Rather than complex rules providing relief, they are complex rules to collect the tax.
Taxpayers will also attempt to recover existing basis explicitly. One basic strategy would be to sell assets to non-taxpayers immediately prior to the transition. The sold property could then be leased back or, after transition, repurchased. Thus, one would expect substantial activity prior to an announced transition date to avoid the tax. Virtually all of this activity would be inefficient and would lower revenues.
From an implementation perspective, the question is whether any rules should be put in place to prevent or reduce pretransition sales. It is not clear that any such rules would be successful. One option is an explicitly retroactive transition date that reaches back to eliminate basis for some unanticipated prior period. This would be reasonably simple but unpopular. The alternative is to try to police sale-lease-backs and sales followed by repurchases. Any such rules would be extremely complex, although, if sufficient revenue is at stake, they may be worthwhile.
Transition without relief would also have the potential to cause significant dislocation for long-term contracts. The most important set of such contracts are debt instruments with fixed interest rates. Debt instruments can have extremely long terms and, if interest rates are set under a system of deductible/includible interest, they may be uneconomic under the Flat Tax rules. The exact effect will depend on how prices and interest rates adjust on transition.(124)
Finally, there is a strong likelihood of transition relief (aside from rules reducing dislocations from the lack of relief).(125) As an article in Fortune magazine aptly stated, "Were Washington to disallow the deductions [for preenactment basis], every CEO-laden corporate jet in America would commence strafing Capitol Hill."(126) Transition relief would be complex. Ron Pearlman explores the various issues in creating such relief. It is not worth repeating Pearlman's analysis here. It is worth adding, however, that we would end up with a complex system of partial transition relief combined with a complex system (the financial products, accounting, and business tax rules) to collect the remaining portion of the transition tax. The implementation costs would be high, and the tax would be far from lump sum.
V. CONCLUSION: EVALUATION AND COMPARISON TO CURRENT LAW
The preceding analysis has important implications for both the simplicity and the efficiency of the Flat Tax. Most students of the tax law generally had the intuition that once the details of the Flat Tax were spelled out, the claims of extreme simplicity would be discredited. The analysis here confirms this intuition. The Flat Tax cannot be as simple as claimed and still both raise revenue and not create adverse incentives. Many of the implementation issues in the Flat Tax would be extremely complex, and one can expect rules close to the level of detail and complexity of those in current law. The Flat Tax would not come close to living up to the prediction of postcard returns.
One reason that the Flat Tax is complex is simply that the economy is complex. A simple concept such as taxing fringe benefits at the business level, which Hall and Rabushka propose for the Flat Tax, is extremely difficult to implement because, in a complex economy, there are a wide variety of ways that businesses can mix compensation and business expenditures. Similarly, disallowing interest deductions but not deductions for other expenditures, as proposed in the Flat Tax, sounds simple but turns out to be complex because interest can be hidden. Implementation of the current income tax in a complex economy is complex, and implementation of virtually any other tax will be as well.
But even given the complexity of the economy, the Flat Tax is complex. The complexity primarily stems from its openness, both domestically because of the lack of invoices and internationally because of its origin basis. An open tax will have inconsistencies and line-drawing problems that are difficult to eliminate. For example, transfer pricing rules and complex rules for the taxation of financial instruments will be needed because the Flat Tax is open. Similarly, the openness of the Flat Tax forces the treatment of losses to be modified from the usual treatment in consumption taxes, which in turn means a host of business tax rules will be necessary. Small businesses should be particularly concerned about the complexity created by the openness of the Flat Tax. This unique feature of the Flat Tax means that many compliance and complexity problems will be completely new.
The significant complexity found here involved only a discrete set of issues. Brief examination of other issues, such as the treatment of financial institutions, the earned income credit, and tax-exempt entities, indicates that further complexities will arise when the Flat Tax is actually implemented. One should also remember that the Flat Tax considered here was pure--political compromises were not generally considered. Thus, tax benefits for powerful constituencies were not included. Such political compromises impose large compliance costs under the current income tax. There is no reason to believe that these compromises would not be repeated in the Flat Tax and impose similar costs.
The claim of complexity, however, should not be overstated. There are some significant simplifications in the Flat Tax. In particular, the international tax rules can be significantly simplified, primarily because of the territorial base. Elimination of capital gains taxation and the classification issues associated with the capital gains tax are great improvements. Capitalization issues and inventory accounting disappear. These are significant simplifications. Even so, the claims of simplicity by proponents of the Flat Tax are wildly overstated. Overall, one should expect a system that is simpler than current law, but not extremely so.(127)
To the extent that the simplifications of the Flat Tax are valuable, it may be possible to achieve them through tax reforms other than the Flat Tax. For example, the Flat Tax will be substantially more complex than a European-style VAT. A reformed income tax, even one that retains the realization requirement, may also be as simple as the Flat Tax. For example, much of the international simplification in the Flat Tax comes from its territorial system, which could be adopted in an income tax. Similarly, a single-level business tax might reduce many of the complexities and adverse incentives of the current corporate tax. The financial products rules under an income tax probably have greater potential to be coherent than those under the Flat Tax because an income tax will not require the same distinction between interest and other flows. Accounting methods, however, will be more vexing under a realization-based income tax than under the Flat Tax, although they are closer under an income tax to book accounting, which is an advantage. Thus, many but not all of the simplifications of the Flat Tax can be achieved in a reformed income tax.
Perhaps more important than complexity, the efficiency claims for the Flat Tax are undermined by the analysis. The transition tax is thought to be efficient only because it is unavoidable. But avoiding the transition tax will be relatively easy. Similarly, the Flat Tax is thought to reduce the distortions of current law created by incentives to structure transactions to avoid tax, but these incentives would remain. For example, the Flat Tax will create incentives to structure service relationships as independent contractor relationships, or to pay wages in the form of returns to capital, or to create offshore subsidiaries, and so on. One can expect the Flat Tax to create significant adverse incentives on businesses. The efficiency claims for the Flat Tax, therefore, are substantially weaker than previously thought.
Without the claim of simplicity, and with the claims of efficiency correspondingly reduced, the case for the Flat Tax becomes extremely weak. A reformed income tax, or a combination of an income tax and a VAT, can probably achieve virtually all of the efficiency benefits of the Flat Tax while retaining the progressivity of current law. At a minimum, advocates for the Flat Tax should be required to demonstrate that its claimed advantages are real and cannot be achieved through a more straightforward method.
(1.) The term "flat tax" could mean any number of different things. Most generically, one might think it refers to any tax that has a proportional rather than progressive rate structure. As used in this article, Flat Tax (capitalized to indicate the specificity of the reference) refers to the specific proposal set forth by Robert Hall and Alvin Rabushka and advocated by prominent policymakers. See ROBERT E. HALL & ALVIN RABUSHKA, THE FLAT TAX (2d ed. 1995). See also The Freedom and Fairness Restoration Act of 1995, H.R. 2060, 104th Cong. (1995).
(2.) A VAT, or value added tax, is a type of consumption tax used widely throughout the world. See text accompanying notes 21-25 infra for an explanation of value added taxes.
(3.) For example, a central assumption in the efficiency analysis of the Flat Tax is that the transition regime is unavoidable. Only by looking at implementation, however, can it be determined if it is indeed unavoidable. (The conclusion below is that the transition regime will be avoidable.)
(4.) There have been several prior articles discussing design issues in the Flat Tax or similar taxes. The most comprehensive is Richard L. Doernberg, A Workable Flat Rate Consumption Tax, 70 IOWA L. REV. 425 (1985) (evaluating the Hall and Rabushka proposal in light of several characteristics of the current tax system). See also STAFF OF JOINT COMM. ON TAXATION, DISCUSSION OF ISSUES RELATING TO "FLAT" TAX RATE PROPOSALS (Joint Comm. Print 1995) (discussing background of flat tax proposals and several technical aspects relating to implementation); Michael Calegari, Flat Taxes and Effective Tax Planning, 51 NAT'L TAX J. 689 (1998) (discussing tax avoidance strategies that would be successful under a Hall-Rabushka model); Alan L. Feld, Living with the Flat Tax, 48 NAT'L TAX J. 603 (1995) (discussing implementation problems with the Flat Tax); Vernon Hoven, Flat Tax as Seen by a Tax Preparer, 68 TAX NOTES 747 (1995) (addressing the implications of the Armey-Shelby flat tax proposal on tax planning); Ronald A. Pearlman, Fresh from the River Styx: The Achilles' Heels of Tax Reform Proposals, 51 NAT'L TAX J. 569 (1998) (discussing tax planning responses to any major tax policy reform including the Flat Tax). See Alan Schenk, The Business Transfer Tax: The Value Added by Subtraction, 30 TAX NOTES 351 (1986), for a discussion of subtraction-method VATs which, as noted in the text, resemble the Flat Tax.
(5.) See note 8 supra for citations to discussions of the merits of income and consumption taxation.
(6.) For a good introduction to the literature, see ECONOMIC EFFECTS OF FUNDAMENTAL TAX REFORM (Henry J. Aaron & William G. Gale eds., 1996).
(7.) For a discussion of this issue, see Colin S. Diver, The Optimal Precision of Administrative Rules, 93 YALE L.J. 65 (1983); Louis Kaplow, Accuracy, Complexity, and the Income Tax, 14 J.L. ECON & ORG. 61 (1998); Louis Kaplow, A Model of the Optimal Complexity of Legal Rules, 11 J.L. ECON. & ORG. 150 (1995).
(8.) The literature is vast. For basic background on the operation of consumption taxes, see DAVID F. BRADFORD, UNTANGLING THE INCOME TAX (1986); William D. Andrews, A Consumption-Type or Cash Flow Personal Income Tar, 87 HARV. L. REV. 1113 (1974); Joseph Bankman & Thomas Griffith, Is the Debate Between an Income Tax and a Consumption Tax a Debate About Risk? Does it Matter?, 47 TAX L. REV. 377 (1992); Michael J. Graetz, Implementing a Progressive Consumption Tax, 92 HARV. L. REV. 1575 (1979); Alvin C. Warren, Jr., How Much Capital Income Taxed Under an Income Tax is Exempt Under a Cash Flow Tax?, 52 TAX L. REV. 1 (1996); Alvin Warren, Would a Consumption Tax be Fairer Than an Income Tax?, 89 YALE L.J. 1081 (1980). For background on value-added taxes, see, among other books and articles, CHARLES E. MCLURE, JR., THE VALUED-ADDED TAX: KEY TO DEFICIT REDUCTION? (1987). For economic analysis, see CONGRESSIONAL BUDGET OFFICE, THE ECONOMIC EFFECTS OF COMPREHENSIVE TAX REFORM (1997); ECONOMIC EFFECTS OF FUNDAMENTAL TAX REFORM, supra note 6.
(9.) Stated algebraically, a taxpayer earns $x and is subject to tax rate t. The return on investments is i. If the investment of $x is not deductible but the yield is exempt, taxpayer is subject to an immediate tax on $x leaving $x (1 - t) to invest. The taxpayer's position after n periods is:
$x(1 - t) [(1 + i).sup.n]
If the taxpayer can deduct the investment, the taxpayer can invest the full amount and the pre-tax position after n periods is $x [(1 + i).sup.n]. The taxpayer is taxed on the return (and the withdrawal of the investment), leaving the taxpayer with:
$x [(1 + i).sup.n] (1 - t)
These two expressions are equivalent.
The assumptions behind the equivalence are listed in Graetz, supra note 8. Other than the assumptions concerning inframarginal returns, discussed in the text accompanying notes 11-13 infra (i.e., that the taxpayer can immediately invest the tax savings at the same rate of return as the original investment), the most important assumption is that rates are not progressive (or, more narrowly, that the taxpayer stays in the same tax bracket) and stay the same during the term of the investment.
(10.) Note that under a consumption tax, and when the assumptions hold, the present value of the taxpayer's tax liability does not change regardless of when it is paid. Similarly, the present value of the taxpayer's consumption bundle is the same regardless of when he consumes. In this sense, the consumption tax is said not to distort the timing of consumption (and correspondingly, the decision to save). This is one reason scholars have argued for adoption of a consumption tax.
(11.) Note that another way to understand the example is to view the government as demanding a portion of the inframarginal investment (equal to the tax rate). This forces the investor to put less money into the inframarginal investment and more into marginal investments, which reduces the investor's returns. For example, using the numbers above, the investor would be viewed as investing $36 in the inframarginal investment, getting a deduction for this investment and grossing up the deduction for a total investment of $60, $24 of which is really the government's. When the investment is sold, the benefit of the deduction offsets the tax leaving the investor with the inframarginal 50% return on $36 (i.e., $54). The investor, however, has $60 to invest, so the investor invests the other $24 at the market rate of 10%, and deducts the investment and is taxed upon sale.
(12.) The only estimates that I am aware of are in WILLIAM M. GENTRY & R. GLENN HUBBARD, DISTRIBUTIONAL IMPLICATIONS OF A CONSUMPTION TAX, 28-29 tbl.6-4, 35-38 (1997). Unfortunately, Gentry and Hubbard based their conclusions on the fact that the ratio of fair market value to book value is higher for the wealthy. This measure, Tobin's q, may indicate the presence of inframarginal returns but it may also reflect returns to risk. See Joseph Bankman & Barbara H. Fried, Winners and Losers in the Shift to a Consumption Tax, 86 GEO. L.J. 539, 546-65 (1998), for a discussion.
(13.) Note that a pure income tax exempts the return from bearing risk. The intuition is that taxpayers can adjust their portfolios to offset the nominal tax. See Warren, How Much Capital Income Taxed Under an Income Tax is Exempt Under a Cash Flow Tax?, supra note 8. This return is also exempt under the consumption tax. Thus, the only component of the return on capital that is taxed under an income tax but not a consumption tax is the riskless return, making the distinction between the income tax and the consumption tax very small. See Bankman & Griffith, supra note 8, at 378 ("The debate between an income tax and a consumption tax may be restated, in large part, as a debate over the manner in which the two tax bases treat risky investments ..."). Of course, the current income tax is far from this pure system (particularly in its limitations on the use of losses) and its affect on the return to risk bearing is uncertain.
(14.) See, e.g., Jane G. Gravelle, The Flat Tax and Other Proposals: Who will Bear the Tax Burden?, 69 TAX NOTES 1517, 1521 (1995) (arguing that "provid[ing] relief to old capital is inconsistent with the fundamental nature of a consumption tax."). See also ALAN J. AUERBACH & LAURENCE J. KOTLIKOFF, DYNAMIC FISCAL POLICY (1987). The best interpretation of their insistence that a consumption tax imposes a tax on existing wealth is that it is an attempt to impose consistency of language within the community rather than an argument that there is something inherent in the words "consumption tax" that require such a result.
(15.) See DAVID F. BRADFORD, FUNDAMENTAL ISSUES IN CONSUMPTION TAXATION 19-44 (1996), for a discussion of transition effects of switching to a consumption tax.
(16.) See, e.g., AUERBACH & KOTLIKOFF, supra note 14, at 55-87 (discussing the selection of the tax base as an integral aspect of tax reform).
(17.) For example, the transition rules are the reason for much of the complexity of the USA tax. See Louis Kaplow, Recovery of Pre-Enactment Basis Under A Consumption Tax: The USA Tax System, 68 TAX NOTES 1109, 1109 (1995) (arguing that "the rules for recovery of preenactment basis under the USA Tax System's individual tax may be deficient"); Alvin C. Warren, Jr., The Proposal for an "Unlimited Savings Allowance," 68 TAX NOTES 1103 (1995) (suggesting tentative conclusions about the operation of the USA Tax system).
(18.) For a similar argument, see Roger H. Gordon & Jeffrey K. MacKie-Mason, Why Is there Corporate Taxation in a Small Open Economy? The Role of Transfer Pricing and Income Shifting, in THE EFFECTS OF TAXATION ON MULTINATIONAL CORPORATIONS 67 (Martin Feldstein, James R.Hines, Jr. & R. Glenn Hubbard eds., 1994).
(19.) This statement is true only in the most abstract sense on an economy wide basis. Actual transition to a yield-exempt tax might be complicated, for example, because long-term contracts and other relationships might not immediately adjust to the tax. For example, a yield-exempt tax would probably eliminate the deduction for interest for the borrower and not tax interest to the lender. The burden on debt is not changed, but, absent immediate changes to interest rates, the switch would create windfall winners and losers.
(20.) This leaves aside many differences between the two systems. In particular, a cash-flow tax is collected at the individual level so that taxes may be tailored to individual circumstances (for example, by allowing deductions for special consumption purchases such as medical services, or by imposing progressive tax rates). The text only attempts to show rough equivalences.
(21.) The reason that avoidance of a VAT is more difficult is that a VAT collects tax at each stage of production. Avoidance at one stage only eliminates a portion of the tax. Avoidance at the retail level in a retail sales tax eliminates the entire tax. In addition, European VATs are designed so that business purchasers have an incentive to ensure that the seller has complied with the VAT, creating a self-policing mechanism. The Flat Tax would not have such a mechanism.
(22.) See BRADFORD, supra note 15, at 15; U.S. DEP'T OF TREASURY, INTEGRATION OF THE INDIVIDUAL AND CORPORATE TAX SYSTEMS: TAXING BUSINESS INCOME ONCE 39-60 (1992) (describing the comprehensive business income tax).
(23.) As noted above, another possible reason for the business tax is to prevent shifting of wage income into capital income. If only wages were taxed, individuals could have firms retain part of the wage income and pay it out as dividends or capital gains. While this is an important function, the only taxes that are supposed to be collected at the business level are the taxes on transition wealth and on inframarginal returns.
(24.) See THE ECONOMIC EFFECTS OF COMPREHENSIVE TAX REFORM, supra note 8, at 39 tbl.3.
(25.) This depends, of course, on how the Flat Tax is announced. If it were a complete surprise, the transition tax would not be avoidable through shifts in pre-transition holdings because taxpayers would have no time to shift portfolios prior to the effective date. If it is expected prior to the effective date, the tax will, however, be avoidable.
(26.) In VAT terminology, these businesses are known as "zero rated." They are treated as paying tax at a zero rate so that they are taxpayers for purposes of the closed system.
(27.) See MCLURE, supra note 8, at 71-79.
(28.) See HALL & RABUSHKA, supra note 1, at 144-45.
(29.) Small businesses operating at the retail level do not have this problem in a closed system as there is no resale by the purchasing consumer.
(30.) See, e.g., William J. Turnier, Designing an Efficient Value Added Tax, 39 TAX. L. REV. 435,459 (1984) (reporting on the British VAT).
(31.) For an excellent analysis of international aspects of consumption taxes, see generally Harry Grubert & T. Scott Newlon, The International Implications of Consumption Tax Proposals, 48 NAT'L TAX J. 619 (1995) (examining the effects of the Flat Tax on cross-border investment and trade, and international tax administration and compliance).
(32.) See Stephen E. Shay & Victoria P. Summers, Selected International Aspects of Fundamental Tax Reform Proposals, 51 U. MIAMI L. REV. 1029, 1047-48 & n.83 (1997) (describing international VAT rules for cross-border activity). The European Union ("EU") has proposed to make European VATs origin-based with respect to commerce internal to the EU. These VATs, however, would remain on the destination basis for external commerce. The proposal, therefore, would effectively treat the EU as a single country for purposes of the VAT. See Anne Murrath, Harmonizing VAT in the EU: A Progress Report, J. INT'L TAX'N, Aug. 1999, at 37.
(33.) Michael Graetz gives the following example from Professor Alvin Warren of switching from an origin-based to a destination-based tax:
Suppose that the U.S. has an origin-based VAT of 10 percent with no border adjustments, and that a U.S. consumer product which costs $100 to produce will sell for $110, including the tax, whether sold in the U.S. or for export. Assume that a comparable product is produced in country Z and sells for 110Z in the local zed currency. Assume further that the exchange rate between the U.S. dollar and the Z zed is $1 = 1Z. Finally, for simplicity, assume that there are no transportation costs for shipping the products. Under these conditions, consumers in the U.S. and Z will choose between the two products on the assumption that they will sell for identical prices. Consumers in Z have the choice of buying the Z product for 110Z or buying the U.S. product for $110, which will require 110Z. Similarly, U.S. consumers can buy either product for $110. A U.S. producer has the choice of selling in the U.S. market for $110 or exporting for 110Z, which will yield $110. In either case, the U.S. producer would retain $100 after payment of U.S. taxes. What will happen if the U.S. replaces its origin-based VAT with a destination-based VAT that exempts exports and taxes imports? Initially, the Z product appears more expensive to U.S. consumers than the U.S. product, because the Z product will sell for $121 (the old price of $110 plus the new 10 percent tax), whereas the U.S. product will still sell for $110. Similarly, the U.S. product now looks less expensive than the Z product to Z consumers, because the tax rebate means that the U.S. product can now be exported from the U.S. for $100. The U.S. producer might therefore think it has an advantage in Z, where the comparable local product continues to sell for 110Z. Hence it is often argued that a destination-based VAT would stimulate exports, and that an origin-based VAT would not. Now consider what happens when the U.S. and Z consumers start to switch from Z products to U.S. products, because the latter appear to be less expensive. That switch would mean that there would be less demand for the Z currency by U.S. nationals (who are reducing their imports of the Z product) and more demand for the U.S. currency by Z nationals (who are increasing their imports of the U.S. product). Given this change in demand, the value of the dollar will rise relative to the zed until there is no longer any advantage to switching from Z products to U.S. products, given the consumers' preferences relating to matters other than price, which preferences are independent of the tax law. In this simple example, the value of the dollar would rise until $1 could be exchanged for 1.1Z. U.S. consumers would then have the choice between buying the U.S. product for $110 (including the tax) or the Z product for $100 (which would exchange for 110Z) plus the 10 percent tax on imports, for a total of $110. Z consumers would have the choice between buying the Z product for 110Z or the U.S. product for $100, which would require 110Z. A U.S. producer could sell the U.S. product at home for $110 (including $10 in VAT) or abroad for 110Z, which would yield $100, on which no U.S. tax would be due. In either case, the U.S. producer would retain $100 after payment of U.S. taxes. Taking into account the change in exchange rates brought about by the change in relative prices of the U.S. and Z products due to the introduction of border adjustments, the destination-based VAT has no advantage over the origin-based VAT in terms of stimulating exports in this example. One of the U.S. products under discussion exchanges for one of the Z products in both the U.S. and country Z under both taxes. The U.S. producer earns the same amount from a sale at home and a sale abroad under either tax.
Michael J. Graetz, International Aspects of Fundamental Tax Restructuring: Practice or Principle?, 51 U. MIAMI L. REV. 1093, 1101 n.15 (1997).
(34.) See General Agreement on Tariffs and Trade, Oct. 30, 1997, 67 Stat. A-11, T.I.A.S. 1200, 55 U.N.T.S. 194.
(35.) See Shay & Summers, supra note 32, at 1048-49 & 1049 n.89 (explaining that tax rebates on exports of goods under GATT are permitted if tax is an "indirect" tax under the Agreement and export rebate does not exceed amount levied on goods sold for domestic consumption).
(36.) See id. at 1054 ("The flat tax ... is not a tax on value added at the business level, because wages are deductible from the base at that level. Thus, under existing GATT rules, the business tax would clearly be deemed a form of direct, i.e., income, rather than indirect, tax.").
(37.) Note that switching to an R-based tax from the current system will create dislocations even when the total tax on a transaction is the same. For example, under current law, interest is deductible to the borrower and includible by the lender. In an R-based tax, it is neither deductible nor includible. The total tax on a lending transaction remains the same and interest rates can adjust to leave the parties indifferent. But a borrower with an existing, long-term obligation must pay interest based on a rate set under the prior regime. See BRADFORD, supra note 15, at 18.
(38.) See, e.g., I.R.C. [sections] 1092 (outlining tax treatment of straddles).
(39.) See, e.g., I.R.C. [sections] 7872 (specifying treatment of loans with below-market interest rates).
(40.) I.R.C. [sections] 1234A ensures symmetrical treatment of financial contracts regardless of how they are settled.
(41.) Note that the property could be fungible. There is some question of whether the business could use T-bills as the property, or whether, because they are financial instruments, the business would be denied a deduction. But the business could use gold or some other fungible, tradable, financial equivalent. If operating business property were used, the property could be leased back during the term of the loan.
(42.) See, e.g., IRC [subsections] 1272-1288, 7872.
(43.) See IRC [sections] 1274.
(44.) See HALL & RABUSHKA, supra note 1, at 64, 144-45 (section 203 (d) of the proposed statute). See also ERNST & YOUNG, VAT IN EUROPE (Adele Boomsma, Anne Ermel, Jim Somers & Jean-Marc Tirard eds., 1989) (showing that all European VATs are refundable).
(45.) They simply state that "[w]henever the government has a policy of writing checks, clever people abuse the opportunity." HALL & RABUSHKA, supra note 1 at 64. The lack of invoices is the primary reason why the opportunities exist for clever people to take advantage of the system.
(46.) The Japanese consumption tax, however, is open (domestically) and allows refunds. One reason they can allow this is that their tax rate is only five percent, reducing the incentive for evasion. Another reason is that the Japanese have retained their income tax, and there may often be tension between avoiding consumption taxes and creating income tax liability.
(47.) Apparently, the interest rate is determined based only on first-year, average daily yields on three-month Treasury Bills and does not thereafter adjust. See HALL & RABUSHKA, supra note 1, at 14445. It is not clear why it does not adjust, particularly as the rate is a short-term rate. As discussed in the text, this creates administrative problems.
(48.) See I.R.C. [sections] 172.
(49.) Some European VATs zero rate or exempt the sale of ongoing businesses, creating the equivalent of nonrecognition under a subtraction-method tax or income tax. See note 26 supra. As zero rating grants relief only from administrative requirements and not from any tax liability that would otherwise be owed on the transaction, the zero-rating rule creates little controversy. See ERNST & YOUNG, supra note 44, for a detailed discussion of the treatment by each of the European Countries.
(50.) See I.R.C. [sections] 351 (discussing tax treatment of transfers to corporations controlled by transferors).
(51.) One question would be whether there should be a control requirement. The current nonrecognition rule for partnership formations has no control requirement, while the analogue for corporate formations requires the contributors to control the new venture. Compare I.R.C. [sections] 721 (partnerships), with I.R.C. [sections] 351 (corporations). The best (and only?) argument for a control requirement would be that it limits loss trafficking' Without a control requirement, businesses could contribute property to a loss corporation to soak up the losses. While current partnership rules do not require control on formation, the partnership rules do prevent transfers of losses or gains to new partners. See, e.g., I.R.C. [sections] 704(c). Nevertheless, given that the problem with losses under the Flat Tax will be less than under current law, and that the control requirement is not a serious attempt to limit the problem, a control requirement may not be optimal.
(52.) See I.R.C. [subsections] 332, 337 (allowing tax-free liquidations of subsidiaries into corporate parents).
(53.) See ERNST & YOUNG, supra note 44, (describing in detail the rules for the sale of a business in each of the European countries).
(54.) See I.R.C. [sections] 338, 368.
(55.) Spin-offs would be very simple. The distribution of stock would be automatically tax free as the Flat Tax is R-based, making all spin-offs tax free.
One interesting question is how much the design should be based on expensing, that is, on the exact offset of the taxable receipts to the seller and deductible payments by the buyer. As future Congresses could easily amend the expensing rules in the Flat Tax, the design of the system should be sufficiently robust to work in a depreciation regime. If the Flat Tax had depreciation rules, nonrecognition would matter even in the absence of losses. (Once the tax law uses depreciation, timing matters and nonrecognition offers deferral.) Arguably, little would change because, given the taxfree sale of stock for cash, it would be difficult to require the sale of assets to be treated differently. Indeed, elective nonrecognition regimes have long been proposed for the current corporate tax. See, e.g., AMERICAN LAW INSTITUTE, FEDERAL, INCOME TAXATION PROJECT: SUB CHAPTER (2 (1980) (proposing elective regimes).
(56.) See, e.g., I.R.C. [sections] 382 (limiting net operating loss carryforwards and certain built-in losses following ownership change); Treas. Reg. [sections] 1.1502-21T(c) (1998) (rules limiting net operating loss carryovers and carrybacks from separate return years).
(57.) Current law also allows certain groups of controlled corporations to consolidate their taxes, which effectively allows a single filing and allows losses of one member of a consolidated group to be used against gains from another member. Consolidation, effectively, allows businesses to put subsidiary operations in separate entities for nontax reasons but to be treated as a single entity for tax purposes. Prohibition of consolidation would force businesses to operate less efficiently as branches, rather than as separate corporate entities. See Andrew J. Dubroff & John Broadbent, Consolidated Returns: Evolving Single and Separate Entity Themes, 72 TAXES 743 (1994) (describing single entity nature of consolidation).
Consolidation has the same benefits in the Flat Tax: Use of losses is important, so consolidation may be necessary to prevent inefficient internal structures. Given that losses increase each year with interest, the need for consolidation might be less than under current law, but nevertheless, it would still be worthwhile. In addition, if most of the contribution and distribution problems discussed above arise in closely held groups of businesses, consolidation may eliminate many of the problems without the need for special nonrecognition rules. The particular scope of the consolidation rules could be based on current law or loosened somewhat. See I.R.C. [sections] 1504 for current definitions. The various complicated rules for consolidated entities under current law would largely not be needed under the Flat Tax. For example, the stock basis adjustment rules of Treas. Reg. [sections] 1.1502-32 (1997) would not be needed as all stock transactions would be tax-free.
Hall and Rabushka would allow completely elective consolidation but imply that it is only available for "subsidiaries." See HALL & RABUSHKA, supra note 1, at 144 ([sections] 203(c) of the proposed statute giving rules for filing units). Rules similar to those of current law defining controlled groups would probably be necessary. See, e.g., id.
(58.) See William D. Andrews, The Achilles' Heel of the Comprehensive Income Tax, in NEW DIRECTIONS IN FEDERAL TAX POLICY FOR THE 1980S 278, 278-85 (Charls E. Walker & Mark A. Bloomfield eds., 1983) (identifying the shortcomings of a comprehensive income tax regime in dealing with unrealized appreciation).
(59.) This has been pointed out previously by Michael Calegari, supra note 4, at 695-96 (outlining two effective tax postponement strategies).
(60.) See ERNST & YOUNG, supra note 44, for a summary of the European VAT rules.
(61.) The accounting period would matter for wages received by individuals. Deferring a year-end bonus until the next year would defer tax liability for a year. But this option is available under current law, and there is no reason why the Flat Tax should police this more than current law does.
(62.) Note that similar problems might occur if businesses are free to choose their tax years. Then, a sale from one business to another will not produce offsetting income and deduction as the timing of their tax payments may vary. The question is whether this effect, which can go either way (either a present value tax receipt or tax loss), is large enough to distort transactions and to require that all businesses have identical tax years, particularly since requiring shorter accounting periods reduces the problem.
(63.) See HALL & RABUSHKA, supra note 1, at 63.
(64.) See I.R.C. [sections] 448 (limitation on use of cash method of accounting).
(65.) See I.R.C. [sections] 451, 461.
(66.) For example, determining exactly when an amount is fixed and determinable is complex because there may be contingencies, doubts about collectibility, or difficulty estimating the amount. See BORIS I. BITTKER & LAWRENCE LOKKEN, 4 FEDERAL TAXATION OF INCOME, ESTATES AND GIFTS, [sections] 105.5 (2d ed. 1992) for examples of some of the complexities of the accrual method.
(67.) See ERNST & YOUNG, supra note 44 for a description of the timing rules in the various European VATs.
(68.) Moreover, the European rules would create some problems under the Flat Tax and might require some modifications. For example, if the supply of goods or services is the taxable event, prepayments cause problems. Unless interest is imputed on the prepayment, so that the taxable mount is the fair market value of the goods at the time they are taxed, too little income will be included. Including prepayments on the cash method would work better. Consider a business that receives an up-front payment to provide services in the future. If the payment is taxable and the cost of providing the services is deductible, the usual pattern of an initial deduction for costs followed by a receipt for sales is reversed. Nevertheless, an up-front tax followed by future deductions should still lead to a present value zero tax.
(69.) See I.R.C. [subsections] 446(e), 481.
(70.) See, e.g., INDOPCO, INC. v. Commissioner, 503 U.S. 79, 83-88 (1992) (demonstrating the capitalization requirement's embodiment in common law principles).
(71.) See the uniform capitalization rules of I.R.C. [sections] 263A.
(72.) See I.R.C. [sections] 471.
(73.) See I.R.C. [sections] 460.
(74.) See, e.g., Reuven S. Avi-Yonah, The International Implications of Tax-Reform, 69 TAX NOTES 913 (1995).
(75.) See id. at 916.
(76.) Reuven S. Avi-Yonah, From Income to Consumption Tax: Some International Implications, 33 SAN DIEGO L. REV. 1329, 1349 (1996). See also Shay & Summers, supra note 32, at 1074 ("Initially, one wonders why a treaty partner would continue to provide relief from its withholding tax at source if the United States did not tax the income in question.").
(77.) See, e.g., COMMITTEE ON FISCAL AFFAIRS, MODEL TAX CONVENTION ON INCOME AND CAPITAL art. 2 para. 2.
(78.) See, e.g., Grubert & Newlon, supra note 31, at 642 (arguing that other nations will not refuse to enter into treaties with the U.S.); Avi-Yonah, supra note 74, at 921-23 (arguing that other countries will refuse to enter into new treaties or abrogate old treaties if the U.S. adopts the Flat Tax); Shay & Summers, supra note 32, at 1074-76 (arguing that adoption of a consumption tax would place pressure on other nations to terminate their treaties with the U.S.).
(79.) Suppose taxing capital were efficient, fair, or otherwise desirable, and suppose that capital is highly mobile. If each country acts on its own, each country has an incentive to have a lower tax rate on capital than other countries do to attract capital, leading to an undesirably low tax on capital. Only through cooperation can the appropriate tax on capital be achieved. The U.S., by adopting a zero tax on capital, would be failing to cooperate.
(80.) See Avi-Yonah, supra note 74, at 915-18.
(81.) See I.R.C. [subsections] 901-08.
(82.) See I.R.C. [subsections] 951-64 (the controlled foreign corporation rules); I.R.C. [subsections] 1291-1298 (the passive foreign investment company rules).
(83.) See I.R.C. [subsections] 861-865.
(84.) See I.R.C. [sections] 367.
(85.) See HALL & RABUSHKA, supra note 1, at 144.
(86.) For example, their 1995 book has a sample business return in which the business has gross receipts of about $47,000. ROBERT E. HALL & ALVIN RABUSHKA, LOW TAX, SIMPLE TAX, FLAT TAX 43 (1983). In their 1985 book, THE FLAT TAX, supra note 1, at 67, they refer to a landlord (of an apartment building) as a taxable business. Note that there is a relatively easy definitional issue lurking underneath the discussion of small businesses. The Flat Tax needs to ensure that irregular profit-making activity, such as the sale of a home or of used property, is not treated as a business. VATs generally solve this problem by requiring regular and consistent activity and explicitly excluding occasional sales. The Flat Tax will need a similar definition regardless of any small business exception.
(87.) See Feld, supra note 4, at 607.
(88.) VATs around the world almost uniformly have some small business exception, but they vary greatly in size. In Europe, thresholds range from just above $75,000 in the United Kingdom to a little under $1,400 in Denmark. See ERNST & YOUNG, DOING BUSINESS IN DENMARK 52 (1992) (registration limit of DKK 10,000, translated using December 1999 currency rate of about .135 dollars per DKK); PRICE WATERHOUSE, DOING BUSINESS IN THE UNITED KINGDOM 288 (1998) (48,000 [pounds sterling] registration limit translated using a December 1999 currency rate of about 1.6 dollars per pound). The EC recommends an exemption level of about $6,000. See Sixth Council Directive, 77/388/EEC, art. 24, 1977 O.J. (L 145). The Japanese VAT, which is open, has a large exemption, covering businesses with gross receipts of less than about $300,000. See ERNST & YOUNG, DOING BUSINESS IN JAPAN 77 (1991) ([yen] 30 million registration limit). Our federal income tax, however, has no small business exception and businesses of any size must file returns.
(89.) See note 87 supra and accompanying text.
(90.) See I.R.C. [sections] 67.
(91.) See CEDRIC SANDFORD & JOHN HASSELDINE, INSTITUTE OF POLICY STUDIES, VICTORIA UNIVERSITY OF WELLINGTON, THE COMPLIANCE COSTS OF BUSINESS TAXES IN NEW ZEALAND 64-65 (1992).
(92.) U.S. Gen. Acct. Off., Tax Policy: Value Added Tax: Administrative Costs Vary with Complexity and Number of Businesses 3 (1993).
(93.) Alan Schenk, Japanese Consumption Tax After Six Years: A Unique VAT Matures, 69 TAX NOTES 899, 904 (1995).
(94.) Credit-invoice VATs should have this same problem, however, and they almost universally have small business exemptions, although they vary greatly in size. Because the Flat Tax is open, however, a small business exception in the Flat Tax would exempt more businesses than it would in credit-invoice VATs. That is, in a credit-invoice VAT, exemption is a mixed blessing and many businesses elect to be taxable. For retail businesses, exemption eliminates tax on the value added at that level. But for a business selling prior to the retail level, exemption doubles taxes as the business would not receive a credit for the purchase of its inputs and the purchaser of its outputs would not receive a credit for the cost. In an open tax, exemption would eliminate tax at both the retail and wholesale level, which may greater expand the size of any exemption.
(95.) Another problem with a large small business exception is the need for aggregation. If the small business exception is sufficiently large, it will become worthwhile to structure businesses in separate entities to avoid crossing the size threshold. Presumably, some sort of aggregation rule would be needed and any such rule would be complex. The larger the exception, the greater the need for aggregation as it will become useful to larger and larger businesses.
(96.) See HALL & RABUSHKA, supra note 1, at 119-20, 142-43.
(97.) See, e.g., Daniel I. Halperin, Business Deduction for Personal Living Expenses: A Uniform Approach to an Unsolved Problem, 122 U. PA. L. REV. 859 (1974) (asserting that such side benefits, when they provide satisfaction to the consumer, ought to be taxed as income); William A. Klein, The Deductibility of Transportation Expenses of a Combination Business and Pleasure Trip--A Conceptual Analysis, 18 STAN. L. REV. 1099 (1966) (calling for a return to objective-allocative approach to the taxation of joint business-pleasure trips, taxing the beneficiary for those days devoted to the latter).
(98.) See Commissioner v. Kowalski, 434 U.S. 77 (1977) (holding that meal vouchers for police officers are taxable); Benaglia v. Commissioner, 36 B.T.A. 838 (1937) (holding that housing and meals provided by a luxury hotel to its manager did not constitute taxable income).
(99.) See I.R.C. [sections] 132 (detailing the exclusion from gross income for certain fringe benefits).
(100.) See, e.g., I.R.C. [sections] 119 (meals and lodging furnished by an employer); I.R.C. [sections] 132 (fringe benefits generally); I.R.C. [sections] 274 (disallowance of certain entertainment, gift, and travel expenses).
(101.) See Avery Katz & Gregory Mankiw, How Should Fringe Benefits Be Taxed?, 38 NAT'L TAX J. 37 (1985) (developing a theoretical framework for the taxation of fringe benefits).
(102.) See I.R.C. [sections] 132(f) (qualified transportation fringe benefits).
(103.) See I.R.C. [sections] 21.
(104.) See Rev. Rul. 87-41, 1987-1 C.B. 296.
(105.) See TAX REFORM, IMPLICATIONS FOR ECONOMIC SECURITY AND EMPLOYEE BENEFITS (Dallas L. Salisbury ed., 1997) [hereinafter TAX REFORM].
(106.) The one additional complexity is that the pension is likely to invest in a business that itself is subject to the cash-flow mechanism. The effect is a double deduction on investment and a double inclusion on withdrawal. As long as the cash-flow mechanism has a present value of zero, the double deduction, double inclusion has no effect. To the extent there is a positive present value tax, say because of inframarginal returns, pensions would be at a disadvantage.
(107.) See Daniel I. Halperin & Michael J. Graetz, Comprehensive Tax Reform and Employee Benefits.' The Case of Employment-Based Pensions and Health Insurance, in TAX REFORM, supra note 105, at 35, 38.
(108.) See Eric M. Engen & William G. Gale, Comprehensive Tax Reform and the Private Pension System, in TAX REFORM, supra note 105, at 65, 72.
(109.) See I.R.C. [sections] 1014.
(110.) See HALL & RABUSHKA, supra note 1, at 126-27.
(111.) Compare HALL & RABUSHKA, supra note 1, at 145 with IRC [sections] 501(c)(3).
(112.) HALL & RABUSHKA, supra note 1, at 145 ([sections] 301 of proposed statute).
(113.) See, e.g., I.R.C. [sections] 501(c)(4)(B).
(114.) See I.R.C. [subsections] 511-15 (stipulating a tax liability to tax-exempt entities for unrelated businesses).
(115.) See HALL & RABUSHKA, supra note 1, at 126.
(116.) See generally David F. Bradford, Treatment of Financial Services under Income and Consumption Taxes, in ECONOMIC EFFECTS OF FUNDAMENTAL TAX REFORM, supra note 6, at 437 (examining the treatment of financial institutions under various tax reform proposals including the Flat Tax).
(117.) See Peter R. Merrill & Harold Adrion, Treatment of Financial Services under Consumption-Based Tax Systems, 68 TAX NOTES 1496, 1497 (1995).
(118.) See HALL & RABUSHKA, supra note 1, at 73-75.
(119.) See Bradford, supra note 116, at 439.
(120.) See HARRY GRUBERT & JAMES MACKIE, MUST FINANCIAL SERVICES BE TAXED UNDER A CONSUMPTION TAX? (on file with the Stanford Law Review).
(121.) See 141 CONG. REC. S8404-03 (June 14, 1995) (statement of Sen. Roth).
(122.) See William G. Gale, Scott Houser & John Karl Scholz, Distributional Effects of Fundamental Tax Reform, in ECONOMIC EFFECTS OF FUNDAMENTAL TAX REFORM, supra note 6, at 281 (providing a comparison of taxation rates by income percentile under current law, the Flat Tax, and the Flat Tax with an EIC).
(123.) The original EIC had a phase-out based on modified adjusted gross income. See H.R. CONF. REP. NO. 94-120, at 5 (1975) (defending the low income tax allowance on the basis of income). In 1995, Congress added the disqualified income test. See I.R.C. [sections] 32(i).
(124.) See DAVID BRADFORD, FUNDAMENTAL ISSUES IN CONSUMPTION TAXATION 36-37 (1996).
(125.) See Ronald A. Pearlman, Transition Issues in Moving to a Consumption Tax: A Tax Lawyer's Perspective, in ECONOMIC EFFECTS OF FUNDAMENTAL TAX REFORM, supra note 6, at 393, 412 (claiming that "a cold turkey transition ... will be impossible").
(126.) Louis S. Richman, The Flat Tax: It's Hot; It's Now; It Could Change the Way You Live, FORTUNE, June 12, 1995, at 36, 44.
(127.) Quantification of the compliance and administrative costs would be extremely helpful to the analysis. It is difficult, however, even to determine the total compliance costs of current law, although estimates put it at about $75 billion per year, and it is even more difficult to estimate compliance costs under the Flat Tax. See Joel Slemrod, Which is the Simplest Tax System of them All?, in ECONOMIC EFFECTS OF FUNDAMENTAL TAX REFORM, supra note 6, at 355, 367, 375. It seems clear that they would be lower, but the magnitude is uncertain. See id. at 375 (determining that it is impossible to forecast the compliance costs of certain parts of the flat tax).
Slemrod estimated the total compliance costs of the Flat Tax to be about $35 billion per year. See id. at 367, 375. Slemrod very likely underestimated the costs because many if not most of the complexities discussed above were not known at the time Slemrod did his estimates. For example, the mischaracterization of interest, the problems with the loss carryforward rules, and the various business tax rules for formations and liquidations of businesses create the need for a host of rules and expensive tax advice. One would expect that most of these rules would be on the same order of magnitude as current law.
David A. Weisbach, Associate Professor, The University of Chicago Law School. I thank Julie Roin, Dan Halperin, and participants in Georgetown Law Center's Ernst & Young Tax Policy Seminar, the Harvard Seminar on Current Research in Taxation, and the University of Chicago Law and Economics Seminar for comments and Nate Cardin for research assistance.
|Printer friendly Cite/link Email Feedback|
|Author:||Weisbach, David A.|
|Publication:||Stanford Law Review|
|Date:||Feb 1, 2000|
|Previous Article:||Corporate risk analysis: a reckless act?|
|Next Article:||Does the First Amendment bar cancellation of Redskins?|