Featured Articles

Financial Transactions and the Border-Adjusted Cash Flow Tax

Posted on January 11, 2017 by Hariton, David P.

Table of Contents


Imposing a Cash Flow Tax

Introducing Border Adjustments

The Blueprint's Argument to the WTO

Financial Transactions & the Blueprint's Proposal

Financial Transactions: No Cash Flow Approach

So Where Does That Leave Us?

The Larger Question

Deducting Interest Expense



The Blueprint for tax reform that House Republicans introduced six months ago proposes important changes in the way we tax business operations.1 This proposal is thought to have considerable traction as the nation's agenda moves toward major tax reform, partly because it is in line with the general platform under which the Republicans won the recent elections. Yet the Blueprint is still just that -- a high-level framework for tax reform comprising concepts that have yet to be fleshed out with statutory substance. Indeed, the outline of proposals that would govern the taxation of business operations consists of only a few pages and says relatively little about how we would tax a business's financial transactions.

This is not surprising. The treatment of financial transactions is one of the most complex and difficult subjects to deal with under any new tax proposal. Needless to say, it is therefore with some trepidation that I offer my own preliminary thoughts on the matter, as they are doubtless mistaken in part or missing important points (or may simply seem presumptuous). Moreover, my assumptions regarding the intentions of the drafters are based solely on the words that have been published to date. That said, time is short, events are unfolding rapidly, and there does not yet appear to be any extant discussion of how financial transactions might best be treated under a border-adjusted cash flow tax of the sort that the Blueprint proposes. I therefore offer my own preliminary thoughts in the spirit of "it is better to light even a single candle than to curse the darkness" and with confidence that these early observations will be improved upon by my fellow tax professionals as they focus more fully on the subject.

Imposing a Cash Flow Tax

The most significant business proposal in the Blueprint is the so-called cash flow tax on business activity -- that is, the amount paid for a business asset would be deducted upfront, rather than incorporated into a basis for the asset. It follows, of course, that tax would subsequently be imposed on the gross receipts generated by the asset (including proceeds from the ultimate sale of the asset) without any basis offset. In other words, no effort would be made to match the timing of the deduction for the expenditure to the receipts that the expenditure ultimately gives rise to. But in theory, at least, the difference would be strictly a matter of timing.

Let's take a very simple example. Corporation A spends $100x on an asset that generates $10x per annum for five years, and then it sells the asset for $90x. Under current law, Corporation A is entitled to depreciate the $100x basis of the asset at some specified rate (e.g., 5 percent per annum for 20 years). Its $10x per annum of gross receipts is therefore reduced to $5x per annum of "net" income, and the basis of the asset is reduced to $75x by the end of year 5. Likewise, its $90x of proceeds from the sale of the asset at the end of year 5 is compared with the $75x basis of the asset at that time to arrive at $15x of taxable gain from the sale. Total taxable income is $40x ($5x per annum for five years, plus $15x of gain).

Under the proposed cash flow tax, Corporation A would simply deduct $100x upfront for its expenditure on the asset. It would then include $10x per annum in income for five years and $90x in income at the end of year 5. Total taxable income, however, would remain $40x ($50x + $90x - $100x).

Introducing Border Adjustments

The Blueprint's principal rationale for switching to a cash flow tax is that it would better accommodate so-called border adjustments, which constitute the Blueprint's second most significant business tax proposal. This change would not be merely a matter of timing. The idea (I think) is that we would (a) not allow any business deductions for assets purchased from outside the United States, (b) impose tax on the direct sale of goods by foreign persons to nonbusiness consumers in the United States,2 and (c) exclude from taxable receipts the proceeds of any sale of goods and services to foreign persons. The net effect of these adjustments (provided that they could be effectively enforced) would be to impose tax on goods that are consumed in the United States, regardless of where they are produced (in contrast with the current business tax, which effectively imposes tax on goods that are produced in the United States, regardless of where they are consumed). This is also referred to as a destination-based tax, as compared with the current origin-based tax.

Let us return to our simple example above, applying now a border-adjusted cash flow tax. If Corporation A above buys its asset from a foreign producer, it will not be allowed its $100x deduction. It will therefore be subject to tax on $140x (rather than only $40x). Alternatively, if Corporation A buys the asset from a U.S. person but sells it to a foreign person, it will be able to exclude the $90x of sales proceeds from income and will therefore have a $50x net tax loss (rather than $40x of taxable income): a $100x deduction for purchasing the asset, partially offset by $10x of proceeds for five years.

Obviously, such border adjustments would be designed to discourage imports and encourage exports as compared with the current tax system (or as the Blueprint puts it, to allow the United States to achieve parity with our major trading partners, which implement similar border adjustments through their VATs).3 Some economists doubt that the tax would have this effect. In a nutshell, I think their argument runs something like this (understanding that I am not an economist): What goes into the United States (i.e., cash proceeds from the sale of exports) must in the long run come out of the United States (i.e., ultimately be invested in imports), or else we would be supplying goods and services to other countries in exchange for nothing. As a result, (a) the increased inducement to export would be offset by the decreased inducement to import, (b) total trade would therefore remain the same, and (c) the U.S. dollar would merely increase in value (in relation to other currencies) to account for the resulting shift in the incidence of the tax (i.e., for the fact that importers would in effect be bearing the exporters' tax burden).4 But such theories notwithstanding, one suspects that border adjustments would have at least some of their intended effects in the here and now5 -- hence their political appeal.

The preceding introduction was doubtless unsatisfactory in many respects, but I lack both the space and the expertise to elaborate further. Instead I will commend to the interested reader a number of excellent background articles (by professors who not only think deeply but write simply and clearly)6 and return to the subject at hand -- the treatment of financial transactions under the House Republicans' proposed border-adjusted cash flow tax, if it were to be enacted.

The Blueprint's Argument to the WTO

There are two arguments for why a shift to a cash flow tax might help accommodate border adjustments. The first is that it might make them easier to impose. Thus, in the example, we can simply disallow the $100x deduction if Corporation A purchases the asset from a foreign producer. If we sought to apply border adjustments under current law, we would have to keep track of and disallow depreciation deductions in later years and track a future basis disallowance into subsequent sales.

But the second, and more important, argument is that a shift to a cash flow tax would make border adjustments more acceptable to the WTO. In the words of the Blueprint:

  • Because this Blueprint reflects a move toward a cash-flow tax approach for businesses, which reflects a consumption-based tax, the United States will be able to compete on a level playing field by applying border adjustments within the context of our transformed business and corporate tax system. . . . The rules of the World Trade Organization (WTO) include longstanding provisions regarding the use of border adjustments. Under these rules, border adjustments upon export are permitted with respect to consumption-based taxes, which are referred to as indirect taxes. However, under these rules, border adjustments upon export are not permitted with respect to income taxes, which are referred to as direct taxes. . . . With this Blueprint's move toward a consumption-based tax approach, in the form of a cash-flow focused approach for taxing business income, the United States now has the opportunity to incorporate border adjustments in the new tax system consistent with the WTO rules regarding indirect taxes.7

But as we shall shortly see, merely shifting to a cash flow tax (by allowing 100 percent accelerated depreciation) arguably does not serve to convert an income tax into a consumption tax. And the taxation of financial transactions may be the biggest fly in the ointment.

Financial Transactions & the Blueprint's Proposal

As for financial transactions, the Blueprint says the following:

  • Under this Blueprint, job creators will be allowed to deduct interest expense against any interest income, but no current deduction will be allowed for net interest expense. Any net interest expense may be carried forward indefinitely and allowed as a deduction against net interest income in future years. . . . The Committee on Ways and Means will work to develop special rules with respect to interest expense for financial services companies, such as banks, insurance, and leasing, that will take into account the role of interest income and interest expense in their business models.8

We will come back to why a disallowance of interest deductions is normally the right answer under a cash flow tax. The Blueprint adds that the "system of immediate cost recovery will apply to both investments in tangible property (such as equipment and buildings) and intangible assets (such as intellectual property)" but not to land.

Little more is said in the Blueprint about financial transactions. This is not surprising. The taxation of financial transactions is a complex and difficult matter, and many new tax proposals therefore leave the subject for later. But this means that the Blueprint does not yet answer an exceedingly fundamental question: What happens when Corporation A invests $100x in a financial asset, such as a bond?

The first possible answer is nothing, because financial transactions are excluded from the tax base. And this is arguably the only answer that is consistent with a consumption tax. Consumption taxes are designed to be imposed solely on the goods and services that individuals consume, rather than on the income derived by businesses or individuals. If a business invests in a stock and then sells it at a gain (or invests in a bond and receives interest), the business may have income, but it has not produced any good or service that individuals will ultimately consume; thus, no tax is due under a consumption tax.9

This is therefore the answer under any VAT, which is merely a modified sales tax on the consumption of goods and services by individuals. In a properly functioning subtraction-style VAT (meaning that the VAT is imposed by reference to cash flows rather than by means of tax credits), each business buyer along the path from raw materials to finished product receives a deduction for its business purchase that precisely equals the upstream business's taxable inclusion of sales proceeds, so that the government collects net tax only when the finished product is sold to the ultimate nonbusiness consumer (just as in a sales tax). A VAT is not designed to tax businesses along the production path on the gains they may have realized from incidental financial transactions. That is reserved for an income tax. This is therefore also the answer under most previously proposed cash flow taxes, such as the "X tax" proposed by the late David Bradford and the "Flat Tax" proposed by Robert Hall and Alvin Rabushka, because these proposed taxes were likewise intended as consumption taxes.

If excluding financial transactions from the tax base (consistent with a consumption tax) was the Blueprint's answer, it would be easy to understand how the taxation of financial businesses and institutions fit into the system. Financial businesses and institutions would be dealing primarily in transactions that were excluded from the tax base. And so there would be only two choices: First, allow such businesses to pay no tax (which is currently the answer under most VATs); or second, require such businesses to disaggregate from their otherwise exempt financial transactions a variety of embedded financial services that they could be deemed to have sold to their customers separately (and upon which tax could then properly be imposed). Both approaches are rife with problems, but this is familiar territory for any VAT or other consumption tax proposal.10

But all this VAT precedent is irrelevant, for this does not appear to be the answer under the Blueprint. Rather, the Blueprint states that it will not allow businesses any deduction for net interest expense, which appears to imply (by negative inference) that apart from any such excess, interest income and expense will be included in the tax base, as will gains and losses from financial transactions and -- for that matter -- such items as cancellation of indebtedness income.

Moreover, excluding financial transactions from the business tax base would be highly problematic, because the Blueprint does not propose to drop the other shoe and eliminate financial transactions from the individual tax base, thereby converting the income tax into a consumption tax as a whole. This means that individuals who would otherwise be subject to tax on their investments and financial transactions could avoid the tax by holding them through a business. It might be difficult to defend the individual tax base in light of the inconsistency between the two regimes. At the very least, property moving from business to nonbusiness status, or vice versa, would have to be marked to market at some sort of "tax border" between the two.

Assuming, however, that financial transactions are to be included in the business tax base, it would seem harder to explain how the proposed business tax is a consumption tax rather than an income tax. For the mere decision to impose tax by reference to cash flow does not itself turn an income tax into a consumption tax. An income tax will by definition be a cash flow tax if it provides for superaccelerated depreciation, but this is merely a matter of timing. Rather, the key distinction between an income tax and a consumption tax is arguably that the former includes financial gains and losses (as well as other non-production-related gains and losses) in the tax base, whereas a consumption tax excludes them, because a consumption tax is not designed to tax income but rather to tax the ultimate consumption of goods and services by individuals.

I note that the Blueprint's proposal also diverges from a consumption tax by allowing businesses to deduct the wages they pay (so that the value added by labor can be taxed at the individual level at progressive rates). But this innovation had already been introduced into the X tax, the Flat Tax, and other recent consumption tax proposals, and these proposals at least argued that an individual-level tax on wages merely substituted for a business-level tax on labor as a factor input. None of these proposals included strictly financial transactions in the business tax base, and certainly a proposal to tax financial gains looks nothing like a VAT.

In any case, the Blueprint appears to concede that its plan might fall a bit short of a consumption tax, for it states that under its proposal, we would be "moving toward" a consumption tax, as opposed to actually having one. But perhaps the Blueprint's argument to the WTO could nevertheless rest adequately on the following refinement: The proposal would reflect the imposition of a consumption tax on business transactions combined with an income tax on their financial investments. I will elaborate on this shortly.11

Financial Transactions: No Cash Flow Approach

This leads us to the second possible answer to the question of what happens when Corporation A invests $100x in a bond, which is that Corporation A deducts $100x, just as it does when it invests in any other business asset. But if that is the answer, then it must follow that Corporation A includes $100x of gross proceeds in income when the bond matures. In other words, the second possible answer is that the amount of tax imposed on financial investment is determined by cash flow, consistent with the tax imposed on business operations. The Blueprint is arguably ambiguous on this point. On the one hand, it generally talks about "deducting investments," rather than about deducting only investments that would otherwise give rise to depreciation deductions. But on the other hand, the Blueprint provides that there would be no deduction for an investment in land, which is similar to a financial asset.

In any case, I do not think this could reasonably be the answer under the Blueprint, because it doesn't make a lot of practical sense. The cash flows associated with financial assets and liabilities often dwarf those associated with business activity, and these financial cash flows do not give rise to income or loss in any economic sense. Thus, if corporations were entitled to investment deductions every time they made a loan (or other financial investment), we could expect them to seek to zero out their tax liabilities with financial investments rather than write current checks to the government. At the very least, corporations would have to likewise be required to include in income their proceeds from the issuance of bonds (otherwise, corporations would quickly wipe out their tax liability by issuing bonds and investing the proceeds). And if that were the case, then they would have to likewise be able to deduct the repayment of their resulting borrowings. I suspect we would therefore quickly find that a corporation's annual tax liability had little to do with its economic profits.

Moreover, as noted above, the Blueprint's primary rationale for shifting to a cash flow tax is that it better accommodates border adjustments. Yet surely we would not want to border adjust financial transactions. Surely we would not, for example, want to provide a business with an enormous permanent net tax deduction (one that swamped the rest of its tax liability) for "exporting" its bonds to a foreign person and investing the proceeds. Or for buying stocks and securities from U.S. persons and exporting them to foreign persons. Nor would we want to do the opposite and saddle a business with a huge tax liability for "importing" the bonds of foreign persons (by lending money to them) or for importing their stocks or securities. Besides, as noted above, financial transactions can be subject to tax only under an income tax, not a consumption tax. There would therefore be little basis for arguing to the WTO that border adjustments were allowed on strictly financial transactions.

(I have little doubt that if this were the plan, the taxation of financial institutions would be all but incoherent. For they generally earn their net profits from the relatively thin spreads between the prices for which they buy and sell financial assets, and the volume of their offsetting cash flows is vastly disproportionate to the net income that they ultimately derive. Thus, a tax imposed on gross cash flows -- along with permanent gross exclusions or inclusions, depending on the direction in which those cash flows crossed the border -- would give rise to arbitrary results that would have little to do with economic income. But I doubt that this will prove to be the answer for financial institutions, because it doesn't seem to be the right answer even for nonfinancial institutions.)

So Where Does That Leave Us?

This leaves us with the third (and most likely) answer to our question, which is that Corporation A takes a $100x tax basis in the bond, and its gain or loss from the ultimate sale or maturity of the bond is limited to the difference between this tax basis and the amount realized. In other words, if we are not planning to shift entirely to a consumption tax and exclude financial transactions from the tax base for both businesses and individuals, then the only alternative seems to involve leaving financial transactions in the place they are in today, with no upfront deductions for investing in them and no associated border adjustments. This means that there would in effect be two simultaneous taxes on businesses: a border-adjusted cash flow tax on regular productive business activity and a regular income tax on their financial investments.

But this would still not be a walk in the park from the government's perspective. (Likewise, from the practitioner's perspective, there would still be numerous traps for the unwary business that did not carefully rethink the structure of its business and financial transactions from a tax perspective.) For the government would still have to draw a detailed and coherent line between business and financial transactions. On the investment deduction side, this might initially be done by reference to what is depreciable today -- e.g., with reference to such current-law code sections as section 168 (dealing with accelerated depreciation) and section 197 (dealing with the amortization of intangibles). But how would we deal with leasing -- i.e., where a corporation constructs or purchases an asset for use not in its own business, but rather to effectively hold as a financial asset for lease to someone else? More broadly, there are many complex arrangements under which corporations and financial investors share in the risks and benefits of business-related property. A long case law history, culminating in the enactment of section 636, grapples with the question of when rights to various "slices" of the proceeds of production from mines or energy reserves should be treated as business property or as financial loans. Similar line-drawing questions arise in connection with the characterization of specified slices of the royalty proceeds of research and development, or of such quasi-business assets as consumer receivables. And the line between business and financial assets is likewise formalistic and thin when it comes to mergers and acquisitions, or private equity, where an investor can choose to acquire either assets or stock.

The Larger Question

This leads us to what is in my view the larger question when it comes to financial transactions and the Blueprint's proposal: Would we be willing to allow taxpayers to use financial transactions to effectively transfer the timing tax benefit associated with deducting business investments, and the permanent tax benefit associated with exporting goods and services, from one taxpayer to another? For unlike the VAT, and unlike most theoretical proposals for a consumption tax, the Blueprint does not propose to write checks to businesses with net deductions in any given year. Rather, it proposes to make them carry net operating losses forward for use in subsequent tax years.

Presumably the concern here is that writing checks would provide too great an incentive and that some taxpayers might therefore find a way to ring the government cash register in cases when no one else was really paying the price. And if that is the concern, then we might likewise be concerned about leasing and other quasi-financial transactions that are designed (effectively with government consent) to allow taxpayers to monetize tax benefits that they cannot currently use (because they don't have enough taxable income) by transferring them to other businesses that can. Under current law, these tax benefits include various tax credits and deductions for depreciation and interest expense. Under the Blueprint's proposal, however, they would include (instead of scheduled depreciation) the lessor's upfront deduction for its investment in the leased asset, which would be economically equivalent to allowing the lessor to earn tax-free income on the equity capital that it invested in the lease.12

Of even greater concern might be the role of financial intermediaries in cross-border transactions. Consider, for example, the case in which a large taxable corporation sets up a special division to assist other businesses in the export of goods and services. The division purchases big-ticket items, like ships and airplanes (or even oil), and then sells them to foreign buyers. The division deducts the purchase price but excludes the proceeds of the sale. Each relevant U.S. seller tells the division which foreign customer is likely to be interested in the relevant ship or airplane (and how much they are likely willing to pay for it), and the division effectively splits its tax benefits with the U.S. seller by paying the U.S. seller more for the ship or airplane than the foreign customer is willing to pay. Perhaps the answer is that we are fine with this because in an ideal world, we would be writing a check directly to the U.S. seller, so the financial middleman is really just helping us out. But if that is not the answer, it might be especially important to clearly define which businesses will be entitled to export exclusions, and when (as opposed to relying on such vague common law concepts as the "economic substance doctrine").13

I note, moreover, that the theoretical argument for why such transactions are not problematic might in some cases seem rather attenuated. Suppose, for example, that a U.S. person currently owns a building in Paris. Our middleman division buys it (fully deducting the purchase price) and then sells it to a French corporation (excluding the proceeds). The French corporation will not be paying U.S. tax on its rental income (as it would if the building were located in New York). What then is the theory under which the government is recompensed for granting the middleman division such a large net tax deduction? The answer (I think) is that somewhere along the line, a U.S. person must have acquired the building and been denied a deduction for doing so (because the building was then deemed to be "imported").

But will the theoretical answer to questions like this always stand up in the hustle and bustle of everyday practice -- with all of its idiosyncratic fact patterns and all of the complex substantive rules and transition rules that do, or do not, coherently apply to them -- given the fact that taxpayers always control the form of the transaction? I suspect it might take some time for the government to learn how best to defend its revised tax base.

Deducting Interest Expense

We might also want to reconsider, in light of this dual tax, the Blueprint's proposal to disallow net interest deductions. One could argue that the proposal, as currently drafted, is somewhat overbroad. What the Blueprint is saying (quite rightly) is that a corporation that borrows money in order to invest in a fully deductible business asset should not be entitled to deduct the interest paid on the borrowing, since the upfront deduction already gives the corporation a tax benefit that takes account of the time value of money. But there might be more limited ways to determine what portion of the proceeds from borrowing should be treated as investment in fully deductible property.

Meanwhile, suppose that a typical corporation borrows long term at high interest rates and invests the proceeds short term at low interest rates because it needs a ready stock of cash to meet its business exigencies. Or suppose it invests them in financial assets -- such as portfolio stocks or various financial derivative positions -- that don't give rise to current interest income. Is it really right to therefore disallow most of that corporation's interest expense?

Arguably doing so would not be in keeping with some of what the Blueprint embraces, which is a further integration of the corporate income tax.14 The Blueprint states that a substantial lowering of the tax on dividend income would help diminish the double taxation of investment income, presumably to the benefit of Americans, their savings, and the economy. And yet, the de facto conversion of most outstanding deductible corporate debt into the tax equivalent of nondeductible preferred equity would seem to run in the opposite direction. If a corporation is indeed going to be subject to an income tax on its net financial income and gains, then it should likewise be entitled to deduct its net financial expenses and losses (including interest expense), except to the extent that interest expense has been incurred to purchase or carry investments that have been fully deducted upfront under the consumption tax for business transactions.

Disallowance of such interest expense could be dealt with by reference to rules that are currently found under section 265(a)(2), which deals with the disallowance of interest expense incurred to purchase or carry tax-exempt bonds. If this more narrow approach is taken, there would be less concern about providing special rules to deal with the net interest expense of financial institutions and businesses, as they would presumably not be making large investments in fully deductible assets. To the extent that they were making such investments, disallowance of their interest expense could likewise be dealt with by reference to rules that are currently found under section 265(a)(2) of the code.


As a practical matter, the Blueprint's proposal would appear to involve two different taxes on businesses: a regular income tax on financial transactions and a border-adjusted cash flow tax on everything else. This might make it harder for us to argue to the WTO that our border adjustments were permissible because our tax resembled a VAT. This would also require us to carefully draw the line between business and financial transactions. But more fundamentally, it would force us to decide whether we are willing to allow taxpayers to use financial transactions to transfer tax benefits from one taxpayer to another. If the answer is no, then Treasury and the IRS would need to play a very active role in promulgating and enforcing rules designed to stop that from happening.


1 Tax Reform Task Force, "A Better Way: Our Vision for a Confident America" (June 24, 2016) (Blueprint).

2 We could either deny businesses a deduction for the purchase of imported goods or require them (or their sellers) to pay tax on such a purchase (as in the case of the purchase of foreign goods by individuals), but we would presumably not do both.

3 Blueprint, supra note 1, at 28.

See, e.g., Alan D. Viard, "Border Adjustments Won't Promote Competitiveness," Tax Notes, Oct. 4, 2004, p. 122.

5 It occurs to me, for example, that in the shorter run, an increase in proceeds from national exports could be "temporarily" invested in outbound savings (i.e., in foreign stocks, bonds and real estate, or in repaying inbound loans and repurchasing inbound foreign investments in U.S. stock and real estate) rather than in more imports. As John Maynard Keynes so aptly put it, in the long run we are all dead. I'm therefore not sure I understand why today's exchange rates would necessarily reflect offsetting purchases of imports that would have to be made in the distant future.

See Daniel Shaviro, "Replacing the Income Tax With a Progressive Consumption Tax," Tax Notes, Apr. 5, 2004, p. 91; David A. Weisbach, "Does the X-Tax Mark the Spot?" 56 S.M.U. L. Rev. 201 (2003); David F. Bradford, "Blueprint for International Tax Reform," 26 Brook. J. Int'l L. 1449, 1452 (2001); Joseph Bankman and Michael Schler, "Tax Planning Under the Flat Tax," in Taxing Capital Income (2007); and Shaviro, Decoding the U.S. Corporate Income Tax (Feb. 2009).

7 Blueprint, supra note 1, at 28.

8 Blueprint, supra note 1, at 25-26.

9 Likewise, if a business borrows money and pays interest, it doesn't deduct the interest, because that has nothing to do with the goods and services it produces for consumption by individuals. Rather, it deducts the investments it makes with the proceeds of the borrowing, and if it takes a long time for those investments to produce taxable receipts, the delay between the investment deduction and the taxable realization plays the same role as a deduction for interest expense. The Blueprint, supra note 1, makes this point at p. 26.

10 Exemption may actually serve to increase the amount of the tax imposed on the value added by financial institutions (as compared with taxing them directly) to the extent that they render services to other business enterprises for use in their own production of goods and services. This is because such other business enterprises are not allowed to deduct the amounts they pay to exempt financial institutions in exchange for their services (as if they were importing such services from foreigners). But the cost of these services is naturally added to the price that such other business enterprises charge their own customers for their own goods and services, and so the value added by the exempt financial institutions is in effect subject to full tax in the hands of these other business enterprises. And in addition, the financial institutions themselves are not allowed any deductions for the amounts they pay to other businesses for the goods and services that they require; yet these other businesses are required to pay tax on the proceeds of their sales to exempt financial institutions. There may therefore be double taxation on the value that other businesses provide to exempt financial institutions.

Let us take a simple example: Computer Co. sells software to Financial Co. for $100x, Financial Co sells a service to Widget Co. for $120x, and Widget Co. sells a widget to an individual customer for $150x. If Financial Co. is taxable (and if the tax rate is 20 percent), then Computer Co. has $100x of taxable income (and pays $20x of tax), Financial Co. has $20x of taxable income (and pays $4x of tax), and Widget Co. has $30x of taxable income (and pays $6x of tax). The government therefore collects $30x of tax in total (20 percent of $150x). But if Financial Co is exempt from tax, Computer Co. still has $100x of taxable income (and still pays $20x of tax), while Widget Co. has $150x of taxable income (because it cannot deduct its $120x payment to Financial Co.) and therefore pays $30x of tax. The government therefore collects $50x of tax, and one can argue that $20x of this tax should not have been collected because Widget Co. should have been entitled to deduct $100x for the expenditure effectively paid "through" Financial Co. to Computer Co.

Of course, we get the opposite result if Financial Co. sells its service to an individual consumer for $120x (or to a business engaged in investing for its own account). In that case the government collects only $20x from Computer Co. (for its sale to Financial Co. for $100x) and doesn't collect an additional $4x of tax from Financial Co. for the $20x of value that it has added. This may lead to economic distortions wherein domestic financial institutions provide their services primarily to individual consumers, while foreign financial institutions provide services to business consumers at lower prices, because they can (i.e., because they are organized in jurisdictions that have no similar tax and they can therefore buy their own "factor inputs" at lower prices).

Moreover, Financial Co. could eliminate the burden of double taxation by developing its own software (paying, say, $105x to do so), rather than by purchasing it from Computer Co. for $100x. Now the government collects only the $30x of tax arising from Widget Co.'s sale of a widget for $150x. This likewise induces economic distortion, encouraging financial institutions to produce their factor inputs in-house.

11 Note that there is arguably no constitutional impediment to imposing a consumption tax on business transactions (even though it would not be an "income tax" within the meaning of the 16th Amendment) because it would arguably not be a "direct tax" that has to be apportioned among the states under Article I, section 8.

12 To the extent that the lessor funded the lease with borrowings, however, the lessor's timing benefit could be effectively offset by disallowing deductions for the interest expense incurred by the lessor to finance its investment. See the discussion below under Deducting Interest Expense.

13 In a well-known line of cases that gave birth to the current tax treatment of "repos," banks that were otherwise financing the tax-exempt bond inventories of various broker-dealers purported instead to be actually purchasing tax-exempt bonds and selling them directly to the broker-dealers' customers. (Side agreements insulated the banks from the risks of actually owning tax-exempt bonds.) This complex arrangement served to allow the banks to claim the relatively minor tax benefit of earning a little bit of tax-exempt interest. Imagine if it had instead allowed them to deduct their purchase of the bonds yet exclude the proceeds of their sale entirely. American National Bank of Austin v. United States, 421 F.2d 442, 452 (5th Cir. 1970); Unions Planter National Bank of Memphis v. United States, 426 F.2d 115 (6th Cir. 197); First American National Bank of Nashville v. United States, 467 F.2d 1098, 1102 (5th Cir. 1970); Rev. Rul. 79-195, 1979-1 C.B. 177; Rev. Rul. 79-108, 1979-1 C.B. 75; Rev. Rul. 74-27, 1974-1 C.B. 24; and culminating with Nebraska v. Loewenstein, 513 U.S. 123 (1994).

14 Blueprint, supra note 1, at 18.