30 Tax Notes 451

Tax Notes

FEBRUARY 3, 1986

LENGTH: 4483 words

DEPARTMENT: Special Reports (SPR)

CITE: 30 Tax Notes 451

HEADLINE: 30 Tax Notes 451 - ERROR IN THE NAME OF INTEREST.

AUTHOR: Johnson, Calvin H.

TEXT:

Calvin H. Johnson is a Professor of Law at the University of Texas at Austin School of Law.

In this article, Johnson defends legislative proposals to attack tax shelters by restricting the deduction of interest. Johnson argues that interest is a cost, just like any other cost, and that interest should not be deducted unless it is a cost of taxable income. Where the income from a debt-financed asset is not fully taxed, as is the case where the asset is depreciated under ACRS, the deduction of interest is a separate error that misdescribes the transaction. George Mundstock of the University of Miami Law School argued in a previous Tax Notes special report that borrowing to finance ACRS property is no problem because ACRS gives identical benefit whether the property is debt- or equity-financed (see Tax Notes, December 23, 1985, p. 1253). Johnson responds that debt financing is different. Debt financing, but not equity financing, allows cost-free liabilities in which the tax benefits alone are greater than the cost of the asset. If interest must always be deducted, moreover, then it is impossible to tax any source of income once one source becomes tax-exempt.

In 'Accelerated Depreciation and the Interest Deduction: Can Two Rights Really Make a Wrong?,' Tax Notes, December 23, 1985, p. 1253, George Mundstock argues that borrowing to finance ACRS property is not a problem. A debt-financed investment in ACRS property gives a negative tax; that is, it bears no tax itself and shelters unrelated income from tax. It has a higher return after tax than before tax./1/ But Professor Mundstock argues that equity-financed and debt-financed investments in ACRS property give identical tax benefits: whatever 'tax preference' is available from accelerated depreciation with debt financing is available with equity financing. He attributes any revenue loss in debt financing to lenders' low tax. He is accordingly critical of proposals limiting tax shelters by limiting deduction of interest.

This article argues that our tax treatment of debt-financed investments in ACRS property is a serious problem and that debt financing is distinguishable from equity financing. Interest should be deducted to compute net income when it is a cost of taxable income, but not deducted when the netting process would not justify it. Deduction of interest when the revenue it was a cost of is tax- exempt or tax-favored leads to a mismatch or misdescription of the taxpayer. The interest deduction then becomes a tax shelter that erases taxable income from high salaries and other sources that we want to tax fully. If deduction of interest is inevitable, it is impossible to tax any source once one source is exempted.

Congress coherently decides to tax some but not all sources. The arguments that were persuading for the adoption of ACRS, for instance, apply only to equity investments made by the taxpayer and do not justify nontaxation of salary nor justify interest deductions on debt used to finance ACRS investments.

A dramatic illustration of the special error in our treatment of debt financing is that debt-financed purchases of worthless ACRS property can be perfectly rational. The tax saved by a liability can exceed the detriment of paying the liability. Equity purchases of the same property under the same circumstances would not be rational; the purchased ACRS property would have to have some underlying value, albeit not as much nontax value as nontax-favored property must have.

This article argues that Professor Mundstock is right about his 'tax preference' point, as he is using the term 'tax preference,' but wrong about its implications. Whatever the wisdom of ACRS, the article concludes, once we have ACRS and other tax-favored investments, interest limits are inevitable.

Netting, the Rationale for Deductions

Interest is a cost or expense, just like any other cost. Like any other cost, it should be netted against the item it is a cost of, and it is from that item that it draws its appropriate tax character. The federal income tax is a tax on profit, on net income, and not on gross receipts. Expenses of getting taxable income must thus be deducted to compute the net income that is intended to be taxed. The expense of effecting a sale, for example, is a current deduction if the sale itself is currently ordinary income; the cost of selling inventory is thus an immediate deduction.

But the process of netting does not require that we give a cost an ordinary deduction if the gross receipts it is related to are not fully taxed. The cost of selling a long-term capital asset just reduces the capital gain from the sale (or increases the long-term capital loss). The cost of issuing corporate stock just reduces the tax-exempt proceeds from sale of the stock and the cost of effecting a section 337 liquidation sale just reduces the tax-exempt proceeds from the section 337 sale; neither cost gets any tax recognition because the sale was not recognized. Selling a used refrigerator will usually generate only a nondeductible personal cost. Netting does not generate a tax deduction unless the sale generates ordinary income.

It would be an error to allow an ordinary deduction for the cost of effecting a sale where the sale was tax-favored. The deduction would strip apart or mismatch revenue and expense, would un-net rather than net, and would misdescribe the transaction.

Suppose, for instance, that we allowed an ordinary deduction for the costs of a corporation issuing its own stock, instead of just allowing net proceeds of the issue to be tax-exempt. Every stock sale would then generate a tax loss that would shelter unrelated corporate income. The issuance of stock would produce a tax loss, not because issuing stock is economically a loss, an irrational act, but because the tax treatment misdescribed the transaction. It would not be that sections 118 or 1032, giving the tax exemption to the proceeds of the issuance of corporate stock, are bad law; it would be the bad netting that would produce the error.

If costs of effecting sales were always deductible, why not go further and allow an ordinary deduction for basis, the cost of the thing sold? A sale of raw land could then generate an ordinary deduction in the amount of the cost basis of the land and long-term capital gain for the gross sales proceeds./2/

Interest is like the cost of effecting a sale. Because borrowed cash gets mixed in with all cash, it is often harder to identify what interest is a cost of than it is to match the costs of effecting a sale with the sale. Interest is an avoidable cost,/3/ but then, on owner-effected sales, so are sales commissions. But all of that does not change the fact that interest is just a cost. It is as bad to allow interest deductions where netting would not allow the deduction as it would be to deduct selling costs while giving exemption or capital gain to the gross sale. Interest may be entitled to a tax deduction from the netting process, but it carries no magic that legitimates deduction in all events.

Can We Tax Anything?

Interest and other costs must be netted against only the related income if we are to maintain a tax on income from any source. The laws of logic seem to allow us to tax income from some sources but not others. As a descriptive matter, the Code certainly seems to want to tax salary, although it does not for instance tax municipal bond income. If interest must be deducted in every case, however, then it is impossible to tax any source, once one source becomes tax-exempt. The interest deduction will just transport the exemption privilege from the favored source to all income.

The Supreme Court explained it in Denman v. Slayton, 282 U.S. 514, 519-529 (1931). The Court held that, although it would not have been constitutional to tax income from a municipal bond, it was still constitutional to disallow interest on indebtedness incurred to purchase tax-exempt municipal bonds (now IRC section 265(2))/4/:

________________________________________________________________________________

' If the interest cost of tax-exempt income were deductible

as the taxpayer in this case argues, then 'A,' a taxpayer with

an income of $ 100,000 from non-exempt securities or salary,

by the simple expedient of purchasing exempt ones with borrowed

funds and paying $ 100,000 interest thereon, would escape all

taxation upon receipts from both

sources. It was proper to make provision to prevent such a

possibility. The classification complained of is not arbitrary,

makes no improper discrimination, does not result in defeating

any guaranteed exemption, and was within the power of Congress

. . . .'

________________________________________________________________________________

At least as a matter of constitutional law, the interest deduction may be deniedif it would otherwise zero out income that is intended to be taxed.

ACRS

Our current tax treatment of debt -- allowing basis for debt and then allowing an interest deduction -- is appropriate only when the purchased property is depreciated under 'economic depreciation.' Professor Mundstock shows this with a hypothetical in which the depreciable property generates 15 percent income before tax and the interest on the debt used to finance the property is also 15 percent. If we must allow our current treatment of debt, then the only way in which the tax accounts can be made to describe the zero economic income the taxpayer has in the hypothetical is by allowing depreciation deductions which match the slow repayment of principal on the debt. If faster depreciation is allowed, then deducting interest in violation of matching means the investment throws off losses that shelter unrelated income.

ACRS depreciation is economically equivalent to tax-exemption or better -- the value of the depreciation deductions and investment credit will normally equal or exceed the value of any tax on the property./5/ From this one can argue that Congress adopted effective exemption of ACRS income (even if it did not literally know or understand that that was what it was doing). But there is no indication that Congress therefore meant to exempt unrelated income.

All of the persuading arguments for ACRS do not go past tax- exemption for ACRS property and do not extend to negative tax or exemption of unrelated income. The spirit of 1981, when ACRS was adopted, was supply side economics: 'If only the weight of taxes can be lifted from the springs of the free market economy, then the springs shall spring up and the desert shall bloom.' That argument does not justify a negative tax, in which the tax treatment is better than freedom from tax for the transaction. Such negative tax is a 'subsidy,' a bad thing, a stretching of the springs, which distorts the free economy. Proponents of ACRS also appealed to consumption tax theory, but the advocates of a consumption tax would correct the negative tax that ACRS gives by disallowing interest or even more draconian cures./6/ ACRS was advocated as an offset to the effects of inflation, but the borrower who buys ACRS property with debt is helped rather than hurt by inflation./7/ Debt-financed ACRS property does not induce capital formation, because the investment and the disinvestment in borrowing cancel each other out and because the negative tax allows an investor to maintain his normal consumption as if he had not made the investment./8/ One cannot defend the negative tax as a contribution to productivity, since the negative tax hurts more than it helps productivity.

But if interest must always be deducted, then exemption for ACRS income is exemption for income from any source./9/

Cost-Free Liabilities

A final distinction between debt and equity is that debt used to purchase ACRS property can be cost-free after tax, whereas equity similarly used cannot. The tax law has traditionally given advanced credit for a liability to be paid long in the future, treating liabilities as if they were already cash payments and ignoring the time value of money. Liabilities, however, commonly have a value less than face, often considerably less than face, especially when the interest deduction is taken into account. Treating a future payment as if it were already paid allows liabilities in which the tax savings generated by the liability is worth more, in time value terms, than the detriment of actually paying the liability. The tax savings from ACRS investments can never exceed 58 percent of basis (in a 50 percent tax bracket, using the optimal eight percent credit and assuming negligible discount rates), but with the interest deduction the present value of the liability can be less than 58 percent of basis. Alternately stated, the value of the interest and depreciation deductions exceeds the detriment of actually paying principal and interest on the liability. Overvalued liabilities were a target of the 1984 Tax Act, but the remedies were insufficient to cure the problem./10/

Where the tax savings are greater than the cost, the liability is cost-free after tax. Cost-free liabilities make investments in worthless property justifiable to the investor. While the property will not return anything, that does not matter because the tax savings alone are sufficient. Where cost-free liabilities are possible it is difficult to see why anyone pays any tax.

Equity investments by contrast will always be worth more than the tax savings. Since equity cash always has its own value, the value of the equity would always be greater than the maximum tax savings from ACRS. Admittedly, ACRS tax savings would justify investments in ACRS property that would not be made without the tax benefits and cause shifts from investments without ACRS, but still the property would have to have some real economic value if the investment is made with equity. Debt-financed investments are different from equity-financed investments then, in part because debt financing, but not equity financing, allows the extraordinary abuse of cost-free purchases of worthless property.

Mundstock Position

In 'Can Two Rights Really Make a Wrong?,' Professor Mundstock argues that the interest deduction on debt-financed ACRS property is no problem because ACRS gives the same 'tax preference' whether an ACRS investment was financed with the taxpayer's own money or with debt. A 'tax preference' is the tax saved because of a departure from some norm of how we ought to tax or at least a departure from how we 'normally' tax. For Professor Mundstock the norm measuring the tax preference is pretax net income. With equity financing, his norm is income using economic depreciation. With debt financing, the norm is a zero income investment, since in his hypothetical the interest costs just offset the income from the investment. The negative tax in debt-financed investments is a departure below zero, but only in the same amount by which ACRS is a departure from the ideal of economic depreciation.

Why do we care that the 'tax preference' is the same? Does the decision to exempt income from some source really imply that all income must similarly be exempted by the same amount? That is what the argument for maintaining constant 'tax preference' means. Is the tax benefit or tax preference really an inviolable right that must be identified and preserved in any transaction even if it is beyond the bounds of the justifications of ACRS? Must then Denman v. Slayton be reversed, at least as a logical position? 'Tax preference' is usually a language setting up a challenge to the underlying privilege or at least calling for a Stockman-like cost-benefit analysis. Professor Mundstock could have used his point to say that those who challenge tax shelters should challenge equity investments too. But he seems to be making a mirror image of that point, i.e., that those who tolerate ACRS must love tax shelters for salary too. That position of course attacks all those who advocate any respectable consumption tax. Does he really mean to say that tax preferenceness is a quality that justifies a tax treatment per se, even where all else fails? May Stanley Surrey turn over in his grave.

To my mind the beguiling aspect of Professor Mundstock's argument comes from the undefended assumption that interest should be deducted. His subtitle, 'Can Two Rights Really Make a Wrong?,' implies that deduction of interest is right, always, and not right just as a part of netting. He measures the tax preferences from a norm in which interest was deducted (to reach zero economic income) and his procedure of measuring the tax preference crosses a line from the pretax sphere, in which interest should have been subtracted, into a sphere, i.e., exemption of the related income, in which netting leads to no deduction. Part of the 'prestige' of subtracting interest to compute economic income carries over to the post-tax position of negative tax. Yet once one concludes that interest should not be deductible, except as netting, then deducting interest becomes a separate error.

Two Rights and Double Dipping

In any event, the answer to Mundstock's subtitle question, 'Can Two Rights Really Make a Wrong?,' is clearly yes. Two rights make wrongs all the time. Tax law abhors a double deduction. Medical expenses may be deducted and an employee may exclude employer reimbursement of medical expenses, but an employee may not both deduct the expense and exclude the reimbursement. Fully depreciated property may not be deducted when stolen, notwithstanding that either depreciation or a theft loss would be appropriate. More subtly, Tufts v. Commissioner, 461 U.S. 300 (1983), held that unpaid liability generated gain when property was disposed of; without the prior deductions caused by the liability, the holding would have been inappropriate; Tufts just prevented double benefits. Negative taxes always arise because provisions that provide for mere partial tax exemption in isolation lead to negative tax when combined with other provisions; there is a synergistic effect. Conversion to capital gain, double dipping for both credit and deductions and many other misdescriptions can be traced to varieties of double deduction.

Either/or situations are common in the tax law. From 'either/or' does not follow 'and both.' Arguably the error in tax shelters from debt-financed ACRS property is solely a double deduction: the costs are deducted where the related income is not taxed. Two rights together often make a wrong.

Taxation of the Lender

Professor Mundstock attributes any revenue loss from the mismatch of interest to our failure to tax the lender who is the recipient of interest. Does it matter whether the other party pays tax on a cost? Well, yes, it might. Return to the hypothetical in which the cost of issuing stock was deductible, even though the item produced by the cost was tax-exempt as a contribution to capital. The mismatch, allowing deduction, would produce a tax loss, where no economic loss existed. But it is useful still that a 'factor of production' that caused the sale bore at least some tax. An excise tax or surtax on stock brokers and underwriters could be an indirect remedy for a misdescription in which a corporation got to deduct its costs of issuing stock. If the tax is high enough and the parties pass it back to the corporation, the misdescription might be remedied. If we were counting on tax on the broker to correct the misdescription, of course, then it would be too bad if the stock broker avoided the tax; one should not count on avoidable paper taxes. Still, taxing another party seems like a backward way to remedy a misdescription of the corporation itself; advocates would have a high burden to show that the indirect remedy really worked. Why not just net costs in the first place?

In the real world, whatever the theory, lenders pay little tax on interest and what they do pay is not passed back to the borrower. The market for interest seems to allow little or no room to cover lender tax./11/ Moreover, a correcting tax would have to be a penalty or excise at a higher rate than the lender normally pays. Mere normal tax would not be passed back very effectively to the borrower. One can thus respond to a suggestion that lender tax might offset the original error by saying, 'Well, it doesn't.'/12/

Second Best

A final way to say that ACRS-related interest should be disallowed is by application of the theory of 'second best.' Let us assume that in an 'ideal' world where all income was taxed, interest would always be deducted because it is 'negative income' or a cost of income. It does not follow that in our 'second best' world, things would be improved by deducting interest. If returns are tax-exempt as ACRS returns effectively are, deduction of the related interest must inevitably be restricted.

Assume, to draw an analogy, that the optimal way to turn a car on a city street is by driving at 10 miles an hour and turning the steering wheel 90 degrees for the first second of the turn./13/ Now let us assume the driver is actually turning his steering wheel 20 degrees in the first second. Does it help to move toward the optimal speed of turning, i.e., toward 90 degrees? Not if the car is going at 56 mph. I suppose one could blame the disaster that would occur with 56 mph and a 90 degree turn solely on the 56 mph speed. Fifty-six mph is too fast. Still, I would offer the following cogent advice: If you are driving at 56 mph, do not turn so fast!

________________________________________________________________________________

FOOTNOTES

________________________________________________________________________________

/1/ See discussion, Johnson, Tax Shelter Gain; The Mismatch of Debt and Supply Side Depreciation, 61 Tex. L. Rev. 1013 (1983).

/2/ Would Professor Mundstock amend section 103 to allow ordinary deduction of the cost of municipal bonds, as well as just exemption of the net profit from the bonds? Or is it a cogent, respectable position to have a section 103 that just allows exemption of the net income, without ordinary deduction of the costs?

/3/ It is said that a deduction for interest is inevitable because without it taxpayers with capital would just withdraw capital from taxed investments, leading to a kind of imputed interest deduction. Klein, 'Borrowing to Finance Tax-Favored Investment,' 1962 Wisc. L. Rev. 608 (1962); White, 'Proper Income Tax Treatment of Deductions for Personal Expense,' H. Ways and Means Comm., 86th Cong., 1st Sess., Tax Revision Compendium 365 (Comm. Print 1959). In fact, however, high bracket taxpayers have already withdrawn from fully taxed investments and changing the rules for interest would have little further impact. Moreover, the tax system cannot match imputed deductions any more than it can match nontaxation of imputed income. If the tax system had to match every nontaxation of imputed income, then we could not tax anyone because carpenters are not taxed when they build their own house and because housewives are exempt on their contribution to the household.

/4/ Some commentators suggest repeal of section 265(2) as a means to restrict ownership of municipal bonds to 50 percent bracket taxpayers. If only 50 percent bracket taxpayers held such bonds, then municipalities would have to pay interest of only half of FMV interest on comparable bonds and the public benefit from tax exemption (lower interest on the bonds) would then approach the cost of the exemption (loss of revenue). The remedy would not, however, restrict ownership to the highest bracket. Tax shelters, which have the full use of leverage, are now directed to taxpayers with 35 percent marginal brackets. There are not enough taxpayers with higher brackets; they have already stripped their higher than 35 percent brackets of reported income by sheltering. If one wants to limit tax preferences to 50 percent bracket taxpayers, one had better do it directly by prohibiting their use except in or against top bracket income.

/5/ See discussion, Johnson, Tax Shelter Gain; The Mismatch of Debt and Supply Side Depreciation, 61 Tex. L. Rev. 1013, 1021-1025 (1983).

/6/ Including all borrowings in income or excluding all borrowings from basis are the alternative cures. See e.g., discussion, Dept. of Treas., Blueprints for Basic Tax Reform 124-125 (1975).

/7/ See, e.g., Aaron, Inflation and Taxes, in Inflation and the Income Tax 9-10 (H. Aaron ed. 1976) ('Whatever lenders lose, borrowers gain').

/8/ See discussion, Johnson, Tax Shelter Gain; The Mismatch of Debt and Supply Side Depreciation, 61 Tex. L. Rev. 1013, 1018 (1983).

/9/ With ACRS or depreciation approaching ACRS, the Act as passed will need to be more savage to interest than is the House bill. Since ACRS is now equivalent to tax exemption, the correlative treatment of interest is to disallow the deduction of the interest permanently, even though the ACRS property is generating income. Under section 163(d) as amended by the House, interest is not disallowed in full; it is allowed as a deduction to the extent there is income from ACRS property and if there is no income the interest may be carried over to future years. Section 163(d) arose in 1969 as a remedy for the excluded portion of capital gain; it is too liberal, a false analogy, when applied to ACRS.

/10/ For a full discussion of cost-free liabilities phenomena, see Johnson, 'Silk Purses From A Sow's Ear: Cost Free Liabilities Under the Income Tax,' 2 Amer. J. of Tax Policy 231 (1984).

/11/ The arguments that tax on the lender remedies the negative tax are discussed at length in Johnson, Tax Shelter Gain; The Mismatch of Debt and Supply Side Depreciation, 61 Tex. L. Rev. 1013, 1039-1049 (1983).

/12/ Frederick Hickman has argued that interest must be deducted because the lender is a partner of the borrower. The interest deduction just serves like section 704 as an allocation of partnership income away from the taxpayer and to the lender. Shepard, 'Tax Incentives and Debt: The Tax Arbitrage Game Explored at Charlottesville Tax Conference,' Tax Notes, October 23, 1985, p. 340. Treating the lender as a partner is a planning disaster (and revenue gain) because the ACRS effective tax exemptions are trapped in an entity that is tax-exempt anyway. The ACRS exemption overlaps with the tax exemptions which the lender gets as entity and are wasted. Why give tax-exempt money to an entity that is tax-exempt anyway? The $1 trillion dollars of ACRS property held by tax-exempt institutions could be held by taxable property who would not otherwise be taxable. In sum, arguing that ACRS investments are like partnerships should mean that the government is entitled to more tax revenue from ACRS than it in fact gets.

/13/ The hypothetical is derived from Markovitz, 'A Basic Structure of Microeconomic Policy Analysis in our Worse-Than-Second- Best World,' 1975 Wisc. L. Rev. 950, 968.

________________________________________________________________________________