Resident  Scholar  Kevin A.  Hassett

Resident  Scholar Alan  D. Viard

I. Introduction

The statutory rate on corporate capital gains is currently equal to the statutory tax rate on ordinary corporate income. Although individual capital gains taxes have received an enormous amount of attention, both in the media and in the academic literature, corporate capital gains have received little attention. In principle, however, the distortions that arise from corporate capital gains taxation are analogous to those that might result from individual capital gains taxation. Corporations might face a higher user cost of capital and they could find that their previous purchases have been "locked in" in the sense that asset sales are avoided because of their tax consequences.

As a general rule, corporate capital gains taxation has been ignored in the investment literature, and corporate capital gains tax rates have been excluded from studies that have attempted to examine the impact of taxation on firm-level investment. For example, a recent review of the literature in this area by Kevin A. Hassett and R. Glenn Hubbard does not mention a single study that addresses the theoretical or empirical impact of corporate capital gains taxation on firm behavior.[1]

More recently, a 2004 paper by Mihir A. Desai and William M. Gentry[2] and a 2006 paper by Desai[3] have begun the inquiry into the impact of the corporate capital gains taxation. The authors highlight a number of important findings:

1. Despite the potential for a lock-in effect, corporate capital gains still account for approximately 20 percent of corporate income in a typical tax year.[4]

2. Using Compustat data, Desai estimates the value of the stock of unrealized capital gains to be, at a minimum, about $839 billion in 2004.[5] That large number suggests that the gains from eliminating the lock-in effect could be enormous.

3. Most countries other than the U.S. already either exempt corporate capital gains from taxation or provide special treatment for those gains. For example, Costa Rica, Hong Kong, Jamaica, Kenya, New Zealand, and Singapore do not tax the income at all. France exempts 95 percent of gains from taxation, and Germany does the same. Elsewhere in Europe--the Netherlands and Sweden, for example--gains are exempted from taxation if they involve at least a 5 percent share of one company in another.[6]

4. The elasticity of realizations of capital gains appears to be quite high. Using time-series regressions, Desai and Gentry estimate that it is about -1.13.[7]

5. Making conservative assumptions, Desai calculates that more than $100 billion of realizations are displaced each year because of the lock-in effect, resulting in a welfare loss to society of about $20 billion per year.[8]

Those findings suggest that the potential consequences of a reduction in the corporate capital gains tax (corporate CGT) are significant. Courtney H. Edwards, Mark H. Lang, Edward L. Maydew, and Douglas A. Shackelford studied a large reduction of the German corporate CGT and found that it had a large and statistically significant effect on stock prices.[9] In particular, firms with large cross-holdings significantly outperformed other firms at the time of the reduction.

Perhaps because of those findings, more attention is beginning to be paid to the corporate CGT in the U.S. However, the analysis to date looks at only one piece of the puzzle. Corporate CGTs certainly have a lock-in effect, but they also have the potential to increase the cost of capital. Because of that, a reduction in the corporate CGT could have additional effects on capital investment.

The purpose of this report is to explore the cost-of-capital effect and to use that exploration to provide some guidance for potential reforms. Our discussion is limited to the taxation of corporate gains on depreciable business property; we do not discuss the taxation of corporate gains on financial assets.

In Part II, we describe the current taxation of corporate capital gains in the U.S. In Part III, we consider a simple model with specific assumptions about why sales of used capital might occur. In Part IV, we apply the simple model to the U.S. tax system. The analysis reveals that current law, under which gains on sales of used capital are taxed at ordinary tax rates and buyers are allowed to depreciate their purchase cost, imposes a tax penalty on sales of used capital. That penalty, which is larger for long-lived property, increases the cost of capital and depresses investment. We identify three reforms that would eliminate the tax penalty in the simple model: zero taxation of capital gains with recapture of excess depreciation, zero taxation of capital gains with the seller's basis carrying over to the buyer, and reduced tax rates on capital gains. In Part V, we discuss extensions to the model and conclude that basis carryover and reduced rates may be the preferable policies.

II. Current Tax Treatment

The corporate tax rate is generally 35 percent for large C corporations.[10] Capital gains are generally taxed at the same rate as ordinary income.[11] As a result, the distinction between ordinary income and losses on one hand, and capital gains and losses on the other, is generally relevant only for corporations seeking to deduct losses.[12]

The gain on the sale of depreciable property is generally the sale price minus the firm's cost basis. The cost basis is generally the nominal purchase price minus the nominal depreciation allowances claimed on the property.

Gains from the sale of depreciable property are not always taxed as capital gains. For equipment and other personal property, all nominal depreciation allowances (up to the amount of the taxable nominal gain) are recaptured as ordinary income under section 1245; only the excess, if any, of the sale price over the original nominal purchase price is treated as capital gain. There is, however, no recapture for gains on structures and other real property; those gains are treated as capital gains.[13] Because both types of income face the same tax rates, the recapture provisions are not relevant for corporations under current law, unless they seek to deduct losses against those gains.

Buyers of depreciable property generally may claim the same depreciation deductions and other tax treatment as buyers of new property.[14]

In some cases, depreciable property may be disposed of without current taxation under section 1031. The firm must engage in a like- kind exchange in which it receives replacement property that is in the same general asset class or the same product class as the original property; the firm then carries over its basis in the original property to the replacement property. Of course, that treatment is available only if the firm is willing to acquire similar property. Moreover, the acquisition and disposition of the two properties must be linked to each other; the favorable tax treatment is denied if a firm engages separately in a sale and a purchase. Among other things, the replacement property must be identified within 45 days of the disposition of the original property and must be acquired within 180 days. The participants in the transaction must comply with complex identification and reporting requirements. Given the tight restrictions imposed on those transactions, we do not further consider like-kind exchanges in our analysis of the tax treatment of sales of depreciable property.[15]

The acquisition of a corporation can sometimes be structured and taxed as the purchase of its stock rather than as the purchase of its assets. In that case there is no corporate CGT on the firms' depreciable property or other assets, but the firm's stockholders are taxed at the individual capital gains rate on the gains on their stock holdings. We also do not examine stock-sale transactions.

In general, then, firms that dispose of depreciable property pay a 35 percent tax on their gains.

III. Model of Corporate Capital Gains Taxation

In subpart A, we review the standard model of the user cost of capital, with no sales of used capital. We introduce those sales in subpart B and subsequently analyze their tax treatment.

A. Standard Framework: No Sales of Used Capital

We sketch briefly how the cost of capital is derived in the traditional framework in which depreciable property is never sold. To facilitate the numerical results presented in Part IV below, we assume that production and tax payments occur at one-year intervals. The algebraic derivations and formulas are not presented here, but may be found in the unpublished version of this article.[16]

At date zero, a new unit of capital is produced, with a production cost of one consumer good. The unit produces no output at that date, but one year later, at date one, it produces C units of consumer goods. One year after that, at date two, it produces C(1-d) units of consumer goods. It continues to produce output thereafter, with its output in year t equal to C(1-d)t-1 units of consumer goods. The parameter d, which is between zero and one, is the true, or economic, depreciation rate of this capital.

Because of the specific assumptions about how capital is produced, one unit of t-year-old capital will yield exactly the same output in every future year as (1-d)t units of newly produced capital. If old and new capital receive the same tax treatment, then, a unit of t-year-old capital must sell for the same amount as (1-d)t units of new capital. We assume that firms incur no additional costs to install the capital.

In this standard framework, capital, once purchased by a firm, is always capable of producing output (of the above-described amount) in the hands of that firm. There is, then, no reason to sell the capital to another firm. To be sure, there is also no reason not to sell, but those sales have no consequences in the framework and can therefore be ignored. In the real world, of course, there are many nontax factors that may stimulate a sale, and we turn to those in subpart B below.

Firms are willing to invest up to the point at which a new unit of capital yields an after-corporate-tax real return of r. In other words, the firm is indifferent between a claim on one consumer good today and a claim to 1+r consumer goods next year. That real return covers the compensation that households demand to supply funds to the firm (rather than consuming, investing abroad, or investing in another sector of the economy), as well as any taxes that households pay on investment returns.

With no taxes, what is the equilibrium level of C? The firm equates the present discounted value (discounted at rate r) of the output produced by the unit of capital to its purchase cost, which is one. The computation reveals that C is equal to r+d. In economic terminology, r+d is the user cost of capital in this no-tax world. The result makes intuitive sense; the product must yield the required real return r, after covering depreciation d.

With corporate taxes, the capital must generally have a higher marginal product to provide the required after-tax real return. We assume the corporate income tax rate is the same for all firms. Firms are allowed to claim depreciation deductions, which are also the same for all firms, although they differ across the various types of capital. Let Z denote the present value of the deductions, discounted back to date zero at discount rate r.

In analyzing tax policy, the standard framework continues to ignore possible sales of used capital. Given its other assumptions, that is not surprising. We have already seen that this framework offers no nontax reason for such sales. Assuming the tax system does not subsidize such sales, they will also not occur with taxes. Again, the framework's neglect of sales is strongly counterfactual, and we will modify it in subpart B below.

As explained in the unpublished version of this article, the corporate tax reduces the present value of the after-tax cash flows. But the firm enjoys tax savings from its depreciation deductions, which reduces the net purchase cost. The equilibrium marginal product, or user cost, is determined by the firm setting the present value of after-tax cash flows equal to this net purchase cost. With taxes, the marginal product C is generally higher than r+d, the value observed without taxes. The marginal product must now cover the firm's tax liability, in addition to depreciation and investors' required return.

It is useful to compare the net-of-depreciation before-tax return C-d to the net-of-depreciation after-tax return r. The effective tax rate is defined as one minus the ratio of r to C-d; it measures the fraction of the net-of- depreciation before-tax return that is paid in taxes rather than received by the investors. For example, if r is two-thirds of C-d, the effective tax rate is one-third, indicating that taxes absorb one-third of the before-tax return and that investors receive the remaining two-thirds.

Note than an increase in the user cost implies a higher effective tax rate and a reduction in investment. Such an increase means capital investment must yield a higher before-tax return to be viable.

B. Introducing Sales of Used Capital

The standard framework's treatment of sales of used capital is clearly at odds with reality. As discussed in Part I above, firms engage in a substantial amount of sales, even though, as we shall soon see, those sales are penalized by the tax system. It is clear that sales take place for nontax reasons; there must, then, be circumstances in which it is economically beneficial for capital to be held by one firm instead of another. We now introduce those sales in a simple manner that illustrates the potential impact of tax policy.

We assume that, T years after each firm purchases a unit of capital, the firm loses its ability to use that unit. The firm's operations no longer "fit" the unit of capital, and the unit can produce no further output in the hands of that firm. Nevertheless, the capital can still be used to produce output (in the amounts described above) if it is transferred to another firm, where it will remain operational for another T years.

In the no-tax world, the equilibrium outcome is clear. The firm that originally held the capital would sell it to another firm, and those sales would be repeated every T years. At each relevant date t (T, 2T, 3T, and so on), the equilibrium sale price is (1-d)t units of consumer goods, because that number of units of new capital would be a perfect substitute for this used capital. No buyer would pay more, since it could produce new capital instead; no seller would accept less, since even the slightest discount would draw many firms eager to pay less than the cost of new capital.

Any tax penalty on sales now has real effects. To be sure, under our assumption that the capital becomes worthless when it is held more than T years, there is no lock-in effect unless the tax absorbs the entire value of the capital. The firm will sell, even at a tax penalty, rather than hang on to worthless capital. Nevertheless, the anticipation of paying that tax penalty when the used capital is sold will increase the user cost and depress initial investment.

C. User Cost of Capital With Sales of Used Capital

We make the following general assumptions about the tax system. These assumptions will generally encompass the current U.S. tax system and the reform options that we consider in Part IV below. First, the sale proceeds, net of the firm's cost basis and any recapture amount, are taxed at a capital gains rate that may, but need not, equal the ordinary tax rate. Second, there may be a recapture amount that is taxed at the ordinary tax rate rather than the CGT rate. Third, purchasers of used capital receive the same tax treatment as purchasers of newly produced capital.

The last assumption (which will be modified for one of the reform options) has the important implication that the real sale price of the unit of t-year-old capital is still (1-d)t, as it was in the no-tax world. The unit of old capital remains a perfect substitute for that many units of newly produced capital, because it has the same tax treatment, as well as the same future output.

As explained in the unpublished version of this article, when the first firm buys the capital at date zero, it equates the present value of the after-tax output generated during its anticipated T-year holding period to its net cost of purchasing the unit. The firm's initial purchase cost for the unit of capital is one, but there are four modifications to the purchase cost. First, as before, the firm enjoys tax savings from depreciation deductions, although only the ones claimed at years zero through T. Second, the firm receives sale proceeds at date T, which must be multiplied by one minus the CGT rate. Third, the firm enjoys tax savings from the basis deduction. Fourth, the firm may pay additional tax at date T, equal to the recapture amount, if any, multiplied by the difference between the ordinary and CGT rates. The same computations, with the same results, apply to the firms buying the used capital at each of the subsequent sale dates.

The resulting equations, which are presented in the unpublished version of this article, generalize the results derived by Alan J. Auerbach in 1981, who performed a similar analysis for a specific tax system, as discussed below.[17] In drawing out the economic implications of those equations, we rely heavily on his cogent analysis. We will also refer to work by Roger H. Gordon, James R. Hines Jr., and Lawrence H. Summers (GHS), who performed numerical computations of this type for structures investment.[18]

D. Economic Analysis

As detailed in the unpublished version of this article, the tax treatment of capital resale has two effects on the user cost and hence the level of investment, a result also noted by Auerbach.[19] We label them as follows:

• The Capital Gains Tax Effect: The sale triggers taxation at the capital gains rate on the difference between sale price and basis, net of any amount that is recaptured as ordinary income.
• The Depreciation Allowance Effect: The sale also changes the tax depreciation allowances available on the capital. By selling the capital, the firm loses the remaining depreciation deductions, those that would have been claimed after date T. The firm may also be taxed on a recapture amount. But it receives a sale price that incorporates the buyer's ability to claim depreciation deductions equal to those on (1-d)T units of new capital. Alternatively, the firm could replace the capital with that many units of new capital, on which it could claim depreciation deductions.

Without more information about the tax system, we cannot say whether either effect, let alone the combined effect, is positive or negative.

Indeed, as demonstrated in the unpublished version of this article, both effects are zero for one textbook income tax system, namely a system of true economic depreciation, involving inflation indexation for both depreciation and basis, with no depreciation recapture. Economic, inflation-indexed depreciation results in an effective tax rate equal to the statutory tax rate with sales at any date or at none, provided that basis is computed in a consistent, inflation-indexed manner. The CGT effect disappears because basis is always equal to sale price, so that no gain or loss occurs on any sale. The depreciation allowance effect disappears because one unit of T-year-old capital receives the same depreciation treatment as (1-d)T units of new capital.

Of course, the U.S. tax system differs significantly from the above textbook system. In particular, depreciation allowances are accelerated, meaning they are more favorable for young assets than for old ones. Because of rapid depreciation, basis is reduced below market value, a disparity that is reinforced by the tax system's failure to index basis for inflation. As a result, the CGT effect discourages sales, because the sale results in a taxable capital gain. On the other hand, the depreciation-allowance effect encourages sales. That occurs because the used capital has relatively little depreciation allowances remaining to be claimed by the selling firm (none if the tax life has ended), while the buying firm is allowed to depreciate its purchase cost in the same manner as if it had purchased new capital.

We now examine the net impact of the two effects in the U.S. tax system.

IV. Application to U.S. Tax System

A. Tax Penalty on Sales of Used Capital

We apply the simple model to the U.S. tax system's treatment of equipment, software, and structures. We set the regular and CGT rates equal to 0.35, in accordance with section 11 of the code.

We first consider three-year, five-year, and seven-year property, as defined in section 168(e). We use the tax depreciation schedules applicable to those types of property, which are computed in accordance with section 168, reflecting the half-year convention allowed by section 168(d)(1). Those schedules are tabulated in recent work by Darrel S. Cohen, Dorthe-Pernille Hansen, and Kevin A. Hassett (CHH).[20]

We also consider nonresidential real property (structures). That property is depreciated on a straight-line basis over a 39-year period under section 168(b)(3)(B) and 168(c); the midmonth convention is allowed under section 168(d)(2)(B). We assume the property is placed in service during the sixth month of the tax year.

We assume the first-year depreciation allowance is deducted at the purchase date. Basis is equal to purchase cost minus the cumulative depreciation deductions, with no inflation indexation. The half-year convention applies to equipment and software in the year of sale, and the midmonth convention applies to structures in the year of sale, with the sale assumed to occur in the sixth month of the tax year.

Under our assumptions, the nominal sale price is always below the nominal purchase price, so section 1245 recaptures all of the gain on equipment and software sales as ordinary income. With the two tax rates equal, however, the recapture is of no significance and can be disregarded for present purposes.

Following CHH and recent work by Robert Carroll, Kevin A. Hassett, and James B. Mackie III (CHM)[21], we set the annual inflation rate equal to 0.03. We set the annual interest rate r equal to 0.05, which is generally consistent with the 0.04 required after-tax return at the household level that CHM assume, plus a markup for household-level taxes. Under those assumptions, the present value of depreciation deductions, Z, equals 0.930 for three-year property, 0.874 for five-year property, 0.825 for seven- year property, and 0.313 for structures.

Following CHM, we take the annual economic depreciation rates d for equipment and software to be the reciprocal of the tax lives--0.333 for three-year property, 0.2 for five-year property, and 0.143 for seven-year property. Following GHS, we set the annual depreciation rate for structures at 0.0247.[22]

For these benchmark parameters, we display effective tax rates for different holding periods T in Table 1:

                Table 1. Effective Tax Rates (percent)
                             (Current Law)
 ____________________________________________________________________
                                             T
                      -----------------------------------------------
                      No Sale       20       10       5       3
 ____________________________________________________________________

 Three-year           22.5          22.5     22.7     24.5    28.1
 Five-year            25.3          25.4     26.6     31.3    35.2
 Seven-year           26.7          27.0     29.5     35.2    38.2
 Structures           35.6          40.0     42.6     44.2    44.9

Allowing for the sale of used capital can significantly increase the effective tax rate, depending on the holding period and the durability of the capital. The effects are larger, of course, for shorter holding periods (more frequent sales). For any given holding period, the effects are larger for longer-lived property, for which a larger portion of the property remains to be sold. Also, the depreciation-allowance effect, which works to reduce the tax penalty, is smaller for longer-lived property, since depreciation allowances are claimed more slowly on that property.

With sales every 20 years, for example, the effective tax rate rises by 0.0 percentage points for three-year property, 0.1 percentage points for five-year property, 0.3 percentage points for seven-year property, and 4.4 percentage points for structures. With sales every 10 years, the increases are 0.2, 1.3, 2.8, and 7.0 percentage points. Not only does the tax penalty on sales discourage investment, it is also likely to distort the allocation of investment because the penalty varies across different types of capital, depending on durability and expected holding periods.

The tax penalty on sales is larger (smaller) with higher (lower) inflation, because basis is lower (higher) relative to sale price. As shown in the unpublished version of this article, however, the impact of inflation is relatively modest.

The impact of the tax penalty on investment can be noticeable. If sales occur every 10 years, the tax penalty on sales increases the user cost of capital (not shown) by 0.04 percent for three-year property, 0.45 percent for five-year property, and 1.30 percent for seven-year property. CHH report that roughly 20 percent of equipment and software spending goes to three-year property and roughly 40 percent each goes to the other two categories.[23] The weighted percentage increase in user cost for equipment and software is then 0.7 percent. Of course, the impact is sensitive to the holding period. With a holding period of 20 years, the weighted increase in the user cost falls dramatically, to 0.07 percent.

In their survey of the literature, Hassett and Hubbard state, "Recent empirical studies appear to have reached a consensus that the elasticity of investment with respect to the tax-adjusted user cost of capital is between -0.5 and -1.0."[24] With a 10-year holding period, that result suggests a 0.35 percent to 0.7 percent reduction in equipment and software gross investment. Since that investment is about $1 trillion annually, those results imply a $4 billion to $7 billion reduction in annual gross investment.

For comparison, note that CHH found a weighted reduction of 3.1 percent for the section 168(k) temporary partial-expensing provision.[25] With a 10-year holding period, the effect of the tax penalty is one-fourth the size of the effect of partial expensing. With a 20-year holding period, however, the impact is only 1/40th as great.

The effect on structures is an order of magnitude larger than the effect on equipment and software. The increase in user cost is 9.3 percent for a 10-year holding period and 5.5 percent for a 20- year holding period.

We now identify and discuss three reforms that would eliminate the tax penalty on sales in this simple model.

B. Three Equivalent Reforms

Under the first policy (the recapture policy), the CGT rate is set to zero, which eliminates the CGT effect. The depreciation allowance effect is then negated through a recapture provision that is detailed in the unpublished version of this article. In intuitive terms, the recapture amount for a sale at date T reflects the excess depreciation deductions that have been claimed through that date. It is computed as the date-T present value of the depreciation that would still be left to claim if the depreciation allowances had been allocated across years in proportion to the depreciation that actually occurred in each year (while keeping their total present value fixed) minus the date-T present value of the depreciation allowances actually remaining to be claimed. If the latter term is smaller, as it is in the U.S. tax system, it indicates that too little of the lifetime depreciation allowance is still unclaimed, which means too much has already been claimed. That excess is recaptured as ordinary income under this policy.

The second policy (the basis carryover policy) sets the capital gains rate to zero, without imposing any recapture obligation on the seller, but requires the buyer to depreciate the capital in the same manner and to the same extent that it would have been depreciated by the buyer. Of course, for any sale that occurred after the tax life had ended, the buyer would take the seller's zero basis and would have nothing to depreciate. If the buyer must follow the same rules as the seller, this effectively places the obligation to pay the recapture tax on the buyer rather than the seller. Under the maintained assumption in this simple model that all firms face the same tax rate, it makes no difference which firm bears the obligation. The sale price simply declines to reflect the buyer's obligation.

To be sure, the basis carryover policy has little precedent in the current code. An unrelated buyer is rarely, if ever, required to carry over the seller's basis. The code does contain provisions, such as sections 1031 and 1033, pertaining to like-kind exchanges and some rollovers, under which CGT is not imposed on the seller and the seller takes a carryover basis in the replacement property. Sections 362 and 723 sometimes require a corporation or partnership to which a taxpayer contributes property (without payment of CGT) to take a carryover basis in the property. Under section 1015(a), a person who receives property as a gift carries over the donor's basis, if that is less than market value. None of those policies, however, require an unrelated buyer to take a carryover basis. The denial of depreciation deduction to a buyer who pays real money to acquire used capital simply because the capital has already been fully depreciated by the seller, may seem counterintuitive, even when the seller receives a price reduction to compensate for that disability.

The third policy (the reduced-rate policy) lowers the CGT rate, with no depreciation recapture or basis carryover. The required CGT rate, which depends on the holding period, is detailed in the unpublished version of this article. Under the assumptions of the current model, this reduced-rate policy is equivalent to the recapture policy and hence to the basis carryover policy. The recapture policy taxes part of the gain at the ordinary tax rate and leaves part of it untaxed, while this policy achieves the same outcome by taxing the entire gain at a single blended rate.

Under any of those three policies, the following neutral outcome emerges:

                Table 2. Effective Tax Rates (percent)
             (Recapture, Basis Carryover, or Reduced Rate)
 ____________________________________________________________________
                                             T
                      -----------------------------------------------
                      No Sale       20       10       5       3
 ____________________________________________________________________
 Three-year           22.5          22.5     22.5     22.5    22.5
 Five-year            25.3          25.3     25.3     25.3    25.3
 Seven-year           26.7          26.7     26.7     26.7    26.7
 Structures           35.6          35.6     35.6     35.6    35.6

C. Simplified Reduced-Rate Policy

Basis carryover may be a relatively novel policy, but it is straightforward. In contrast, the recapture formula and the reduced capital gains rate are complex. As described in the unpublished version of this article, the recapture amount and the capital gains rate vary by both the holding period and the property's depreciation rate.

It is possible, however, to adopt a simplified form of the reduced-rate policy. As described in the unpublished version of this article, the capital gains rate can be set equal to Z times the ordinary tax rate. Recall that Z is the ratio of present value of depreciation allowances to the purchase cost. Using the values of Z previously reported, the simplified reduced-rate policy applies tax rates of 32.5 percent to gains on three-year property, 30.6 percent to gains on five-year property, 28.9 percent to gains on seven-year property, and 11 percent to gains on structures. As with the exact reduced-rate policy, there is no recapture and no basis carryover.

Although the rate still varies by type of property, it can be stated in a simple form and does not depend on the holding period. As shown in the unpublished version of this article, the simplified reduced-rate policy is identical to the exact reduced-rate policy (and therefore achieves neutrality) for sales that occur after the tax lifetime has ended and all depreciation has been claimed. The simplified policy is similar to the exact reduced-rate policy for sales within the tax lifetime and therefore results in approximate neutrality. In his 1981 analysis, Auerbach demonstrated that this simplified policy achieves neutrality for sales at any date if depreciation allowances follow a particular pattern, but U.S. depreciation allowances do not precisely follow that pattern.[26] The Auerbach result is discussed further in the unpublished version of this article.

Under the simplified reduced-rate policy, the following effective tax rates emerge:

                Table 3. Effective Tax Rates (percent)
                       (Simplified Reduced Rate)
 ____________________________________________________________________
                                             T
                      -----------------------------------------------
                      No Sale       20       10       5       3
 ____________________________________________________________________
 Three-year           22.5          22.5     22.5     22.5    23.2
 Five-year            25.3          25.3     25.3     25.8    27.8
 Seven-year           26.7          26.7     26.7     28.8    29.5
 Structures           35.6          36.5     37.4     37.9    38.1

The tax penalty on sales is largely removed for sales within the tax life and fully removed for later sales. Sales at date one (not shown) are subsidized under that policy, so it would probably be necessary to modify the policy for those sales.

D. More Modest Reforms

Under the recapture and basis carryover policies, the CGT rate should be zero. A more modest reform would combine a positive CGT, perhaps 15 percent, with either the recapture or the basis carryover policy. Table 4 displays the resulting effective tax rates:

               Table 4. Effective Tax Rates (percent)
           (15 Percent Capital Gains Rate, With Recapture or
                         Basis Carryover)
 ____________________________________________________________________
                                             T
                      -----------------------------------------------
                      No Sale       20       10       5       3
 ____________________________________________________________________
 Three-year           22.5          22.5     22.6     23.4    25.0
 Five-year            25.3          25.3     25.9     28.0    29.9
 Seven-year           26.7          26.8     27.9     30.6    32.1
 Structures           35.6          37.5     38.8     39.6    40.0

That policy removes much of the tax penalty on sales that appears in Table 1, but leaves a significant penalty in place.

A similar reform lowers the CGT rate to 15 percent, while leaving section 1245 in place. The recapture rule in section 1245 is more stringent than the neutral recapture rule, but it applies only to equipment and software, not to structures. The resulting effective tax rates are shown in Table 5:

                Table 5. Effective Tax Rates (percent)
           (15 Percent Capital Gains Rate, With No Change to
                             Section 1245)
 ____________________________________________________________________
                                             T
                      -----------------------------------------------
                      No Sale       20       10       5       3
 ____________________________________________________________________
 Three-year           22.5          22.5     22.7     24.5    28.1
 Five-year            25.3          25.4     26.6     31.3    35.2
 Seven-year           26.7          27.0     29.5     35.2    38.2
 Structures           35.6          37.5     38.8     39.6    40.0

For the three categories of equipment and software, the results of Table 5 are the same as the current-law results in Table 1. The reason is that, under the assumptions of the simple model, the nominal sale price is always lower than the original purchase price, so that section 1245 continues to recapture all of the gain as ordinary income. For structures, on the other hand, current law imposes no recapture requirement, so those results are the same as the corresponding results in Table 4.

Another proposal to mitigate capital gains taxation is to index basis to inflation. The effects are shown in Table 6:

                Table 6. Effective Tax Rates (percent)
              (Current Law, With Inflation-Indexed Basis)
 ____________________________________________________________________
                                             T
                      -----------------------------------------------
                      No Sale       20       10       5       3
 ____________________________________________________________________
 Three-year           22.5          22.5     22.7     24.5    27.3
 Five-year            25.3          25.4     26.6     30.8    31.5
 Seven-year           26.7          27.0     29.5     32.4    33.0
 Structures           35.6          36.7     36.0     35.4    35.1

The effects are modest for equipment and software. A key reason is that basis indexation matters only when basis is positive, which is true only for sales within the tax life of the property. Basis indexation is, however, beneficial for structures. With a 20-year holding period, for example, the tax penalty on sales is reduced from its current-law value of 4.4 percentage points to 1.2 percentage points; with a 10-year holding period, the penalty is reduced from 7 percentage points to a mere 0.5 percentage points. Note that basis indexation actually results in a small subsidy to sales after short holding periods, which could result in churning of structures.

With no recapture and no basis carryover, a rate reduction larger than that prescribed by our reduced-rate policy would have harmful consequences. The policy would artificially encourage sales because the CGT effect would be smaller than the depreciation allowance effect.

The unpublished version of this article considers a 15 percent CGT rate, with no depreciation recapture and no basis carryover. That rate is significantly lower than the neutral rates for equipment and software; recall that their approximate neutral rates range from 28.9 percent to 32.5 percent. (Of course, a 15 percent rate with no recapture and no basis carryover does not create those problems for structures, for which the approximate neutral rate is 11 percent.) As a result, firms would engage in tax-motivated sales; all equipment and software would be turned over every year, at least in the absence of transaction costs.[27] Antiabuse rules might prevent sales that were purely tax-motivated with no business purpose. But even those sales that have some business purpose should not be artificially encouraged by the tax system. Although subsidizing those sales lowers the cost of capital, it is a distortionary way to do so, compared with a policy that is neutral toward sales.

Fortunately, there has been little or no interest in those policies. Proposals for reduced capital gains rates generally retain a recapture provision for software and equipment.

V. Extensions

The basic conclusions from Part IV hold in a more general context, but some complexities are introduced. On the whole, the basis carryover and reduced-rate policies seem preferable in the more general context.

A. Lock-In Effect

The simplest modification is to assume that capital does not become worthless after it has been held for T years, but merely becomes less productive. That has little real impact on the analysis. A tax penalty on sales remains undesirable, but the form of the inefficiency changes and its magnitude becomes smaller. The difference is that, with a sufficiently large tax penalty on sales, the tax will cause the capital to not be sold. That is the familiar lock-in effect.

The tax penalty on sales of used capital always continues to raise the user cost of capital and to depress initial investment. The penalty results in either a tax payment on sale or in the inefficient retention of less productive capital, either of which reduces the profitability of the initial investment. When the capital is retained, the resulting loss of output must be less burdensome than the tax payment (otherwise the firm would not have chosen to hold on to the capital), so the availability of that option does reduce the initial burden on investment. Note, however, that because no revenue is collected from the CGT in that case, the burden rises as a fraction of revenue.

All of the key results still apply with that modification. It is still desirable to eliminate the tax penalty on sales of used capital, regardless of the form or mix of distortions that are caused by such a penalty. And if the other assumptions of the above simple model are retained, any of the three reforms would eliminate the tax penalty.

Things become more complicated, however, when we allow fluctuations in the price of used capital.

B. Price Fluctuations

In the actual economy, the sale price of T-year-old capital may be higher or lower than (1-d)T. For example, the production cost of the perfectly substitutable new capital may have changed, which would automatically change the value of used capital. Or the used capital may remain fully productive for the selling firm, but have extra value for the buying firm, which for some reason cannot produce perfectly substitutable capital at the original cost. In general, the U.S. tax system ignores any fluctuations in the value of capital until and unless it is sold; we assume any reform continues to do the same.[28]

A desirable policy has two parts. First, it sets the depreciation schedule in a way that delivers the desired expected user cost of capital. That criterion can be satisfied by any of the policies that we have considered by choosing the correct depreciation schedule. Second, a desirable policy neither penalizes nor subsidizes any sales that may occur, regardless of what the sale price ends up being. Some of the policies that we have considered do not result in such neutrality.

The basis carryover policy achieves neutrality. For any sale price, the selling firm pays no tax and receives no deduction; there is also no change in depreciation deductions, except they are claimed by the buyer rather than the seller. Everything continues as if the capital had never changed hands, under our maintained assumption that all firms face the same tax rates.

The unpublished version of this article examines the implications of price fluctuations for current law and for the recapture and reduced-rate policies. As shown in Table 1 above, current law penalizes sales in the simple model. That remains true with price fluctuations; the penalty is larger (smaller) at higher (lower) values of the sale price.

As shown in the unpublished version of this article, the reduced-rate policy continues to achieve neutrality at any sale price, for sales after the tax lifetime has ended and all depreciation has been claimed. (Recall that the exact and simplified reduced-rate policies are identical for sales after the tax lifetime has ended.) Neither the exact nor the simplified policy is precisely neutral for sales within the tax lifetime, but approximate neutrality is generally attained.

In contrast, the recapture policy does not achieve neutrality. If the sale price exceeds the value set forth in the simple model, the sale is subsidized, because the seller pays no tax on the additional gain (remaining liable only for the fixed recapture amount), but the buyer depreciates the higher purchase price. Conversely, if the price is lower than that value, the sale is penalized, because the seller remains liable for the fixed recapture amount, but the buyer's depreciation allowances are smaller because of the smaller purchase price.[29] Under the recapture policy, therefore, resale decisions would be affected by the tax system.

This analysis indicates that the basis carryover and reduced- rate policies are the better policies. As noted in Part IV.B above, however, the basis carryover policy has little precedent in the current code.

C. Differences in Firms' Tax Treatment

Another complication is that firms may not face the same tax rates. Some may be passthrough entities rather than C corporations, some may be subject to the corporate alternative minimum tax, and some may be in loss-carryforward situations. Those differences in tax treatment already cause problems under current law. The basis carryover policy is not appropriate for this situation, since the movement of property between firms results in a change in tax treatment. None of the policies work perfectly in this environment, however, so it is difficult to make a clear comparison.

VI. Conclusion

The taxation of corporate gains on sales of depreciable property is an important topic that has received little attention in the literature. Our analysis reveals that current law, under which gains on sales of used capital are taxed at ordinary tax rates and buyers are allowed to depreciate their purchase costs, places a tax penalty on sales and raises the cost of capital. In a simple model, we identify three policies that are equivalent to each other and that eliminate the impact on user cost and investment: a zero tax rate on capital gains with recapture of excess depreciation allowances at ordinary tax rates, zero taxation of capital gains with the seller's basis carrying over to the buyer, and reduced tax rates on capital gains. In more general models, the basis carryover and reduced-rate policies appear to be preferable.

FOOTNOTES

1. "Tax Policy and Business Investment," Handbook of Public Economics, vol. 3, ed. Alan J. Auerbach and Martin Feldstein (Elsevier Science B.V. 2002), pp. 1293-1343.

2. "The Character and Determinant of Corporate Capital Gains," Tax Policy and the Economy, vol. 18, ed. James M. Poterba (MIT Press 2004), pp. 1-36.

3. "Taxing Corporate Capital Gains," Tax Notes, Mar. 6, 2006, p. 1079, Doc 2006-3505 [PDF], 2006 TNT 44-39.

4. Desai and Gentry, supra note 2, at 21; Desai, supra note 3, at 1081.

5. Desai, supra note 3, at 1081.

6. Desai, supra note 3, at 1083-1084; Desai and Gentry, supra note 2, at 8-9.

7. Desai and Gentry, supra note 2, at 24, report an elasticity of -1.3. Desai, supra note 3, at 1088, corrects the number to -1.13.

8. Desai, supra note 3, at 1088-1089.

9. "Germany's Repeal of the Corporate Capital Gains Tax: The Equity Market Response," 26 Journal of the American Taxation Association, 2004 supplement, pp. 73-97.

10. Section 11 imposes a 35 percent rate on corporate income in excess of $10 million. (Unless otherwise noted, section references are to the Internal Revenue Code of 1986, as currently in effect). Under section 199, 6 percent of net income from qualified production activities may be deducted in 2007 through 2009 and 9 percent may be deducted thereafter, lowering the effective tax rate for those activities to 32.9 percent or 31.85 percent.

11. Section 1201 allows firms to choose a flat rate of 35 percent tax on their net capital gains, but that option does not lower tax liability.

12. Corporations generally may not deduct capital losses against ordinary income. Complex rules govern the combination and netting of short-term and long-term gains and losses. See Desai and Gentry, supra note 2, at 6-7, and Desai, supra note 3, at 1090-1091.

13. Although section 1250 includes a recapture provision for real property, it applies, under section 1250(b)(1), only to depreciation in excess of that allowed under the straight-line method. Under section 168(b)(3)(B), however, structures are now depreciated using the straight-line method.

14. Although the statement holds for general depreciation allowances, some special incentives are restricted to new property. For example, under section 168(k)(2)(A)(ii), partial expensing, now expired, was offered only to property for which "the original use . . . commences with the taxpayer." The investment tax credit, now repealed, also featured an original-use requirement in some cases.

15. If anything, those restrictions may be further tightened. A Senate Finance Committee aide recently said Congress may want to curtail like-kind exchanges. See Dustin Stamper, "Like-Kind Exchanges Possible Target, Tax Aide Warns," Tax Notes, Mar. 12, 2007, p. 971, Doc 2007-6063 [PDF], 2007 TNT 47-3.

16. The unpublished version of this article is available on the American Enterprise Institute Web site (http://www.aei.org/publication26301) or from the authors. The details of the numerical calculations for this article are also available from the authors.

17. "Inflation and the Tax Treatment of Firm Behavior," 71(2) American Economic Review 419-423 (May 1981). Auerbach assumed that production and tax payments occur continuously, rather than at one-year intervals, but that that difference has no impact on the substantive results.

18. "Notes on the Tax Treatment of Structures," in The Effects of Taxation on Capital Accumulation, ed. Martin Feldstein (Chicago: University of Chicago Press, 1987), pp. 223-254.

19. Auerbach, supra note 17, at 420. See also Mark H. Robson, "Measuring the Cost of Capital When Taxes Are Changing With Perfect Foresight," 40(3) Journal of Public Economics 261-292 (Dec. 1989) at 269-273.

20. "The Effects of Temporary Partial Expensing on Investment Incentives in the United States," 55(3) National Tax Journal 457-466 (Sept. 2002) at 460.

21. "The Effect of Dividend Tax Relief on Investment Incentives," 56(3) National Tax Journal 629-651 (Sept. 2003).

22. Gordon, Hines, and Summers, supra note 18, at 233.

23. Cohen, Hansen, and Hassett, supra note 20, at 460.

24. Hassett and Hubbard, supra note 1, at 1325.

25. Cohen, Hansen, and Hassett, supra note 20, at 463, Table 3. The results are taken from the "new law" column of the table and from the intermediate-adjustment-cost assumption that omega is 0.5.

26. Auerbach, supra note 17, at 420.

27. Gordon, Hines, and Summers, supra note 18, concluded that individuals, but not corporations, had an incentive to churn structures under the laws in place before the Tax Reform Act of 1986.

28. If property is destroyed or damaged by fire, storm, shipwreck, or other casualty, however, the resulting decline in market value--but no more than the cost basis--may be deducted as a casualty loss under section 165(a) and reg. section 1.165-7. Also, under reg. section 1.167(a)-8, a firm may be able to deduct its remaining basis in a property if the firm permanently withdraws the property from use in the trade or business.

29. The sale is still penalized, although to a lesser extent, if the recapture amount is capped at the amount of taxable gain.

Kevin A. Hassett is a senior fellow and director of economic policy studies at AEI. Alan D. Viard is a resident scholar at AEI.