CORPORATE TAX REVENUE GAINS, LOSSES ABOUT EVEN, SAYS CRS.

Jane G. Gravelle Senior Specialist in Economic Policy

June 14, 1991

SUMMARY

The corporate income tax has been criticized for taxing income from capital used in the corporate sector twice -- once at the corporate level and once at the individual level. The results are that higher pre-tax returns in the corporate sector are forgone in favor of lower pre-tax returns in the noncorporate sector. There are other distortions -- economic inefficiencies -- that occur as well. On the other hand, the corporate tax is a source of revenue which has been viewed as contributing to the overall progressivity of the tax system. Full integration of the corporate tax also presents some administrative difficulties and would cause windfall gains and losses in different economic sectors.

The effective tax rates calculated in this study demonstrate the pronounced disparities produced by the current tax system. Effective tax rates on corporate equity capital are seventy percent higher than rates on noncorporate equity capital. Corporate debt is subject to a NEGATIVE tax rate even with modest levels of inflation. That is, use of corporate debt is effectively subsidized by the current tax system. Corporations which pay out all of their earnings in dividends subject their stockholders to individual taxes three times those paid by firms which pay no dividends. The calculations in this study suggest that, because of these tax differentials, most of the revenue yield from the current tax system is wasted, primarily in inefficient production, but also in financial distortions. The cost of economic distortions is estimated at 1.34 percent of consumption, while the extra yield from the corporate tax is only 1.38 percent of consumption. Although these measures of waste cannot be calculated with precision, economic waste from the current tax is so large as to present a strong case for reforming the tax.

The corporate tax does, however, contribute to progressivity of the tax system since higher income individuals have higher shares of capital income and since the burden of the corporate tax falls primarily on capital income. Financing the cost of corporate integration through increases in the progressive personal income tax or through increases in other capital income taxes would limit any redistributional effects.

Integration, particularly "full" integration, involves administrative and revenue costs. Administrative difficulties do not appear to be crippling, however, particularly if full integration were combined with elimination of capital gains tax on corporate stock. There are many partial and limited proposals, such as dividend relief and restrictions on interest deductibility which present no serious administrative cost and involve more limited revenue costs.

                              CONTENTS

OVERVIEW

The corporate income tax, introduced in 1909, has been the subject of considerable interest throughout its history. Economists have criticized the imposition of income taxes twice -- at the corporate and the individual level -- for causing inefficient resource allocation. On the other hand, the corporate tax is a source of revenue which has been viewed as contributing to the overall progressivity of the tax system. This latter effect is true to the extent the tax falls on capital income and as long as capital income makes up a larger fraction of income at higher income levels. Full integration of the corporate tax -- taxing corporations in the same manner as partnerships -- would also present some administrative difficulties.

Despite considerable attention, there is still a substantial amount of uncertainty about both the allocative and distributional effects. Some recent studies which rely on the "new view" of dividends suggest that the efficiency losses arising from differential taxation of capital income are smaller than traditionally thought. Other studies which account for the presence of both corporate and noncorporate production within industries suggest that these losses may be much larger. Moreover, there has been more attention devoted to financial distortions, such as those which affect leveraging.

The effects of the corporate tax on income distribution have also been the subject of debate. Closed economy models of the corporate tax suggest that the tax falls on capital income, but these results have been questioned for open economies engaged in international trade and investment. These results have also been questioned for growing economies, where the tax could affect the total savings rate; the standard models assume a fixed total capital stock. In addition, views of the distributional issue which take a life-time perspective suggest that the progressivity of the tax may be overstated by measuring it with reference to annual incomes.

There have been occasional attempts to enact policy changes which would reduce the distortions caused by the separate corporate tax. The Carter Administration explored the possibilities of a major revision. Revisions were also considered during the Reagan Administration. The Treasury's original proposals for tax reform in 1984 included a deduction for half of dividends paid and a proposal for indexing interest. These provisions were eroded and gradually dropped in the process of enacting the proposal, so that the Tax Reform Act of 1986 only reduced the tax rate. Lately, interest has surfaced again in this area, in part due to the increase in debt taken on by the corporate sector and accompanying concerns about failures of highly leveraged firms. The Treasury's Office of Tax Analysis has undertaken an extensive study of corporate tax integration.

The first section of this study explains how the corporate tax causes differential effective tax rates for different forms of investment. While the current system imposes a second layer of personal tax on profit remaining after the corporate tax, this effective personal level tax varies considerably depending on the dividend policies of the firm, and the frequency of realizations of gains on corporate stock. For a taxpayer in the 28 percent bracket, the personal level effective tax rate can vary from virtually zero to over 50 percent. Low tax rates are associated with investments in firms which pay low dividends and where assets are held until death. The higher tax rates are associated with high dividend payout, high inflation rates, and frequent sale of assets. Using a representative investment (allowing for some assets to be held through tax exempt forms such as pensions) and a 4 percent inflation rate, the overall effective tax rate on corporate capital is estimated at 42 percent. This rate is 70 percent higher than the estimated 25 percent rate on noncorporate investment.

Debt is taxed much more favorably than equity because nominal interest can be deducted and because the tax rate of the lender is lower than the corporate statutory rate. Even at a relatively low inflation rate of 4 percent, corporate debt is taxed at a NEGATIVE ten percent rate. Thus, the corporate tax actually provides a subsidy to debt financed investment. The total tax rate on corporate sector capital, reflecting a weight of the equity and debt tax rates, is estimated at 34 percent. This rate is some 60 percent higher than the overall rate in the noncorporate sector of 21 percent.

The next section of the paper examines some possible options to reduce the tax rate differentials, and estimates effective tax rates under these options. These options include full integration or partnership taxation, dividend relief at the corporate level, dividend relief at the individual level, and exclusion of both dividends and gains on corporate stock at the individual level. Since many of these approaches can use the corporate tax as a withholding device, the refundability of the withholding tax in the case of tax exempt investors can be important. Various methods of limiting the benefits to debt finance are also discussed, such as allowing partial disallowance of interest deduction or indexing of interest.

How important might the current tax distortions be? The next section of the paper addresses the magnitude of the economic inefficiencies created by the tax. This section examines five types of distortions: distortions in the allocation of resources between the corporate and non-corporate sectors, distortions in the debt/equity financing mix in the corporate sector, distortions in the dividend payout rate, the lock-in effect of the capital gains tax on corporate stock, and the savings distortion. Although there are many difficulties in estimating these inefficiencies, the results suggest that the waste induced by the corporate tax may well be as large as the extra revenues collected by the tax (both are slightly over 1 percent of output). Such a finding is not that surprising, since the corporate tax, while increasing taxes on corporate equity, decreases them on corporate debt. Thus, it leads to large differences in tax rates without necessarily increasing the overall tax rate very much. These large measures of waste suggest that the corporate tax is a very inefficient method of raising revenues.

One major concern about integration options is the role that the corporate tax plays in tax progressivity. The corporate tax is progressive if its burden falls primarily on capital income, since higher income individuals have larger shares of capital income than lower income individuals. The standard static closed economy modeling of the tax clearly indicates that the burden of the corporate tax is on capital in general. This view has been challenged for open economies engaged in international trade and investment and for growing economies. The analysis of these models still suggests that the burden of the tax falls generally on capital. The progressivity of the tax may, however, be less pronounced if lifetime income and tax are considered. It would not be difficult, in any case, to offset any important distributional consequences, since the revenue loss could be offset by other progressive tax increases, such as increasing individual income tax rates.

The administrative issues surrounding the tax are primarily of concern if a full integration -- partnership treatment -- approach is considered. Any administrative complexities could largely be avoided if the corporation is used to withhold the tax and if changes in liability on audit are allocated to the current year's owners. The full partnership treatment would also be simplified if the individual tax on capital gains were eliminated. Other methods of relief have few implications for administrative costs or could actually simplify tax administration and compliance.

The revenue cost of various reforms to the corporate tax, while large relative to budgetary deficit targets, is not large in comparison to output. A straightforward full integration proposal could cost close to $100 billion at FY 1992 income levels. A more neutral system could be introduced at a smaller cost by restricting deductions for interest paid by firms, and limiting credits for tax exempt entities, which could cut the cost by 40 percent. Such a change could be financed by a small increase in tax rates. Other less costly proposals could also yield efficiency gains.

THE CORPORATE TAX AND DIFFERENTIAL TAXATION

By imposing a separate corporate tax rather than taxing income directly to shareholders, the current corporate tax structure has a variety of differential effects on different types of income: corporate sector income is taxed differently from income in the non- corporate sector; dividends are taxed more heavily than retained earnings; and, because it can be deducted, debt is favored over corporate equity financing. Moreover, because so much of income from corporate equities accrues as capital gains on stock, the current corporate tax system contributes to the lock-in effect on capital gains (creates barriers to sale) by making the tax burden higher the shorter the period of time assets are held.

We consider first the tax treatment of corporate versus noncorporate equity income, and related issues: the differential taxation of dividends and retentions and variations in tax rate depending on holding periods. Secondly, we consider the effects of the tax treatment of corporate vs. noncorporate debt. Effective tax rates presented here take into account the timing of tax payments and the effects of inflation; they measure the difference between the pre-tax and post-tax real returns divided by the pre-tax real return. Technical details of the computation of corporate and noncorporate effective tax rates are contained in Appendix I.

TAXES ON CORPORATE EQUITY INCOME

Corporate profits are subject to tax at two levels -- once at the level of the firm and once at the personal level through dividends and capital gains taxes. Although this treatment, particularly of dividends, is commonly referred to as "double taxation," the degree of excess taxation compared to taxation of noncorporate income depends on several factors. For example, under current law the corporate tax rate is 34 percent and under current rules the effective tax rate is roughly consistent with that rate. $100 dollars of corporate income before tax will result in $66 of corporate income after tax. If the entire amount is distributed, an individual in the 28 percent bracket will pay a tax of $18.48 (28 percent of $66). The total tax will be $52.48. The tax on a similar dollar of noncorporate income will be $28. Thus, the tax rate on corporate sector income, if distributed, is not quite double the rate of the individual income tax. If the taxpayer is in the 15 percent bracket, the tax on corporate source income will be $43.50, well over twice the rate applied to noncorporate source income.

In some respects the corporate tax is regressive, since the extra tax imposed per dollar actually becomes larger for taxpayers with lower tax rates and incomes. The additional tax for the 28 percent individual is $24.48 (the difference between the total tax of $52.48 and the noncorporate tax of $28). For the 15 percent individual it is $28.50 ($43.50 minus $15), and for the zero tax rate individual it is $34.00. Although these amounts of extra tax rise as the individual's tax rate and income fall, the corporate tax is frequently viewed as progressive because higher income people tend to receive a larger share of their total income from capital income.

There are other factors which tend to mitigate the differences between tax burdens on corporate and noncorporate sector equity income. First, normally only a part of earnings are distributed; reinvested earnings can escape tax until realized as part of capital gains. These taxes are deferred and thus have a lower present value; moreover, earnings are not subject to tax at all if assets are held until the individual owner's death. When an asset is sold, however, the capital gains tax applies to the appreciation in value that is due to inflation as well as real growth. This capital gains tax, however, also applies to appreciation arising solely from inflation in the value of noncorporate assets, such as real estate. Thus, the effective tax rate can vary depending upon the share of earnings that is paid out in dividends, the part of the portfolio that is actually sold, the number of years that the portion sold is held, and the inflation rate.

Indeed, in the past, the generous tax treatment of retained earnings could lead high income individuals to use the corporate form as a shelter. When individual tax rates were as high as 70 percent, as they were before 1981, and the corporate tax rate was 46 percent, high income individuals could actually pay a lower tax by investing in corporations which retained most of their earnings, particularly if the shares were held until death. Since the corporate tax of 34 percent is now almost always higher than any of the individual tax rates, the tax in the corporate sector is always higher than that in the noncorporate sector.

Table 1 shows how the effective tax rate at the personal level varies as different assumptions are made about dividend payout and holding periods. If two thirds of assets are held until death, and enough income is paid out in dividends to allow steady state growth of 3 percent (i.e. 57 percent), and the holding period for assets sold is seven years, the overall effective tax rate at the personal level is about 23 percent with a four percent inflation rate and 19 percent at a zero inflation rate. Thus, the overall personal tax rate for this investment is about two thirds of the statutory rate. Note also the substantial variation in tax rates. For stocks which pay no dividends and which are held until death, the tax is zero. For stocks that are either sold frequently or which pay a lot of dividends the tax rate can be quite high -- in excess of 40 percent in some cases. Also, these tax burdens can be very heavily affected by inflation.

Table 2 accounts for the fact that investments in the corporate sector can be held indirectly by tax-exempt or largely tax-exempt entities such as pension funds and life insurance companies where the personal level tax does not apply on either dividends or capital gains. ¹ We estimate that about thirty percent of equities are held by the non-taxed entities. The statutory individual tax rate employed in this table is the estimated average marginal tax rate on capital gains (25 percent); with the adjustment for tax-exempt holdings the rate is 18 percent. In this case the effective tax at the personal level is cut in half on average. In such a case, corporate equity capital on average is taxed more than half again as much as noncorporate capital (165 percent of the noncorporate rate).

                               TABLE 1:

Finally, these tables show that the amount of deferral and forgiveness of tax through capital gains treatment can be quite important. If no assets are held until death and all assets are sold every seven years with a 57 percent dividend payout, the personal tax rate rises to 23 percent if all gains are realized, compared to 14 percent when 2/3 of assets are held until death. The increase is larger for firms which have low dividend rates. For gains that are realized, tax rates are higher the more frequently the asset is sold. For example, Table 2 indicates that if all assets are sold and there are no dividends, the effective tax rate rises from 18 percent to 25 percent as the holding period falls from 10 years to 3 years. These capital gains tax effects contribute to a lock-in effect, which is an incentive for individuals to retain existing assets.

Using reasonable assumptions (two thirds of capital gains held until death, the 57 percent dividend payout, and a seven year holding period, a four percent inflation rate, and thirty percent of assets held through tax exempts) the effective tax rate at the personal level in the aggregate is only 14 percent rather than the full statutory 25 percent rate. ²

The effective tax rate on overall equity capital in the corporate and noncorporate sectors is also influenced by incentives and penalties at the firm level. First, the tax law does not allow economic depreciation which would require that the proper rate of real decline be used and that depreciation be indexed for inflation. On one hand, the law does not allow indexing; on the other hand, it does allow accelerated recovery of costs. In addition, physical assets are frequently sold in the noncorporate sector and attract capital gains taxes due to inflation and accelerated depreciation.

The presence of these factors makes the effective tax rate at the firm level different than the statutory rate. These tax rates vary by asset, and therefore by industry, depending on the importance of the offsetting factors of accelerated depreciation and inflation. For a composite of total business investment in the economy, we estimate that the 34 percent statutory corporate tax rate results in an effective rate of 27 percent for a zero inflation rate, 33 percent for a 4 percent inflation rate, and 37 percent for an 8 percent inflation rate. Thus, the effective tax rate is close to the statutory rate for a four percent inflation rate, considerably less with no inflation, and somewhat more with an inflation rate of 8 percent. These relationships are similar for the lower tax rates which are likely to apply in the noncorporate sector; taking capital gains taxes into account increases tax rates in the noncorporate sector slightly. We estimate that a 25 percent statutory tax rate results in an effective rate of 19 percent at a zero inflation rate, 25 percent at a 4 percent inflation rate, and 28 percent at an eight percent inflation rate.

At the 4 percent inflation rate, the overall tax rate on corporate equity is estimated at 42 percent (.33 plus .14 times .66), about seventy percent higher than the noncorporate rate of 25 percent.

CORPORATE DEBT

The corporate tax also influences the level of corporate debt relative to corporate equity, and relative to debt in the noncorporate sector.

If the income base were economic income, debt financed investments are simply subject to the personal income tax rate because debt service is a deduction under the corporate income tax. Thus, they do not bear any double tax and the effective tax rate on debt, whether in the corporate or the noncorporate sector, is the same as the tax rate on noncorporate equity. This treatment of debt favors corporate debt over equity, but at the same time reduces the differential between the aggregate tax burden on noncorporate and corporate capital, since some corporate capital will be financed by debt.

However, two factors can cause the tax burden on corporate debt to be lower than that on noncorporate equity as well as different in the corporate and the noncorporate sector. First, as noted above, there are incentives and penalties at the firm level which cause the effective tax rate to depart from the statutory rate; these benefits and penalties also affect debt financed capital.

Secondly, and more importantly in the case of debt, firms are allowed to deduct not only the real portion of the interest rate (which represents a real cost) but also the inflation portion. These deductions would not matter if borrower and lender were subject to the same statutory tax rate, since the value of excess deductions to the borrowers would be offset by excess inclusion of income by the lender. If, however, the lender's tax rate is below the borrowers, as is the case in corporate borrowing from individuals, this feature provides a tax benefit to debt financed investment. Moreover, this benefit can accrue to noncorporate debt as well if some of the interest income is received via tax-exempt entities such as pension funds.

Table 3 illustrates the differential taxation of capital financed with equity and debt, and calculates the total effective tax rate assuming one third debt finance. The first segment of the table again illustrates the tax burden for a 28 percent individual tax rate in the noncorporate sector corresponding to the tax rate assumptions in table 1. The second segment is based on an estimated average marginal tax rate of 25 percent, but also adjusts for the holding of about thirty percent of financial assets by tax exempt institutions, for a tax rate of 18 percent on financial income. These assumptions correspond to the assumptions in table 2. (Both tables assume, for purposes of measuring the tax on equity, a 57 percent dividend payout ratio, two thirds of capital gains held until death, and an average holding period of seven years).

The results in this table show that the tax rate on corporate and noncorporate equity rises with the inflation rate but also that the tax rate on debt falls with inflation. Moreover, the tax rates can be negative; i.e., debt financed investment can actually be subject to a subsidy rather than a tax. For example, with 4 percent inflation and with tax-exempt entity participation, the tax rate on corporate equity capital in the economy is 42 percent, while the tax rate on corporate debt capital is MINUS 10 percent. The tax rate on noncorporate equity capital is 25 percent, while the tax rate on noncorporate debt is 19 percent, the latter reflecting tax-exempt entity participation. A weighted average of debt and equity indicates that overall corporate sector capital is subject to a 34 percent tax while overall noncorporate sector capital is subject to a 21 percent tax. These tax rates show that the distortion between corporate debt and corporate equity is pronounced, that there is also a distortion between noncorporate debt and equity, albeit smaller, and that corporate capital overall is taxed at a rate 60 percent higher than that of noncorporate capital.

Before turning to the question of the magnitude of the efficiency losses induced by the corporate tax, we consider methods of relieving double taxation and their effects on these differential taxes.

                               TABLE 3:

Because of the many imperfections in the current tax system, none of the standard methods of relieving double taxation would lead to a uniform tax treatment of all business capital income, but many would come rather close. ³

The main commonly cited methods of relief of double taxation are: partnership method or classical integration, dividend relief at the corporate level, and exclusion of dividends to individuals. There are various permutations of these basic forms that have to do with the treatment of tax exempt shareholders and whether tax is withheld at the corporate level. Some of these proposals are much simpler to implement than others. For example, a dividend exclusion or a deduction would be relatively straightforward. Most full integration systems and certain types of dividend relief systems would be more complex to administer.

There are also other types of options which might reduce tax differentials. For example, one could add to a dividend exclusion an elimination of the personal level tax on corporate capital gains. There are also various approaches to modifying the tax treatment of debt, such as disallowing a portion of the debt deduction or indexing interest payments.

Several different approaches are described below. Then, the effects of these alternative approaches on differential tax burdens are discussed.

METHODS OF DOUBLE TAX RELIEF

This section describes several approaches to revisions: full integration, dividend relief at the corporate level, dividend relief at the shareholder level, full relief at the shareholder level, dividend relief coupled with interest disallowance, and dividend relief coupled with interest indexation.

Full Integration

The purest method of relieving double taxation is the partnership method, in which income would be attributed to the shareholders as if they were partners. In the simplest variation, the corporate tax would simply be eliminated and the shareholders would include in their income a proportionate share of the corporation's earnings.

It is important, however, to note that even with this system, corporate and noncorporate capital would still not be taxed at equal rates since large tax exempt intermediaries such as pension funds could hold either debt or corporate stock. Since tax exempt intermediaries do not own noncorporate businesses, corporate equity would be thus favored over noncorporate equity.4 Other consequences are that corporate debt would be taxed at the same rate as corporate equity while noncorporate debt would be taxed at a lower rate than either, since interest would be deducted at a higher overall tax rate of the noncorporate equity owner.

Aside from eliminating tax exempt privileges, the other method of dealing with the tax exempts would be to allow a credit imputation system which withholds tax at the corporate level and credits it at the individual level, but does not allow refundability. The credit imputation system also deals with the problem of individuals having to pay tax on income which they do not actually receive (retained earnings). For taxable individuals, both systems have the same result. For example, suppose the individual tax rate is 25 percent and the corporation distributes half of its pre-tax income and invests half.

Under the straight partnership system, for each dollar of earnings, the firm would pay a dividend of $.50 and reinvest $.50. The individual would receive $.50 of income and also count as income the $.50 of retained earnings for $1 of income, which would generate $.25 in tax. The dollar would be split up into $.50 of reinvestment, $.25 of tax, and $.25 of after tax disposable income of the individual. The basis of the stock would be increased by the amount of retained earnings so that these earnings would not be taxed again when the stock is sold.

Under a credit imputation system with a corporate tax rate of 34 percent, the corporation would reinvest $.50, pay a tax of $.34 and distribute the remainder of $.16 in dividends. The individual shareholder would include in his taxable income after tax income (both dividends and retained earnings) of $.66 grossed up by the tax. (Technically post tax corporate income would be multiplied by u/(1- u), where u is the corporate tax rate, and that amount would also be included as income). This amount would total to $1. Individuals would then be charged a tax of 25 percent but receive a credit of $.34, for a net credit of $.09. The final outcome is not changed: the firm reinvests $.50, the government collects $.25 (the $.34 collected from the corporation minus the net credit of $.09 paid to the individual), and the individual has after-tax income of $.25 (the sum of $.16 in cash dividends and $.09 in net tax credits).

The outcome of the credit imputation system would be different from a classical system if the credit were nonrefundable. A tax- exempt entity would pay zero tax under the straight partnership system. It would pay a 34 percent tax under the imputation credit system if the credit is nonrefundable since the full tax at the firm level would apply. Thus, with a credit imputation system which is nonrefundable, and a 34 percent corporate withholding rate, corporate equity would be taxed at higher rates than noncorporate equity assuming pensions and similar institutional investors still wanted to hold corporate stock. With a straight partnership approach or a credit imputation system with refundable taxes, corporate equity is still favored since, in a practical sense, pensions and institutional investors can hold corporate equity and not noncorporate equity.

There is, therefore, no ideal solution to this issue which would fully equalize the treatment of corporate and noncorporate firms, as long as tax-exempt entities maintain their preferential tax treatment. Moreover, no system of integration could completely deal with the potential differential distortions between debt and equity as long as the tax system does not index for inflation and remains progressive, since interest may not necessarily be deducted and included at the same tax rate. Some forms of integration would mitigate these differences however, and indexation of interest would be a potential revenue enhancing adjustment.

Full integration would equalize the treatment of dividends and retentions. Moreover, it would mitigate the effect on tax rates of holding assets versus selling them frequently since retained earnings would be taxed currently and the basis increased, leaving smaller amounts of income to be taxed as a capital gain. Indeed, a case could be made for eliminating the capital gains tax on corporate stock entirely under full integration, since most of the gain would be due to inflation and since taxing gains would be much more complicated under an integrated system. ⁵

Dividend Relief at the Corporate Level

Since dividends tend to be more heavily taxed, double taxation relief proposals often concentrate on relief for dividends rather than capital gains. One relatively straightforward method of providing for dividend relief would be to allow corporations to deduct all or a portion of dividends from taxable income. Indeed, this is the method of double taxation relief which is used in many other countries.

Under a dividend relief approach, retained earnings would be subject to both the corporate tax and to the personal capital gains tax. Thus, unlike full integration, the dividend deduction does not eliminate the distortion between dividends and retentions. Indeed, while current law favors retentions over dividends, dividend relief would favor dividends over retentions.

Like full integration, dividend relief at the corporate level can also be accomplished through a credit imputation system which would retain the corporate tax for tax-exempt recipients. Credit imputation systems are frequently used in other countries, in part to impose taxes on foreign investors. The credit imputation system works just as it does in the case of full integration, except that it applies only to dividends.

Unlike full integration, dividend relief does not prevent the significant tax advantages to debt financed investment, as long as sufficient taxable income remains to cover interest deductions. Interest is still deducted at the full statutory tax rate. Thus, while a dividend relief proposal narrows the differential between debt and equity, it does so only by reducing the tax rate on equity; full integration lowers the rate on equity and raises it on debt.

In addition, the dividend relief proposals, unlike full integration, do not relieve the taxation of capital gains and the holding period distortions associated with the capital gains tax.

Thus, while dividend relief can substantially reduce the differential between corporate and noncorporate investment, it does not perform as well in relieving the distortions between corporate debt and equity or between retentions and dividends. One advantage is that it avoids many of the administrative problems associated with full integration.

Dividend Relief at the Shareholder Level

Another way to provide dividend relief is to eliminate the tax on dividends at the individual level. Because individual tax rates are, on average, below the corporate rate, such a proposal would be less generous than dividend relief at the corporate level, even when a credit imputation system which does not allow benefits to tax exempts is included.

Simply eliminating the dividend tax at the individual level is perhaps the simplest approach administratively; indeed, it would simplify rather than further complicate the tax law. Because it tends to be less generous it would be less likely to create much of a distortion in favor of dividends, but would do less to reduce the debt/equity distortion.

As compared to a credit imputation system, dividend relief at the individual level would have somewhat different effects on distribution. The current corporate plus personal level tax on dividends is slightly progressive with respect to dividends, being composed of in part a flat rate tax at the corporate level and a progressive individual rate. A zero tax rate individual pays an approximate rate of 34 percent; a 15 percent tax rate individual a rate of 45.8 percent, a 28 percent tax rate individual a rate of 52.5 percent. With a dividend deduction at the firm level, the rates are lower but progress at their statutory rates: 0, 15 percent, and 28 percent. (For an imputation credit system without refundability, a zero tax rate individual would pay 34 percent). With an exclusion at the individual level, the rates are all set at 34 percent. This point may not be of great importance, however, since dividends are highly concentrated at the top rates and since most zero rate taxpayers are holding assets on behalf of taxable individuals (via pension funds for example).

Full Relief at the Shareholder Level

Another option is to have relief on tax on corporate income at the individual level, by excluding dividends and capital gains on corporate stock. This approach essentially eliminates the tax at the individual level while retaining it at the firm level, and thus effectively integrates the tax by providing a single level of tax. Such an approach would be more generous than the dividend exclusion alone, would provide more neutrality among debt and equity, would be neutral between retentions and dividends, and would eliminate the holding period distortions for capital gains on corporate stock. It would also be simple. Like the dividend exclusion, however, it would result in a less progressive tax than standard integration, and it would result in an overall higher level of tax than classic integration.

There are numerous variants on this idea. For example, extraordinary capital gains and losses on corporate stock could still be subject to tax by increasing the basis of corporate stock for real retained earnings and by indexing the basis for price changes. The average results would be the same but unexpected gains and losses would still be subject to tax. Other variants could be considered, such as adjusting the tax basis but not indexing for inflation. One of the drawbacks of this type of change is that it would require some complex accounting for basis adjustment which would increase administrative complexity.

Dividend Relief and Interest Disallowance

One of the problems with all of these changes is that they would tend to be costly in terms of revenue. It is nevertheless possible to improve the neutrality of the tax between debt and equity and dividends and retentions without changing the overall tax burden by simultaneously allowing partial dividend relief and disallowing a portion of interest deductions. The revenue raised from disallowing a share of interest deductions would be used to make up the revenue lost from the share of dividends deducted. This type of change would do relatively little for the aggregate distortion between corporate and noncorporate production but would reduce other distortions (that between debt and equity and between dividends and retentions).

Dividend Relief and Interest Indexation

This approach is related to the previous notion in that it would simultaneously reduce the tax on corporate equity income, but increase the tax on corporate debt. Indexing of interest deductions could extend, however, beyond corporate debt to other forms of debt. At the same time it would be appropriate to include only the real return in income. For example, if the nominal interest rate is 10 percent and the inflation rate is four percent, only sixty percent of interest paid would be deductible and only sixty percent of interest received would be taxable. This indexation of interest could be extended throughout the economy, to apply to noncorporate debt and perhaps to mortgage interest as well. It could be coupled with a variety of dividend relief approaches.

It might, of course, be difficult to institute a broad indexation proposal. An alternative would be to disallow a small portion of interest which, on average, would make up for the advantage of deducting the inflation portion of interest at a high rate and taxing it at a low rate. For example, assuming that the effective tax rate at the firm level is close to the statutory rate, the portion excluded would be (u-t)/(u(1-t)) times the share of the nominal return reflecting inflation. For the tax rates of u = .34 for the corporation and u = .25 for the noncorporate firm, t = .18, and for forty percent of the return representing inflation, 23 percent of corporate interest would be excluded and 13 percent of noncorporate interest would be excluded.

EFFECTS OF ALTERNATIVE OPTIONS

These alternative approaches would have markedly different effects on the types of tax rates experienced by each type of asset. Table 4 reports effective tax rates on debt, equity, and aggregate tax rate, assuming current patterns of shares of income realized as dividends, capital gains realizations patterns and inflation rates. These rates don't account for replacing lost revenues if any, but what is important is the relative tax rates.

Under current law, the estimated effective tax rate on noncorporate capital is 21 percent and the rate on corporate capital is 34 percent, with the corporate rate some 61 percent higher. Moreover, corporate equity pays a rate of 42 percent, while debt pays an effective rate of MINUS 10 percent. Partnership taxation would roughly equalize the treatment of corporate equity and debt and would actually lower the overall tax rate to a lower rate than noncorporate treatment -- 18 percent. This lower rate occurs because of the holding of corporate equity, as well as debt, through tax-exempt financial intermediaries.

Not allowing the financial intermediaries their zero tax rate, via a non-refundable credit imputation system would result in an overall tax rate of 23 percent, slightly above the noncorporate rate, and while it would reduce the differences between debt and equity, it would not eliminate them. Intermediate positions between these two approaches could be obtained by lowering the corporate tax rate. For example, by lowering the corporate tax rate to 25 percent, the noncorporate results (listed in an addendum) would be obtained -- 21 percent tax rate, with a 25 percent rate on equity capital and a 10 percent rate on debt capital. (Recall that noncorporate debt is subject to a lower tax rate because of the holdings of tax-exempt entities. When there is inflation, the effective rate is further lowered, due to the advantage of deducting the inflation portion at a high rate and taxing it at a lower rate) While equalizing the tax treatment of corporate and noncorporate firms, it would maintain some tax differential between debt and equity.

                               TABLE 4:

Revisions could be proposed which would affect debt without the complex administrative apparatus of full integration. Indeed, one proposal made by Congressman Gradison was to consider a revenue neutral change which would combine a dividend deduction with a disallowance of interest payments. The amount of interest disallowance would depend on the share of dividends that could be deducted. As the table indicates, a full dividend deduction would go too far, and result in a higher tax rate on debt than on capital. It appears that a 70 to 80 percent dividend deduction, combined with disallowing one half of interest, would tend to more or less equalize the taxation of both debt and equity at roughly current tax levels. Of course, if revenue losses were allowed, one could lower the overall tax rate. For example, allowing a full dividend deduction, but disallowing a third of interest deductions would produce a 25 percent tax rate on both debt and equity. One could also allow restrictions on debt in concert with other types of dividend proposals such as the credit imputation approach or the dividend exclusion. This type of proposal could be made very flexible and has the advantage of simplicity.

Another approach to dealing with the debt issue is to index interest. Much of the benefit of deducting interest at a high tax rate relates to the ability to deduct the full nominal interest at high income levels. With indexation, only the real portion of the interest rate would be deducted -- and only the real portion would be included in income. As shown in table 4, this treatment would produce a 17 percent tax rate on corporate debt, and it could be combined with many other types of dividend proposals. Noncorporate debt could be covered as well, with tax rates rising on debt.

Table 5 examines the effect of these potential revisions on the dividend retentions choice, assuming that two thirds of gains are held until death and a seven year holding period. Under current law, if all real income is retained by the corporation, the tax rate at the personal level is 7 percent, while if all income is paid out the tax rate is 20 percent, suggesting a substantial incentive in favor of retentions. Under classical integration, there is no differential regardless of refundability, so that neutrality between dividends and capital gains is achieved. (Tax rates are all personal level tax rates holding the corporate rate constant for comparability across proposals; for that reason tax rates can be negative reflecting the credits for corporate taxes paid).

                               TABLE 5:

Assessing these alternatives is difficult. If the revenue cost and administrative simplicity were not of concern, the proposal which would appear most neutral would be integration via a credit imputation system with no refundability to tax-exempt shareholders, but with the corporate tax rate lowered to 25 percent, coupled with interest indexation or a partial disallowance of interest. If administrative simplicity were a high priority, but revenue loss not important, the elimination of taxes on both capital gains and dividends at the personal level along with a lower corporate tax rate and/or a restriction on the deductibility of interest would certainly lessen many of the economic distortions. If revenue were important, relief to corporate equity capital could be offset by restrictions on tax deductions for interest payments, perhaps in the noncorporate sector as well as the corporate sector.

CORPORATE TAXATION AND ECONOMIC EFFICIENCY

It is on the grounds of economic efficiency that economists usually argue in favor of relieving the double taxation of income. As shown earlier, the corporate tax results in disparate taxes on corporate vs. noncorporate capital, debt vs. equity finance, and dividends and retained earnings, and stocks which are sold frequently relative to those held for a long time or until death. The higher taxes on capital income as a result of the additional corporate tax may also contribute to a distortion between debt and equity, although the distortionary effects of alternative taxes to replace revenues would need to be considered.

The cost of distortions is usually referred to as a welfare loss or an "excess burden" of the tax. For a tax which interferes with productive efficiency, it is the value of lost production. For taxes which distort prices or financial choice it is a measure of how much individuals would need to make them as well off as they would be with a nondistorting tax. Generally the measure of welfare loss is proportional to the square of the tax; thus, the distortion per dollar of revenue rises as the tax rises.

Distortions are larger the greater the response of altering investment decisions to differential tax rates. These responses are often expressed in terms of elasticities. An elasticity is a percentage change in quantity (or ratio of quantities) divided by a percentage change in price (or ratio of prices). High elasticities are, therefore, associated with large percentage changes in quantity relative to a give percentage change in price. Empirical estimates of these elasticities are important to assessing the welfare effects of tax distortions.

CORPORATE VS. NONCORPORATE CAPITAL

The potential waste of economic resources from distorting the allocation of capital between corporate and noncorporate sectors has historically been the major issue of concern about the efficiency effects of the corporate tax. Some thirty years ago a formalized treatment of this effect was proposed by Arnold Harberger, and this analysis remained the central focus of concern about the corporate tax for many years. ⁶ In Harberger's model, the corporate tax causes capital to flow out of the corporate sector, driving down the pre-tax rate of return in the noncorporate sector and driving up the return in the corporate sector. This process continues until the after tax returns are equated in the two sectors. The tax results in inefficient production because the sectors use a distorted mix of capital and labor. It also results in inefficient consumption because the prices of corporate products are increased relative to the prices of noncorporate products.

A larger, more sophisticated version of the Harberger model has been constructed and refined over the years. ⁷ Despite many innovations in modeling the tax, static welfare losses continued to be estimated at about the same magnitude as the earlier models, around .5 percent of GNP or less.

Some economists began to question the applications of the Harberger model because the model is based on tax differentials across industries while the corporate tax is based on legal form of organization. ⁸ It was increasingly clear that the Harberger model had overlooked the coexistence of corporate and noncorporate firms in the same industry, in some cases producing completely identical goods (e.g. crude oil). In order to have these firms coexist both with and without taxes, there must be some advantages and disadvantages to each form. Gravelle and Kotlikoff proposed an explanation for this behavior which was based on scale advantages for corporate firms and entrepreneurial skill for noncorporate firms; their model suggested much larger distortions from the corporate tax. ⁹ In an application to a complex numerical model, this type of model produces a welfare gain from establishing uniform taxes of over one percent of consumption, twice the estimate which would be derived from the Harberger model. ¹⁰ Most of this gain derived from uniform treatment of corporate and noncorporate firms within an industry; a small part derived from providing more equal treatment of owner occupied housing with business capital. ¹¹ As a share of GNP this gain from uniform taxation is about .85 percent.

These estimates suggest that the potential welfare gains from more uniform taxation of capital in different sectors are significant. This is particularly important given that the extra revenues collected from the corporate tax (over the amounts collected through a partnership treatment) is currently less than 2 percent of consumption (as discussed below). Thus, the cost of collecting the tax in wasted physical resources may be more than one half the revenue from the corporate tax.

THE DISTORTION BETWEEN DEBT AND EQUITY CAPITAL

Another type of distortion which is affected by the corporate tax is the distortion between debt and equity capital within the corporate sector. It is not clear what leads corporations to choose shares of debt and equity. The most common argument for limiting the amount of debt is the risk of bankruptcy; if firms are committed to paying out all or much of their earnings they have little or no cushion to protect them against hard times. One argument for having some debt, however, is to discipline the managers by requiring them to meet some continual commitments. In addition, individuals may have preferences for holding different kinds of assets with varying degrees of risk, and the mixture of financial claims allows risk to be allocated in accordance with preferences.

The tax system tilts the preference towards debt; indeed, the largest differentials in tax treatment are in the differential treatment of debt versus equity in the corporate sector, where the effective tax rate on equity is 42 percent and the rate on debt is MINUS 10. Thus, even with a limited amount of substitutability of debt for equity, the degree of distortion imposed by the tax system could be quite large. Although there has been considerable attention focused on the debt vs. equity issue, there has been little attempt to measure the cost of these distorting effects. Gordon and Malkiel ¹² estimated that the welfare loss from this distortion based on measures of bankruptcy cost was $3.2 billion in 1975. This amounts to about .24 percent of consumption and about .2 percent of GNP. If their estimates are reasonable for current law, another sixteen percent of the revenue collected from the tax may be wasted in the distortion between debt and equity.

Moreover, this excess burden from the debt/equity distortion is like an additional tax on corporate sector capital, suggesting that the cost of the distortion between corporate and noncorporate capital is understated.

DIVIDENDS VS. RETENTIONS

The lack of attention to measuring the potential efficiency gains from reducing the distortion between debt and equity can also be found in examining the differences between dividends and retentions. Part of the reason for this lack is that attention has been diverted to a long standing disagreement in the literature regarding the effect of dividend taxes. One strand of the literature argues for a traditional view of dividends, where the tax on corporate equity reflects the taxes paid on dividends and capital gains as well as the corporate tax. Another strand of the literature argues for a "new view" which essentially says that dividend taxes do not matter for the cost of capital. ¹³ If this theory is correct, there would be no distortion between dividends and retentions. Moreover, the effective tax rates would be lower on corporate equity, reducing the distortions between corporate and noncorporate taxes and between corporate debt and corporate equity. Moreover, proposals which provide only dividend relief would have no effect on the cost of capital. Thus, this issue could be crucial to evaluating policy options.

The basic notion of the new view of dividends is that the burden of an additional tax on dividends for existing firms is capitalized into the incomes of existing shareholders. (This tax could be imposed at any level, so long as it only applies when income is paid out in dividends; it could be thought of as the excess of the dividend tax over the effective capital gains tax). For example, suppose the corporation earns $100 after corporate tax and the shareholder pays a tax of 25 percent. The corporation has two choices -- to pay out the dividend, with the taxpayer receiving $75 net of tax, or to invest the dollar. Suppose the dollar is invested and earns an after- corporate-tax rate of return of 10 percent, for an income of $110 in the following year. The dividend can then be paid out, with the taxpayer earning $82.50. If the 25 percent tax really had an effect the shareholder should receive a 7.5 percent rate of return. But the shareholder's rate of return on the forgone net-of-tax dividend is 10 percent since $82.50 is 1.1 times $75.00.

This treatment is similar to the treatment of an Individual Retirement Account (IRA). By making the investment via retained earnings the individual avoids the current dividend tax, just as if he had taken a deduction on his tax return; but he has to pay the tax later when he does realize a dividend. The present value of these taxes cancel out so that the net return is not taxed. If the individual instead received the dividend, and reinvested this after- tax dividend in a corporate share, he would receive a 7.5 percent return on his investment.

The basic notion underlying this analysis is that the individual's dividends are trapped in the firm at the time a tax is imposed or lifted; for this reason the new view of dividends is also called the "trapped equity" view. Proponents of this trapped equity view would argue that repealing the dividend tax would not increase the after tax return (which would remain at 10 percent) and not encourage additional investment; rather it would cause the value of the stock to rise. Thus, any tax change would be capitalized in asset values.

This theory does not mean that firms never pay dividends. In an economy where the rate of return is higher than the growth rate some share of corporate income will be consumed in the aggregate, and corporations must eject some cash out of the corporation. Rather it means that the dividend tax does not affect the marginal return to an investment financed via retained earnings. This new view does, however, rely on the notion that individuals are indifferent between receiving their incomes in the form of dividends versus sale of shares.

The major difficulty with the new view is that it is incompatible with observed behavior. First, a corporation should never simultaneously issue new stock and pay dividends, because, under the new view, it is cheaper to reduce the level of dividends and finance with retained earnings. Income on capital raised through issuing stock is affected by the dividend tax. Yet, firms have been observed to issue stock while still paying dividends. Secondly, firms have always been able to increase the flow of funds to stockholders by repurchasing their own shares on the open market rather than paying dividends. Individuals can obtain the same income by selling shares back to the corporation. Indeed, corporations have engaged in a lot of share repurchase during the past few years.

This behavior suggests that individuals do not view dividends and capital gains in the same way. A number of reasons have been advanced to explain this phenomenon. One view is that dividends act as a signal to apprise individuals as to the profitability of the firm. Another is that regular expected dividends impose constraints on the behavior of the managers. This concern may be important given the separation of ownership and control in large corporations. A final one is that individuals may prefer either dividends or capital gains income to limit transactions costs and variability in the market. An individual who prefers a steady stream of income would have to be continually selling assets to realize that stream if few dividends are paid, while an individual who is increasing the size of his portfolio would have to be continually buying assets if dividends are too large. U.S. corporations offer a wide diversity of income oriented and growth oriented investments.

A heavier tax on dividends then, would distort these choices and result in an efficiency loss. Most econometric research has suggested that the response of dividends to tax changes is fairly pronounced, and James Poterba, in a recent paper, estimated an elasticity of 1.57 for dividends with respect to the ratios of after tax shares from dividends and capital gains. ¹⁴ A rough calculation of the implied excess burden suggests that it is about .04 percent of consumption. ¹⁵

THE LOCK-IN EFFECT OF CAPITAL GAINS TAXES

Another distortion to which the current tax system contributes is the lock-in effect of capital gains taxes on corporate stock. Because a capital gains tax is assessed only if the stock is sold, there is an incentive to keep one's current portfolio of stocks even if other investment opportunities appear more attractive either because of higher return, less risk, or some combination of risk return attributes. If corporate sector income were taxed on a partnership basis, reinvested income would not be subject to a capital gains tax, since basis would increase by these amounts. Capital gains in excess of these increases could still be subject to tax, and as long as inflation is present the typical investment would still appreciate. Thus, part of the lock in effect would remain if capital gains are still subject to tax, due to the tax on inflationary gains.

The behavioral response to the capital gains tax cut has been extensively addressed through more than a dozen econometric studies. Unfortunately, these studies have produced a dramatic range of realizations elasticities (percentage change in gains divided by percentage change in tax), ranging from below .5 to as high as 4. ¹⁶ A recent study has, however, pointed out the limits to this feed back effect, which is constrained by the level of accruals; that is, many of the econometric studies produce realizations which would far exceed accruals. ¹⁷ In this calculation, we use a value halfway between zero and the absolute upper limit to estimate the magnitude of these effects as calculated in that study.

Our estimate of the excess burden arising from capital gains taxation of corporate stock is derived in Appendix II. It is calculated only for capital stocks held by individuals, since no lock-in effect due to income taxes occurs for tax exempt entities. The results indicate that the excess burden of the full capital gains tax on corporate stock is .06 percent of consumption. About .04 percent of consumption could be gained from more efficient sale of corporate stock if partnership taxation were adopted while still retaining the capital gains tax on inflation.

THE INTER-TEMPORAL DISTORTION

The excess corporate tax, like any other capital income tax, contributes to a decreased rate of return. This decreased rate of return causes a distortion in the choice between consumption today and consumption in the future -- the inter-temporal choice. The evaluation of this distortion is, however, a little more complex than the other distortions because correcting for this distortion will lose revenues. With the allocative distortions, one could imagine that revenues could be raised by increasing taxes on other types of capital income. The inter-temporal distortion can only be reduced by reducing marginal tax rates on capital income.

In considering inter-temporal efficiency effects, it is necessary to consider how the revenues are to be made up. The other source of income in the economy is labor. One source of increased revenue for a large tax change may be an increase in income taxes, of which labor income constitutes the larger proportion.

Thus, an interesting question to ask is what would be the efficiency effects of substituting labor for capital income taxes. Theoretically, this substitution may produce either a gain or a loss, since labor income taxes distort the choice between labor and leisure (or, to be more precise, between leisure and other goods purchased with money).

Gravelle considers a variant of the Auerbach-Kotlikoff overlapping generations, endogenous labor model, and finds that full replacement of wage for income taxes using their choice of elasticities produces a small efficiency gain -- .1 percent of consumption. ¹⁸ Despite the large shift in tax bases, there is only a small effect because the gain from reducing the inter-temporal distortion is largely offset by a loss from increasing the distortions between labor and leisure due to higher taxes on wage income. Full corporate tax integration would replace only about a third of the capital income tax; this proposal would actually yield a larger gain of .23 percent. (Replacing only part of the capital income tax can be more efficient than replacing it entirely, because tax burdens rise with the square of the tax rate. These results simply suggest that the optimal capital income tax is less than the current tax but not zero.)

There are, however, two reasons that this number may be too large. First, the "inter-temporal elasticity of substitution" used, .25, may be too high. ¹⁹ More recent studies suggest that the value may be considerably lower, perhaps less than .1. ²⁰ Even more compelling, however, is the stability of the savings rate over long periods of time; the savings rates shifts, particularly in the short run, predicted by the models with the .25 elasticity are not very consistent with this history. At this lower inter-temporal elasticity, there is essentially no gain in welfare from corporate tax integration (the gain of .01 percent of consumption).

The second reason that the gain may be high is that the gain from reducing the inter-temporal distortion is considerably less for individuals already alive and with shorter time horizons while the loss from the labor/leisure distortion, at least for the non-retired part of the population, appears in full force. Thus, the appearance of gains in the long run steady state, which may take fifty years or so to reach, may not represent an overall gain to society.

The basic thrust of these findings is that the role of the corporate income tax in raising overall taxes on capital relative to labor income has uncertain but probably small effects on economic efficiency. These efficiency issues, therefore, should probably not be considered paramount in a choice of how to finance corporate tax integration. It is primarily the distributional issues which are of concern in determining methods to finance corporate tax integration.

SUMMARY OF EFFICIENCY ISSUES

This survey of efficiency issues suggests that there is potentially a lot of waste associated with the corporate income tax. From numbers suggested here, for example, the corporate-noncorporate distortion is measured at 1 percent of consumption, the debt/equity distortion at .24 percent of consumption, the dividend-retained earnings distortion at .04 percent of consumption and the capital gains distortion at .06 percent of consumption, for a potential total of 1.34 percent of consumption.

These numbers do not compare favorably with the extra revenue yield from the corporate tax. For example, in 1988, the corporate tax raised about $91 billion at the corporate level (corporate receipts net of State corporate taxes and Federal reserve earnings). $78 billion of individual dividends and an estimated $58 billion of capital gains, for a total of $136 billion, produced individual tax revenue at a 25 percent tax rate of $34 billion. Thus total revenue collected was $125 billion ($91 billion plus $34 billion). If corporations were to be taxed the same as corporate equity, the corporate tax rate should be 25 percent rather than 34 percent, and the tax should occur at only one level. Total corporate collections would have been only $67 billion ($91 billion times .25/.34). Thus the extra revenue yield from the corporate tax is $58 billion ($125 billion minus $67 billion), which amounts to 1.38 percent of consumption. That means that the cost of economic distortions from the tax was 97 percent the size of the tax revenue.

It is this type of comparison between the welfare effects and the revenue yield which constitutes the basic case for altering the corporate tax. The ratio might be somewhat more favorable because corporate tax collections for 1988 were low because the accumulation of large tax depreciation deductions from the 1981-1986 liberalized depreciation rules. Moreover, debt shares had climbed somewhat in the eighties and might subside in the future. And the welfare effects estimated could be smaller. But even if the ratio were only half as large, the implication is that for every dollar of additional tax collected, half is thrown away through economic distortions.

CORPORATE TAX INCIDENCE AND DISTRIBUTIONAL ISSUES

THE GENERAL VIEW

The distributional effects of the corporate tax have been the subject of considerable debate. First, because the corporate tax rate is proportional and the personal rate applies to earnings net of the corporate level tax, and because the corporate form allows for deferral of the personal tax on earnings retained in the corporate form, the additional burden of the tax declines relative to corporate sector income as income (and personal tax rate) rises. For example, for a taxpayer with a statutory 31 percent marginal tax rate (where the effective personal rate on dividends and capital gains is about 25 percent), the combined corporate and personal tax rate is 51 percent (.34+(1-.34).25). This rate is some 20 percentage points higher than the 31 percent tax rate which would apply to noncorporate capital. For the taxpayer in the 15 percent bracket (where the effective personal rate is about 12 percent), the effective combined rate is 42 percent, or a rate about 27 percentage points higher. For the zero percent taxpayer, the extra tax is 34 percentage points. In this sense, even though the corporate rate is a proportional tax on corporate income, the EXTRA tax tends to be slightly regressive relative to corporate equity income. If all individuals had the same proportion of corporate income to total income and there were no behavioral responses to the tax, the effect of the corporate tax system would be regressive.

Neither of these conditions obtain. Indeed, the corporate tax may be progressive, and is frequently characterized as so. Higher income individuals tend to earn much more of their income from capital income, especially corporate equity capital, and a tax on capital is therefore a progressive tax. While it results in a smaller additional tax relative to capital income as income rises, it involves a larger share of total income as income rises. For example, the Congressional Budget Office reported that capital income constituted about 17 percent of overall income, but accounted for only 7 percent of income in the lowest income decile and 29 percent of income in the highest income decile, representing 48 percent of income of the top one percent. ²¹

This observation of the relationship of the corporate tax to income, via shares of corporate equity income, is not, however, a complete analysis of the effects of the corporate tax for several reasons. First, there are price effects which occur when a corporate tax is imposed. As the rate of return falls, individuals move their investments into the noncorporate sector where the rates of return are higher. This movement of capital tends to raise the rate of return in the corporate sector where capital is now scarcer, and lower it in the noncorporate sector where it is more abundant. This response tends to spread the burden of the tax to owners of noncorporate capital as well. These shifts in capital allocation also result in shifts in the allocation of labor and the composition of corporate and noncorporate goods. The tax can potentially affect rates of return to capital before tax (in both corporate and noncorporate sectors), wage rates, and relative prices of products. Tracing these changes in income is necessary to determine the incidence of the corporate tax -- that is, whose income is reduced (either via taxation or changes in prices) by the tax.

There are two strands of literature on the subject of the incidence of the corporate tax. First, in many distributional studies, alternative allocations of the corporate tax to capital income, labor income, or prices were used. These alternative allocations yield quite different depictions of the distributional effects of the tax. ²²

These particular choices are, however, not those which would be suggested by the other major strand of literature, which involves assessing the incidence of the tax through a general equilibrium model of the economy. The Harberger analysis of the tax showed how the incidence of the tax depended on the ability of firms to substitute labor and capital in the production process (the factor substitution elasticity) and the willingness of consumers to substitute products in consumption (the product substitution elasticity). ²³ An important result of this analysis is that under reasonable assumptions about these responses, the corporate tax appears to fall on capital in general (both corporate and noncorporate). Varying assumptions within reasonable parameters suggest that capital can bear more or less than 100 percent of the tax (and labor can either gain or lose). In general, however, the deviation from the assumption of capital bearing exactly 100 percent of the tax is not too great. ²⁴

There have been some problems with the Harberger model in that the model does not allow corporate and noncorporate firms to produce in the same industry. Gravelle and Kotlikoff propose two alternative ways to deal with this problem -- through differentiated technology (the Mutual Production Model) or differentiated products (the Differentiated Product Model). ²⁵ In their studies of a two industry economy, they found that the differentiated product model produces results quite similar to the Harberger Model, while the Mutual Production Model produces somewhat different effects. In both cases, the basic conclusion of the Harberger model -- that the corporate tax probably falls on capital in general -- is maintained. ²⁶ Thus, this type of analysis suggests that the corporate tax is progressive since capital income is more concentrated at higher income levels. ²⁷

Although these types of findings are relatively straightforward, there have been numerous issues raised about whether one can close the book on the incidence of the tax. Some concerns, such as the presence of monopoly power, have no important implications for the incidence analysis; indeed, they make it even more likely that the tax will fall on capital income in the long run.

Of more concern are the assumptions of fixed capital stocks. The fixed capital stock assumption can be inappropriate either because the imposition of the tax discourages the flow of capital from abroad or because the reduction in tax increases savings. Both the open economy and the dynamic economy assumptions have the potential for altering the expected incidence and distributional effects of the tax.

INCIDENCE IN AN OPEN ECONOMY

It has been asserted that the corporate income tax would, in an open economy engaged in worldwide trade and investment, largely fall on labor. Certainly, in a country where capital investment can be imported from foreigners and where domestic firms and individuals can invest abroad, the domestic capital stock is no longer fixed. Thus, the incidence results derived from the fixed capital stock models discussed above could change considerably.

Joseph Pechman and Arnold Harberger have both suggested that the tax would be largely borne by labor in an open economy. ²⁸ The argument is roughly as follows. In an open economy, imposition of a corporate tax causes capital to migrate abroad until the net return in the economy is restored to the worldwide rate. The burden of the tax will therefore fall on the fixed labor input which is unable to migrate, since the smaller size of the capital stock will make labor less productive. Harberger's discussion extends to considering corporate and noncorporate sectors and tradable and nontradable goods, but reaches the same conclusion.

An exploration of this view in a simple two country model suggests that such a view of the corporate tax in an open economy is unwarranted. First, the U.S. corporate income tax is not imposed solely on the basis of location; U.S. firms wishing to repatriate income or operating in branch form must pay any U.S. tax in excess of the foreign tax rate. Indeed, with a fully residence based tax and with the taxing country a capital exporter, we might expect similar results as in a closed economy. In practice, the U.S. tax is partly residence based and partly source based, since taxes on income reinvested abroad can be deferred.

Even with a fully source based tax system, the view that the burden of the tax falls on labor is not justified. The burden of the tax falls entirely on domestic labor only when (1) capital is perfectly substitutable across countries, (2) the country involved is very small relative to the rest of the world, and (3) there is an infinite substitution elasticity between imports and domestic production of the good. Clearly, none of these assumptions are correct. Relaxing these assumptions can produce cases where there are no effects on labor or where labor income can increase.

Indeed, the analysis in Appendix III suggests the following. If the substitution elasticity between domestic and foreign investment of capital is zero, the corporate tax seems likely to fall on U.S. capital income. If the substitution elasticity is infinite, the tax falls on worldwide capital income; U.S. capital income falls by an amount proportionate to its share of capital. Since the U.S. owns about 30 percent of the worldwide capital stock, about 30 percent of the tax falls on U.S. capital income. The effects on labor income net to zero, but the effects in different countries depend on the substitution elasticity between domestic and imported goods. If that substitution elasticity is infinite, U.S. labor income falls by seventy percent of the tax and foreign labor income rises by seventy percent of the tax, with a net zero effect on total foreign income. As the substitution elasticity between products falls, the tax burdens U.S. labor less, and can indeed benefit it. In this latter case, the corporate tax may be largely exported abroad, would redistribute income from high income to low income individuals in the U.S., but could redistribute from lower income to higher income individuals abroad. Thus, the open economy assumption does not produce a strong case for an incidence on labor income.

Finally, this analysis does not consider the possibility of a reaction from the foreign country. If other countries respond to the imposition or removal of a tax by enacting similar policies, the incidence of the tax will be likely to fall on worldwide capital, wage rates will be likely unaffected, and the incidence will be the same as in the closed economy.

There are several reasons, therefore, to expect that the corporate income tax is unlikely to fall on labor income in an open economy. First, not all of the corporate income tax is applied on a source basis. Secondly the United States is not a small country; indeed it owns close to a third of the world's capital stock. Thirdly, there is clearly a less than infinite substitution elasticity between capital flows. For example, the elasticity of substitution for equity capital appears to fall in the range of 1 to 4. ²⁹ At an elasticity of 4, U.S. capital in our calculations would bear 64 percent of the tax rather than 30 percent; at an elasticity of 1 domestic capital would bear 85 percent of the tax. Moreover, recall that the corporate tax increases the burden for equity but decreases it for debt. If debt is more mobile than equity internationally, as we might expect, then the overall response of international capital flow to the imposition of the corporate tax seems likely to be even smaller. These observations suggest that the tax incidence is quite similar to the incidence in a closed economy model -- i.e. it falls primarily on domestic capital income. Finally, even with some exporting of the burden of the tax to other than domestic capital, it is not likely that the residual tax burden will fall on domestic labor. Rather it seems more likely to fall on foreign capital and labor.

INCIDENCE IN A GROWTH CONTEXT

The second important modification of the incidence and distributional results has to do with growing (dynamic) economies. In the Harberger and similar models, capital is fixed. The argument is then made that increasing the corporate tax (and thus the overall tax on capital) causes the capital stock to contract and causes the wage rate to fall, thus causing the tax to fall on labor as well as capital.

It is not clear, however, that imposing a corporate tax depresses savings. Most evidence suggests that the savings response is relatively small, and indeed that savings could go up. In a simple constant savings elasticity model, even with a relatively large savings response, a calculation of these effects suggests that it is unlikely that corporate integration financed by higher income tax rates would increase the economy's output in the long run by very much.

To explore these effects, consider some simple calculations with such a model, outlined in Appendix IV. The total tax rate on capital income is estimated at 30 percent. ³⁰ Substituting partnership taxation would lower the overall tax rate to approximately 20 percent. That is, the excess of the corporate tax over the partnership alternative accounts for about a third of the total tax. With a capital income share of .25, this assumption would set the excess corporate tax at 2.5 percent of NNP (Net National Product). The standard static incidence result would be that capital incomes rise by approximately ten percent. Of course, the total distributional effects in the economy would depend on what other changes occur when the tax is reduced -- government spending, other taxes, or debt. If the savings elasticity is zero, the static results are those which would obtain.

Calculations with this model suggest that, even with a relatively large savings response, it is unlikely that long run income would be increased by much over one percent due to the savings response. This dynamic effect is not large relative to the direct effect on capital income (which is an increase of ten percent). This increase would also accrue to labor income, since the calculations use the common assumption of a fixed income shares Cobb Douglas production function. ³¹ And, of course, theory and evidence suggest that the savings effect might well be negative. In other words, both capital and labor income rise when savings increases due to the corporate tax change.

Moreover, more sophisticated life cycle models suggest that the effects of this tax replacement are quite modest, and likely to be negative. For example, simulations using life cycle models indicated that income would actually FALL by about .2 to .3 percent with corporate integration financed by income tax increases (See Appendix IV). This effect occurs because there is a windfall benefit to older individuals from repealing the tax on existing capital and these individuals are less likely to save. Other things equal, old individuals are characterized by a higher ratio of capital income to total income, while young individuals have a relatively low ratio. Thus, replacing the corporate tax with an increased income tax would increase the tax rate on wages and decrease it on capital, shifting income from the young to the old. The effect on the redistribution of income from high savers to low savers tends to offset the incentive effects of the tax. If the tax were replaced by a consumption tax, such as a value added tax, the effects on long run income would be likely positive (but also small) because a consumption tax tends to burden older individuals who may be spending out of assets. If the tax were financed by government borrowing, the effects would be a more pronounced negative.

While this life cycle analysis suggests that the savings response is probably not very great, it does raise some important issues about distribution. The first issue is whether the income distribution from a cross section is a good representation of lifetime income. When we look at annual incomes, there is always the possibility that the increasing share of capital income as income rises may be reflecting, in large part, life cycle effects. For example, income from wages as well as capital tends to be low earlier in life when little capital has been accumulated, and higher in the later working years. Thus, the rising share of capital income as income rises (which leads to the characterization of the corporate tax as progressive) could become less pronounced, or even absent, if one could compare individuals with different incomes of the same age.

It is difficult to assess this effect because of lack of data. An assumption is occasionally made that a flat rate capital income tax as well as a consumption tax would be proportional rather than, respectively, progressive or regressive, if account were made of life cycle earnings. This is surely far too simplified. There are a number of reasons to expect that individuals with higher levels of permanent lifetime income receive proportionally more of that income from capital. One is that very wealthy families are likely to transmit large stocks of wealth from one generation to the next. Indeed, wealth is far more highly concentrated than could be explained by a life cycle model; the top one half of one percent of the income distribution own 19 percent of assets. ³² In addition to bequests, individuals with lower wages would be expected to save less for retirement because public income security programs are likely to replace more of their income than is the case for high income individuals.

This dynamic analysis does, however, suggest that there is another distributional aspect of the corporate tax which should be considered, and that is the distribution across generations. Thus, a reduction in capital income taxes will benefit older individuals relative to younger individuals. The overall inter-generational redistribution depends on how revenues are to be made up. If they are made up through an increase in income tax, then there will be a net benefit to the old and a net burden on the young. This distributional effect across generations is a transient effect, but is an important consideration when evaluating a change in tax rules.

SUMMARY OF DISTRIBUTIONAL ISSUES

The analysis here suggests that the corporate tax is progressive, although its progressivity with respect to lifetime incomes may be more modest than its progressivity based on a single cross section. This distributional aspect of the corporate tax would not, however, prevent integration from being considered if maintaining current progressivity were paramount. If offsetting revenues are raised through increases in taxes on other capital income or increases in income taxes, the current distribution of the tax burden could be maintained. The distributional effects become, however, more problematic if other sources of tax revenue which are not in themselves progressive, are tapped, or if spending is cut. Social security and excise taxes, the other remaining revenue sources, tend to be regressive, and many spending programs tend to primarily benefit lower income individuals.

A change in tax rules would also induce some generational shifts, again depending on how revenues were to be made up. These generational shifts are quite difficult to offset unless the corporate reform were financed by raising taxes on other sorts of capital income (such as indexing or partially disallowing interest deductions). Raising income tax rates would be the same as substituting wage for income taxes and would tend to burden young and future generations relative to older ones. Raising taxes on consumption would tend to burden the old relative to a tax on wages or income at lower income levels, but not necessarily at higher levels. Data suggest that the assets of lower income individuals tend to be lower on average during retirement than in the ten years preceding retirement (retirement defined as age 65), suggesting that individuals are consuming their assets and that consumption is higher than income. The assets of higher income individuals tend to rise. ³³

It is also important to consider the distribution of the benefits of relieving the excess burden of the corporate tax, particularly if these benefits are large relative to the direct cost. The benefits associated with reallocation of physical capital would be distributed proportionally to consumption, and thus would tend to favor lower income individuals. The benefits from lessening the financial distortions would be associated with capital income and would tend to favor higher income individuals. Since most of the estimated gains have to do with physical reallocation of investment, these excess burden gains tend to be more concentrated towards lower and moderate income individuals than the direct revenue benefits. In addition, the efficiency gains would tend to accumulate slowly over time, given an adjustment process; thus, these benefits are largest for young and future generations. Thus, while the direct reduction in revenues may favor higher income individuals and older individuals, the efficiency gains tend provide substantial benefits to lower and moderate income individuals. Moreover, these efficiency gains will be greater for younger and future generations.

ADMINISTRATION AND COMPLIANCE

One of the standards by which a tax change is evaluated is the administrative and compliance costs of a change. These administrative issues are most problematic with full integration, where all income is attributed to the individual. It would not be practical to treat such income as a partnership in the fullest sense. For example, if the corporate tax were revised on audit, it would be very costly to adjust the tax payments of the thousands of prior year shareholders; these changes would simply have to be attributed to the shareholders at the time of the change. There is also a potential problem of imputing income not actually received to individuals and requiring a tax, which would be the normal partnership treatment. For these reasons, the administratively practical way of providing for full integration would be to keep the corporate tax in place as a withholding device and provide an imputation of income and a credit to individuals. The only administrative change would be that attached to a dividend payment would be a statement of the total earnings to be reported and the amount of tax withheld, to allow individuals to figure their credit.

If the personal level capital gains tax on corporate stock were repealed, there would be no further administrative issues involved. If this capital gains tax were kept in place, there would be an additional computation -- basis would have to be increased by the amount of the imputed pre-tax income not paid out as dividends. The retention of a capital gains tax could involve significant costs in administration and compliance.

The dividend relief proposals present little in the way of complexity; indeed, many countries have adopted dividend relief methods. For that matter, eliminating the personal level tax on dividends would simplify administration and compliance as would eliminating personal level tax on dividends and capital gains from corporate stock. ³⁴ Similarly, disallowing a portion of interest deductions would be straightforward. Indexing interest deductions would be somewhat more complex since the share of interest that reflects inflation varies across interest rates. In practice, it would probably be more practical to prescribe a percentage disallowance for the year based on projected or actual inflation and interest rates. The approach of disallowing interest deductions entirely and not taxing them at all would be relatively simple.

REVENUE CONSEQUENCES

One of the underlying issues surrounding potential corporate tax integration approaches is the problem of how to pay for the revenue loss. Approaches, however, vary considerably in their costs.

If we desire to avoid income redistribution, the optimal way of paying for corporate tax integration is to raise taxes on other types of capital income. The efficiency analysis does not suggest that cutting overall capital income taxes per se would necessarily be more efficient. If taxes were reduced in the corporate sector, they would have to be raised in the noncorporate sector or on owner occupied housing. While some revenue might be raised in the noncorporate sector by restricting the deduction of interest in some fashion, this source of revenue increase is limited. The other major source of capital income is owner-occupied housing. In theory, taxes could be increased by imputing rent to owner-occupants. Restrictions on interest deductibility might be more practical.

The revenue consequences of various alternative options depend on the type of proposal considered. Given the presence of tax exempts it is difficult to obtain a full non-distortionary tax. The following provide general magnitudes of various alternatives. These estimates are based on data for 1988 (although they are inflated to 1992 income levels); they might be somewhat larger in the future because 1988 was a year when cumulated excess tax depreciation was quite high. The long run effect also depends on how debt shares settle down; they rose rapidly during the eighties and it is not clear whether they will subside or increase further. These estimates should be considered quite rough; more precise estimates could be obtained through a micro simulation. They are presented for the purpose of providing some general notion of magnitude.

Table 6 reports these estimates in FY 1992 levels of income, which requires an increase of 25 percent in the magnitudes of the losses. The most costly of the alternatives is a simple partnership treatment, with no personal level capital gains tax on corporate equity. In this case the single level tax on corporate equity would be $48 billion ($91 billion times .18/.34), and the revenue loss would be $77 billion, adjusted to $96 billion at FY 1992 levels. Partnership taxation while retaining the capital gains tax in place would lose slightly less. Assuming that the basis adjustment would reduce the effective tax rate on individual nominal gains of $58 billion from .25 to .1425, total taxes would be $56 billion for a total revenue loss of $69 billion, or $86 billion at FY 1992 levels.

The imputation credit system which provides relief only to taxable stockholders would result in a larger single level tax on corporate equity of .2757 (assuming the thirty percent of equity associated with tax exempts is taxed at 34 percent, and the remainder at 25 percent). This treatment would result in a loss of $56 billion ($70 billion at FY 1992 levels). If a capital gains tax remained in place, the loss would be $44 billion ($55 billion at FY 1992 levels). Such a system could also be coupled with an indexation of interest (assuming $100 billion of nominal interest and half of that disallowed on the deduction side and not taxed on the individual side), for a total gain of $4 billion. This approach would reduce the revenue loss to $40 billion ($50 billion in 1992 levels); indexation of noncorporate interest would reduce the overall loss to slightly over $38 billion ($48 billion in 1992 levels).

                               TABLE 6:

Another option which has the merit of being very simple would be to eliminate all personal level taxes on both dividends and capital gains. This approach would cost $34 billion ($43 billion at 1992 income levels).

One of the criticisms of the dividend relief option and of relief at the personal level is that it leaves tax advantages to debt intact. Some offsetting provisions which increase taxation of debt could be considered, such as indexing debt (raising $8 billion, assuming half of interest reflects inflation, or $10 billion at 1992 levels), or disallowing some portion of interest deductions. For example, if the tax rate for deducting debt were set at 18 percent by disallowing a portion of interest deductions or substituting a credit, $16 billion would be raised ($20 billion at 1992 levels). That proposal, for example, coupled with exclusion at the individual level, would result in a $23 billion overall revenue loss ($43 billion minus $20 billion). It would keep the overall effective tax rate in the corporate sector at around 34 percent on average. Additional revenues of about $4 billion would be raised by extending the same treatment to noncorporate firms. This approach would equalize the tax treatment of debt and equity in both sectors, equalize the treatment of dividends and capital gains, and eliminate the lock-in effect. It would, however, not go very far in equalizing the treatment of corporate and noncorporate firms; in order to achieve that result, the corporate tax rate would also have to be lowered to 25 percent on average.

Indeed, the closest we could get to a neutral system would be to use an imputation credit method for full integration, setting the corporate rate at 25 percent and deducting interest at a rate of 18 percent for both corporate and noncorporate businesses. This deduction rate could be achieved by either allowing a tax credit for interest rather than a deduction, or by disallowing part of the interest deduction. This approach should on average tax all sources of business income (debt and equity in both corporate and noncorporate sectors) at roughly equal rates. At 1992 income levels, the imputation credit system with a 25 percent corporate withholding rate would cost $73 billion, with $9 billion offset through corporate debt deduction restrictions (for a total of $64 billion). Another revenue gain of about $4 billion could probably be obtained by restricting interest deductions for noncorporate firms. Another possible source of revenue offset might be obtained by converting the mortgage interest deduction into a credit at a lower rate or disallowing it entirely.

We can clearly find on this list of proposals items which would cost very little; indeed the combination of a partial dividend deduction and disallowance of interest could be chosen to be revenue neutral. This type of change would improve some of the efficiency costs associated with financial costs, particularly the distortion against debt and equity. But any proposal which is financed solely out of offsetting receipts within the corporate sector cannot do much for the major source efficiency loss -- the corporate noncorporate distortion.

These revenue costs are, however, reasonably small relative to GNP. Consider, for example, the imputation credit system with the corporate tax rate set at 25 percent, and restrictions on interest to set the deduction rate equal to the average lenders rate. This scheme would cost $60 billion, but it is only one percent of GNP. If we wished to offset these revenues through higher income tax rates an increase of 2 percentage points at the 15 percent rate and of 3 percentage points at the top rate would provide an adequate revenue offset. Although this rate would increase labor income taxes it would decrease overall capital income taxes.

There are other sources of revenue which might provide a closer distributional offset. For example, eliminating interest deductions on owner occupied housing would raise about $40 billion. Since capital income taxes favor older, higher income individuals, increasing the share of social security taxed might be a distributionally related offset. Raising the share taxed to 85 percent would increase revenues by about $13 billion. While these changes might not be entirely feasible, more modest approaches such as capping the mortgage interest deduction or increasing the share of social security benefits taxed might be possible revenue sources.

There are also some approaches that are related as far as generational distribution although not necessarily because of income distribution. Excise taxes or value added taxes tend to fall on older people who are benefitted by the reduction in capital income taxes; if some way of countering the regressivity of these taxes with respect to income level could be found, such taxes might be a reasonable offset for the cost of corporate tax integration.

CONCLUSIONS

Perhaps the central conclusion suggested by this analysis is that the present system of corporate taxation results in highly variable tax rates, depending on the physical and financial form of investment. These distorted tax rates in turn lead to efficiency losses for the economy that appear to be quite large relative to the revenue collected. If these measures of welfare cost are correct, then each dollar of revenue collected by the corporate tax results in a waste of approximately a dollar in welfare loss, making the current tax system highly inefficient.

The corporate tax does, however, contribute to progressivity in the tax system, although perhaps not to the degree indicated by the basic cross section income data. Thus, to obtain the efficiency gains without sacrificing progressivity, revenue offsets would need to be obtained by raising capital income tax rates elsewhere or through increased individual income taxes, particularly at the higher income levels.

Administrative issues do not appear to be a major barrier, although full integration schemes would be much simplified if capital gains taxes on corporate stock were eliminated entirely.

[Note: Appendix I, Appendix II, and Appendix III contain equations which are unsuitable for on-line production and, therefore, have been omitted.]

² These assumptions are similar to those used in other studies and are consistent with a realizations to accruals ratio of about 25 percent. Recent evidence suggests that the fraction of assets never sold may be somewhat smaller historically. See Jane G. Gravelle, Limits to Capital Gains Feedbacks, Congressional Research Service, Library of Congress, Report 91-250 RCO. Changing this would have only small effects on the tax rates.

³ Note that even with a uniform treatment of business capital income, consumer durables such as owner occupied housing would still be favored because the imputed rent is not subject to tax.

⁴ It is possible for tax exempt entities, such as pension funds, to own certain types of passive noncorporate investments, such as real estate. There is no evidence that this kind of activity is common.

⁵ Capital gains taxes on noncorporate capital may be largely due to inflation as well. The capital gains tax is, however, partially offset by the restarting of depreciation deductions.

⁶ See Arnold C. Harberger, The Incidence of the Corporation Income Tax. Journal of Political Economy, vol. 70, June 1962, pp. 215-40 for the basic model of the tax, which addressed incidence. The calculation of welfare effects appeared in his Efficiency Effects of Taxes on Income from Capital, in the volume Effects of the Corporation Income Tax, ed. Marian Krzyzaniak, Detroit, Wayne State University Press, 1966.

⁷ See John B. Shoven and John Whalley, A General Equilibrium Calculation of the Effects of Differential Taxation of Income from Capital in the U.S., Journal of Public Economics, Vol. 1, November 1972, pp. 281-322; John B. Shoven, The Incidence and Efficiency Effects of Taxes on Income from Capital, Journal of Political Economy, Vol. 84, No. 6, December 1976, pp. 1261-83; Don Fullerton, John B. Shoven, and John Whalley, Replacing the U.S. Income Tax with a Progressive Consumption Tax, Journal of Public Economics, Volume 20, 1983, pp. 3-23; Charles L. Ballard, Don Fullerton, John D. Shoven, and John Whalley, A General Equilibrium Model for Tax Policy Evaluation, Chicago, University of Chicago Press, 1985; Don Fullerton and Yolanda K. Henderson, A Disaggregate General Equilibrium Model of the Tax Distortions Among Assets, Sectors and Industries, International Economic Review, Vol. 30, pp. 391-413; Don Fullerton, Yolanda K. Henderson, and James Mackie, Investment Allocation and Growth Under the Tax Reform Act of 1986. in Compendium of Tax Research, Office of Tax Analysis, Washington D.C., U.S. Government Printing Office, 1987.

⁸ See Jane G. Gravelle, The Social Cost of Nonneutral Taxation: Estimates for Nonresidential Capital, in Depreciation, Inflation, and the Taxation of Income from Capital, edited by Charles Hulten, Washington, Urban Institute, 1981; Liam P. Ebrill and David G. Hartman, On the Incidence and Excess Burden of the Corporation Income Tax, Public Finance, 37, No. 1, 1982, pp 48-58.

⁹ Jane G. Gravelle and Laurence J. Kotlikoff, The Incidence and Efficiency Costs of Corporate Taxation When Corporate and Noncorporate Firms Produce the Same Goods, Journal of Political Economy, August 1989, vol. 97, 749-790.

¹⁰ These results are reported in Jane G. Gravelle, Differential Taxation of Capital Income: Another Look at the 1986 Tax Reform Act, National Tax Journal, Vol. 62, December 1989, pp. 441- 464. A slightly smaller gain of .9 percent of consumption would occur with an alternative model which disaggregates the real estate industry into single family and multi-family.

¹¹ A loss of about .1 percent of consumption can be attributed to the benefits to owner occupied housing relative to noncorporate taxation.

¹² Roger H. Gordon and Burton G. Malkiel, Corporation Finance, In How Taxes Affect Economic Behavior, in Henry J. Aaron and Joseph A. Pechman, Washington, D.C., The Brookings Institution, 1981.

¹³ The underlying theory of the new view can be found in several articles. See Alan J. Auerbach, Wealth Maximization and the Cost of Capital, Quarterly Journal of Economics, vol. 93, August 1979, pp. 433-46 and Share Valuation and Corporate Equity Policy, Journal of Public Economics, Vol. 11, 1979, pp. 291-305; David Bradford, The Incidence and Allocation Effects of a Tax on Corporate Distributions, Journal of Public Economics, February 1981, pp. 1-22; and Mervyn A. King, Public Policy and the Corporation, London, Chapman and Hall, 1977. For a general overview of the traditional and the new views, see James M. Poterba, Tax Policy and Corporate Saving, Brookings Papers on Economic Activity, Vol. 2, 1987, pp. 455-515.

¹⁴ See James M. Poterba, Tax Policy and Corporate Saving, Brookings Papers on Economic Activity, vol. 2, 1987, pp. 456-457. The after-tax share is 1 minus the tax rate.

¹⁵ This calculation uses the triangular approximation for welfare: .5 * E * (dp)(squared) Q/p, where p is the ratio of one minus the tax rate on dividends to one minus the tax rate on capital gains. As reported in table 5, the tax rate when all income is paid out as dividends is 20 percent; when all income is paid as capital gains it is 7 percent. The value of p is .86, dp is .14 and E is equal to 1.57. The measure is calculated for 1988 when dividends reported on tax returns (Q) were $78 billion (and assumed to represent about three quarters of dividends), and consumption $4214 billion.

¹⁶ See Jane G. Gravelle, Can A Capital Gains Tax Cut Pay for Itself?, Library of Congress, Congressional Research Service Report 90-161 RCO, March 23, 1990, for a survey and critique of these studies.

¹⁷ Jane G. Gravelle, Limits to Capital Gains Feedback Effects, Library of Congress, Congressional Research Service Report 90-250 RCO, March 15, 1991.

¹⁸ Jane G. Gravelle, Income, Consumption, and Wage Taxation in a Life Cycle Model: Separating Efficiency from Redistribution, forthcoming American Economic Review, September 1991.

¹⁹ The inter-temporal substitution elasticity is the percentage change in relative consumption amounts in any two time periods divided by the percentage change in the relative prices. The relative price for consumption T years into the future is 1/(1+r)(to the T power), where r is the after tax rate of return and 1 is the price of current consumption.

²⁰ Robert E. Hall, in Inter-temporal Substitution in Consumption, Journal of Political Economy, Vol. 96, April 1988, pp. 339-357, argues that the inter-temporal substitution elasticity is probably below .1. He also finds some flaws in the studies which found much higher elasticities.

²¹ Congressional Budget Office, the Changing Distribution of Federal Taxes: 1975-1990, October 1987, p. 67.

²² For example, Joseph Pechman in Federal Tax Policy (Brookings Institution, Washington, D.C., 1987) used two allocations: all of the tax to capital income and half to capital income and half to prices. The Congressional Budget Office has used two cases: all of the tax allocated to capital and all of the tax allocated to labor. See The Changing Distribution of Federal Taxes: 1975-1990, October 1987.

²³ Arnold C. Harberger, The Incidence of the Corporation Income Tax, Journal of Political Economy, vol. 70, June 1962, pp. 215-140.

²⁴ For example, 100 percent of the tax falls on capital when there is a unitary substitution elasticity between labor and capital and a unitary substitution elasticity between products. Gravelle and Kotlikoff report in their study of a simplified two sector model the following incidence results. If the factor substitution elasticity is doubled, 112 percent of the tax falls on capital (and labor gains by 12 percent); if the factor substitution elasticity is halved, 82 percent of the tax falls on capital and 18 percent on labor. If the factor substitution elasticity is 1 in the corporate sector and .5 in the noncorporate sector, the share falling on capital is 108 percent; if the reverse is true the share falling on capital is 73 percent. Lower product substitution elasticities tend to increase the shares. In general, elasticities in the noncorporate sector would be likely, if anything, to be lower than those in the corporate sector because much of noncorporate production is in housing. Moreover, it seems unlikely that any of the elasticities are much above one. Thus, a range of 82 percent to 119 percent seems to cover most reasonable elasticity assumptions in the Harberger model. (In these models, to allocate the corporate tax to price is not very meaningful, since it is equivalent to allocating the tax proportionally to labor and capital according to their income shares; since such an approach would allocate the tax primarily to labor, it is not consistent with the results.) See Jane G. Gravelle and Laurence J. Kotlikoff, The Incidence and Efficiency Costs of Corporate Taxation when Corporate and Noncorporate Firms Produce the Same Good. Journal of Political Economy, Vol. 97 (August 1989)

²⁵ See Jane G. Gravelle and Laurence J. Kotlikoff, The Incidence and Efficiency Costs of Corporate Taxation when Corporate and Noncorporate Firms Produce the Same Good. Journal of Political Economy, Vol. 97 (August 1989); Does the Harberger Model Greatly Understate the Excess Burden of the Corporate Tax? -- Another Model Says Yes. National Bureau of Economic Research Working Paper No. 2742, 1988.

²⁶ As noted above, the range in the percentage of tax falling on capital for reasonable assumptions of elasticities for the Harberger model is 82 percent to 119 percent for a simple two good model. The range, for the same data, in the mutual production model is somewhat larger -- 58 percent to 147 percent. For the differentiated product model the range is 78 percent to 117 percent, similar to that of the Harberger model.

²⁷ In the traditional Harberger model as applied to U.S. data, a lower product substitution elasticity tends to raise the share of the burden falling on capital, while lower factor substitution elasticities lower it, although the former effect is relatively unimportant. The share of the tax borne by capital is smallest when factor substitution elasticities are low in the corporate sector and high in the noncorporate sector. This same pattern applies in the differentiated product model, where corporate and noncorporate firms within an industry produce slightly different goods. In the mutual production model, the share of tax borne by capital tends to be larger when factor substitution elasticities are low and when product substitution elasticities are low. Like the other models, however, a large (more than 100 percent) share of the tax falling on capital is associated with a lower elasticity in the noncorporate sector relative to the corporate sector.

²⁸ Joseph A. Pechman, Federal Tax Policy, The Brookings Institution, Washington, D.C., 1987, p. 144; Arnold C. Harberger, "The State of the Corporate Income Tax," in New Directions in Federal Tax Policy for the 1980's, Ed. Charles E. Walker and Mark Bloomfield, Cambridge: Ballinger Publishing Co., 1983, p. 166.

²⁹ See Kan H. Young, The Effects of Taxes and Rates of Return on Foreign Direct Investment in the United States, National Tax Journal, vol. 41, March 1988, p. 113 and N. R. Vasudeva Murthy, The Effects of Taxes and Rates of Return on Foreign Direct Investment in the United States: Some Econometric Comments, National Tax Journal, v. 42, June, 1989, p. 207.

³⁰ Jane G. Gravelle, Differential Taxation of Capital Income: Another Look at the Tax Reform Act of 1986, National Tax Journal, Vol. 62, December 1989.

³¹ This production function assumes a unitary elasticity of substitution between labor and capital; that is the percentage change in the ratio of capital to labor is equal to the percentage change in the ratio of wage rate to rate of return. One result of this common assumption is that the shares of income are fixed; when the capital stock goes up the rate of return falls, and the wage rate rises just enough so that the share of total income accruing to capital stays fixed.

³² See Financial Characteristics of High Income Families, Federal Reserve Bulletin, March 1986, pp. 165-175.

³³ This pattern is shown in Financial Characteristics of High Income Families, Federal Reserve Bulletin, March 1986, pp. 163-212.

³⁴ This discussion does not deal with the problems of eliminating the capital gains tax in the noncorporate sector, where the adjustment could be complex because of the repetition of depreciation.

³⁵ The loss from the dividend deduction takes account of both the loss of corporate revenues and the additional individual taxes assuming that all pre-tax dividends are distributed. This total loss is $111 billion divided by (1-.34) to obtain pre-tax dividends, multiplied by .34 for the corporate revenue loss. Dividends to individuals will rise by the amount of the tax savings and this increase will be taxed at an average rate of 18 percent. Thus, the overall revenue loss is $111 billion times .34 times (1-.18), divided by (1-.34).

³⁷ Jane G. Gravelle and Laurence J. Kotlikoff, The Incidence and Efficiency Costs of Corporate Taxation When Corporate and Noncorporate Firms Produce the Same Good. Journal of Political Economy, August 1989.

³⁸ Jane G. Gravelle, Income, Consumption and Wage Taxation in a Life Cycle Model: Separating Efficiency from Redistribution, forthcoming, American Economic Review, September 1991.

³⁹ Lawrence Summers, Taxation and Capital Accumulation in a Life Cycle Growth Model, American Economic Review, Vol. 71, September 1981.

⁴⁰ The results employ the elasticities chosen from the literature by Alan Auerbach and Laurence J. Kotlikoff in their book, Dynamic Fiscal Policy, Cambridge: Cambridge University Press, 1987. The inter-temporal substitution elasticity (the elasticity of substitution between present and future consumption) is set at .25.

⁴¹ Alan Auerbach, Laurence J. Kotlikoff, and Jonathan Skinner, The Efficiency Gains from Dynamic Tax Reform, International Economic Review, Vol. 24, February 1983. This model requires an intra-temporal substitution elasticity (elasticity of substitution between consumption and leisure), which is set at .8.