CRS REPORT WEIGHS EFFICACY OF INVESTMENT INCENTIVES.

                          Jane G. Gravelle

There has been recent interest in providing a new tax subsidy for investment, primarily motivated by concerns about reviving the economy. These proposals include investment tax credits and various types of accelerated depreciation, in some cases on a temporary basis and in others on a permanent one. This study describes current law and the various proposals, provides a brief history, and assesses the economic issues of tax neutrality, efficacy as a fiscal stimulant, and effects on long-term growth. Some special issues associated with an incremental tax subsidy are also addressed.

The investment credit produces tax distortions across assets depending on coverage and durability of the assets covered. Under current law, tax rates are reasonably even across most depreciable assets, with the extremes ranging from 22 to 39 percent for corporate investments. With a ten-percent credit, tax rates would range from MINUS 56 percent to 39 percent. These distortions cause a misallocation of capital which can cause some efficiency losses. In some cases, these distortions would be magnified further by limiting the credit to certain industries. Tax benefits are more likely to produce more evenhanded tax rates if they are allowed in the form of deductions rather than credits. There are other types of incentives which are more neutral across investments, such as partial expensing.

Overall, the economic research does not give very high marks to the use of investment incentives as counter-cyclical tax devices, because of the long lags in responding to these incentives. Although a temporary provision would have more impact than a permanent one, this planning lag would also cause a delayed response to a temporary credit.

Investment incentives are often argued to promote savings and growth. Yet there is little economic evidence to support expectation of a large effect on capital formation. Moreover, the degree of capital formation obtained will, even under assumptions conductive to a large effect, have a modest impact on future standards of living and growth rates over the next few years.

While incremental credits have attracted some interest because they can stimulate investment at a much smaller revenue cost, there are a number of serious problems with these credits if enacted on a permanent basis. The incremental credit's effect on economic markets has not been studied extensively, but a preliminary assessment suggests that the credit could cause large declines in the value of on-going firms and create powerful incentives to avoid the incremental nature of the credit by eliminating firms from the market, creating new firms, and engaging in lumpy investment. The credit also tends to magnify business cycles. An incremental credit can be unfair to firms that have recently made large investments and overly generous to firms with unusually low recent investment levels. There are some serious administrative problems with the incremental credit including its relative complexity, and how to deal with divestitures, start-up firms, partnerships, leasing, and used assets. Some of these problems would occur with a temporary credit as well.

                              CONTENTS

TAX SUBSIDIES FOR INVESTMENT: ISSUES AND PROPOSALS

There has been recent interest in providing a new tax subsidy for investment, primarily motivated by concerns about reviving the economy. These proposals include investment tax credits and various types of accelerated depreciation, in some cases on a temporary basis and in others on a permanent one. This study describes current law and the various proposals, provides a brief history, and assesses the economic issues of tax neutrality, efficacy as a fiscal stimulant, and effects on long term growth. Some special issues associated with an incremental tax subsidy are also addressed.

DESCRIPTION OF CURRENT LAW AND PROPOSALS

The two most common ways of providing tax subsidies for new investment have been through accelerated deductions for depreciation and through investment tax credits. Accelerated depreciation speeds up the rate at which the cost of the investment is deducted. Investment tax credits provide for a credit against tax liability for a fraction of the purchase price of assets.

In order to tax investments in depreciable assets at an effective tax rate which is consistent with the statutory tax rate, tax depreciation must correspond to economic depreciation, the latter being the decline in the value of the asset. Such a depreciation deduction allows the original cost of the investment to be recovered tax-free and measures, and taxes, the return to investment in each period of the asset's use.

In practice, tax depreciation is based on a set of rules which apply to specific sets of assets (either of a particular type or used in a particular industry). These rules specify the period of time over which deductions should be taken and the method of allocating those deductions over the time period. Equipment assets, under current law, are depreciated over three, five, seven, ten, fifteen, or twenty years.

Although economic depreciation is difficult to estimate, our best estimates indicate that tax depreciation is considerably more accelerated than economic depreciation -- that is, too much of the flow of gross income from the investment is deducted in the initial years of the assets economic life and too little in the later years. Speeding up the recovery of costs is beneficial to the taxpayer because of the time value of money (a dollar of tax savings today is worth more than a dollar in the future).

This beneficial treatment arising from tax lives that are shorter than the economic lives of the assets and recovery of costs that are quicker than economic depreciation is, however, offset under current law by the failure to restate depreciation deductions in current dollars. During periods of general price inflation, relying on historical cost for measuring depreciation understates the real decline in the value of the asset.

As a result of these offsetting benefits and penalties, the overall value of tax depreciation, at current interest and inflation rates, is relatively close to economic depreciation in most cases -- effective tax rates on most depreciable assets are relatively close to the statutory tax rates.

There are several proposals to provide some type of permanent investment subsidy. Several bills were introduced in 1991 which would provide for these investment benefits. S. 1865 (Roth), S. 1921 (Bentsen et al.) and H.R. 3810 (Guarini and Levin) would provide an incremental investment credit -- a credit for expenditures in excess of a base amount. The credits proposed in these bills appear to be limited to investment in manufacturing, mining, and agriculture. H.R. 960 (Delay) would index depreciation deductions initially, and eventually phase in expensing of investments. Expensing would allow all costs to be deducted when the investment is made. Several other bills would provide an increase in expensing which is currently allowed for small businesses.

There are also a number of proposals to provide temporary investment incentives, in order to stimulate investment and help end the recession. The President has proposed a one-time deduction of fifteen percent of the cost of investing in equipment if the equipment is acquired before January 1, 1993 and placed into service before July 1, 1993. (The remaining 85 percent of the cost would be depreciated over the normal useful life.) This provision is in the form of accelerated depreciation. There have also been proposals to restore the ten-percent investment credit, either on a temporary or a permanent basis.

HISTORICAL SUMMARY OF TAX SUBSIDIES FOR INVESTMENT

Over the postwar period, depreciation allowances became gradually more generous. In 1954, accelerated methods of depreciation were allowed, which provided for a faster recovery than the common straight-line method (where costs were deducted in equal amounts over the life of the asset). These methods (double-declining balance and sum-of-years digits) allowed a larger fraction of the recovery to take place in earlier years and provided benefits for all assets.

Tax lives were shortened in 1962 and again in 1971, primarily for equipment. The 1969 Tax Reform Act restricted depreciation methods for non-residential real estate to a slower 150 percent declining balance method. In 1981, tax lives were reduced substantially. Most equipment was depreciated over five years and structures over 15 years, the latter considerably shorter than the lives of thirty years or so which appeared to prevail. Depreciation methods (how deductions were allocated over these useful lives) were made temporarily less accelerated for equipment (150 percent declining balance). Depreciation methods were allowed based on 175 percent declining balance, slightly more accelerated for non- residential structures, and made slightly less accelerated for residential structures. Overall, the 1981 changes made depreciation much more generous for structures and certain long-lived equipment assets, but had a negligible effect on short-lived equipment. In 1982 the temporary less accelerated methods for equipment were made permanent. Tax lives for structures were increased slightly in 1984 and 1985.

The Tax Reform Act of 1986 made depreciation less generous, particularly for structures which were required to use the straight line method over 31.5 years for nonresidential structures and 27.5 years for residential structures. Equipment was classified into lives of three, five, seven, ten, fifteen, and twenty years, with the latter two allowed 150 percent declining balance and the other categories allowed double declining balance.

The investment tax credit was first enacted in 1962 at a rate of 7 percent (3 percent for regulated utilities). It was envisioned as a permanent investment subsidy. The original investment credit included a basis adjustment -- the amount of the cost of the property which could be depreciated was reduced by the amount of the credit. This basis adjustment was repealed in 1964.

In 1966, the President recommended the temporary suspension of the investment credit as an anti-inflationary measure. The credit was to be suspended from October 10, 1966, through the end of 1967. In March 1967, the President recommended that the credit be restored, and Congress approved the restoration.

The investment tax credit was repealed as part of the Tax Reform Act of 1969, again as an anti-inflationary measure. It was decided to repeal rather than suspend the credit because the previous suspension became a deterrent to investment as the end of the suspension period approached, and because of administrative complexity.

The credit was reinstated in 1971 again as a measure to stimulate economic growth. The rate remained at 7 percent but was increased to 4 percent for regulated utilities.

The investment credit was temporarily increased to ten percent (including property of regulated utilities) in the Tax Reduction Act of 1975, apparently in an attempt to provide a short run stimulus to the economy. The Tax Acts of 1976 and 1977 extended the credit. A main reason was the feeling that capital formation in the economy was lagging. The credit was made permanent in 1978.

In 1982, a basis adjustment for half the credit was introduced. The investment credit was repealed by the Tax Reform Act of 1986.

ECONOMIC ISSUES

There are several major issues surrounding the effect of investment subsidies: the effects on the allocation of capital to different uses, the short-run stimulus effects, and the long-run effects on capital formation. There are also some special issues associated with incremental investment credits (or other subsidies). These issues include the effects on asset values, the effects on the timing of investment, possible administrative problems and incentives to recombine firms, and a built-in pro-cyclical bias.

NEUTRALITY ACROSS ASSETS

One of the major issues surrounding the investment tax credit which received attention in the early 1980s was the distorting effect of the credit on different kinds of assets. This issue of tax neutrality had not figured in earlier debates over the investment credit or over accelerated depreciation, in part because of the slow development of the tools used to assess these distortions.

A tax incentive is neutral (beginning from a neutral system) if it results in equal reductions of tax rates across different assets. A tax incentive will be most efficient if it tends to equalize tax rates across different investments.

The investment credit, which was a popular incentive, does not perform very well in terms of its effects on different investments. First, of course, the investment credit was not neutral because it only applied to equipment assets. Even for those assets that it applied to, it tended to favor short-lived assets over long-lived ones. For shorter lived properties, which must be replaced more frequently, the credit may be repeated more often. Thus, the credit produces differing tax rates across different kinds of assets even when it is uniformly applied.

Other forms of subsidies can be more neutral. For example, a proportional shortening of tax lives would be relatively neutral. Allowing a fraction of the asset to be deducted when purchased (partial expensing), while depreciating the remainder, is a neutral subsidy as long as tax and economic depreciation are the same. An investment credit can be made neutral across assets if it rises the more long-lived the asset is or if it is allowed only for investment in excess of depreciation. There are also subsidies which would be even more distorting across assets than the investment credit -- such as allowing more than 100 percent of the asset to be depreciated.

The investment credit also tended to be a less efficient form of investment subsidy because it was allowed in the form of a credit rather than a deduction. Overall, corporate sector investments are taxed more heavily than investments in the non-corporate sector. A credit provides the same dollar amount of offset for corporate and noncorporate investments; thus, the credit does relatively little to narrow the tax differentials between these two types of organizational forms. An extra deduction, by contrast, would be more valuable for the corporate investment.

These differential effects of investment subsidies across asset types and across organizational forms are illustrated in table 1. Consider first the differences in effects between overall equipment assets (which are more short lived) and structures assets in the corporate sector. Under current law, these assets are taxed at similar rates (30 percent and 32 percent) which are very close to the statutory rate.

If a simple 10 percent investment credit were enacted, it would produce a NEGATIVE tax rate of 12 percent for equipment and a 19- percent tax rate for structures. Moreover, these aggregates shown in the table conceal even greater differentials when assets are further disaggregated. Under current law, the range in effective tax rate by asset type is from 22 percent to 39 percent. With a 10 percent credit for all assets the range would be from MINUS 56 percent to a positive 28 percent. And, of course, the disparities would be even larger if the credit were not allowed for buildings, in which case rates would range from MINUS 56 percent to a positive rate of 39 percent.

The distortions arising from an investment credit can be lessened somewhat if a basis adjustment is provided (that is, if the amount of the property eligible for depreciation is reduced by the amount of the credit), but the credit will still be non-neutral. As indicated in the table, allowing an investment credit only for investment in excess of depreciation, or allowing partial expensing is neutral, while allowing depreciation of more than 100 percent of the asset cost is more distorting.

The second two columns present effective tax rates for the noncorporate sector based on the average marginal tax rate of 23 percent. Comparing the investment credit across organizational types, one can see that the credit does not necessarily move the tax rates for a particular investment in each form closer together, as would accelerated depreciation or an additional deduction. In the case of the ten-percent credit, while the tax credit reduces the effective tax rate on corporate investments from 30 percent to MINUS 12 percent, it reduces the rate on noncorporate investments from 23 percent to MINUS 35 percent. (As in the case of the corporate sector, these aggregations do not show the range of effects. Noncorporate tax rates range from 17 to 28 percent without a credit, but from minus 150 percent to 13 percent with a ten-percent credit.) If instead of a credit, an additional deduction were allowed of equivalent value to the corporate sector (e.g., a deduction of .10/.34, or approximately 32.4 percent), the effects on the noncorporate sector would be much more even.

        TABLE 1: EFFECTIVE TAX RATES, STRUCTURES AND EQUIPMENT

Although most economic analysis suggests that investment is allocated most efficiently when tax rates are equalized across assets, there are occasionally arguments for favoring certain types of investment over others. These arguments suggest that investments in equipment may be more "productive" than investments in structures and other assets, or that investments in certain types of industries, such as manufacturing, may be more "productive."

Economic theory indicates that assets will be allocated to their most productive uses if the market is allowed to allocate resources. There is no inherent benefit in, for example, producing goods with capital in the form of machinery as opposed to capital in the form of buildings or in the form of inventories; rather, these assets should be invested in to the degree that they earn a return which justifies an investment. Similarly, there is no inherent benefit in production of goods as opposed to transportation and distribution of those goods, or the provision of services. All types investments should be made as a function of production technology and the tastes of consumers.

The exception to this rule is when there is some sort of external benefit which is not captured in pricing. Such external benefits are recognized in some cases -- for example, inadequate investment in research and development normally occurs in a free market because of the appropriability of the innovation by other firms; hence, tax subsidies or regulations (such as patent protection) may be in order. There is no similar identification of an external benefit associated with investment in capital goods of a particular type or with manufacturing as opposed to other industries. ¹

INVESTMENT INCENTIVES AS A FISCAL POLICY INSTRUMENT

Many of the recent proposals for an investment subsidy, including the proposal made by the President in his budget, are designed to stimulate the economy in the short run. How effective might such a stimulus be?

There have been a number of studies which considered the effect of tax subsidies for investment on investment demand. One of the earliest of these studies was based on a model developed by Hall and Jorgenson. ² They examined the effect of the depreciation revisions in 1954 and 1962 and of the investment credit in 1952. The model assumes that the elasticity of substitution between labor and capital (the percentage change in the ratio of capital to labor divided by the percentage change in the relative prices of capital and labor) was 1. Their results suggested a significant impact of investment tax incentives on investment.

Their approach was criticized on several grounds. First, there were questions as to whether the responsiveness of investment to the change in relative price is as large as one, particular in the short run. Criticisms along these lines were made by Coen and Eisner. ³ Eisner suggests that the impact was only a sixth as much as that predicted by Hall and Jorgenson. Another criticism was that interest rates in the Hall-Jorgenson analysis were held constant whereas they might be expected to rise if there is a greater demand for capital. ⁴

A number of other studies of investment related tax incentives have been done, some of which suggest that these incentives are effective and others suggesting that they are not. Bischoff, using a model with a constant but unspecified elasticity of substitution found that the effects of investment incentives were significant and, for the investment credit, exceeded the revenue loss. ⁵ Coen, however, found that these investment incentives were not very effective in relation to the revenue foregone. ⁶

Klein and Taubman modeled the suspension of the investment credit in 1967 using the Wharton Econometric Model and found relatively small effects. ⁷ They also found that the suspension would have an effect about a third larger than a permanent repeal. Aaron, Russek, and Singer also found relatively small effects of the 1969 repeal of the credit and the 1971 reinstatement using the Federal Reserve Board-MIT econometric model. ⁸ Brimmer and Sinai simulated the effects of several tax proposals with the DRI econometric model and found that the investment credit produced only $0.68 of investment for every dollar of revenue loss. ⁹

Taubman in examining the impact of the restoration of the investment credit in 1971 argued that the investment credit had limited usefulness as a counter-cyclical device because of the slowness with which it had an effect on investment. ¹⁰ Similarly, Gordon and Jorgenson concur with Taubman's findings that the investment credit is not a good stabilizing tool. ¹¹ They conclude that there is about a two-year impact lag associated with the credit, a fact which makes investment subsidies a poor choice for counter- cyclical policy. Eckstein agreed with this assessment that the credit is a poor stabilizing device. ¹²

Other investment studies have tended to concur with this view that the effects of investment incentives appear to be modest. Clark ¹³ found little evidence that investment was very responsive to price as did Hendershott and Hu. ¹⁴ Auerbach and Hassett concluded that the repeal of the investment credit in 1986 had little effect on spending on equipment, given the strength of equipment spending in 1987-89. ¹⁵ The same authors, however, suggest in another paper that equipment spending might have grown even faster without the changes in TRA. ¹⁶

There are reasons to expect that tax incentives for equipment might have limited usefulness in stimulating investment in the short run, primarily because of planning lags and because of the slowness of changing the technology of production. Essentially, there are two reasons that firms may increase investment. First, they may expect output to increase. This response, called the accelerator, is a result of other forces which increase aggregate demand and simply requires making more of the same type of investment (along with hiring more workers). The second reason is that the cost of investment has fallen. Part of this effect may be an output effect -- since the overall cost of investment is smaller, output can be sold for a smaller price and thus sales would be expected to rise in the future. But part of this effect has to do with encouraging more use of capital relative to labor.

The accelerator or output effect requires a reaction by the firm to order more capacity, and so there is a lag associated with this effect. The price effect is subject to this general planning lag but also a more fundamental lag reflecting a change in technology. That is, it is a much simpler and quicker process to add more production using existing technology than to change the mode of technology to use a different, more capital intensive process.

In general, an investment incentive would perform poorly relative to other alternative tax cuts or spending increases if the amount of investment induced falls below the revenue cost. If it does, then an additional dollar of government revenue loss would induce, initially, less spending than an alternative use. ¹⁷

A temporary incentive, such as the President has proposed, should be more likely than a permanent one to stimulate investment in the short run, since it will provide only a brief window of opportunity to make investments that are eligible. Moreover, such a temporary subsidy provided in the form of an investment credit rather than a rate cut can provide more price reduction for a given revenue cost (a greater "bang for the buck"), since investment is only a fraction of the total capital stock. ¹⁸ Nevertheless, such a temporary incentive is still subject to planning lags. Moreover, there is no incentive in such a case to change the technology of production. In the one experience we had with an explicitly temporary provision -- the increase in the ITC in 1975 -- investment remained quite low.

On the whole, most of the literature on the investment credit and other investment subsidies suggests that investment incentives inherently affect investment with long time lags because of the lengthy time periods involved in business investment projects, making these subsidies ineffective counter-cyclical devices.

EFFECTS ON SAVINGS AND ECONOMIC GROWTH

One of the main arguments for a PERMANENT new tax incentive to investment is to encourage additional investment and productivity, and hence an increase in the future standard of living. The major problem with this policy objective is that there appears to be no strong evidence that cutting taxes, or increasing the rate of return, causes individuals to save more.

There is no a PRIORI reason from economic theory to expect that increasing the rate of return will increase savings. If the rate of return goes up, individuals would wish to consume more in the future (because the "price" of future consumption in terms of foregone current consumption is less). This effect is called the price effect. This price effect might lead to an increase in savings to finance future consumption. It is, however, possible to consume more in the future and save less, since current savings will grow to a larger future amount. This effect is called the income effect. Indeed, it is quite easy to find examples where savings will decrease with a deficit financed tax incentive.

The effect on savings remains ambiguous even if the government finances the tax incentive with an offsetting tax, such as an increase in tax rates. The overall effects on savings depend on who receives the tax increases and tax cuts, and the sum of these income effects.

Since economic theory does not provide a clear answer to the effect of tax incentives on savings, researchers have turned to empirical evidence. A number of studies have been done which have examined the relationship between savings rates and after tax rates of return. Of the studies which have accounted for inflation and taxes, the first done by Boskin indicated a positive, but small relationship. ¹⁹ Subsequent studies have generally failed to find a statistically significant relationship. ²⁰

Recently, there has been attention paid to constructing more sophisticated models of individual savings behavior which model individuals over the life cycle, account for many generations at one time, and can allow offsetting tax increases to achieve revenue neutrality. ²¹ Unfortunately, these models tend to predict much larger savings responses than tend to be observed in the economy. This result may reflect use of responses to price which are too large and the difficulties in modeling the evolution of savings over time. In these models, older individuals tend to dissave and consume their wealth. Investment subsidies cause the value of existing capital to fall and thus reduce the income of the old who save little and increase the income of the young and new generations. These, income effects act to encourage savings. For a rate cut, the incomes of the old increase and the incomes of the young decrease (due to higher taxes on labor income); the income effects tend to decrease savings. ²²

These models suggest that investment subsidies are more likely to increase savings than reductions in tax rates on capital. Investment subsidies, such as investment credits and accelerated depreciation, only apply to new investment and do not provide tax benefits for existing capital. Rate reductions (such as a corporate or capital gains rate cuts) benefit existing as well as new investment. This is the permanent version of the "bang for the buck" issue; but the advantages of restricting subsidies to new investment are much more modest in the long run. For example, with no depreciation a subsidy for new investment would be only about three percent of the cost of a benefit for all capital in the first year, but the cost would be fifty percent as large for a permanent incentive. In the case of equipment, which has a high depreciation rate, an investment subsidy would be about 15 percent of the cost in the short run, but 86 percent of the cost in the long run. ²³

Although the life cycle models suggest that investment subsidies are likely to increase savings, it is not clear at this point how realistic these models are. These models do not account for depreciation, which, as noted above, can be important in evaluating the effects of investment incentives. The method of financing (via deficits or alternative taxes) can be important. Moreover, any advantages which occur due to choosing an investment subsidy over a rate cut is not the result of creating price incentives to save but rather the result of transferring income from the old to the young via a lump sum tax on the old.

In sum, the evidence of the effect of tax subsidies on savings is unclear, even if the revenue is made up in other taxes. Thus, a more certain and more effective route to savings might be to reduce the deficit and increase government savings.

Even if investment subsidies act to increase investment, the long term effects on future standards of living are likely to be relatively small. For example, consider the possible effects of a ten-percent credit ²⁴ on equipment with two assumptions which would maximize the effects -- a unitary factor substitution elasticity and an infinitely large savings elasticity, so that the after tax return remains fixed. While this tax credit would increase investment in equipment by about 12 percent, its effect on output in the long run would be only 1.5 percent, because equipment investment is only one of many inputs into the productive process.

Moreover, most of this output would not be available to increase the standard of living. Rather, a large fraction would be required for increased investment, in order to maintain the value of capital relative to GNP. As a result, the percentage change in long-run consumption levels would be only 0.3 percent. That is, the ten percent credit would increase the long run level of consumption by only 3/10 of one percent even under very generous assumptions about behavioral responses (on both the savings and investment side).

Note that this percentage change is a one-time change in the level of consumption, not a change in the rate of growth. The rate of growth in the long run is not dependent on the rate of savings or investment; rather, it reflects the rate of population growth and technical change. During the intermediate period when output is adjusting to its long run equilibrium, the rate of growth would increase. This change in the rate of growth would, however, be negligible. For example, suppose it took ten years to adjust to the new capital stock. In this case the rate of growth would be higher for each of the ten years by one tenth of the total change, or about 3/100 of one percent.

SPECIAL ISSUES SURROUNDING AN INCREMENTAL CREDIT

There are several issues which are particularly relevant to an incremental investment credit (or other incremental approach to providing a tax incentive). With an incremental credit (modeled on the credit allowed for research and development), firms would be eligible for credits only for investments in excess of a base, which might be averaged over three or four years. The initial base would be increased over time, based on growth of sales.

Other types of incremental credits or benefits, including a credit in excess of depreciation, differ from this type of credit and will be discussed is the final section.

The attraction of the incremental credit is that it could be used to provide a stimulus with relatively little revenue cost, since only a small part of investment would qualify. Thus, it would be likely to induce more investment per dollar of revenue loss. The investment credit already has this advantage over a rate cut -- more marginal incentive can be provided with a smaller revenue cost. The incremental credit would be even more cost effective than a regular investment credit.

Despite this apparent advantage of an incremental investment credit, there are a number of serious problems with the credit. Of course, an incremental credit would still involve the allocational distortions across assets which were discussed above, unless a format other than a credit were used. (It would be possible to allow a neutral incremental benefit, such as partial or full expensing.) In addition, the proposals would limit the credit to assets used in industries that produce tangible personal property (apparently referring to agriculture, extractive industries, and manufacturing).

Unless noted, these problems relate to a permanent incremental credit, although some of the administrative issues would still have to be faced even with a temporary incremental credit.

Effects on Asset Values and Market Structure

As noted earlier, an investment credit can result in the same marginal incentive to invest at a smaller cost than a rate reduction because the credit applies only to new investment, while the rate reduction applies to all capital. But this stream of saved revenues does not simply disappear; rather, it shows up as a change in asset values of private individuals. In the long run, a rate reduction has no effect on asset values -- a dollar of capital is still valued at a dollar. An investment credit, however, will cause the value of assets to fall relative to the actual replacement cost of the capital -- by exactly the amount of the credit (see the appendix for a mathematical exposition of this point).

Another way of explaining this effect is to assume an instant adjustment in earnings (the incentive causes an increase in investment and a fall in the pre-tax rate of return). The value of the existing capital stock will fall through lower earnings: if the rate of the investment credit is ten percent, the return will fall so that existing capital is valued at only 90 percent of its previous value. Any reinvestment necessary to replace capital or to provide for growth will have a present value of zero -- the investment will produce a stream of net revenues worth 90 cents on the dollar, but the cost of a dollar is only 90 cents due to the credit. Moreover, at any future point in time, the current capital stock is worth 90 cents on the dollar, and the present value of future reinvestment is equal to zero.

A ten-percent credit provided to all investments will cause the value of assets in the steady state (for example, shares of stock or value of a firm) to fall by ten percent if all investment were financed by equity. ²⁵ If the asset is debt financed, the fall in value will be larger since all the decrease for the entire capital stock will fall on the equity share. For example, if the investment were financed one-half by debt, the effect of a ten-percent credit would be to reduce asset values by 20 percent; if the investment were financed one-third by debt, asset values would fall by 15 percent.

In practice, this effect on asset values may not occur immediately because some time will be required for the capital stock to adjust. For example, if the adjustment occurred linearly over a five year period and one third is debt financed, the effect of the adjustment period (an gain which offsets the 15 percent loss) would cause the asset value to initially fall by 7 percent, and then decline another 8 percent from the original value over five years. If the adjustment period were ten years, there would be virtually no effect in the short run, but the value of stock would fall by 15 percent over the next ten years.

Of course, these effects would not be as large for an equipment investment credit because only a fraction of the firm's assets are eligible. In manufacturing, about one-third of assets are equipment, and the fall in asset value would be only 5 percent in the long run. In construction, equipment is much more important (about 68 percent of assets) and the value would fall by ten percent in the long run. In general, the effects would depend on the debt share and the share of equipment, as well as rates of return, growth, and depreciation.

The incremental investment credit and its effects on asset values and markets has never been carefully explored. A preliminary assessment suggests, however, some troubling potential consequences for asset values and market structures.

Consider an on-going firm, or an industry, which invests enough to receive the incremental credit. The value of the firm's future reinvestment (to maintain the capital stock and normal growth) will no longer be zero. Rather the infra-marginal investment will have a negative value. Although the return to this investment has a present value of 90 cents on the dollar, the outlay for the firm is a dollar because this investment does not receive a credit. Thus, each dollar of inframarginal investment loses ten cents. If an existing firm is assumed to survive into the equilibrium, the asset value of the firm will fall dramatically, particularly when assets are depreciating quickly and the present value of the firm's reinvestment is very large. For example, in the case where only ten percent of investment qualifies for the credit, for the part of capital reflecting equipment, asset values would fall in the long run by 76 percent. Even in the short run, asset values would fall substantial. For example, assuming even a lengthy ten-year adjustment period the initial fall in the value of the share represented by equipment would be over 60 percent, implying a fall in the value of manufacturing assets of about 20 percent and the value of construction assets by about 40 percent. Essentially, the substantial savings as a result of using an incremental investment credit shows up as a fall in asset values -- and it will affect the stockholders at the time the tax credit is introduced.

It seems difficult to imagine that firms will make this negative investment, since they have the option of ceasing to operate or to invest. Indeed, in a perfectly competitive equilibrium, with no barriers to entry and no base for new firms, existing firms might disappear and be replaced by new firms. The credit might simply lead to the dissolution of existing firms in some industries. ²⁶ (These effects will depend on what rules are provided to new firms.) In many industries, however, there are barriers to new firms. If new firms cannot enter, the incremental credit tends to create a situation of declining average cost, and interferes with a competitive market equilibrium. If a market is competitive to begin with, this type of credit pushes the industry towards a monopoly.

But it also seems difficult to imagine that the incremental credit will lead to the disappearance of large U.S. companies. These firms may be earning excess profits, and if these excess profits are large enough, they will continue to operate and experience a decline in asset value.

In general, the incremental credit may lead to a combination of effects -- declines in asset value, pressure to move towards fewer firms (where the amount of credit will be maximized), evasion of inframarginal amounts of investment by various mechanisms, or pressure time investments so that more investments will be eligible for the credit. The magnitude of these various effects is highly uncertain. Moreover, there would be uncertainty as to how much investment might be eligible and how long the incremental credit is expected to continue, as well as possible myopia on the part of individuals which could affect asset values and investment decisions in the short run. But the effect on assets values and firms could be serious.

Procyclical Nature of an Incremental Credit

One of the problems of introducing a permanent incremental tax credit is that, once in place, it may act to magnify business cycles. During times of rapid growth, when investment is relatively high, most firms will be affected by the investment incentive which lowers their tax burdens and encourages new investments. When economic growth slows or reverses, investment tends to fall and firms may find their investment inadequate to qualify.

For example, suppose a credit based on a four-year average had been put into place in 1989. Using the growth in corporate product to increase the base, equipment investment in 1989 would have exceeded the base by 7.9 percent, equipment investment in 1990 would have exceeded the base by 4.8 percent and equipment investment in 1991 would have fallen short of the base by 3 percent. During 1991, when the economy was suffering from a recession, the incentive to investment would have been reduced. This is similar to placing a temporary penalty on investment in bad times when investment should be encouraged and providing a temporary incentive in goods times when investment is booming.

Administrative Issues Surrounding an Incremental Credit

There are a number of issues which cause considerable complexity in providing an incremental investment credit.

Unlike the R&E credit, the investment credit is widely available to many types of firms, including many small businesses. For example, in 1985, 9 thousand corporations claimed the R&E credit, while 800 thousand claimed the investment tax credit. Moreover, the many unincorporated businesses run as sole proprietorships and partnerships will claim the credit. These individuals will not only have to calculate the credit, but also to calculate the base period amount and how it is changing. One solution is to allow the credit on a non-incremental basis for firms under a certain size, but this approach would also require a phaseout which has its own complications, or a notch problem. ²⁷

There are complications of allowing the credit for partnerships, since the determination of eligibility could either apply at the partnership level or the individual level, or both. Applying the limits at the entity (firm) level could allow individuals to reorder their partnership arrangements to avoid the base level requirements, but applying them at the individual level would be very complicated.

A decision would have to be made on whether the credit would be recaptured on sale, and how to treat used versus new assets. If such rules are not applied, individuals could, in theory, increase their eligible investments by buying and selling assets. If used assets are excluded, then a decision would have to be made to determine how much refurbishing would be necessary in order to identify an asset as new.

Another issue is how to treat acquisitions and dispositions. When a firm merges or splits, the basis for the credit needs to be allocated, but there is no way of ensuring that the allocation of basis is correct based on the future activities of the two parts of the firm. That is, one component of the firm could be heavily engaged in equipment investment and splitting up might lower the base and allow more of the credit to qualify.

Another issue is whether the lessee or the lessor would receive the credit with leased property and how the base would be figured. Unless the credit is passed through to the lessee, firms could avoid the limits of the fixed base by leasing assets, and firms (such as financial institutions) could easily set up leasing operations with little or no base. Disallowing the credit to the lessor, however, would interfere with the use of leasing to allow firms with no tax liability to take advantage of the credit.

There is also the issue of what to do with new, start-up companies. If they are given a basis of zero, all of their assets will qualify indefinitely. If they are given a fixed base as a fraction of investment, the marginal incentive will be diluted.

Unfairness Across Firms

Another issue with the incremental credit is that the amount of the credit will depend on the initial base period selected. For firms with lumpy investments, the base period might be abnormally high if the firm has undertaken a recent expansion program or bought a major asset. These firms may never be able to take advantage of the credit and the firms will be unable to take advantage of either the cash flow benefits, or even the marginal incentive effects, of the credit. The extreme case of a high base would be a firm which had just started up for a few years and whose investment in the future would never be as large relative to receipts. For other firms, the base might be unusually low, because the firm had not been engaged in an investment program during the investment period. Firms with a higher than average base might eventually be forced out of the market.

Incentives To Make Lumpy Investments

The incremental investment credit also encourages firms to time investments and make lumpy investments in order to qualify for the credit. If expenditures which were originally planned to be made over several years could be saved up and made in a single year, much more of the expenditures could qualify for the credit.

Tax Proposals Related To Incremental Credits

There are several possible tax proposals which are related to incremental credits. First, a credit might be temporary, so that most of the issues relating to non-neutral investments, fall in asset values, pro-cyclical nature of the credit, the tendency for lumpy investments, and many of the administrative problems would disappear.

Incremental incentives could be designed with are neutral across assets, but otherwise are incremental, such as incremental partial expensing. The criticisms of incremental credits discussed in this section would still apply to these neutral incremental incentives.

Another incremental incentive which has slightly different economic effects is one allowing investment credit for investments in excess of depreciation. This modification is a way of making an investment credit a neutral incentive, rather than achieving an incentive at a smaller revenue cost, since the incentive at the margin will be reduced because the new investment will cause the depreciation basis to go up in the future. Thus, most of the special issues associated with the incremental ITC would not be applicable to this type of credit. Some of the administrative problems, however, would persist, such as the treatment of new businesses and the treatment of divestitures. (These problems would, however, be self correcting in the long run.)

Another proposal is to allow expensing in excess of a base. ²⁸ This proposal would keep the base fixed, with the ultimate objective to move to full expensing of all assets, and to allow a smoother revenue loss pattern over time. Full expensing would have a large effect on asset values, not because it is incremental but because it is a very generous tax benefit and because of its timing. (For an equity financed investment, asset values in the long run would fall by the tax rate, or 34 percent).

CONCLUSION

This analysis has suggested that investment incentives can be very distorting across different types of investments, unless they are designed to be neutral. The investment credit is not neutral across assets.

There is little evidence that investment subsidies constitute an effective tool of counter-cyclical fiscal policy. While these credits may increase savings in the long run, their effects are modest and can be achieved by a reduction in the deficit.

The incremental investment tax credit is a provision whose economic effects have not been carefully explored. Although such a credit would cost less than a regular credit, it has potentially troubling consequences for asset values and market structure. The credit also appears to be procyclical. There are some extremely serious problems of tax administration associated with the incremental credit.

[appendix omitted]

² Hall, Robert E., and Dale W. Jorgenson, Tax Policy and Investment Behavior, American Economic Review, Vol 58, No. 3, June 1967, pp. 391-414. Results of a later study were reported in Tax Policy and Investment Behavior: Reply and Further Results, American Economic Review, Vol. 59, June 1969, pp. 388-400 and Application of the theory of Optimum Capital Accumulation, In Tax Incentives and Capital Spending, Gary Fromm (Ed.), Washington, D.C.: The Brookings Institution, 1971, pp. 9-60.

³ Coen, Robert M., Tax Policy and Investment Behavior: Comment, American Economic Review, Vol 59, June 1969, pp. 370-379; Eisner, Robert, Tax Policy and Investment Behavior: Comment., Ibid, pp. 379-388. For a survey of the literature on this elasticity see Gerard M. Brannon, The Effects of Tax Incentives for Business Investment: A Survey of the Economic Evidence, In The Economics of Federal Subsidy programs, Joint Economic Committee, 92nd Congress, 2nd Session, July 15, 1972, p. 251.

⁴ Brannon, Gerard M. The Effects of Tax Incentives for Business Investment: A Survey of the Economic Evidence, op. cit.

⁵ Bischoff, Charles W., The Effective of Alternative Lag Distributions, In Tax Incentives and Capital Spending, op. cit., pp. 61-130.

⁶ Coen, Robert M., The Effect of Cash Flow on the Speed of Adjustment, In Tax Incentives and Capital Spending, op. cit., pp. 131-146.

⁷ Klein, Lawrence W. and Paul Taubman, Estimating Effects within a Complete Econometric Model, In Tax Incentives and Capital Spending, op. cit., pp. 197-242.

⁸ Aaron Henry J., Frank S. Russek, and Neil M. Singer, Tax Changes and the Composition of Fixed Investment: An Aggregate Simulation, Review of Economics and Statistics, Vol 54, November 1972, pp. 343-356.

⁹ Brimmer, Andrew and Allen Sinai, The Effects of Tax Policy on Capital Formation, Corporate Liquidity and the Availability of Investible Funds: A Simulation Study, The Journal of Finance, Vol. 31, No. 2, May 1976, pp. 287-308.

¹⁰ Taubman Paul. The Investment Tax Credit, Once more. Boston College Industrial and Commercial Law Review, Vol. 14, May 1973, pp. 871-890.

¹¹ Gordon, Roger and Dale Jorgenson, Policy Alternatives for the Investment Tax Credit, at joint seminars on Encouraging Capital formation Through the Tax Code, before the Committee on the Budget, United States Senate, September 18, 1975.

¹² Eckstein, Otto, et al., Tax Reform Studies, the Data Resources Review of the U.S. Economy, August 1977, p. 110-1.19.

¹³ Peter K. Clark, Investment in the 1970s: Theory, Performance, and Prediction, Brookings Papers on Economic Activity 1: 1979, pp. 73-124.

¹⁴ Hendershott, Patric and Sheng-chung Hu, Investment in Producers Equipment, In Henry J. Aaron and Joseph A. Pechman, Eds., How Taxes Affect Economic Behavior, Washington, D.C., The Brookings Institution, 1981, pp. 85-126.

¹⁵ Auerbach, Alan and Kevin Hassett, Investment, Tax Policy, and the Tax Reform Act of 1986, In Do Taxes Matter: The Impact of the Tax Reform Act of 1986, Ed. Joel Slemrod, Cambridge, Massachusetts: The MIT Press, 1990, pp. 13-49.

¹⁶ Auerbach, Alan J. and Kevin Hasset, Recent U.S. Behavior and the Tax Reform Act of 1986: A Disaggregate View, National Bureau of Economic Research Working Paper 3626.

¹⁷ A spending increase would be spent dollar for dollar. The effect on aggregate demand depends in the amount of underemployed resources in the economy.

¹⁸ For a temporary incentive, investment as a fraction of the capital stock would the sum of the depreciation rate plus the growth rate. For assets that did not depreciation, this fraction might be only two or three percent; for equipment which depreciates at an average rate of about 15 percent, the share might be 17 or 18 percent.

¹⁹ Michael Boskin, Taxation, Savings and the Rate of Interest, Journal of Political Economy, Vol 86, January, 1978, pp. s3-s27.

²⁰ See Barry Bosworth, Tax Incentives and Economic Growth, Washington D.C.: Brookings Institution, 1984; Irwin Friend and Joel Hasbrouck, Saving and After Tax Rates of Return, The Review of Economics and Statistics, Vol.65, November 1983, pp. 537-543; E. Philip Howry and Saul H. Hymans, The Measurement and Determination of Loanable Funds Savings, Brookings Papers on Economic Activity, No. 3, 1978, pp. 655-705; John Makin and Kenneth A. Couch, Savings, Pension Contributions, and the Real Interest Rate, The Review of Economics and Statistics, Vol. 71, August 1989, pp. 401-407. Occasionally, the study done by Lawrence Summers, Capital Income Taxation and Accumulation in a Life Cycle Model, American Economic Review, Vol. 71, September, 1981, pp. 533-44 is cited as evidence of a high savings response. This paper is not, however, an statistical analysis but a simulation model which can be made consistent with any savings response. See Owen J. Evans, Tax Policy, the Interest Elasticity of Saving, and Capital Accumulation: Numerical Analysis of Theoretical Models, American Economic Review, Vol. 73, June 1983, pp. 398-410.

²¹ Auerbach, Alan and Laurence J. Kotlikoff, Dynamic Fiscal Policy, Cambridge: Cambridge University Press, 1987; Summers, Lawrence, Capital Income Taxation and Accumulation in a Life Cycle Model, American Economic Review, 71, September 1981, pp. 533-44. The Auerbach and Kotlikoff study contains a survey of the literature on the magnitude of behavioral responses.

²² These effects may be overstated, however, since the life cycle models do not incorporate bequests or disaggregate individuals into high and low lifetime incomes. There is some evidence that older individuals, particularly those with substantial wealth, do not dissave in old age.

²³ These measures assume a three percent growth rate and a 15 percent depreciation rate for equipment. In the short run, new investment is equal to .03/(1.03) as a fraction of current capital stock and new investment without depreciation and (0.03+.15)/(1 + 0.03+0.15), or fifteen percent, with depreciation. For the long run, assuming a 6 percent real discount rate, the present value of new investment for every dollar of existing capital stock is equal to the sum of the depreciation rate and the growth rate, divided by the difference between the discount rate and the growth rate. With no depreciation, this value is 1, and the share of the discounted value of new investment to the sum of current stock and new investment is fifty percent. With depreciation, the present value of investment is 6, and the share of the discounted value of new investment out of the sum of current stock and discounted value of new investment is 86 percent (6/(1+6)).

²⁴ For this application, the illustration is more easily made with an investment credit with a full basis adjustment, since such a credit changes the rental price of capital proportionally for all assets.

²⁵ These illustrations assume that additions to investment are financed by the issuing of new stock, which simplifies the calculations. One could also measure values based on additional debt finance or use of dividends, but these types of calculations are more complicated.

²⁶ Totally new firms do not initially have enough tax liability to use the credit, but they would in the long run.

²⁷ With a phase out, the firm would gradually restore its base as income rises. A phaseout increases the tax rate because as assets (which depend on investment) rise, the base will be phased in and credits will be lost. To avoid a phase-out, all firms below a dollar cut-off in asset value could get the credit for all investment. If there is an abrupt cut-off, however, then firms which rise just above the asset limit will have a large increase in tax liability.

²⁸ This notion is not reflected in and legislative proposal, but has been raised as a possible investment incentive in statements by Congressman Gradison.