CRS REPORT GIVES ECONOMIC ANALYSIS ON ISSUE OF TAX POLICY AND RENTAL HOUSING.

CRS REPORT FOR CONGRESS

Pre-1987 tax law favored equipment over structures. The new tax changes eliminate that favoritism, and as a consequence rental structures are unlikely to be adversely affected by the changes. Many studies which found large adverse effects used taxpayer characteristics which are not consistent with statistical aggregates or with economic theory regarding the marginal investor in real estate.

                                   by

TAX POLICY AND RENTAL HOUSING: AN ECONOMIC ANALYSIS

Following the announcement of a major tax reform proposal by the Treasury Department in November 1984, there were a number of studies which predicted or hinted at adverse consequences of the proposed changes for real estate investment. ¹ The message of these studies was that the sector of the economy most likely to suffer adverse consequences of the proposals was the real estate industry. Such a message was in considerable contrast with certain other analyses which suggested that the proposal would tilt the balance away from investment in equipment to investment in structures. ² Contrasting assessments of the effects of the tax bill continued to appear during the debate and after final passage of the act, with some assessments stressing the negative effects on equipment ³ and others the adverse consequences for structures. ⁴ This confusion and disagreement about the effects of the tax reform proposals reflect a more fundamental disagreement about the basic effects of the current tax system on different types of assets.

Briefly, those who stress the adverse consequences of a more even tax system on investment in rental housing are implicitly suggesting that rental housing was a relatively tax preferred investment while those who stress the smaller adverse consequences or eventual benefits are implicitly suggesting that real estate is heavily taxed relative to other investments. The notion that real estate investment was heavily taxed derives from the fact that it involves a business investment (making it more heavily taxed than consumer durables such as owner occupied housing) and from the fact that the investment credit and depreciation rules were more generous for equipment than for structures. The new tax law would reduce this relatively heavy burden compared to business equipment largely through repeal of the investment tax credit. (Its position relative to consumer durables would depend on the trade off between lower tax rates and the less generous depreciation rules; rough calculations suggest these effects are offsetting).

The notion that rental housing was lightly taxed derives from the argument that rental housing was favored by the tax law not so much because of provisions directed at it, but because such investments are able to benefit from generally available provisions in the tax code. These provisions include the lower taxes on noncorporate investment, tax advantages to debt finance, and the benefits of converting ordinary income into capital gains. ⁵ These aspects of the tax treatment play an important role in the typical "tax shelter" investment which is the model used in many of the studies which predicted highly adverse consequences for structures. Those studies which suggested relatively small effects for real estate frequently model structures as corporate investments. The new tax provisions will narrow the relative advantages for noncorporate investment and debt finance through lower tax rates (although those advantages will still remain in reduced form) and eliminate the tax advantage for capital gains which are to be taxed in full.

The notion that the tax system favors real estate is also apparently due to the prevalence of real estate as a tax shelter investment and the very small amount of investment in equipment tax shelters by comparison. If equipment were favored by the tax law, so the argument goes, it would have been a more common tax shelter. ⁶ This reasoning seems persuasive, as high bracket individuals would normally earn their greatest return on the most tax preferred investment.

This argument, however, fails to recognize the deliberate intent of the tax law to constrain equipment leasing in the noncorporate sector, through numerous restrictions on the ability of the lessor to receive depreciation and investment credit benefits. The limited investment in equipment shelters is more readily explained when the full ramifications of the tax law for leasing activities are considered. When the investment credit was restored in 1971, rules were adopted limiting eligible equipment leasing activities to short term leases, and providing other constraints, in order to prevent the growth of tax shelters in equipment leasing. ⁷ Moreover, certain general longstanding rules regarding whether a lessor has a true economic interest in a lease acted as a barrier to leasing industrial equipment, since leasing of property which could only be used by the lessee was not regarded as a true lease. In addition, at risk rules which limited deductions were imposed on all tax shelters except real estate in 1978.

As further evidence of the flaw in judging favoritism by the prevalence of tax shelters, the most popular tax shelter historically has been oil and gas, not real estate. ⁸ Moreover, given the smaller aggregate importance of oil and gas assets in the economy as compared to real estate, oil and gas tax shelters would account for an even larger fraction of that investment and would seem to be the most "popular" tax shelter. Yet, oil and gas shares tax characteristics similar to equipment (very generous capital recovery), but would seem unlikely to benefit from either an ability to carry debt or significant tax advantages on resale. ⁹ Indeed, the primary difference between oil and gas and equipment is that tax shelters in the latter were not constrained by tax regulations. Thus one cannot judge the overall tax favoritism afforded to investments by their popularity as tax shelters.

The "tax shelter" characteristics ascribed to rental housing investments in many studies are frequently not bolstered by statistical evidence or by economic reasoning. To try to assess the consequences of tax policy for rental housing, it is first important to determine whether the "tax shelter" model, the corporate model, or some other model is appropriate to determining the size of the rental housing stock and the nature of the ownership of that stock.

Section II of this study considers the statistical evidence which can be pieced together from a number of different sources to describe the basic characteristics of the ownership of rental property which affect the determination of tax effects. The evidence gathered suggests a very heterogeneous market where many types of investors -- individuals, partnerships (both passive and active), corporations, tax exempt entities, and others supply rental housing. Moreover, the statistical data provided suggests that many of the common perceptions about leveraging rates, turnovers of property, and high marginal tax rates which are associated with tax motivated real estate investment are not correct for the statistical aggregates.

Section III discusses how these characteristics of the rental housing market might be incorporated into a model of economic behavior which can, in turn, provide information about the relative tax treatment in the industry. In particular, it is important in economic analysis to determine, for purposes of estimating effects on rents and supply of rental housing, the marginal investor. This modeling of the real estate supply function incorporates some of the more recent theoretical developments in modeling corporate and noncorporate sector investment allocation. This theoretical analysis points to the corporate sector as the marginal -- although not average -- investor. In this model, the rules which apply to other investors, including tax shelter investors, affect the distribution of ownership in the long run (that is, the shares held by the corporate sector, partnerships, individuals, and so forth), but not rents.

Section IV of the study discusses some consequences of the Tax Reform Act of 1986 for the long run stock and supply of rental housing and the amount of resources allocated to this sector of the economy. The estimated effects of these proposals vary considerably depending on the modeling of the marginal investor. Use of the corporate model as the marginal investor generally leads one to the conclusion that rental housing is currently taxed at heavy rates compared to other assets and is unlikely to suffer from the tax reform changes. Use of the "tax shelter" investor suggests major adverse consequences. Averaging effects across all investors produces relatively small effects which are closer to the corporate investor. The simulation results clearly demonstrate that the large negative effects estimated by many studies are a direct result of using assumptions inconsistent with observed characteristics, particularly the high marginal tax rates employed. ¹⁰

CHARACTERISTICS Of RENTAL HOUSING OWNERSHIP

The investment modeled in many studies of tax reform and rental housing is a prototype of what one might call the "tax motivated" investor, or tax shelter, in real estate. The investment is characterized by an extremely high rate of leverage with 80 to 90 percent of the funds borrowed. The investor is typically an individual in the top tax bracket, typical of a large syndicated partnership which is a tax shelter arrangement. The property is held for a short period of time, typically five to ten years. These assumptions heavily influence the level of tax burden, with perhaps one of the most important aspects being the treatment of debt finance. The choice of holding period also has an influence, particularly in that the debt to asset ratio remains high over the entire life of the investment.

How representative is this large, tax shelter partnership? Table 1 presents some data from the Bureau of Census survey of Residential finance on ownership characteristics of tenant occupied residential properties. The striking impression from these statistics is that it is very difficult to choose any representative investment. The rental housing stock is composed of highly diversified housing, with the small 1-4 unit structures representing a large fraction of the units. Many of these units may have filtered into the rental housing stock out of owner-occupied housing, where tax rules could have played a very different role. ¹¹ Moreover, the partnership arrangement is not the predominant form of organization and indeed only barely becomes the majority of the properties in the 50+ unit properties.

Corporate sector holdings which may become crucial to an economic analysis (for reasons noted below) are in the minority, but are not negligible. While corporate ownership is extremely small in the 1-4 unit category of properties, it becomes significant in the multi-family properties representing about 12 percent in the largest category. In addition, corporations are also represented among partnerships, as partnerships can be of individuals, corporations, or corporations and individuals. ¹²

In addition to the individuals, partnerships, and corporations which hold rental property, there is an "other" category which contains a mix of types. These include real estate investment trusts, financial institutions, housing co-operatives, churches, and other unspecified types. This data source is restricted to privately owned housing, but there is a substantial amount of housing owned by governments.

             TABLE 1. Summary Characteristics of Privately

Also of interest in the table are the mortgage to asset value ratios. These ratios, which range from 26 percent for the 1-4 unit properties to 47 percent for the 50+ unit properties, are considerably lower than those implied by the tax shelter models. Indeed, this degree of borrowing is not greatly out of line with corporate sector debt to asset ratios, which are around 40 percent. ¹³

These conclusions regarding debt to asset ratios can also be confirmed for corporate sector investment with data from corporate tax returns. These data include physical assets which are not adjusted for inflation, and therefore, the overall debt to asset ratio tends to be overstated. In 1982, the aggregate debt/asset ratio for all nonfinancial corporations was 60 percent, while the debt to asset ratio for lessors of buildings was 71 percent. ¹⁴ The leveraging ratio for lessors of buildings is, however, more overstated than that for corporations in general because the assets of lessors are almost entirely composed of buildings carried on the books at outdated acquisition prices. The assets of corporations in general include a large fraction of inventories which are mostly valued at current prices. If one grosses up depreciable assets of non-financial corporations by the factor necessary to obtain the 40 percent ratio estimated from other sources, and uses the same factor to adjust depreciable asset values for lessors of buildings, the two ratios are identical.

Of course, these debt ratios appear after a period of rapid price inflation and may be at an unusually low level since asset prices rise during inflationary periods while mortgage debt outstanding does not. This same effect would occur for corporate debt in general, but corporate debt in general has shorter maturities. The census survey also, however, reports initial borrowing rates. In the case of corporations, the initial share borrowed is 66 percent. The average degree of leveraging over the life of the property will, of course, be smaller because the asset value rises with inflation and the outstanding mortgage does not. These initial borrowing rates appear reasonably consistent with the stock measures reported above. ¹⁵

The Census survey also provides data on the distribution of properties by construction time and time of acquisition. By combining these two sources of information, it is possible to determine what the average holding period for properties is. This exercise was performed for the larger unit properties which would be expected to change hands more frequently and was estimated at approximately 20 years.

On the whole, therefore, the data on statistical aggregates do not provide obvious support for modeling the rental housing market as one driven by the typical "tax shelter" arrangements. Leveraging and turnover rates are much lower in these statistical averages and the ownership of rental properties is spread through an array of ownership types.

As noted earlier, an important question is the extent to which corporations own rental housing. The census data shows corporations to be participants in the market. Weighting the categories, the corporate sector holds 5.1 percent of the total housing stock and 9.3 percent of the multifamily housing stock.

This data source, of course, represents the stock of assets and not the investment amounts. Since syndicated tax shelter partnerships largely developed in the 1960s, there is some possibility that corporate shares of the stock of properties represent old investments. These old investments might, in turn, be trapped in the corporate form, but new investments (by construction or acquisition) may be made by partnerships. In fact, comparison of the 1971 census with the 1981 census shows that the corporate form declined in ownership share of 50+ unit properties from 29 percent to 12 percent. Corporate ownership also declined, although by smaller proportions, in the other two categories.

Yet, the Census data do not support the absence of the corporate sector investment. Other data from the Census survey show that the corporate sector share of mortgaged properties acquired between 1977 and 1981 was 9.9 percent and thus the corporate sector was clearly recently active in the market, even though its share might have declined over the seventies. Moreover, there are certain drawbacks to the Census data. The estimates are in terms of properties (which in some cases might not be contiguous) rather than units or value, and thus the corporate share of properties could change if there were a shift in the average size of properties even if the share of units remained constant.

Indeed, there appeared to be some major shifts in the characteristics of properties, with 50+ unit properties growing faster than other categories. Thus, it is very difficult to determine exactly what caused the decline in the corporate share or how good a proxy the observed stock share might be for the long run equilibrium. Another drawback is that corporate ownership through a partnership form is not represented in the Census data. To the extent that some real estate corporations were involved in syndicated partnerships, but maintained an ownership share, such a share would be reported as partnership rather than corporate ownership.

Similarly, the mortgage to value ratios reflect characteristics of the existing stock rather than the expected mortgage to value ratio over the life of the property for a new investment. As noted earlier, however, the census study does report initial borrowing rates. (Averaged over all properties, these rates for 1977-1981 are 57 percent for individuals, 68 percent for partnerships, and 66 percent for corporations.) These initial borrowing rates do not, however, provide a proper estimate of leveraging, because the ratio of the mortgage to asset value will decline over the holding period, since the mortgage principal declines and the asset value increases during periods of inflation. In any case, the important issue in debt finance and taxes is not the level of debt finance but how that level compares to other types of investments, and values based on the stock of assets are the only values which are comparable across diverse investments.

Finally, the holding period estimate relates to the existing stock rather than the expected holding period and we have no way of knowing how the two compare or the nature of the distribution of properties across holding periods. Some analysis of the optimal holding period will be undertaken in the next section which suggests that longer holding periods are not only observable but desirable for tax reasons.

The final issue we wish to consider in this survey of statistical characteristics of rental housing is how important the "tax shelter" is in the total investment in rental housing. Answering this question involves certain subjective (e.g., what is a tax shelter?) judgments and shortcomings in available data. Nevertheless, such information as can be derived may be useful.

The Census data presented in table 1 indicate that individuals own roughly 69.8 percent of residential rental properties, partnerships own 18.5 percent, and corporations own 4.6 percent. Tax shelter operations are in the form of partnerships; however, all partnerships are not tax shelter investments. Tax shelters tend to have a fairly large number of partners. Yet the vast majority of all partnership income is in firms with a very small number of partners. According to tax statistics, a third of partnership gross receipts in the industry category "lessors of buildings" is earned by partnerships with two partners. ¹⁶ Choosing any particular dividing line is somewhat arbitrary, but if we assign all partnerships with more than ten partners as "tax shelters" (which may be overstating the case somewhat), only 22 percent of gross receipts in real estate is earned by these partnerships. Multiplying this 22 percent share by the 18.5 percent total partnership share yields a "large partnership" share of about 4 percent.

This measure probably understates the share of assets held by large partnerships because there are different definitions of partnerships in the Census and tax data. Only legal partnerships would be classified as such in the Census data. Some informal partnerships will, however, file partnership tax returns. Moreover, the tax data include both residential and nonresidential lessors. Poterba uses data from tax shelter filings to estimate that tax shelters are responsible for nine percent of new residential investment. ¹⁷ He notes that this share is overstated because tax shelters sales include used buildings, gross investment includes only new construction. If, however, we use an average turnover rate of 20 years, and assume a real growth rate of three percent, gross investment should be increased by about 30 percent to include sales of existing properties. This calculation results in a share of seven percent.

Thus tax shelters appear to be responsible for only a small percentage of the supply of rental housing in the United States. Indeed, their share is similar in magnitude to the corporate sector. Adjusting these numbers roughly for multi-family housing implies that both corporations and tax shelters have a similar share of the multifamily market at around ten percent.

The share of investment represented by tax shelters is important not only because of the leveraging and holding period assumptions, but also because of the level and change in marginal tax rates. Virtually every analysis assumes that the individual owners of rental housing are in the top bracket or in a very high bracket. ¹⁸ Indeed, tax shelters are typically marketed only to these groups. These high bracket taxpayers would typically be the investors most sensitive to tax changes. If most investors in the real estate market are not high bracket taxpayers, changes in the tax rules would have a smaller effect on them. In fact, estimates by the Congressional Budget Office indicate that the average marginal tax rate on rent losses was 20.3 percent weighted by losses and 22.3 percent weighted by units before changes in the Tax Reform Act of 1986. The average on rent income was 30.2 percent weighted by income and 22.8 percent weighted by units. (These rent items would include some income from small partnerships not involved in a trade or business, as well as direct individual owners of real estate properties). These rates are dramatically lower than the rates assumed in most studies of the real estate market. ¹⁹

The information presented in this section suggests several conclusions. First, the ownership of residential rental property is quite mixed as to type of ownership and one is hard pressed to actually find a "typical" investment. Secondly, many of the assumptions about typical arrangements appear questionable. The extremely high debt to asset ratios often assumed are not typical of the statistical aggregates. Indeed, there appears to be no more leveraging going on in the area of multi-family rental than in corporate business enterprise in general. Thirdly, there does not appear to be a great deal of evidence that properties turn over so frequently, and in the following section we will suggest some reasons for expecting that such turnovers might not be frequent. Finally, investments in real estate are not dominated by high tax bracket individuals.

ECONOMIC EFFECTS OF TAX POLICY: AN ANALYTICAL OVERVIEW

The consequences of tax policy on prices and quantities depend on the effects of these policies on the marginal supplier. Thus, a first step in assessing the economic effects of taxes on rental housing, when faced with the heterogeneity of supply, is to aggregate those suppliers into a market supply function. One approach would simply be to add up all of suppliers and weight any estimated change in required price by the share of original output.

Such an approach would normally be considered rather naive for economic analysis. The more appropriate approach is to try to determine how the various suppliers react not only to the initial tax changes, but also to other changes which occur in the economy. Frequently such an analysis can lead to the identification of a marginal investor whose circumstances ultimately set the price and/or quantity of output. The marginal investor may not be the dominant supplier in the market and thus statistics on the shares of production may not tell us very much about the long run price and output effects of tax policy changes. (As a simple illustration of this point, one has to take into account the actions of a single oil producer in the Middle East as the major determinant of the price of oil in the United States, even though this producer accounts for only a small share of oil consumed).

Economic analysis of marginal behavior may also be used to try to determine the effects of leveraging and the influence of taxes on holding periods. Certain aspects of the marginal investor and the behavior of that marginal investor are crucial to determining the effect of taxes. These important aspects include: (1) whether the marginal investor is an individual, a corporation, or an entirely tax-exempt entity (such as government), (2) the debt/asset ratio and how the investor responds to changes which alter the relative "price" of debt and equity, and (3) how frequently the property is turned over and how tax burdens are altered with turnover ratios. We turn first to the determination of the marginal investor.

WHO IS THE MARGINAL INVESTOR?

If one holds that the marginal investor is the tax shelter investor, then prior tax law may be viewed as favoring residential rental construction and, by the same token, a general base broadening reform which lowers rates could reduce the degree of preference. If one holds that the marginal investor is the corporation, the opposite conclusion is likely given that previous rules disfavor structures within the corporate sector. And, if one holds that the marginal investor is a tax exempt entity, then tax rules have effects only by affecting taxes on other forms of activity.

Standard microeconomic theory would suggest that the marginal investor has to be the high cost investor. Assuming that the basic technology of operating rental housing is not differentiated by legal form of operation, this marginal investor would be the corporate sector, which bears the heaviest tax. Despite this rather straightforward finding, much analysis assumes that the marginal investor is the individual, and that, moreover, it is the individual investor in a tax shelter. To explain how these notions developed, a brief history of economic analysis of the corporate income tax is in order.

A model which has dominated this analysis since the early sixties is the Harberger model. ²⁰ This model was developed to explain the presence of a heavily taxed sector -- the corporate sector -- and to assess the effects of the tax. In this model, there was a corporate sector and a noncorporate sector. Investors in either sector required the same after-tax return and the effect of the corporate tax was to drive capital out of the corporate sector and into the noncorporate. The effect was a misallocation of capital, and a distortion of prices, with prices in the corporate sector rising relative to those in the noncorporate sector.

The model was applied to corporate versus noncorporate forms of legal organization, even though the model required that each sector produce different products. The fact that industries producing similar or identical products contained both a corporate and a noncorporate production sector was generally not addressed in the use of the model. ²¹ Instead, those industries which were dominated by the corporate sector (such as manufacturing) tended to become thought of as corporate while those industries which were dominated by the noncorporate form (such as agriculture and real estate) came to be thought of as noncorporate.

It is less clear, however, how the high bracket tax shelter investor came to be thought of as the primary investor in real estate, since the data on ownership of rental housing do not support such a thesis if one naively judges marginal investors to be simply the most predominant ones, or aggregates of suppliers. It seems likely that the publicity surrounding tax shelters in real estate led to an assumption that these investors were, indeed, the important ones.

Recently, the Harberger model has been challenged. ²² Harberger's model relies on price differentials between the corporate and the noncorporate sectors, which implies that the sectors produce different products. This assumption is counterfactual. While certain types of output are specialized in the corporate sector (e.g. automobiles), the corporate sector produces the range of goods and services produced by the noncorporate sector. Since only one price can prevail in the marketplace, the price differentials outlined in the Harberger model cannot exist.

In other words, application of the Harberger model requires that renters pay higher prices for apartments owned by corporations than for identical apartments owned by unincorporated individuals. Similarly, for other sectors where corporate and noncorporate activity occur one would have to argue that a barrel of corporate produced oil would sell for a higher price than a barrel of noncorporate produced oil for this type of model to be valid.

An alternative model eliminates the need for price differentials by allowing different suppliers in the market to earn different aftertax returns. ²³ Individuals are willing to earn different returns because the investments differ in their basic characteristics (for example, a corporate stock investment is likely to be less risky, and more liquid, than an individual investment). The shares of the market held by individuals versus corporations will depend on many factors, including tax rules.

The basic assumption of this model is that direct (noncorporate) investments and investments in corporate stock are not perfect substitutes and the shares individuals are willing to hold of each depend not only on return after tax but on riskiness and other factors. Such a theory allows both the corporate and the noncorporate sector to coexist and produce identical products facing a common market price. Such a model is not counterfactual.

The identification of the marginal investor in such a model can be as straightforward as elementary microeconomic theory which suggests that the marginal investor is the high cost investor. One can also, however, test for the marginal investor by reasoning through the process of the tax change. Suppose that there are three investments facing an individual: corporate stock, direct investment in real estate, and direct investment in an alternative composite asset (representing other assets in the economy). Since corporate stock is a homogeneous financial investment, earnings from corporate investment must be the same after tax (net of risk) whether invested in real estate or the composite investment. Consider now a tax increase confined to the tax on individual investment in real estate. If individual real estate investment is not marginal, as the theory suggests, price and output would not be affected in the long run.

The increased tax, in a fixed capital stock world, would cause individuals to reduce their direct holdings of real estate and increase their holdings of other assets. Initially, real estate rents would increase and prices of other products would decline. The corporate sector is now, however, in disequilibrium, with real estate earning too high a rate of return and the composite asset earning too low a return. Thus, corporate sector investment will flow into the real estate sector until in final equilibrium the two aftertax returns are equalized. With a fixed capital stock there is a unique set of prices and quantities which will lead to corporate sector equilibrium -- the initial starting point. A change in the tax on the individual will not affect price and quantity. A change in the corporate tax, by contrast, will change the equilibrium prices and quantities. Thus, the marginal conditions are determined by the corporate sector and, in the case of rental housing, we would find rental housing to be disfavored because it is not eligible for the investment tax credit.

Other investors in the sector are basically price takers. Changes in tax rules could, therefore, alter the ownership allocation, with the corporate sector, for example, displacing individual investors in the face of general tax reform.

In short, the identification of the corporate sector as the marginal investor depends on two basic requirements: (1) that a more heavily taxed corporate sector produces the same product as the noncorporate sector, and (2) that corporate aftertax returns (net of risk) must be equated. When these conditions are met, the results follow mathematically.

In the case of real estate, it has been suggested that the corporate presence reflects holdings of older properties, or holdings of Subchapter S corporations which are taxed as partnerships, or corporations which are small and subject to the lower rates which applied to small corporations under prior law (where tax rates might be lower than individual rates). The first argument was shown to be incorrect in the previous section. The Subchapter S form is of minor importance, representing only about 5 percent of corporations (operators and lessors of real estate). Similarly, corporations with taxable income low enough to tax them at the lower corporate rates represent only about 20 percent of asset holdings in that industry.

Another, and perhaps more important, argument is that the corporate sector and the noncorporate sector are not really producing a homogeneous product. The corner store and the supermarket may sell products for different prices because of location, hours, speed of service, or whatever. One could similarly argue that rental housing is differentiated in type and location, and the corporate sector might not be present in all basic types and locations. Thus, individual investors may be marginal in some markets. This may be a particularly important argument for detached housing in small towns and rural areas, and for smaller rental properties in urban areas. The marginal individual investors in this market are not, however, likely to be the large tax shelter partnerships, but rather the small scale individual investors who hold the lion's share of small properties.

Most large multifamily properties are located in metropolitan areas and both tax sheters and corporations are likely participants in this market in most urban areas. There are also some additional comments which might be made about this theory. First, even if each rental unit is differentiated, the prices constrain each other. That is, the presence of the corporate sector renting apartments of some type does affect the rents of other apartments which may be highly substitutable. In the case of real estate investment, however, there is an even more compelling argument against this heterogeneity argument. Although the product may be heterogeneous to consumers, there is a third market here -- the tax shelter market itself. Since this market is a national market and one real estate tax shelter should be substitutable for another, the returns should be constrained to equate across types of housing and locations net of differential risk. ²⁴

By contrast with this analysis of the corporate sector as the marginal investor, at least in multifamily housing, there is little to suggest that one should view the top brackt tax shelter investors as marginal. Indeed, one would be most likely to view them as inframarginal investors much as is commonly assumed in the case of holdings of tax exempt bonds. That is, high bracket individuals are in somewhat inelastic supply and are able to earn excess returns on their investment because of that fact.

THE QUESTION OF DEBT FINANCE

A second reason that rental housing (and other real estate) is often viewed as subject to preferential taxation is based on tax preferences for debt as opposed to equity finance. The argument is made that structures are more amenable to being heavily debt financed than other assets (for example, because their values are easier to determine, they are fixed in place, and there is a resale market). Debt is preferentially treated under the tax law. In the corporate sector, interest is subjected to only one level of tax (the individual level), as it is deductible. In the case of equity, both corporate and individual taxes apply. In the case of individuals, debt would appear to carry no special tax advantage as compared to foregone equity investment of the same risk class. However, high income individuals face a lower implicit tax rate on tax exempt State and local bond interest, so that the leveraging ratio would appear to be relevant for some individuals.

If the corporate sector is the marginal investor, then the focus should be on the corporate use of debt. Current economic theory, however, does not necessarily hold that firms benefit from the full value of tax savings due to debt finance. If there were a tax benefit to debt, but no cost, firms would tend to increase the share of debt without limit, so as to maximize tax benefits. But use of debt is risky and increases the possibility of a costly bankruptcy. Thus, at the margin, the tax benefit of additional debt just offsets the increased risk of bankruptcy so that the marginal tax advantage to debt disappears. ²⁵ For any investment, however, the cost of debt is the average risk of debt on the entire investment, rather than the increased risk of borrowing an additional dollar. Thus the advantage of debt finance is not the actual tax benefit to debt but the difference between the average and the marginal risk of bankruptcy, a smaller number and one which may be negligible if the marginal cost of debt rises very slowly.

Of course, this does not mean that individuals cannot use leveraging to increase their aftertax return on rental housing but these higher returns at the margin also compensate them for risk. (There may be substantial benefits for inframarginal investments, however).

One of the most interesting aspects of this debate, however, is that the statistical data reported in Section II suggest that rental housing is no more leveraged than other types of investment. Thus, even if there is some tax benefit to leveraged investment, there would be no empirical support for translating this benefit into a tax subsidy for residential structures.

THE QUESTION OF OPTIMAL TURNOVER PERIOD

A third reason for the claim that the tax system favors structures, including rental housing, is that one can convert ordinary income into capital gains, which were subject to lower taxes under pre-1987 law. That is, since depreciation is accelerated relative to economic decline, investors could take depreciation deductions which offset ordinary income, but repay those deductions upon sale at the lower capital gains tax rate. For these reasons, the tax law has generally contained rules which require a portion of gain to be taxed at ordinary rates (recaptured).

These rules have always been less stringent for housing than for nonresidential property. Under pre-1987 tax law, for example, for rental housing ordinary rates applied only to the difference between depreciation figured at straight line rates and that figured at accelerated rates. (No recapture applies to the acceleration due to the short write-off period). For nonresidential structures all depreciation is recaptured, unless straight line depreciation is originally used. The same rule applies to equipment. Even with recapture rules, it is widely thought that structures enjoy considerable advantage from frequent resale. This "churning" aspect was one of the stated reasons for extending the write-off period for structures from 15 to 18 years in the Deficit Reduction Act of 1984.

Any investment could be sold and potentially benefit from the conversion of ordinary income into capital gains, although recapture rules were more favorable for residential structures. But all structures are generally thought to have benefited because there is a more developed resale market for these assets. Of course, there is also a well-developed resale market for some other types of assets, such as automobiles; many types of assets are not easily resold, however.

For many types of investments, the tax law is presumed to encourage a longer holding period, because sale induces a capital gains tax. In the case of depreciable property, this capital gains tax penalty is offset by the advantage of increasing depreciation deductions on the new investment. When a property is held by the original owner, depreciation deductions decline over time under current depreciation rules and fail to keep pace with inflation. On resale, the basis becomes adjusted for inflation and depreciation starts over with more rapid rates. If capital gains are not taxed at full rates, it is possible to reduce total tax burdens through resale.

Whether there was an advantage to turning over property from a tax standpoint, and therefore whether the taxation of capital gains at full rates under the 1986 Tax Reform Act eliminated this advantage, depends on the relative advantages of the increased depreciation deductions and the capital gains taxes. Appendix B presents a model used to estimate these relationships. One of the first outcomes of this modeling is that the optimal time to sell -- if it was ever optimal to sell -- was after depreciation is completed. ²⁶ In the case of prior law, this optimal period was 19 years, a period of time far greater than that which is often assumed for tax shelter operations.

Such a finding presents a puzzle if tax shelter prospectuses are normally set up to involve much shorter holding periods. The most likely explanation is that investors do not wish to commit themselves to very long term investments, since in the typical tax shelter arrangement the partnership interest terminates with the sale of the property. Even ten years is a substantial holding period for an investment. Individuals engaged in putting together tax shelter arrangements might wish to encourage shorter holding periods in order to increase the frequency of commissions. Sale might be associated with refinancing to increase the leveraging rate, although there is no reason that refinancing could not be undertaken without a sale.

An alternative explanation is that such behavior was based on misinformation about the advantages of sale. For example, individuals may have believed that sale was optimal when the properties began to generate taxable income rather than tax losses. ²⁷ This decision rule would, however, have led to higher tax burdens.

It may still not be desirable to sell at 19 years. Several factors influence the presence and degree of tax advantage to sale. Since the advantage depends on the exchanging of a capital gains tax payment for the right to a stream of larger depreciation deductions, the nominal discount rate applied to that stream will affect the value of selling. Lower nominal discount rates -- due to either lower inflation rates or lower real discount rates (that is, yield on alternative investments) -- make it more likely that selling will increase rates of return. Lower real discount rates also increase the value because the advantage of sale occurs in the future and this value must be discounted. This discount rate can be very important.

One way to translate the effect of sale into simple terms is to express it as a percentage change in the net rentals required to earn a given return on the investment. If net rentals decline, then there is an advantage to selling which would ultimately lead to more investment, more housing, and lower rents. The size of the percentage change would provide some notion of how valuable the sale is. Another way of expressing it is to measure it as an effective tax rate. If sale is beneficial, it will cause the effective tax rate on the return from the investment to fall.

To illustrate this effect at its maximum we calculate these effects for the highest tax bracket (any effects will be smaller, the lower the tax bracket of the investor). For example, assuming an individual under pre-1987 tax law is in the 50 percent tax bracket, and an inflation rate of five percent, the effect of selling using a five percent real discount rate (a nominal rate of ten percent) would be to reduce net rentals by 1.6 percent and reduce the effective tax rate from 36 to 35 percent. These effects are quite modest. (This example, rather unrealistically, assumes that there is no land component and no transactions cost.) If the discount rate were one percent, the effect of selling on rental price would be a reduction of 41 percent and the tax rate would fall from 41 percent to minus 61 percent. These effects are quite large, although they would be smaller for lower tax rates. If the discount rate were 10 percent, it would never be advantageous to sell.

The appropriate discount rate is, therefore, of great importance. A five percent discount rate reflects levels commonly used in corporate tax analysis and associated with historical levels of corporate discount rates. Individual aftertax returns are normally higher and the returns in tax shelter investments often exceed a ten percent real return, which would imply no advantage to selling even under the generous circumstances of no land component and no transactions costs. Hendershott and Ling used very low discount rates (and found important consequences for trading). ²⁸ Their argument was that tax benefits are certain and should be discounted at a low rate. These tax benefits from selling are not certain, however, because they depend on the property's sales price which is a risky component of return. Thus this argument, while applicable to initial tax benefits which are certain, would not apply to tax benefits from trading. ²⁹

The advantages to trading, if any, become considerably smaller if one incorporates land components and transactions costs. Using the values assumed in Hendershott and Ling, ³⁰ the advantage to selling becomes negligible (a reduction in net rentals of .3 percent) when the land component is included with the five percent discount rate. When transactions costs are also added, it is no longer advantageous to sell at all. ³¹ In addition, any benefits to or costs of trading for individuals represent the maximum effects because they employ the top marginal tax bracket. As the individual tax bracket declines, all tax influences on investment become smaller.

Finally, the advantages for trading under prior law were much less for the corporate sector as compared to the high bracket individual because the differences between the ordinary and capital gains tax rates are smaller. For example, the individual investor in a 50 percent tax bracket faced a 20 percent capital gains tax rate. The corporate investor in the top 46 percent tax bracket faced a 28 percent capital gains rate. There were no cases, within reason, where there was an advantage to trading for the corporate sector. ³² Thus, if the corporate sector is the marginal investor, it would seem unlikely that turning over of property constitutes a tax advantage which affects the allocation of resources to housing.

There is one aspect of the sale of property which could change this analysis considerably. That aspect is the possibility of installment sales. In an installment sale, the payments are deferred and capital gains tax is not due until payment is actually received. By deferring the capital gains tax payment, the penalty for selling is reduced. ³³

A closer examination of this issue suggests that the installment sales option is probably of minor importance. Data on capital gains realizations through installment sales indicates they represent only about ten percent of capital gains from sales of capital assets (excluding securities.) ³⁴ Moreover, there are some important constraints on the use of this option which cause its value to be lessened.

Consider first the fairly straightforward option of exchanging the property for a mortgage and simply delaying the recovery of the original investment. This action would force the investor to convert funds which could have otherwise been used in real estate to investment in interest bearing assets (i.e. the installment obligation). The tax on interest bearing assets can be, however, quite high, especially in periods of inflation. Thus, even if both assets earned the same pre-tax rate of return, there would be no tax advantage. The proportional savings from not paying the capital gains tax would be more than offset by the loss from the heavy taxation of interest income. Further complicating this issue is the fact that the assets probably do differ in riskiness and the individual would probably also undertake reduced risk and return which might not otherwise be desirable.

An investor could, in theory, enjoy the benefits of installment sale without any of these drawbacks by borrowing money to provide cash to continue investing in real estate. (The possibilities for doing this are limited by new restrictions in the 1986 Tax Reform Act). Indeed, if borrowing were in identical amounts with identical terms, taxpayers could enjoy the cash effects of a sale, without paying the capital gains tax. There are a number of constraints on this activity which make it not entirely straightforward. First, such an arrangement might be disallowed by the Internal Revenue Service if there were a direct link between the installment obligation and the borrowed funds (such as pledging the installment obligation as security). Yet, without such a direct link, the taxpayer might in some cases have difficulty obtaining funds at a reasonable rate. Moreover, such arrangements may involve incurring additional risk (because the investor is obligated to repay the loan, but holds a risky note) and additional transactions costs, as well as the risk that the Internal Revenue Service will disallow installment sales treatment.

These observations suggest that even when installment sales are employed, the investor does not benefit from the full value of the tax payments due to these constraints on the use of the installment method. Thus, both by virtue of lack of quantitative importance and because of the constraints which apply to the investor, installment sales would appear to be of little importance.

Note also that the corporation, which has been argued to be the marginal investor and which might be most suited to engaging in borrowing to preserve cash flow, finds selling costly in the absence of installment sales. When considering installment sales, we find that even ignoring the problems associated with borrowing and assuming that no additional risk occurs, trading would provide a negligible benefit even under very generous assumptions. ³⁵ Thus, while installment sales provisions are relevant to the analysis, they do not suggest that turning over property provides a significant marginal subsidy to rental housing. The installment sales issue may, however, be relevant to issues of tax equity and simplification.

EFFECTS Of TAX REFORM ON RENTAL HOUSING

The Tax Reform Act of 1986 made a number of sweeping changes in the tax treatment of different types of assets. Tax rates were lowered considerably for high bracket individuals, where the maximum rate fell from 50 percent to 28 percent (although the tax rate could rise effectively as high as 33 percent due to the phaseout of certain exemptions). The corporate rate was reduced from 46 percent to 34 percent. The investment credit, which applied to equipment, was repealed and the depreciation rules were made less liberal, particularly for real estate. In the case of individual investors, a passive loss restraint was enacted, which disallowed deductions in excess of income. (An exemption was allowed for certain investors with direct involvement in managing property, however).

It is in estimating the effects of these tax reforms on rental housing that one can see the importance in tax policy analysis of the basic issue of modeling the marginal investment. One's expectation of results is straightforward. If rental housing tends to be a relatively heavily taxed investment, tax reform will reallocate resources into rental housing and reduce rents in the long run. If rental housing is, however, a relatively lightly taxed investment, tax reform will contract the housing stock and raise rents in the long run.

Several studies which were partial equilibrium in nature and focused on the real estate industry predicted rather pronounced price increases for rental housing as a consequence of the Tax Reform Act of 1986. The National Apartment Association estimated rents would rise from 14 to 35 percent; Hendershott and his coauthors found similar increases of 19 to 33 percent with fixed interest rates, with a decline to a 10 to 15 percent increase with a percentage point drop in the interest rate (although a more recent paper suggests an increase of 6 percent as most likely); Poterba found an 18 to 27 percent increase, falling to a 9 to 16 percent increase with a percentage point drop in the interest rate. ³⁶ Other studies using the tax shelter investor found similar results.

These studies suggest what would have to be considered massive increases in the levels of rents as a result of the tax act. And they all produce results which are of the same general magnitude. Each of the studies, however, basically treats the marginal investor as an individual in an extremely high tax bracket -- that is, the marginal investment is typically driven by the tax shelter investors who account for about 7 percent of the market.

The estimates of the effects of the Tax Reform Act of 1986 in this study are done separately for each type of investor -- corporation, individual, non-tax shelter partnership, and tax shelter. The calculations use a standard discounted cash flow model just as in the case of the other studies, where rents are set at the amount necessary to earn a specified rate of return.

There are several parameters which can affect the results of these calculations. Several basic measures reflect those used in the Price Waterhouse study of tax shelters, including economic depreciation at 1.5 percent and the return to equity for tax shelter investors at 12.5 percent. ³⁷ In general, the percentage increase in price tends to be smaller the larger are these values. The estimates also employ an assumption of 20 percent of the cost in land, following Hendershott and Ling. ³⁸ The interest rate is set at 10 percent, and the inflation rate at 4 percent, values which seem reasonable in the current economic climate. Operating expenses and property tax rates are set at 4.5 percent of asset value. The mortgage is a level payment 30 year mortgage, and the final sales price reflects appreciation at the inflation rate and depreciation at the rate of economic depreciation.

Table 2 shows the basic assumptions as to marginal tax rates and initial borrowing rates of the various classes of suppliers. The average marginal tax rates used for individuals were provided by the Congressional Budget Office and are weighted by unit (although weighting by income and losses yields virtually identical results). The same tax rate was used for the small partnerships outside of tax shelters. The tax shelter investors are assumed to be in the top marginal tax brackets. (As a result, the average marginal tax rate for all partnership investments is 33 percent). The percentage of purchase price borrowed is taken from Census data, except for tax shelters which are set by assumption, consistent with values typically used for tax shelter analysis. Since the leveraging rate for tax shelters was above the aggregate partnership leveraging rate (68 percent) the rate on the remaining partnerships was lowered to produce the observed average for all partnerships.

    TABLE 2. Investor Characteristics in the Rental Housing Market

Finally, the equity return was calculated endogenously for all investors other than tax shelters (which were set by assumption) in order to yield a common rent for all investors.

In the case of individuals, small partnerships, and other, the 20 year holding period derived from Census data was used; in the case of tax shelters a shorter period of ten years consistent with values typically used in tax shelter analyses was used. In the case of corporations, properties were assumed to be held indefinitely, as there would seem no reason for corporations to have a finite investment horizon as in the case of individuals.

Three simulations of the possible effects of the Tax Reform Act of 1986 were done for these classes of investors. (Details of the methodology are presented in Appendix A). The first, termed "Tax Shelter Methodology" estimates the required increase in rents holding the investor's aftertax return on equity constant, similar to much of the tax shelter modeling done. In the case of corporate investment, the equity return to the corporation, before imposition of personal taxes, is held constant. These estimates include the passive loss restriction where it becomes a binding constraint, a result which is highly sensitive to initial borrowing rates.

A second approach follows that of Poterba, which is a more consistent approach to the debt/equity issue. Poterba's analysis acknowledges that there is no tax favoritism of debt for individual investors and thus, when tax rates fall, the opportunity cost of equity finance rises as well. This approach produces somewhat higher effects on rents. (In the case of the corporate sector the return to stockholders after tax is adjusted; in this particular simulation, there is no difference because individuals earn a higher after tax return on corporate sector income in any case).

Finally, the third simulation incorporates a one percentage point reduction in the interest rate, using Poterba's method. A reduction of this magnitude is consistent with estimates from the Congressional Research Service tax model and with work done by Galper, Lucke and Toder. ³⁹

The bottom line of this table provides a weighted average of these rent increases, assigning a zero tax rate value to the remaining suppliers (primarily tax exempt entities). In addition, government owned housing which is not included in the table (representing about 7 percent of the total stock), was assumed to be unchanged. The number in parentheses for tax shelters represents the only case where the passive loss restriction would become a binding constraint.

            TABLE 3. Effects on Rents, by Type of Investor

The effects for the tax shelter investors are somewhat below Poterba's results for the top bracket (as he found a 27 percent increase). The major reason for this difference is that the analysis here includes the 20 percent land component while the analysis in Poterba's study did not. With no land component the estimate, with no change in the interest rate, would be about 23 percent, quite close to his estimate.

The tax shelter estimates tend to fall below those reported in some other studies. The estimated rent increases for tax shelters using the tax shelter methodology are, however, extraordinarily sensitive to assumptions, particularly if the passive loss restriction is considered. For example, altering the aftertax rate of return from 12.5 percent to 11 percent would cause rent increases to rise by 11.3 percent rather than 10.3 percent without the passive loss restriction, while it would rise from 12.1 percent to 15.1 percent with the passive loss constraint. By contrast, the effect on individuals of changing the aftertax return by 1.5 percentage points would be to increase rents from 4.1 percent to 4.7 percent, and the passive loss constraint is not binding in any case. Even more important is the leveraging rate for the tax shelters. For example, if the leveraging ratio increased from 80 percent to 90 percent, the tax shelter model approach would produce an increase of 20 percent without the passive loss restraint and an increase of 33 percent with the passive loss restraint.

These estimates suggest that the long run effect on residential rents from the Tax Reform Act of 1986 would be relatively small if one judges by the corporate sector equilibrium, as the theory presented in this paper suggests. If interest rates fall as the general equilibrium models suggest, there would basically be no effect. Similar results are obtained if a weighted average of the suppliers is used. Certainly, nothing in the neighborhood of the 20 percent to 30 percent price increases could be justified short of using the tax shelter model as the representative investor. This approach, however, is not supported either on the grounds of dominance in the market place or from any theoretical rationale.

² An earlier version of this study made this argument: U.S. Library of Congress. Congressional Research Service. U.S. Tax Policy and Rental Housing: An Economic Analysis. Report 85-208 E, by Jane G. Gravelle, November 6, 1985. In addition, see Behravesh, Nariman and Kurt Karl. An Analysis of the Treasury's Tax Reform Proposals. Wharton Econometric Forecasting Associates, December 17, 1984; Brinner, Roger E., et.al. The Treasury Tax Proposals: Steps Toward Neutrality. Data Resources, Inc., January, 1985; U.S. Library of Congress. Congressional Research Service. Assessing Structural Tax Revision With Macroeconomic Models: The Treasury Tax Proposals and the Allocation of Investment. Report 85-645 E, by Jane G. Gravelle, December 27, 1984.

³ Economic Report of the President. January 1987, pp. 79-96, Fullerton, Don, Yolanda Henderson, and James Macke. Investment Allocation and Growth under the Tax Reform Act of 1986, preliminary, unpublished.

⁴ National Apartment Association (Housing and Development Reporter, January 26, 1987, pp. 744-745); Hendershott, Patric, James R. Follain, and David G. Ling. Real estate and the Tax Reform Act of 1986. National Bureau of Economic Research, Working paper 2098, December, 1986; Poterba, James M. Tax Reform and Residential Investment. December, 1986. Follain, Hendershott, and Ling. Understanding the Real Estate Provisions of Tax Reform: Motivation and Impact, May, 1987. A number of studies are summarized in Crone, Theodore, Housing Costs after Tax Reform, Federal Reserve Bank of Philadelphia Business Review, March/April 1987. Roger H. Gordon, James R. Hines, Jr., and Lawrence H. Summers suggest in Notes on the Tax Treatment of Structures, Harvard Institute of Economic Research Discussion Paper 1239, May 1986, that structures may not be disfavored relative to equipment because of their greater ability to carry debt and because of tax advantages on resale.

⁵ For example, Roger H. Gordon, James R. Hines, Jr., and Lawrence H. Summers suggest in Notes on the Tax Treatment of Structures, Harvard Institute of Economic Research Discussion Paper 1239, May 1986, that structures may not be disfavored relative to equipment because of their greater ability to carry debt and because of tax advantages on resale.

⁶ Lawrence H. Summers. A fair Tax Act That's Bad for Business. Harvard Business Review, Vol. 65, No. 2, March-April, 1987, p. 55.

⁷ U.S. Congress. Joint Committee on Internal Revenue Taxation. General Explanation of the Revenue Act of 1971. December 15, 1972, p. 40.

⁸ Public Citizen. Running for Shelter. February, 1985.

⁹ The advantage to debt is generally associated with lower riskiness, which would not likely be true of oil and gas. The advantage to selling is the ability to repeat generous depreciation deductions. While initial investment in oil and gas is treated fairly generously, through immediate deduction of a large fraction of investment, cost recovery on an existing property would be recovered through cost depletion which is not very generous; and not allowed at all when percentage depletion was allowed.

¹⁰ This study does not deal with the special case of low income housing, whose treatment in the Tax Reform Act was substantially different from that of rental housing in general.

¹¹ The effect of tax rules on owner-occupied housing which is subsequently converted to rental housing is more complex. Tax rules regarding rental housing affect the decision to convert, but tax rules affecting owner-occupied housing influence the original construction of these properties.

¹² Unfortunately, there are no data on how important this aspect of corporate ownership might be. Large, limited liability, partnerships usually have a corporate general partner who must have some significant ownership right (typically ten percent) in order to satisfy certain requirements to be treated as a partnership (Revenue Procedure 72-13).

¹³ U.S. Congress. Committee on Banking, Finance and Urban Affairs and the Subcommittee on Housing and Community Development. House of Representatives. Housing -- A Reader. Prepared by the Congressional Research Service, p. 84.

¹⁴ U.S. Department of Treasury. Internal Revenue Service. Statistics of Income. Corporate Sourcebook, 1982.

¹⁵ These results are different from those reported for real estate based on Compustat data in Fullerton, Don and Roger H. Gordon. A Re-examination of Tax Distortions in General Equilibrium Models. In Behavioral Simulation Methods in Tax Policy Analysis. Chicago, University of Chicago Press, 1983. They reported an aggregate ratio of 40 percent, but a ratio of 80 percent for real estate. The high estimated debt share was one of the two major reasons (the other being repeated resale) for claims made in a subsequent paper co- authored by Gordon with James R. Hines and Lawrence H. Summers, Notes on the Tax Treatment of Structures, that equipment may not necessarily be favored relative to structures. This 80 percent number is, however, greater than the overstated number for real estate in the corporate statistics of income. Moreover, it is considerably greater than the 66 percent initially borrowed reported in the Census data (which is in turn an overstatement of the effective leveraging rate over the life of the property, as the asset value increases with inflation while mortgage principal declines). It seems likely that their data for real estate reflects ratios for subdividers and developers, who tend to be very highly leveraged. The debt to asset ratio for subdividers and developers in the statistics of income data was 84 percent and is not likely to be seriously overstated because their assets are not primarily composed of understated depreciable property.

¹⁶ U.S. Department of Treasury. Internal Revenue Service. Statistics of Income. Partnership Returns. 1978-82.

¹⁷ Poterba, Tax Reform and Residential Investment Incentives, p. 10.

¹⁸ The use of very high marginal tax rates has been incorporated in the academic studies which purport to examine the entire market as well as in the studies aimed explicitly at the tax shelter investor. For example Hendershott, et al, Real Estate and the Tax Reform Act of 1986, employs the top tax rate and a rate of 25 percent, plus State and local taxes. Their subsequent paper, Understanding the Real Estate Provisions of Tax Reform: Motivation and Impact, employs a 42 percent rate. Poterba, Tax Reform and Residential Investment, uses the top tax rate and a rate of 40 percent. Gordon, Hines, and Summers, Notes on the Tax Treatment of Structures, employ the top rate in making their case for the advantage of churning for structures.

¹⁹ The tax rate data are inconsistent with the popular notion that real estate is held primarily by wealthy individuals in high tax brackets. Further evidence of widespead ownership of real estate across the income classes can be found in the census survey data, which indicated that a quarter of rental real estate is held by families with household income less than $24,000 and over half is held by families with household income less than $48,000. See U.S. Department of Commerce. Bureau of the Census. Household Wealth and Asset Ownership: 1984. Series P-70, No. 7, July, 1987.

²⁰ Harberger, Arnold. The Incidence of the Corporation Income Tax. Journal of Political Economy. Vol. 70, June, 1962, p. 215-240.

²¹ Harberger did acknowledge this issue, but considered it not to affect the results. The analysis which follows here, however, suggests that this is a crucial aspect of the corporate tax model.

²² Gravelle, Jane G. The Social Costs of Non-Neutral Taxation: Estimates for Non-Residential Capital. In Depreciation, Inflation and the Taxation of Income from Capital. Hulten, Charles R. (ed.). Washington, D.C., The Urban Institute, 1981, p. 239-250; Ebrill, Liam P. and David G. Hartman. On the Incidence and Excess Burden of the Corporation Income Tax. Public Finance. Vol. 37, 1982, p. 48-58.

²³ The analysis presented here applies to the large scale partnerships versus the corporate sector, both of which represent passive investments of the individual. For small scale directly managed investments of individuals, additional factor such as closer managerial control in non-corporate investment versus economies of scale in large investments may also differentiate the two investments.

²⁴ A more technical and subtle point is whether corporate stock invested in different assets is perfectly substitutable. In fact, portfolio theory suggests that the shares and returns on a portfolio are determined by the covariance of the asset with the rest of the portfolio. If the portfolio were confined to corporate stock, the aftertax return on corporate stock invested in real estate ought to rise relative to other returns. In this case, however, the share of real estate in the portfolio is actually remaining constant and the consequences for covariance of corporate stock invested in real estate are not clear. At any rate, the effects should be small, could conceivably be in the opposite direction, and are in need of further study.

²⁵ Gordon, Roger H. and Burton G. Malkiel. Corporation Finance. In How Taxes Affect Economic Behavior. Aaron, Henry J. and Pechman, Joseph H., eds., Washington, The Brookings Institution, 1981, p. 131-198.

²⁶ The same result was obtained by Patric H. Hendershott and David C. Ling. Trading and the Tax Shelter Value of Depreciable Real Estate. National Tax Journal. June, 1984, p. 213-224.

²⁷ Such an objective is considered in Alberts, William W. and Richard P. Castianias, II. The Impact of Changes in Tax Depreciation Rates on Holding Periods for Real Estate Investment. National Tax Journal. March, 1982, p. 43-54.

²⁸ Hendershott and Ling, Trading and the Tax Shelter Value of Depreciable Real Estate.

²⁹ Technically, one could argue that the trade-off between capital gains and depreciation is certain at the time of sale but that the value of that trade-off is uncertain because both depend on the sales price. Unfortunately, there is really no way to decompose the return into these two components. The effect would tend to make an advantage to selling more likely, but to make the advantage smaller in value.

³⁰ Land is assumed to constitute 20 percent of acquisition cost and transactions costs are five percent of sales price. Hendershott and Ling. Trading and the Tax Shelter Value of Depreciable Real Estate.

³¹ An advantage would still occur in the case of low discount rates. At a one percent rate, selling at 18 years would reduce net rentals by 6.2 percent after incorporating land costs and transactions costs.

³² A small advantage would occur in the case of the low discount rate with no land component and no transactions cost. In addition, for some types of corporations, sale might be accomplished via sale of stock rather than sale of the property itself, which would involve a lower, individual, capital gains tax.

³³ The installment sales options was the main device cited by Gordon, Hines, and Summers, Notes on the Tax Treatment of Structures, in reducing effective capital gains taxes and creating an advantage for structures through churning. The advantages they found were also increased by the use of the top marginal tax bracket and the lack of a land component, since there is always a tax cost to turning over land. Their analysis was done, however, for commercial real estate where churning is somewhat less valuable because of the restrictions on installment sales for recaptured depreciation.

³⁴ U.S. Department of Treasury. Internal Revenue Service. Sales of Capital Assets, 1981 and 1982, by Bobby Clark and David Paris. Statistics of Income Bulletin. Vol. 5, No. 3, Winter 1985-86, pp. 65-89.

³⁵ If we assume standard land components and transactions costs, a 12 percent interest rate and that the firm provides a 30 year level payment mortgage with a balloon payment after 18 years (when it becomes optimal to sell again), the effect is to reduce rentals by only about 3.7 percent and reduce the effective tax rate from 37 percent to 36 percent. This example assumed no down payment. If a down payment of 20 percent were included, there would be no advantage to selling. Since equipment tends to be taxed at rates which approach zero and owner occupied housing is subject to negative tax rates, this installment sales treatment is far from providing beneficial treatment for rental housing relative to other major assets.

Note that there has been considerable concern over the use of installment sales by builders, with bonds issued to provide them with cash. This phenomenon is not directly relevant to the issue at hand, as builders normally would sell their properties and are in the business of doing so.

³⁶ National Apartment Association; Hendershott, et al, Real Estate and the Tax Reform Act of 1986; Follain, et al, Understanding the Real Estate Provisions of Tax Reform; Poterba, Tax Reform and Residential Investment. The ranges for the Hendershott and Poterba papers represent different marginal tax rate assumptions -- the top rate and a somewhat lower rate. Both studies employ, even for their lower tax rate investor, relatively high income taxpayers (rates of 40 percent or more).

³⁷ Price Waterhouse, The Economic Impact of Tax Reform on Commercial Real Estate.

³⁸ Hendershott and Ling, Trading and the Tax Shelter Value of Real Estate.

³⁹ Galper, Harvey, Robert Lucke, and Eric Toder. Revenue Effects and Portfolio Behavior: A Simulation of the Tax Reform Act of 1986. Brookings Discussion Paper, September, 1986. Hendershott, et al, Real Estate and the Tax Reform Act of 1986, report a decline of 1.4 percentage points. The reason that interest rates would decline is that the tax changes would initially cause demand for investment, and thus financial capital, to contract, requiring a drop in the interest rate to equilibrate supply and demand.

APPENDIX A: The Discounted Cash Flow Model

The basic model used in this and other analyses of the effects of tax changes on new investment is a discounted cash flow model. The equations themselves are very complex, and therefore only the basic outline of the model will be presented here. The equilibrium is of the form:

 E = PVR*(1-t) + F*PVD*t - (1-E)PVI*(1-t) - (1-E)PVP + PVS - PVM-

The model can be set at a given rent and be solved for internal rate of return or set at a given internal rate of return and be solved for rent. The latter is employed in this study as in many of the other simulation studies. Note that a change in tax rate could either raise or lower the rent. The larger is depreciation, interest rates, and the fraction borrowed, the more likely a tax rate reduction is to produce an increase in rent. The larger the stream of rents, the more likely a tax reduction is to decrease rents. Since the size of the rent stream is affected by the economic depreciation rate, it is not only the level of tax depreciation but its relationship to economic depreciation which affects the role of depreciation in causing tax rate reductions to either increase or decrease rents.

In preparing the simulations reported in this paper, we calibrated all of the other investors by estimating rents for tax shelter investor and then setting equity returns for the remaining investors so that rents would be identical for all investors. This exercise produced a return of 8.2 percent for individuals, 8.6 percent for small partnerships, and 7.9 percent for corporations, as contrasted with a 12.5 percent return for tax shelter investors.

How does the passive loss restriction fit into this model? Basically, whether and when the passive loss restriction exerts a negative effect on the investment depends on the relationship between the positive cash flow -- rent, and the tax offsets -- interest and depreciation. The tax depreciation rules are set by law, although they are much smaller in the new law than in the pre-existing one. Hence the passive loss restriction is less important in the context of the new provisions than in the old. Moreover, the tax rate is lower so that the effect of restriction is also lessened.

Several other factors will also influence the constraint imposed by the passive loss restriction. The size of interest offsets depends on the amount borrowed. It also depends on the level of the NOMINAL interest rate. This level as compared to rents is strongly affected by inflation. Since rents tend to rise over time with the inflation rate but mortgage instruments charge a nominal payment depending on the nominal interest rate, inflation will make the passive loss restraint more likely to be important. Again, the more moderate inflation environment we find ourselves in today will mitigate the effect of the passive loss restriction. Finally, the larger the initial rent level, the less likely the passive loss restriction. This rent level depends on the projected depreciation in real rents and the level of the equity return.

Thus, the factors associated with the passive loss restriction playing an important role are high inflation rates, low economic depreciation rates, high real interest rates, low returns on equity, and a large fraction of purchase price borrowed.

The mechanics of solving the model with a passive loss restriction require the determination of when cumulated rents become as large as cumulated offsets. Until that point, no taxes are paid or credited and after that point tax payments apply to the excess of rents over offsets as in the standard case. If the passive loss restraint is binding over the entire holding period, cumulated losses are deducted when the property is sold and netted out against capital gains taxes. In actually finding a solution for rents, both equilibria must be satisfied. That is, the point at which the passive loss restraint is no longer binding is a function of the rent which in turn is a function of the effect of the passive loss constraint, so that two equations must be solved simultaneously, by an iterative process.

APPENDIX B. A Model of the Optimal Turnover Period

[This appendix contains mathematic equations which cannot be reproduced electronically. It is available through the Tax Notes Microfiche Data Base.]