CRS ADDRESSES IMPACT OF FLAT TAX, OTHER REFORM PLANS ON HOUSING.

                  THE FLAT TAX AND OTHER PROPOSALS

[1] Several proposals for major reform of the federal income tax system, including replacement of the current tax with a new type of tax, have been introduced or considered in the 104th Congress. Among the most widely discussed are the flat tax, a value added tax, a national sales tax, a proposal for a direct consumption tax (called the USA tax), and income tax reform. Most of these new taxes convert the tax base from an income to a consumption base, most eliminate deductions for mortgage interest and property taxes, and most flatten the rate structure -- in some cases by adopting a single tax rate. While these tax revisions touch on all aspects of the economy, this paper focuses on the effects on housing.

[2] Some recent studies have suggested a significant reduction in the price of houses -- perhaps in excess of 15% -- would occur if one of the major proposals, the flat tax, were adopted. Similar results might be expected from closely related proposals such as a value added tax.

[3] The principal finding of this study is that effects of the flat tax on housing prices are likely to be limited in the short run and very small in the long run. These findings derive in large part from expectations that the supply of housing for sale is not fixed, contrary to the apparent presumption in other studies, and indeed is likely to be very responsive to changes in demand in the long run. As a result, the effect of the tax over time will be to alter the amount of housing rather than the price. Short-run impacts on housing prices are less predictable but both theory and empirical evidence support some supply response. There are other factors as well, including the possibility of increased savings rates, that could mitigate any contraction in demand.

[4] Owner-occupied housing seems the asset that merits less concern with respect to price, since it is an asset normally protected, in both the flat tax and in other consumption taxes, from the lump-sum tax on old capital that will otherwise be imposed on firms. Indeed, current owners could actually gain under some types of tax revision.

[5] Demand for owner-occupied housing and home-ownership should contract relative to other asset demands in response to a shift to a consumption base, but that occurs because a burden on competing investments is lifted, and thus a relative preference to owner-occupied housing lifted. It is this increase in return on competing investment, and not the restriction in itemized deductions, that is the main source of shifting asset preferences from housing to business investment. By eliminating the distortion in choice due to the tax system, economic efficiency and welfare should be improved. Economists would normally regard this reallocation of resources as a long-run benefit rather than a cost of tax reform; the cost is in any undesirable distributional effects that cannot be offset with other policies.

[6] Rental housing demand would probably be encouraged with a shift to a consumption tax base.

                              CONTENTS

INTRODUCTION AND OVERVIEW

[7] Several major federal tax reform proposals, including replacement of the current income tax with a new type of tax, have been introduced or considered in the 104th Congress. ¹ Perhaps the most widely discussed proposal is the flat tax introduced by Representative Armey and Senator Shelby. Interest has developed in the possible impacts of the flat tax, and other proposals, on residential real estate, and particularly on the potential adverse effects on the price of owner-occupied housing.

[8] Most proposals would convert the tax base from an income to a consumption base, and most would flatten the rate structure -- in some cases by adopting a single tax rate. A number of these proposals would eliminate deductions for mortgage interest and property taxes, a treatment consistent with the treatment of business interest, which is neither deducted nor taxed under these consumption tax proposals.

[9] The issue of housing prices was highlighted by a study by DRI/McGraw-Hill (hereafter the DRI study). The DRI study estimated that eliminating mortgage interest and property tax deductions would be equivalent to losing 15% of current housing asset value. The study also estimated an overall fall in housing prices with the enactment of a flat tax in the neighborhood of 15%, most of which would occur immediately. ² These results are dependent on interest rate changes; without a fall in interest rates, the price reductions are estimated at 34%. A subsequent study by the same organization indicted that the drop in house prices would triple mortgage losses and double foreclosure rates. ³ In addition, a preliminary paper by Green, Hendershott, and Capozza that examines the effects of various tax revisions estimates that there would be: (1) substantial effects of the flat tax on housing prices in metropolitan areas; (2) more modest, but still significant, effects from a disallowance of mortgage interest and property tax deductions; and (3) virtually no effects from another proposal, the USA tax. ⁴ These effects vary from one locale to another, but are very large in magnitude.

[10] The first section of this study briefly describes the proposals and how they treat housing. The current proposals fall into five categories: the retail sales tax, the value added tax (VAT), the flat tax, the USA tax, and income tax reforms. The first four proposals have a consumption base, and the first three are levied at a flat rate, although the flat tax proposal, by taxing wages at the individual rather than the firm level, allows an exemption. The USA tax includes both a VAT and a direct consumption tax on individuals. Income tax reforms can include restrictions on itemized deductions, including mortgage interest and property taxes; current proposals disallow property tax deductions.

[11] Most of the analysis in the paper is of the flat tax, which was the subject of the DRI study, and whose rules have been specified in bills introduced. Findings relating to this tax are then used to examine the effects of other types of proposals.

[12] The next section of the paper clarifies the two different types of effects on assets of switching to a consumption tax base. A consumption tax allows a deduction for purchases of assets, but imposes a tax on the sale of assets. For this reason, a consumption tax is characterized as a tax on wages plus a lump-sum tax on old capital. Firms pay taxes on the sale of assets directly, and in the case of incorporated businesses this lump-sum tax on old assets is reflected in stock market prices, which fall. Owner-occupied housing is exempt from this tax on old assets under the flat tax -- indeed it is the only asset that is exempt.

[13] For new business investments, the deduction of cost when sold effectively eliminates the tax on the return to new investment, even though earnings from the investment, including its sales proceeds, are taxed when consumed. The elimination of taxes on the return to new investment causes investment to flow out of the currently tax favored owner-occupied housing sector (where returns are already effectively tax-exempt, because the implicit rent from owning a house is not taxed) into the business sector; this shift in demand can cause owner occupied housing prices to fall. It is this effect on owner-occupied housing prices, which may be largely transitory, that is the focus of the recent studies on asset prices.

[14] The following section of the paper examines what this effect on asset prices in likely to be. The effect on asset prices will depend on the magnitude of the increased cost of housing, which is affected by the loss of itemized deductions for mortgage interest and property taxes, but largely arises from the higher returns available on alternative assets. This cost increase may be diminished by any fall in pre-tax returns, and interest rates. Some fall in interest rates would be expected even in the absence of savings increases, because of the current tax subsidy for debt finance at the firm level; approximately a 1% fall in interest rates would occur for this reason. Additional reductions in rate of return will occur if savings increases, an outcome that is generally predicted by economic theory but whose magnitude is difficult to estimate. Several other factors could also moderate the effects on the cost, or on the shift in demand due to the change in cost. The increased cost of mortgages to cash-constrained first-time home-buyers should be largely offset by the automatic decline in the interest rate that would occur even without an increase in savings.

[15] Ignoring these factors, and assuming no increase in savings, the overall price of housing would fall by 22% as a result of the flat tax IF the supply of housing were fixed. Only a third of this effect is due to the loss of itemized deductions for mortgage interest and property taxes; the remainder is due to the possibility of earning higher returns elsewhere on equity that would be invested in the home. The reduction could be considerable less if savings increases. In general, these effects are smaller than those estimated by DRI for a given savings rate assumption. Larger properties would have larger price declines than smaller ones, in part because the current tax benefits are larger for the higher income individuals who purchase these properties.

[16] These price changes are the upper limit of expected effects (given the estimates of increased costs) because they assume a fixed supply. Even a limited amount of response in supply could strongly affect the price change. In the long run there are reasons to believe that supply response is such that there would be virtually no price change -- the price of housing largely reflects its cost of construction. Nor are prices in urban areas likely to be affected much in the long run since urban land prices are influenced by the availability of substitutes outside the cities, and because urban land has alternative uses.

[17] While there is more possibility of asset prices being affected in the short run, there are reasons, even in this case, to expect many factors to smooth and constrain the adjustment process. Historical evidence does not support the expectation of large price changes. Indeed, data presented in that section shows that in the early 1980s when the cost of owner-occupied housing rose substantially, there was only a negligible effect on prices.

[18] Rental housing should become more attractive under the flat tax, although the effects appear to be modest. Current owners of rental housing would, however, have to pay a tax on sale, even though they can deduct acquisitions; hence, rental housing is subject to the lump-sum tax on old capital.

[19] In the case of a VAT or retail sales tax, the effects will depend on how housing in treated. New owner-occupied housing can be included in the base (the normal presumption); it can be excluded from the base, or existing housing can be included. Assuming that only new houses will be included in the tax base, the effects would be similar to the flat tax, with one caveat. Since VAT and sales taxes are likely to require a price accommodation, current home- owners may actually receive a windfall from the lump-sum tax on old assets, since the increased construction cost will cause their prices of housing to rise, while the value of outstanding debt will be fixed. This permanent effect will cause their equity values to rise by more that the tax rate. If however, new housing is excluded, this price effect will not occur and owners of homes will experience an effective burden on their equity asset because other consumption goods will rise in cost. Owner-occupied housing will, however, retain its preferred status and there will be a much more limited effect on asset prices due to the shift of investment out of housing and into business use. Finally, existing housing could be included. In this case, there is both a lump-sum tax on housing (although housing prices will rise, the increased price will be needed to pay the tax and the sales proceeds will still purchase fewer consumer goods).

[20] The effects on rental housing demand would be similar to the effects of the flat tax.

[21] The USA tax has effects similar in many ways to the flat tax, although the ability to deduct mortgage interest will modify any effects by retaining part of the preferential treatment of owner- occupied housing. Most income tax reforms would be likely to have much more modest effects on owner-occupied housing.

[22] The likely outcome of the tax change would be to contract the amount spent on housing, and cause some individuals to make different tenancy choices. Economists would typically view this shift as a benefit of tax reform, because the current tax system encourages an inefficient overinvestment of funds in housing and consequent underinvestment in business assets. Such a shift would constitute a change of tax policy which has long favored home-ownership.

A BRIEF DESCRIPTION OF THE PROPOSALS

[23] Consumption tax proposals have included sales taxes, value added taxes (VATs), flat taxes, and direct taxes on individuals that use consumption rather than income as a base. There have also been proposals for income tax reform. In some cases, details of the tax proposals have not been specified.

SALES AND VALUE ADDED TAXES

[24] Straightforward consumption tax proposals include a retail sales tax or a value added tax (VAT). Retail sales taxes would impose a tax on products at the final stage of consumption. Senator Lugar and Representatives Schaefer and Tauzin have proposed a national sales tax. Sales taxes can either exclude or include purchases of housing in the base; and they can differentiate between new and existing housing.

[25] An alternative to the retail sales tax is the value added tax, or VAT, which has been suggested by Representative Gibbons. The VAT, in theory, is the same as the sales tax, except that it is collected at each stage of production. It is termed a value added tax because the tax base for any given firm is receipts minus purchases from other businesses, or value added. ⁵

[26] Unless a special deduction or credit is provided, either through a refund to the seller or a benefit to the purchaser, newly constructed owner-occupied housing would be included in the tax base as a consumption item. Practically speaking, however, it is probably impossible to provide full consumption treatment to the owner- occupied housing sector (which would require including the implicit rental value of the house in income while deducting the acquisition cost, and including proceeds from sales in the tax base). Rather, the consumption value is taxed by taxing the major inputs into housing services -- the structure itself and materials and services used in maintenance and repair. ⁶

THE FLAT TAX

[27] The flat tax proposal is among the most publicly visible of the tax reform proposals. This proposal has been advanced by Representative Armey (and introduced as H.R. 4584; with a companion bill, S. 1050, introduced by Senator Shelby; and a similar bill, S. 488, introduced by Senator Specter). A flat tax was also proposed by former presidential contender Steve Forbes.

[28] The flat tax would be imposed on wage income of individuals and in the form of a modified VAT on firms, where wages would be deducted from the base. Thus, the firm tax would be imposed on business receipts minus purchases, including capital goods, and wages paid. There is an exemption from the wage tax portion, so that the effective tax rate is graduated on wages.

[29] The flat tax is fundamentally a VAT with the part of the tax base reflecting wages taxed to individual workers rather than the firm. This treatment allows a wage deduction. Current versions do not allow a mortgage interest deduction, a treatment that is consistent with the treatment of business interest (which is also not deductible); nor is interest taxable to creditors. Because of the value-added-tax nature of the proposal, new housing construction would be effectively included in the base, but purchases of old housing would not be subject to tax on, or deducted by, the purchaser.

THE USA TAX

[30] Also under discussion is the Nunn-Domenici-Kerrey proposal for a combined direct consumption tax (with graduated rates) on individuals and a VAT on firms (S. 722), called the USA tax (for unlimited savings account). The direct consumption tax imposes the tax on income used by individuals for consumption purposes, rather than on the products. This tax is graduated, but has the same effect as an indirect products tax that rises with the overall level of consumption. The USA tax is modified in a number of ways, however, through a series of transition rules that allow firms and individuals to recover the basis of existing assets, in some cases, and by other rules that cause the proposal to differ substantially from a pure consumption tax. Specifically, it continues to allow mortgage interest deductions. The USA tax substitutes for both the income tax and the employer's share of the payroll tax.

INCOME TAX REFORM

[31] Finally, there are several proposals for individual income tax reform, which would broaden the base, including elimination of the property tax deduction, and flatten the rate structure. Representative Gephardt has proposed a flatter but graduated rate structure; Senator Gramm and Presidential contender Buchanan have proposed a flat rate with an exemption. Unlike the other proposals discussed, these proposals would retain income as the base for the tax. It would also be possible to formulate an income tax reform that disallowed the mortgage interest deduction.

CONSUMPTION TAXES AND ASSET PRICES

[32] There are two distinct asset-price effects of switching to a consumption tax base. First, the tax on old assets that is fundamental to a consumption tax may manifest itself in a fall in stock market prices. Second, asset prices, both physical and financial, can be affected in the course of altering the allocation and size of the capital stock. Because of the subject matter of this study -- housing -- the focus is largely on the second category of asset-price effects.

[33] Since the flat tax is a consumption tax, it imposes a tax on existing assets. This tax on existing assets is most clearly seen for self-employed individuals running an unincorporated business. Depreciation on existing assets cannot be deducted, and assets and inventory are subject to full taxation when sold. Purchases are deductible. In the case of incorporated businesses, the value of the stock should fall (permanently) to reflect the tax on old capital paid by the corporation. ⁷ This effect will be permanent and it is essentially a one-time tax on the owners of capital at the time the tax is imposed. (Those who purchase capital, either new or used, after the tax is enacted, will be compensated for the tax or lower asset value by the deduction or discount when they originally purchase it).

[34] While rental housing will be subject to this asset tax, owner- occupied housing is actually spared from the tax, because owner occupied housing is outside of this full consumption tax treatment. Indeed, it is the only major physical asset that is spared such treatment. That is, owner-occupied housing is not subject to tax when sold nor deducted when acquired. This treatment is probably the only practical way to tax the flow of imputed consumption from owner- occupied housing.

[35] Any potentially negative effects on owner-occupied housing proceed from the second source -- arising from the reallocation of capital. Under the current tax system, the return from investment in owner-occupied housing unlike other investments, is not subject to tax. If owning a home were treated like a business, the rental value would be included in income, and depreciation and maintenance costs deducted along with property taxes and interest paid. In that case, the return to the investment would be taxed.

[36] Since the return to other assets is taxed, the implicit price of housing (the rental value) is lower than it otherwise would be. By removing all taxes on new investment, the flat tax raises the rate of return and the implicit price (foregone returns on investment) of owner-occupied housing, causing contraction in demand. This effect may be straightforward, as in the loss of mortgage interest and property tax deductions, or more subtle due to the possibility of higher returns on alternative investments.

[37] This contraction in demand can show up either in less construction or lower prices, although an overall increase in the savings rate which lowers interest rates could offset the contraction, perhaps completely. If the supply could expand or contract in the long run without affecting price, these price effects would be transitory in nature. This same process could temporarily raise the prices of other assets, thus temporarily offsetting the direct taxes on the sale of assets and slowing the fall in permanent stock market prices.

[38] Rental housing, like other business assets, is subject to the permanent tax on sale. It is less affected by the contraction in demand, and indeed, may be subject to increased demand. Owner- occupied housing and rental housing are considered in turn.

[39] This reallocation of capital from owner-occupied housing into business assets and rental housing may be an improvement in economic efficiency, since current tax law favors housing. Because housing is subsidized, individuals tend to consume too much of it relative to other goods, and may even choose to own rather than to rent, when they would prefer, were the costs reflective of true costs, to rent. From an economic efficiency perspective, a shift out of owner-occupied housing is generally seen as a positive, not a negative, aspect of tax reform. ⁸

THE FLAT TAX: OWNER-OCCUPIED HOUSING SERVICES

[40] The analysis of tax revision and owner-occupied housing is complicated by a number of factors. First, housing services can be obtained through owner-occupancy or rental, and demand for owner- occupied housing can be altered by both changes in tenure choice (owning vs. renting) and the quantity consumed. There are clearly links between the two markets. Changes in tenure choice largely affect the number of owner-occupied units vs. the number of rental units, while quantity consumed largely affects the size, quality, and location. Obviously, also, the stock of existing housing can be converted between rental and owner-occupancy without fundamental physical change. ⁹

[41] In the case of owner occupancy, housing becomes jointly a consumption and an investment choice. The flow of housing services must be differentiated from the stock of structures. An individual buys a structure as an input into the provision of housing services, which requires other expenditures such as maintenance. He then consumes this flow of housing services which has a cost that is comprised of servicing the mortgage, forgone earnings on his equity position, depreciation, property taxes, and maintenance.

[42] In addition, the housing stock is very durable and existing housing can only be consumed in its particular location. The quality and size of existing housing cannot be easily changed. In the long run, however, the housing stock can be relocated and altered. The links between local housing markets are crucial to the analysis of a change in tax characteristics as is the time frame being considered. The heterogeneity of ownership of both types of housing, with respect to tax and other characteristics, must be considered. Higher income individuals tend to live in larger, more expensive, houses.

[43] Owner-occupied housing is more common than rental housing. For the third quarter of 1995, there were 65 million owner-occupied units and 35 million rental units. In general, owner-occupants tend to have higher incomes, and owner occupancy is very pronounced at the higher income levels. Owner-occupants who itemize (about half do) can deduct mortgage interest and property taxes.

[44] To assess the consequences of a tax revision, the first step is to determine how the changes alter the cost of housing services. This change will then have impacts on the quantity and price of houses, an effect that will depend on behavioral responses.

OWNER-OCCUPIED HOUSING, TAXES, AND THE USER COST

[45] Some of the discussion of the effect of the flat tax on owner-occupied housing has focused on the loss of the mortgage interest deductions, and, to a lesser extent, the property tax deduction. The loss of these deductions, were there no other changes, would increase the price of consuming owner-occupied housing. There is, however, an additional effect to be considered: because the flat tax is a consumption tax it effectively eliminates the tax on the return to new investment, thereby, absent a savings response, causing the rate of return to new investment to rise. As a result, the lost earnings on equity investments that could have been made elsewhere -- which are part of the price of owning a house -- are greater.

[46] Although many investors begin with a relatively high mortgage compared to the purchase price of their house, over time, as the mortgage declines and the nominal value of the house rises, an increasing cost is the return on equity. Our calculations of user cost reflect a weighted average of these costs, with equity accounting for two-thirds of the asset over time. (We discuss new purchasers who may have cash flow constraints, and to whom mortgage interest costs matter more, in a subsequent section).

[47] At this point, it is important to distinguish between two types of prices. The price of housing services is the rent that must be paid to earn a return on capital, cover depreciation, pay taxes. and cover maintenance costs. This rental price, or user cost, is the sum of all costs which, in some cases, are directly linked to the price of purchasing or constructing the structure.

[48] The price of acquiring the house itself is based on those parts of the price of housing services that are connected with the asset price -- the net-of-tax mortgage cost, the lost return on equity investment, and the property tax. In assessing the possible quantity and price effects in the market for housing, we calculate the tax-induced percentage change in cost excluding depreciation and maintenance, which will be referred to as the net user cost. When considering the potential effect on the price of houses, we adjust the demand function to compensate for the fact that the percentage change in the cost of housing services is smaller than the percentage change in the net cost of housing services and that the demand curve will shift by a smaller amount. ¹⁰

Overall Effects

[49] For the average homeowner, the user cost net of depreciation and maintenance is estimated to increase by 29% with the flat tax, holding pre-tax rates of return constant; the full increase in user cost would be 18%. Details of the assumptions behind this calculation are provided in Appendix A. This analysis takes into account a tax-induced fall in interest rates that keeps the overall marginal product of capital constant (as discussed below). ¹¹ These calculations do not include any interaction with state and local income taxes. ¹²

[50] This 29% increase in user cost would require a 22% decrease in price of the structures to offset the cost effect and keep quantity demanded (and supplied) constant. This is a maximum effect which would occur only with a perfectly inelastic supply of owner-occupied housing, no savings response, and a frictionless shift in demand. If price of structure times cost per dollar is to remain constant, and cost is multiplied by (1+x), where x is the proportional increase in cost, price must become original price times (1-y). In that case. (1-y)(1+x)=1. The proportional decline in price is therefore x/(1+x), or in this case, 0.29/(1+0.29), or 0.22.

[51] The DRI\McGraw-Hill 1995 equivalent estimate was 50% higher than this estimate. Their equivalent estimate for the user cost (additional steps are required to translate this change into a housing price effect) was 43%. There are several reasons for these differences discussed in appendix A.

[52] The 29% increase in cost in this study (assuming no savings response) would require a 22% increase in price. DRI's 43% increase in user cost, if reflected completely in the price of structures to keep demand fixed, would be 30%; DRI actually estimates 34%. ¹³

[53] This effect on user cost is not solely due to the loss of the mortgage interest deduction and the property tax deduction; rather, it occurs also because the return to equity investment has risen in general. For housing, which is not subject to current taxation of receipts, that effect means that the implicit cost of investing in housing has risen. Indeed, this is the more powerful of the effects, accounting for more than two-thirds of it. Merely eliminating the mortgage interest deduction and the property tax deduction would produce a much smaller effect on user cost: 9% net of maintenance and depreciation, and 6% overall, holding interest rates constant, or about one-fourth of the total. ¹⁴ That is another way of saying that the important effect of the flat tax on the price of housing services is not the loss of itemized deductions but the reduction of taxes on new business investments.

[54] Note also that for making a tenure choice (to own rather than to rent), the tax benefits of these deductions are smaller because they should reflect only deductions in excess of the standard deduction.

Effects on Different Types of Taxpayers and Homes

[55] The largest effects are likely to apply to individuals with the higher incomes, and, thus, to larger, more expensive houses, for several reasons. Higher-income taxpayers who tend to own more expensive homes have higher marginal tax rates than others and, for a given marginal tax rate, more of these individuals will itemize deductions. Although home ownership is one of the major reasons that deductions are itemized, the tendency to itemize deductions is also influenced by the size of other potential deductions, such as state income taxes, medical expenses, and charitable contributions.

[56] For that matter, individuals may itemize deductions at some period in their home ownership but not at others. Mortgage interest deductions decline over time as the loan is paid down, making it less likely that it will be advantageous to itemize deductions later in the home ownership period.

[57] Table 1 shows the percentage change in user cost, and in user cost net of depreciation and maintenance for the five current tax categories. The cost increases can become quite significant for high-income itemizers, suggesting a much larger shift in demand. It is important to note, however, that these larger effects apply to only a small fraction of taxpayers. About three-quarters of taxpayers pay tax at the zero or 15% rate; only the top 3% or so pay at rates higher than 28%, and most of those would pay at 31%. The price effects necessary to restore original demand would also be much smaller with larger exogenous changes; a 71% increase in cost would require only a 42% decline in price to keep cost times price constant. Note also that there are a number of other factors that are likely to mitigate these effects at high income levels.

       TABLE 1: PERCENTAGE CHANGE IN USER COST AND USER COST NET

[58] The calculations presented here include a certain type of interest rate change. Actually, there are two forces that could affect interest rates. One is a direct tax-induced effect arising from the presence of a general subsidy due to the deduction of the inflation premium in the interest rate at the firm level, which has already been included in the estimates. Holding the capital stock constant, this effect could also show up, in part, in a lowered equity return, through a portfolio effect, depending on the substitutability of debt with equity. The second effect is a savings effect -- if savings increases, the rate of return will fall.

THE DIRECT TAX INDUCED EFFECT AND PORTFOLIO CHOICE

[59] To understand this direct reduction in interest rates, it is necessary to explain the general tax treatment of investment at the firm level and the difference in tax treatment of debt and equity.

[60] The tax burden on equity is, on average, heavier than the tax burden on debt because of the current taxation of corporate profits at a higher rate than the average personal tax rate, and the additional tax on dividends and capital gains. Holding current interest rates and costs of equity constant, after tax returns to equity rise proportionally more than returns to debt, causing individuals to want to hold more equity. At the same time, firms will only be willing to hold more equity and less debt if the cost of equity before tax falls relative to debt, which would require a rise in the interest rate. In a simple world of no inflation and a fixed capital stock, there are two likely effects. First, the overall pretax return -- the weighted average of the return on debt and equity -- will not change. (There is no tax on debt financed earnings at the firm level, because earnings are included in income but then deducted; the before-tax cost of debt is the interest itself). If owner-occupied housing has the same debt/equity shares as the economy in general, there will be no consequences for the overall user cost. Secondly, there is likely to be a shift away from debt and into equity, and the most likely result of that shift is to have the interest rate rise and the pre-tax equity return fall, but in such amounts that the ratio of the after-tax return to equity and the after tax return to debt is higher than it was before the tax change. ¹⁵

[61] When inflation is present, however, this outcome becomes less likely. That is because debt at the firm level is currently subsidized because the full nominal interest rate, which includes an inflation premium, is deductible. ¹⁶ When all taxes on new investment are removed, as in the case of the flat tax, the cost of debt finance automatically rises, allowing the interest rate (at some levels of inflation) to decline while still satisfying simultaneously the preferences of debtors and creditors. Of course, inflation also creates a benefit in the deduction of mortgage interest including the inflation premium, which makes the initial exogenous change larger, but that effect is already incorporated in the direct calculation of the percentage change in user cost. Depending on how closely substitutable debt and equity are, some of this effect might be reflected in a lower pre-tax return to equity.

[62] Appendix B presents a model to explore the effects of this portfolio shift on interest rates. Simulations suggest a reduction in the interest rate on the order of 1%, with a small reduction in the equity return, at current inflation rates. Incorporating those changes in the user cost calculations would reduce the effects of the tax revision on net user cost from 36% to 29% and on gross user cost from 23% to 18% on average. The reduction is relatively small because it applies largely to debt financed capital, which over the entire life of the asset is assumed to account for about one-third of asset value.

CHANGES IN THE SAVINGS RATE

[63] Rates of return -- both interest rates and returns to equity -- could also be affected by changes in the savings rate. Here, too, the outcome is not entirely clear and depends on the supply and demand for funds. If savings increased, the supply of funds would increase, driving down the rate of return. This effect would, other things equal, tend to restore demand for owner-occupied housing and, for other investments, be transmitted in the short run into offsetting asset price increases if there were insufficient capacity to expand investment. As time passed, the capital stock would expand and a new equilibrium would develop with potentially more capital goods of all types.

[64] If the labor supply increased, however, some of that capital would be absorbed in providing capital for additional workers (or hours worked by existing workers); this process could actually raise rates of return in the short run. If the labor supply expands faster than the capital stock, the rate of return will rise, not fall, as capital becomes relatively more scarce.

[65] Nevertheless, an overall drop in all rates of return could have a powerful offsetting effect on the user cost of capital for owner-occupied housing.

[66] Several different approaches have been used to estimate the effect of consumption taxes on the rate of return. One approach is to assume that after-tax rates of return are fixed and, thus, the pre-tax return will fall to produce the same after tax yield. That is, if the interest rate is 8% and the tax rate is 25%, the after-tax interest rate is 6% and, thus, the interest rate should decline to that level. This view does not necessarily address the effect on equity returns although presumably they should decline as well.

[67] This assumption is not based on any particular modeling, however, that is based on behavioral responses to tax changes, and it implies a large savings response. ¹⁷ In fact, our knowledge about the response of savings to changes in tax regimes is very limited. Simple reductions in tax rates on capital income could, in theory, reduce savings. A shift to a revenue neutral consumption tax should, however, increase savings. ¹⁸ Some stylized models do suggest a very large positive response, particularly in the short run, to a shift to a consumption tax. Yet, it is difficult to discern large responses to any types of variables in the historical record; indeed, empirical studies of the response of savings rates to changes in the rate of return have, overall, found very small effects, or even negative responses.

[68] It is important to distinguish between the short run and the long run. The illustration above is a short-run one, but the adjustment process to a new long-run equilibrium can take a very long time. First, however, consider the long run, which is more straightforward to discuss theoretically, although quite difficult to assess empirically.

LONG RUN.

[69] The assumption that rates of return would decline to offset the tax changes is consistent with the long-run results of a formal model of the economy, which presumes that the economy is characterized by a single infinitely lived consumer. Regardless of any behavioral parameters, this model predicts a fall in interest rates (and returns to equity) until the rate returns to the old net of tax equilibrium. ¹⁹ Using the numbers in our model, all returns would fall by 1.9% (in addition to the portfolio effect that reduces the interest rate by 1%). If such a change in rates of return occurred, the effect on owner-occupied housing would be completely eliminated, because the effective tax burden on owner occupied housing is approximately zero and, thus, the cost of capital is approximately the net after-tax return. With this model, there would be an expansion of business investment and rental housing, with little or no change in investment in owner-occupied housing, in the long run.

[70] This outcome is pre-determined, however, by the nature of the model and, while such a model is mathematically tractable and has been popular for that reason, it is not a very realistic model of the economy. It is also very difficult to make this type of model consistent with open-economy assumptions if capital is very mobile.

[71] An alternative model is the life-cycle model. A life-cycle model takes account of the fact that individuals save over their lifetime and of the fact that many generations are alive at any one time. In these models, the long-run effects on interest rates (and savings) depend on the willingness of individuals to substitute consumption over time (the inter-temporal substitution elasticity) and the distribution of the tax burden across the generations. Both Auerbach, and Rogers and Fullerton, have simulated an Armey-Shelby- type consumption tax; the reductions in interest rates in these models range from 0.5% to between 1% and 2%, depending on the elasticities used. ²⁰

[72] Despite the greater apparent realism of the life cycle model, these models tend to predict changes in savings rates, especially in the short run, that are far larger than those that have been observed in the economy. A new study by Engen and Gale suggests that the response would be considerably less -- leading to a decline of about 0.5% in the rate of return -- if one takes account of precautionary savings, particularly in a system where much of retirement savings is already subject to consumption type treatment. ²¹

[73] One of the great challenges associated with examining the savings response is determining what magnitude of response is realistic. Many of the models produce results that are outside the typical observed response. One could also model this response in terms of a straightforward savings elasticity. This model is presented in Appendix C. In this model there is an elasticity of savings with respect to the rate of return that would determine the ultimate change in interest rates. Evidence from times series data has suggested that the savings elasticity, even if positive, is modest, although even fairly small savings elasticities could yield non-trivial changes in rates of return. ²² An elasticity of 0.2 could yield a reduction of 0.4%, and elasticity of 0.4 could yield a reduction of 0.6%. ²³

[74] These declines in rate of return would reduce the percentage change in user cost. These effects are shown in Table 2, which calculates percentage change in user cost with different changes in rates of return.

     TABLE 2: PERCENTAGE CHANGE IN COST OF OWNER-OCCUPIED HOUSING,

[76] It remains difficult to determine the magnitude of potential interest rate offsets; if they occur, however, they would reduce the effect on housing demand, perhaps significantly.

SHORT RUN.

[77] In the short run, it is difficult to predict not only how the savings rate might change but also what effect that change would have. Short-run adjustment is a complicated process, which is very difficult to assess, even if we had knowledge of initial changes in the savings rates. We cannot simply examine the change in the overall return as capital stock is accumulated because of adjustment costs, including capacity constraints. These adjustment costs tend to moderate changes in the reallocation of consumption and investment in the short run.

[78] The loss of the mortgage interest and property tax deduction has a direct, but modest, initial effect on user cost which would reduce demand for owner occupied housing. The consumption tax makes business capital goods relatively cheaper, by allowing a direct tax deduction for unincorporated businesses and rental housing and by causing corporations to wish to issue more stock to buy cheaper capital goods and driving down stock prices. The cost of debt finance initially increases. These processes attract capital into equity markets and drive down interest rates.

[79] Overall, these effects reduce the demand for owner- occupied housing in order to supply funds to business where higher returns are now available, but they may also induce an overall increase in savings. The overall increase permits a moderation of the contraction of demand for owner-occupied housing and increased spending from both sources may drive up the price of business capital goods. A higher asset price further limits investment in that market and directs it back to owner occupied housing. If this price increase is perceived to arise from a short-run capacity limit, businesses will likely delay their purchases (that is, their own demand will be quite responsive to asset price changes). Thus, while the rate of return may adjust slowly, changes in asset prices in the short run could lead to significant consequences for the demand for owner- occupied housing. Another way of saying this is that limits in the ability of business investment to absorb the new flow of savings directed towards it would offset the initial higher rates of return and increase the attractiveness of owner-occupied housing relative to the initial contraction in demand.

OTHER FACTORS AFFECTING THE DEMAND FOR HOUSING

[80] There are several other factors that could affect the way in which the flat tax affects the demand for housing. In general, the factors tend to mitigate the effects of the shift, and, in particular, they tend to do so for effects that occur at high income levels. At the same time, however, it is important to note that a general contraction in the quality and size of housing demand tends to exert more contraction in demand for larger and more expensive homes.

[81] Potentially the most important of these, and yet the least understood in some ways, is the relationship between an income tax and risk-taking. Although an income tax imposes an additional burden of tax on investors, it also reduces the riskiness of returns. When returns rise, the government shares in the higher return and when returns fall the government also shares in the lower returns. When individuals are free to rearrange their investment portfolios and when the riskless return is zero, it can be shown that the income tax has no effect at all, since individuals can restore their pre-tax levels of both risk and return by increasing the risky share of assets in their portfolio. ²⁴

[82] There are several complicating factors. If returns are negative and the firm or portfolio is not large enough, not all losses will be deducted. In practice, all individuals cannot expand their risky investments without limit; and the riskless return is not likely to be zero. Nevertheless, this effect of riskiness does lighten in some fashion the burden of the current income tax, and makes the implicit change in the return on business investments smaller -- perhaps considerably smaller. It may be of particular importance to higher-income individuals who are more likely to be engaged in risky investment.

[83] Most of the other factors tend to particularly favor higher-income households. First, there is an income effect. Individuals with higher incomes will increase their demands for all commodities, including housing. The flat tax tends to increase incomes of high income individuals because of its flat rate. According to a recent study, the above $200,000 income class has a net increase in income of almost 9%. ²⁵ If the income and price elasticities are the same, this effect would offset a quarter or more of the price induced reduction in demand among very-high-income individuals. ²⁶ These income effects would not otherwise be very important: lower-income individuals may pay more taxes, but they do not tend to invest in owner occupied housing, while middle-income individuals experience small effects in general.

[84] The shift in the tax base to a consumption tax also has distributional effects across age groups, generally benefitting (on a lifetime income basis) higher-income individuals who expect to do considerable individual saving over their lifetimes and who tend to save in financial form. These individuals would be expected to increase their lifetime consumption overall and hence might increase their purchases of owner-occupied housing. They are more likely to be the purchasers of more expensive homes.

[85] Also, individuals purchasing more expensive owner occupied housing will have some slight benefit from the repeal of the capital gains tax; using typical assumptions, the present value of future capital gains taxes is about 3% or so of current asset value, assuming that no exclusion applies, and somewhat less if there is. ²⁷

[86] One qualification would reduce the effect on moderate- income purchasers. Recall that there is both a decision about whether to rent or to own a home and a decision about the quantity of housing to purchase. For moderate income individuals the relative price of owning a home, when making the tenure decision, is higher than it would appear from a simple analysis, because an individual who itemizes deductions loses the standard deduction. If the individual would take the standard deduction as a renter and the itemized deduction as a homeowner, the value of the housing benefits is the difference between those two amounts, which will be smaller than the value simply assuming that all deductions are net reductions in tax liability. It is difficult to know the effect of this factor with current law, but a previous study by Hendershott and Slemrod estimated that the tax rate for determining the effect for tenancy choice was about three-quarters of the rate for quantity. ²⁸

[87] Another qualification is that the individual investor may have tax-preferred methods of investing, such as pension plans and individual retirement plans. To the extent that these options are available and operate at the margin, the current tax burden is smaller and the effects on housing demand more limited. These options might be available to both higher and moderate income individuals. Assuming that 30% of financial assets are held in tax exempt form, the net user cost would fall to about 26% if all such investments were marginal. In many cases, of course, the individual investor has no direct control over these decisions.

[88] All of the factors mentioned here tend to moderate the contraction of demand for housing.

EFFECTS ON QUANTITY AND PRICE

[89] The analysis of the effect of the flat tax on housing quantities and prices begins with a supply and demand framework. Then the effect of the shift in the demand curve on quantity and price can be ascertained based on several parameters -- the size of the shift, the slope of the demand curve, and the slope of the supply curve.

[90] In analyzing the market one must distinguish between the short run and the long run, and also distinguish between local and national markets. We begin, however, with the discussion of the fundamentals of supply and demand.

[91] Figure 1 shows such a supply and demand model. The demand curve traces out the quantity of housing consumers are willing to purchase at each price. As price falls, more is purchased. The supply curve traces out the quantity owners are willing to supply at each price. The intersection of the two solid curves shows the original price and quantity, denoted as P and Q.

[92] The change in the user cost of capital, discussed in the previous section, creates an exogenous shift in the demand curve, shown by the dashed line. In comparing the new demand curve to the old, at every point the demand will remain the same only if the acquisition price falls by a fixed amount to restore the old price. The intersection of the new dashed-line demand curve with the original supply curve shows the new price and quantity, P' and Q'.

FIGURE 1: SUPPLY AND DEMAND FOR HOUSING

[Figure 1 omitted]

[93] One can see by this simple analysis that the shapes of these two curves determine the extent to which any shift will be translated into price effects vs. quantity changes. Imagine the supply curve rotating counter-clockwise around its original intersection; as it does, the effect on quantity becomes smaller and the effect on price becomes large. If the supply curve is actually vertical, no change in quantity can occur, and the entire effect will show up in price. On the other hand, suppose the supply curve becomes flatter; as it rotates clockwise, the effects on quantity become larger and the effects on price smaller. If the supply curve is horizontal, no effect on price can occur, and the entire effect will show up in quantity.

[94] These effects can also be shown with some simple elasticity formulas. (The elasticity is the percentage change in quantity divided by the percentage change in price). There is a demand elasticity that traces out how movements in quantity are related to movements in price along the demand curve, and a supply elasticity that does the same for the supply curve. A vertical supply or demand function has a zero elasticity, since quantity changes are zero; and a horizonal function has an infinitely large elasticity. The formulas are for a small change, but they can be used to illustrate the outcomes. As shown in Appendix D, the percentage decline in price is equal to the percentage increase in user cost, multiplied by the demand elasticity divided by the sum of the demand elasticity and the supply elasticity (all elasticities are in absolute value; that is, if the slope is negative the negative sign will appear in the formula), or:

Percentage change in price = -Ed Times Percentage Change in User Cost

[96] Before turning to a discussion of the short-run and the long-run equilibria, however, it is important to deal with which price -- the user cost, or the net user cost -- should be used in the analysis. Basically, the actual shift in the demand curve should be based on the gross user cost, rather than the net, since most empirical estimates of demand are based on this cost. If the cost is to be capitalized in price, however, a full capitalization would be based on the percentage change in net user cost; since this is how much the price would need to fall to restore aggregate demand at the current quantity. ²⁹ Therefore, in our analysis, we use the net user cost, but adjust the elasticity to make it smaller, reflecting the ratio of gross to net percentage change.

[97] In actually making calculations, we assume that demand and supply curves are characterized by a constant elasticity, making them curves rather than straight lines; these assumptions also affect the magnitude of shifts in quantity and price, when a large change occurs. These formulas are shown in Appendix D.

LONG RUN

[98] The long run is generally easier to assess than the short run. For most commodities the presumption is that production can take place with constant returns to scale; that is, in the long run, it is possible to increase the quantity without altering price. Even if increased supply requires a higher price, that price rise is likely to be very modest, leading to a very flat supply curve.

[99] Similar assumptions would normally be made about construction costs, where labor and supplies can, in the long run, be attracted from one industry to another. Housing does, however, differ from other assets in that it needs to be consumed in a fixed location and that it uses a substantial amount of land, which is itself fixed in given locations. That is, the quantity of land cannot be altered in the city center. This creates the possibility of some upward slope to the supply curve. That is, if the quantity is fixed, then a decrease in demand would in part lower the price.

[100] Despite that possibility, there are many reasons to expect that price is unlikely to be altered much in the long run even in urban areas.

[101] There should be little dispute that prices at the fringes of the city where land is readily available should not be much affected by the expansion of demand. At these spots, there is a reasonable amount of land whose value is low enough that an expansion or contraction in housing can be accomplished without altering price. Moreover, since on average land accounts for about a quarter of the value of owner-occupied housing ³⁰ -- and presumably less on the city fringe -- this value constrains the amount by which price can fall.

[102] For land at both the fringe and in the city center, there are many other possibilities in the long run. First, methods of construction can be used to economize on land or to use more land. In addition, land can be used for other purposes. Land released from owner-occupied housing use can be converted into rental housing use or business use. Indeed, there is a limit to the fall in land prices in the long run if demand for alternative purposes (e.g., business use or rental housing) remains strong, which it presumably would. Indeed, if the savings rate goes up, the price of land might well rise.

[103] Prices in the city and in the fringes are also related. The simplest model of this relationship is that the difference in price of housing varies with commuting costs. In the urban center, commuting costs are very low and prices are higher; at the suburban fringes, commuting costs are high. In this simplified model, the price differentials are entirely set by commuting cost differentials. In such a model, there would be no long run change in the price of urban dwellings due to a shift in demand induced by interest rates or tax burdens that are common to all properties. If demand for housing fell, construction at the fringe would contract, initially driving up prices in the fringe relative to the city center. (With growth, this contraction would simply cause housing not to keep up with the pace of growth; this effect could also be accomplished through natural depreciation, filtering of properties from higher to lower income uses, and conversion to other uses). As prices in the fringe rose, demand would shift back to the city center, until a new equilibrium is reached, with a smaller quantity of housing, a smaller average commuting distance, and no change in price. ³¹ Large price changes in the long run in the city center would have to be traced to changes in commuting costs.

[104] Of course, this is an oversimplification of the model. Clearly, there are specific tastes for housing in given areas and houses, adjusted for commuting costs, are not likely to be perfect substitutes. Housing in different political jurisdictions may also be affected by the taxes and benefits provided in those jurisdictions. However, if houses in different locations are very close substitutes for some individuals, any effects on relative prices arising from a change that affects all housing would likely be modest. This would be equivalent to depicting a less than horizontal, but nevertheless very flat, aggregate supply curve.

[105] For example, suppose the elasticity of the supply curve is 10 (rather than infinity), while the elasticity of the demand curve is 1. Given the average exogenous change in price of 29% corresponding to a user cost of 18%, in the case of no fall in the interest rate, one would calculate the change in price as averaging under 2%. ³² It would be half as big, approximately, with a 1% decrease in rate of return.

[106] Table 3 illustrates the effects on prices and quantities of various assumptions regarding the supply curve and interest rate changes.

[107] It seems clear that it is possible to expand the housing stock and accommodate to different tastes without a secular rise in housing prices because such a rise has not been observed over time. ³³ This evidence points very strongly to a relatively flat supply curve.

  TABLE 3: PERCENTAGE CHANGE IN PRICE AND QUANTITY OF OWNER OCCUPIED

[109] On the whole, therefore, there is no reason to expect much decline in prices of owner-occupied housing in the long run, although there could be a decline in the average size and quality and a reduction in owner-occupied housing tenure choice.

Short Run

[110] The short run is more difficult to assess, and both supply and demand response may be different. Let us begin with the supply function. (The short run cannot be precisely defined; but in this context, where the consumption of service flows from very durable goods is the subject of analysis, the period of time would be over several years. There is also an intermediate period that could cover a number of years during which short run effects gradually develop into the long run equilibrium.)

[111] In a simple analysis, the effects on price would appear to be considerably larger in the short run because the quantity of housing tends to be fixed. New construction, which can be altered in the short run, is only about 1% of the stock of housing, which suggests that it is very difficult to adjust the quantity of housing in the short run. In that case, the supply elasticity would appear to be very low and the effect would mostly occur through changes in prices (although an increased savings rate could moderate the effect).

[112] Viewing the supply as roughly fixed may be inappropriate for several reasons when considering the market of owner-occupied housing. For one thing, in considering the supply curve, it is useful to distinguish between the total supply and demand for housing and the supply and demand on the market.

[113] The market for housing sales relies more heavily on new construction, which comprises about 15% of total units. ³⁴ That difference might be fairly meaningless if the purchase and sale of housing was viewed as a choice easy to reverse, without significant transaction costs. Current homeowners, however, who might now prefer less housing, or even to change their tenancy choice, would find themselves unwilling to undergo the transaction costs of selling their current property and purchasing another. They might tend to maintain their current consumption, or to adjust by making fewer improvements over time -- a much more gradual and limited process of adjustment. In the short run, therefore, it is largely the flexibility of adjusting the supply of houses normally on the market that defines the relevant supply curve. While there are limits on that adjustment, they are much less severe than the limits on the adjustment of the total capital stock. (Of course, homeowners who had already planned to sell and purchase can make such adjustments, but they are already part of the existing sales market.)

[114] In addition, the supply of housing on the market can be altered if individuals elect to take houses off the market when the price declines (either to delay moving to a rental unit, to shift the property itself to rental which would become more attractive because rentals are deductible acquisitions, ³⁵ or simply to leave a property vacant for a period of time). This behavior might also take the form of simply leaving a house on the market without lowering the price. This sort of behavior, while burdensome to sellers, tends to flatten the supply curve and allow a contraction in sales without lowering price. It would also be a logical choice to make if the fall in price were perceived, or even suspected, to be transitory.

[115] There is ample evidence that the volume of sales of existing single family homes can fluctuate considerably, presumably owing largely to changes in interest rates. For example, from 1979 to 1982, the sales of existing homes fell by almost 50%, presumably reflecting the rise in interest rates and the recession. Some of that effect was, of course, due to individuals who were considering trading up failing to do so; so it is difficult to draw precise conclusions from this observation, except to indicate a significant degree of adjustability in the supply of houses on the market.

[116] The demand response can also be somewhat different in the short run -- and in particular the shift in demand may be smaller. It is first important to consider exactly how the flat tax alters the relative price. It produces a direct effect on net user cost estimated at about 9% due to inability to deduct mortgage interest and property taxes. The remainder of the contraction in demand, however, proceeds from the diversion of funds into other investments. Some of this demand will be financed by the cash flow savings from deducting assets. But, for large corporations, there may not be much cash flow reduction, due to the loss of depreciation deductions and deductions for interest. These firms should still respond, but the response may occur with a lag and can only attract capital from individuals through either a reduction in dividend payments or the sale of new stock shares (or reduction of current repurchases). It is only through the driving down of stock prices (asset prices), which is the equivalent of the deduction for acquisition of capital, that individuals should be encouraged to increase their investments in equity.

[117] Moreover, the short-run investment demand response is inhibited by two other factors. First, the ability to expand production through the addition of capital may be limited in the short run by technology. Secondly, there may be short-run capacity constraints in the investment goods industries -- both for manufacturers of equipment and for commercial and industrial structures. A capacity constraint will tend to drive up prices and curtail the initial expansion of investment, especially for a temporary expansion, thereby restoring demand to the owner-occupied housing sector.

[118] At the same time, a perceived transitory fall in the price of owner-occupied housing should also encourage purchasers to maintain demand, thereby choking off any fall in demand. In other words, there are many forces that tend to moderate transitional effects, especially if these effects are perceived, or even suspected, to be temporary.

[119] If the supply curve becomes more elastic, the price effects would be moderated -- an important issue since much of the discussion that has occurred about housing in existing studies seems to assume that the supply curve is relatively, or even perfectly, inelastic. With even a low elasticity of 0.5, the price effect will be cut close to half -- from 22% to 13%. With an elasticity of 1, it will be cut to 9%. If the demand shift is moderated by increased savings or an inability of other sectors to absorb demand, the effects will be even smaller.

[120] While it is difficult to assess this outcome from a theoretical standpoint, the empirical evidence strongly suggests that the large price effects discussed in the DRI study and in the Hendershott, et al., studies are unlikely. Figure 2 presents the series of real price changes in a constant quality house juxtaposed with the changes estimated for the flat tax by DRI and in this study (denoted CRS), with and without interest rate changes (an A suffix denotes constant rates of return; the CRS numbers are for a 1% decline in the overall rate of return), based on inelastic supply assumptions. ³⁶

[121] In addition, while there are occasionally fluctuations in housing prices, the period covered in the historical series, as noted above, included periods when tax and/or interest rates caused user costs to vary significantly. For example, between 1980 and 1982, Poterba finds the user cost rose by 14% for moderate income taxpayers and by 58% for high income individuals -- effects similar to those predicted for the flat tax. Despite these effects, the price shown in figure 2 fell by only 2%. ³⁷ Of course, one must be cautious in drawing conclusions from examining data as simple as these. Many potential influences, from both demand side and supply side forces, could be operating and even be offsetting. Nevertheless, if changes in user cost exerted such a powerful short run effect on prices, we would expect to see significant fluctuations over time.

FIGURE 2: PERCENTAGE CHANGE IN HOUSING PRICE: HISTORICAL VS. EFFECTS OF FLAT TAX CALCULATED WITH A FIXED SUPPLY CURVE

[figure 2 omitted]

LIQUIDITY CONSTRAINED AND PSYCHOLOGICAL MODELS

[122] In some cases, demands for owner-occupied housing may be constrained by liquidity. That is, individuals may be limited in their ability to purchase a house, or to purchase a house of a desirable size, by their ability to borrow. While this effect does not occur for all purchasers, it does occur for some, perhaps particularly for first time purchasers. In this case, what will matter is the cash-flow effects of the tax revision.

[123] It is likely that these liquidity-constrained purchasers would be less affected by the flat tax changes than those responding to the classic user-cost effects. The portfolio-induced decline in the interest rate will be more important for these purchasers than for those in which equity investments, either now or in the future, play an important role. Once debt is the focus, it is important to note that debt financed investments are favored at the firm level because they are deducted in full, including the inflation premium. This effect is particularly pronounced in the corporate sector, where tax rates are relatively high compared to the average creditor. Thus, with inflation, a significant fall in the interest rate, earlier predicted at around 1%, is likely even without an additional savings response.

[124] If the interest rate is 8%, the value of the tax deduction for an individual in the 15% bracket is 1.2% off the interest rate -- an amount almost entirely offset by the interest rate decline arising from portfolio effects. Recall that 75% of taxpayers are in the 15%) bracket or lower. Although one-third of taxpayers, who tend to be lower-income, are renters, these data imply that slightly over half of home-owners are in the 15% bracket, and the liquidity constrained purchasers are especially likely to be in that bracket. Since any tax increases in the middle-income groups are small, ³⁸ a 1% decrease in the interest rate would probably leave these families with little effect on their cash flow, particularly if they were not able to take full advantage of itemized deductions (i.e., their net gain in deductions due to itemizing is less than their total housing-related deductions).

[125] Another argument that might be made is that the symbolic value of deducting mortgage interest and property taxes encourages home-ownership. While such an argument is possible, one would need evidence to demonstrate this phenomenon. It would seem unlikely to be the dominant force in affecting individual preferences for homes, and many individuals who own homes do not itemize their deductions. Changes in sales techniques might also mitigate any of these effects.

THE FLAT TAX: RENTAL HOUSING

[126] The rent on tenant-occupied housing would also be affected by the flat tax and other proposals that shift to a consumption base. Rental housing is difficult to assess because it is not entirely settled how rents are determined at the margin. There is a small corporate presence in rental housing (about 10%), and if rents between corporate-owned and non-corporate-owned buildings are to be equated (when that is the only difference) then only one rent can prevail. There are models where this rent is thought to be the corporate rent, with non-corporate investors earning additional returns. One could also weight the shares and produce an average rent, or try to differentiate the market into detached and multifamily rentals.

[127] How this effect is modeled, as well as the assumptions about interest rate changes, would affect these consequences for rental housing. Table 4 shows the change in user cost and net user cost, for corporate and noncorporate investments, using both gross and net user cost measures, and varying the interest rate. The effects could be either positive are negative, but it is clear that rental housing is unlikely to have its demand reduced and indeed may experience an expansion in demand. (The noncorporate landlord is assumed to have the same marginal tax rate as the individual owner- occupant; landlords with higher tax rates would have smaller increases or larger decreases in user cost.)

    TABLE 4: EFFECT ON USER COST OF RENTAL HOUSING, DUE TO FLAT TAX

[129] Rental housing, unlike owner-occupied housing, is not outside of the consumption tax system. Therefore, the full sales price of an existing unit, when sold, will be subject to tax. Since only real assets, and not financial ones, are subject to the tax and deduction, the repayment of any loan is not deductible; hence, it is possible for properties with high loan-to-value ratios to yield inadequate revenues to both pay the tax and repay the loan. This treatment is true of all assets. Those individuals continuing to rent properties will forego deductions for depreciation and interest. These consequences stem from the nature of a VAT type consumption tax and the lack of a need to accommodate the tax with a price increase, which would drive up nominal asset prices.

OTHER TAX PROPOSALS

[130] With the analysis of the flat tax as a background, this section discusses the effects of alternative tax proposals.

VALUE ADDED AND SALES TAXES

[131] Although the flat tax can also be described as a value added tax with a wage exemption, there are some potentially important differences between the flat tax and the VAT or retail sales tax. A value added tax imposes the consumption tax at each stage of production. A retail sales tax imposes the tax on consumption goods at the final stage of sale.

[132] In discussing these effects, it is crucial to keep in mind the two different responses to the tax. One is the reallocation of capital into business use that tends to depress housing prices but only in a largely transitory fashion, as discussed in the bulk of this paper for the flat tax. The second is the lump-sum tax on old assets, which is permanent, and which arises from the tax on the sale of old assets and the deduction for the purchase of new ones. Owner- occupied housing is typically exempt from this permanent lump-sum tax because the owner-occupant is not treated as a business (paying tax on sale of assets, deducting acquisitions, and including the flow of consumption in the tax base).

[133] We discuss several cases that illuminate the role of price accommodation and the possibility of excluding housing sales from the tax.

[134] Consider first a VAT or retail sales tax that includes the sale of new owner-occupied housing in the base as a consumption good.

[135] If the VAT or retail sales tax includes new housing construction in its base and there is no price accommodation to the introduction of these taxes, the effects would be quite similar to the effects of the flat tax -- in the long run, housing prices would be largely unaffected, but in the short run there could be some effect, depending on a variety of factors affecting the supply and demand curve. These effects derive from the removal of taxes on alternative investments.

[136] There would likely be pressures, however, to accommodate these taxes with an expansion in the money supply and an increase in the overall price level. If there were a full accommodation, then nominal housing prices would be likely to rise because the price of new construction would increase. Under a value added tax, the price of business assets and rental housing would also rise, but that increase would be offset by the ability of the investor to deduct the acquisition. Hence, the differences between business assets and owner-occupied housing would remain. Under a retail sales tax, capital goods prices would not rise because they would not be included in the tax base, but the price of housing would rise, accomplishing the same effect.

[137] Set against this backdrop of a general nominal price rise could be some partial offset based on supply and demand in the housing market, as discussed in the case of the flat tax -- the effects that derive from the rise in rates of return and the effect on user cost.

[138] The nominal price rise that occurs with a VAT or retail sales tax that is accommodated by price is not, however, very meaningful. The purchasing power of assets has fallen by the amount of the tax. Thus, ignoring debt for the moment, the owner of housing is in no different position than in the case where no price change has occurred; indeed, the outcome would be quite similar to the outcome in the case of the flat tax. Although the value of the house has increased, the cost of purchasing consumption goods has also increased.

[139] The rise in price does, however, benefit sellers who have some existing debt to repay, since the outstanding debt is fixed in dollar value. Thus, the proportional rise in the equity value of the home is greater than the overall price rise. For example, if the current mortgage is half the value of the home, and the price level rises by 20%, the owner's equity has increased by 40%. For individuals who are selling their properties to obtain funds for consumption this effect may well offset any supply and demand induced relative price effects that have been discussed with respect to the flat tax, and that would also occur in this case. Consumption goods have risen in price by 20%, but the homeowner's equity has increased by twice that amount. This is just another way of saying that, with a price accommodation, all owners of capital -- including debtors -- bear the lump sum tax on old capital that is intrinsic in a consumption tax. Once again, the exception is the owner of housing that is outside the system and the owner of equity in that housing receives a windfall.

[140] While it is highly unlikely that the flow of consumption from owner-occupied housing could be taxed directly, it is possible under the VAT or retail sales tax to exempt owner-occupied housing from the tax. It is also possible to impose the tax on the sales of old housing. Either of these alternatives would affect the current owner-occupant's effective price. If new housing is excluded from the tax, then the price of housing will not rise, even while the price of other goods does. Thus, there will be a burden imposed on the owner-occupant that is similar to that of individuals purchasing other capital goods. Of course, there may be some temporary offset, since the cost of new housing is lower relative to other consumer goods and housing will be more attractive for new purchasers. This effect would offset the normal contraction in demand that occurs from the reallocation of capital. It is, in effect, equivalent to allowing individuals to deduct their interest and foregone equity returns at the new consumption tax rate, and it restores the preferred status of housing.

[141] Another possibility is to impose the sales or VAT on the sale of old housing as well as new. In that case, although the nominal price of housing would rise, the entire increase will be needed to cover the taxes and the individual bears a burden similar to that of other assets -- since the costs of consumer prices has risen. This approach would impose the lump sum tax that applies to other capital goods to housing, as well as making new housing less favored.

[142] The effects on rental housing would be similar to the effects under a flat tax.

THE USA TAX

[143] The USA tax is more complicated to analyze. At the firm level, it imposes a value added tax offset by a payroll credit, which is a combination of a value added tax and a lump-sum tax on old capital, except that the tax is partially offset by some transition rules. At the individual level, it imposes a direct consumption tax, which allows deductions for financial savings but includes dissaving in income. Owner-occupied housing is left out of the system entirely. There is no deduction for acquiring a house, and no tax on selling it.

[144] The direct part of the USA tax creates no pressure for a price accommodation and the payroll credit at the firm level removes most of the pressure for most of the VAT. (The pressure to raise prices arises from the need to pass on a tax on wages to the individual by lowering the nominal wage; the inability to do so in the short run can cause an economic contraction). Essentially, the treatment under the direct consumption would be the same as the treatment under a flat tax, except that the lump-sum tax is imposed directly on financial claims to assets rather than real assets and so is shared by both debt and equity claims. When an asset is sold, the tax to the owner is imposed only on the equity share. The USA tax modifies this effect, however, by excluding mortgage borrowing (and some additional borrowing as well), which is part of the means of leaving housing our [sic] of the system. The USA tax, however, retains the mortgage interest deduction, retaining part of the preferential treatment for owner-occupied housing.

[145] The effects on rental housing would be slightly less beneficial than under the flat tax.

INCOME TAX REFORMS

[146] Proposals to simply restrict the mortgage interest deduction and/or the property tax deduction would have much more modest effects than the consumption tax proposals, as indicated earlier. Eliminating the property tax deduction alone would increase the net user cost by about 2.5% and the user cost by 1.5%; eliminating both would increase the net user cost by about 9% and the user cost by about 6%. Neither of these numbers contains any change in the interest rate or rate of return. Such changes are likely to be relatively small, since there is no broad change in the overall tax burden of great significance and no certainty as to the direction of a change. ³⁹

[147] Rental housing would be largely unaffected by an income tax reform, unless some direct change were to be made in its treatment; restriction of benefits to owner-occupied housing could, however, increase demand for rental housing services. These effects would likely be small.

CONCLUSION

[148] The principal finding of this study is that effects of the flat tax on housing prices are likely to be limited even in the short run and quite small in the long run. These findings derive in part from expectations that the supply curve is not inelastic, contrary to what appears to be presumed in other studies, and indeed is likely to be very elastic in the long run. Both theory and empirical evidence support some significant supply response in the short run. There are other factors as well, including the possibility of increased savings rates, that could mitigate any demand response.

[149] Indeed, owner-occupied housing seems an asset that limited concern should be devoted to, since it is an asset normally protected, in both the flat tax and in other consumption taxes, from the lump-sum tax on old capital. Indeed, current owners could actually gain under some types of tax revision.

[150] Demand for owner-occupied housing and home-ownership should contract somewhat in response to a shift to a consumption base; but that occurs because a burden on competing investments is lifted and, thus, a relative preference to owner-occupied housing lifted. By eliminating the distortion in choice due to the tax system, economic efficiency and welfare should be improved.

[The Appendices contain equations which are unsuitable for online reproduction and are omitted.]

² DRI/McGraw-Hill, Residential Real Estate Impacts of Flat Tax Legislation. Summary Prepared for the National Association of Realtors, May 1995. Principal Investigators: Roger E. Brinner, David Wyss, and Mark Lasky.

³ DRI/McGraw-Hill. The Impact of the Flat Tax on Mortgage Foreclosures and Losses. Summary Prepared for the National Association of Realtors. January 1996. Principal Investigators: David Wyss and Cynthia Latta.

⁴ Richard K. Green, Patric H. Hendershott, and Dennis R. Capozza, Taxes, Mortgage Borrowing, and House Prices, Preliminary Draft prepared for Brookings Conference on the Economic Effects of Fundamental Tax Reform, January 18, 1996.

⁵ Two forms of VAT are commonly discussed: the European style VAT where firms pay the tax and take a credit for tax on purchases, and a subtraction-method VAT (also sometimes referred to as a business transfer tax), where the cost of purchases from other businesses is deducted from the base before applying the tax. VAT proposals in the U.S. generally use the subtraction method. Such a proposal was introduced in the previous Congress by Senators Boren and Danforth.

⁶ This treatment excludes the imputed value of the owner- occupant's labor in providing maintenance and repairs and general household management.

⁷ See Jane G. Gravelle, "The Flat Tax and Other Proposals: Who Will Bear the Tax Burden?" Congressional Research Service, Library of Congress, November 29, 1995, Report 95-1141 E for a further discussion. This asset price effect also occurs with a VAT or sales tax. Normally these taxes would require an increase in the price level, because taxes on labor income must now be paid by the firm and, since it is difficult to lower wages, it is difficult to immediately pass the tax backwards through lower wages. When prices rise to accommodate the tax, it can be passed forward in price and both equity and debt claims to assets bear the tax through reduced purchasing power when assets are sold to finance consumption goods. In the case of the flat tax, where wages are still taxed to individuals, there is little need to accommodate the tax change with an increase in the price level. The tax on assets must be borne by those assets (passed back) and, therefore, the tax is born only by the equity owners. Debt holders do not bear the tax.

⁸ There are two caveats to this assessment. First, State and local jurisdictions tend to impose taxes on land and real property, which burden housing (both rental and owner-occupied) more than other assets. There are arguments, however, that individuals are compensated for their property tax by the direct benefits of local government services and can choose their location -- vote with their feet -- in a way that optimizes the tradeoff between tax and benefits, rendering the property tax not a true burden. Secondly, some who favor subsidies suggest that there are arguments that home- ownership confers additional benefits to the community, such as more care in upkeep and a greater interest in the community, that justify tax subsidies. Such claims have not been quantified, however; nor is it clear that the actions of home-owners in maximizing their own welfare are necessarily beneficial to others in the community.

⁹ There may be restrictions on modifications, such as refitting housing to include more units in an existing structure. Some of these modifications are constrained by zoning regulations.

¹⁰ The depreciation rate is not included in this net rental price, even though it is linked to sales price. The most likely way in which housing prices could change would be through a fall in the value of land, which would make the physical depreciation of the structure independent of the tax induced price change, as land as a share of value would be altered. That is, if only the value of the land falls, then the cost of depreciation, which relates to the structure, will not be altered.

¹¹ Simply holding interest rates fixed would yield a change, net of depreciation and maintenance of 36 percent, and a change of 23 percent in the overall user cost of capital. This outcome is not consistent with a fixed capital stock and would, in fact, require a decrease in savings.

¹² While these effects are relatively small, the direction is difficult to determine. If State and local income taxes are not altered, the percentage changes will be slightly smaller because the loss of itemized deductions for these taxes will raise their effective rates; if however, these taxes are converted to a consumption base, the percentage changes will be slightly larger. While it is not possible to predict what will likely occur, State and local governments would find it very difficult to continue with an income tax base if the Federal government switches to a consumption base; if so, the effects will be slightly larger.

¹³ If x is 0.43, this value is 0.30, exactly equal to the tax rate assumed. The 34 percent change occurs because DRI's methodology keeps the absolute amount rather than the rate of property taxes fixed.

¹⁴ These numbers are smaller than those estimated in the DRI study. One difference is the base -- the DRI base effectively reflects a percentage change in required rate of return without accounting for property taxes; our 9 percent number would be 12 percent if property taxes were excluded. Both methods can encompass measurement errors. However, the particular method used by DRI can produce significant errors simply because the measurement of the value of deductions is estimated from one source, the value of the capital stock from another, and the discount rate from yet another. In particular, the discount rate used to capitalize these costs is probably too low. In addition, there is a difference between estimating the present value of deductions relative to the value of a new property as an effect on user cost and capitalizing the current flow of deductions relating to existing properties and dividing by the existing capital stock.

¹⁵ Portfolio adjustment has also been discussed in Martin Feldstein, The Effect of a Consumption Tax on the Rate of Interest, National Bureau of Economic Research Working Paper 5397. Feldstein concludes that the portfolio effect causes interest rates to rise, but that has to do with different assumptions regarding the role of inflation in the firm's tax rate. Note that even in the absence of inflation, it is possible for the interest rate to fall, although that is not a likely outcome. Alan Auerbach, in a more recent version of his paper Tax Reform, Capital Allocation, Efficiency and Growth that was presented at the Brookings Institution in February 1996 concludes that the interest rate will fall, as does this study.

¹⁶ Inflation also creates a penalty by causing depreciation and inventories to be understated, but this does not have an effect on the outcome because it applies to both debt and equity financed capital. Moreover, this effect is generally offset, on average, by accelerated depreciation and LIFO inventory accounting.

¹⁷ Such a conclusion is reached through the use of supply and demand curves analysis in an article by John Golub, "How Would Tax Reform Affect Financial Markets? Federal Reserve Bank of Kansas City, Economic Review, Fourth Quarter 1995, Vol. 80, No. 4. The author of this article, however, has apparently failed to incorporate the taxes paid by businesses on profits generated by debt financed capital.

¹⁸ The differences relate to income effects and the timing of tax burdens. For a direct reduction in tax rates, the individual has more income (in that he can consume more in the present and the future) and could actually consume more in the future by saving less. In a consumption tax, with a stylized life cycle model, the income of older individuals who are mostly dissaving is reduced, which discourages saving and consumption. The income of younger individuals who are saving is increased, but taxes are also shifted to the future when consumption occurs, thus encouraging individuals to save more currently to pay future taxes.

¹⁹ Such a model has been advanced by Dale Jorgenson and Robert Hall. For a discussion of Hall's application to the flat tax see Robert E. Hall, "The Effects of Tax Reform on Prices and Asset Values" in Tax Policy and the Economy, 1995, James M. Poterba, ed., National Bureau of Economic Research. For a more detailed discussion of this type of model, references to the literature, and a comparison with others, see Jane G. Gravelle, The Economic Effects of Taxing Capital Income, Cambridge, MA, MIT Press, 1994, pp. 34-43.

²⁰ See Don Fullerton and Diane Lim Rogers, Lifetime Effects of Fundamental Tax Reform, January 1996, presented at the Brookings Conference, 1996, who use a higher and lower elasticity scenario that shows a range of reductions between about two percentage points and one/half a percentage point in the long run. Their model is a myopic model where individuals expect current interest rates to prevail; such a model tends to predict a very large savings response in the short run. Alan Auerbach, Tax Reform, Capital Allocation, Efficiency and Growth, January 1996, (also presented at the Brookings Conference) finds rates of return unaffected in the short run (there is actually a negligible increase because of an increase in labor supply); after 5 years the return has fallen by 0.2%, after 10 years, by 0.7%, and in the long run by 1.4%. Both of these papers were presented at a Brookings Conference on February 15-16, 1996.

²¹ Eric Engen and William Gale, The Effects of Fundamental Tax Reform on Saving, February 1996, also presented at the February 1996 Brookings Conference.

²² For a survey of the evidence on the interest elasticity of savings, see Jane G. Gravelle, The Economic Effects of Taxing Capital Income, Cambridge, MA., MIT Press, 1994, pp. 25-28. Not all of these results indicated a positive savings response; however, a consumption tax should, in theory, increase savings.

²³ There is, however, not a linear relationship. As the elasticity increases the interest rate reductions become smaller, until they asymptotically approach two percentage points at very large values.

²⁴ The argument has been made that this characteristic is important in evaluating the distributional consequences of consumption taxes. See William M. Gentry and R. Glenn Hubbard, Distributional Implications of Introducing a Broad Based Consumption Tax, Presented at a Conference of the American Enterprise Institute, March 22, 1996.

²⁵ William G. Gale, Scott Houser, and John Karl Scholz. "Distributional Effects of Fundamental Tax Reform." Presented at a Brookings Conference, February 15-16, 1996. These results are also consistent with a previous study by U.S. Department of the Treasury. Office of Tax Analysis, Preliminary Analysis of a Flat Rate Consumption Tax, March 10, 1995.

²⁶ There is some evidence that the overall income elasticity of demand for owner-occupied housing services is slightly below the price elasticity. Rosen estimates the price elasticity at 1.0 and the income elasticity at 0.76. He also notes that price elasticity tends to fall with income. See Harvey S. Rosen, Housing Subsidies, Handbook of Public Economics , Vol. 1, Ed. Alan J. Auerbach and Martin Feldstein, Amsterdam: North Holland, 1985, pp. 395-396. Most researchers tend to find income elasticities of less than one. See, for example, the summary of findings in Randall Johnston Pozdena, The Modern Economics of Housing, Quorum books, New York, 1988, p. 24. This summary also reports price elasticities, which tend to be below one, but these are for housing in general and not for owner-occupied housing. There is also a tenure choice response, which makes the owner-occupied housing demand more elastic than the demand for housing overall. Follain and Ling report simulation results that suggests that the tenure choice elasticity is closer to one at lower incomes, but declines to close to zero at higher incomes. See James R. Follain and David C. Ling, Another Look at Tenure Choice, Inflation and Taxes, in The Journal of the American Real Estate and Urban Economics Association, Vol. 15, no. 3, Fall 1988, pp. 207-229.

²⁷ Under current law, the first $125,000 of gain may be excluded once for taxpayers 55 and over and gain is not recognized if another home is purchased within a specified time period. Ignoring these effects, at a 28% rate, an inflation rate of 3%, and depreciation of 1.4%, the savings after 10 years is 2.9%. After 20 years, the savings is 3.9% and, after 30 years, 4.8%.

²⁸ Patric H. Hendershott and Joel Slemrod, "Taxes and the User Cost of Capital for Owner-Occupied Housing," Journal of the American Real Estate and Urban Economies Association, Vol. 10, No. 4, Winter, 1983, pp. 375-393.

²⁹ This method presumes that the other components of housing services are in a fixed relationship to the housing structure itself -- that is, that maintenance cannot be substituted for asset purchase. While some substitution is actually likely, it is probably small.

³⁰ See Jane G. Gravelle, The Economic Effects of Taxing Capital Income, Cambridge, MA, MIT Press, 1994 and Don Fullerton and Diane Lim Rogers, Who Bears the Lifetime Tax Burden?, Washington, D.C., The Brookings Institution, 1993.

³¹ For a review of the literature on housing markets, see Lawrence B. Smith, Kenneth T. Rosen, and George Fallis, Recent Developments in Economic Models of Housing Markets, In The Journal of Economic Literature, March 1988, Vol. XXVI, No. 1.

³² The demand elasticity for housing purchase is 1 times .18/.29. Appendix D provides the formulates for estimating effects from a discrete change.

³³ See James M. Poterba, Taxation and Housing Markets: Preliminary Evidence, In Do Taxes Matter?, In Do Taxes Matter?, ed. Joel Slemrod, Cambridge, MA, MIT Press 1990, p. 154.

³⁴ These shares vary by region -- from 9% in the Northeast, 11% in the Midwest, 17% in the South, and 18% in the West. See U.S. Department of Housing and Urban Development. U.S. Housing Market Conditions. 3rd Quarter, 1995.

³⁵ Of course, there are often zoning barriers to turning an existing owner-occupied home into a rental property, if some physical alterations, such as subdividing into apartment units, is needed.

³⁶ This series is taken from James M. Poterba, Taxation and Housing Markets: Preliminary Evidence. In Do Taxes Matter? Ed. Joel Slemrod, Cambridge, MA, MIT Press, 1990.

³⁷ A similar pattern of small annual changes is found for the price of the median home.

³⁸ See Gale, Houser and Scholz, Distributional Effects of Fundamental Tax Reform. Tax increases as a percent of income are less than 2% in the middle income levels; itemizers who own homes would have a somewhat larger increase, and renters a somewhat smaller one (or a benefit).

³⁹ While switching to a consumption tax should increase the savings rate, changing the overall tax burden on capital income has uncertain effects in direction. This change is so small that it is unlikely to have significant consequences in any case.