CRS: CUT IN MORTGAGE INTEREST DEDUCTION BOOSTS HOME OWNERSHIP, PROGRESSIVITY.

DENNIS ZIMMERMAN SPECIALIST IN PUBLIC FINANCE ECONOMICS DIVISION

August 9, 1991

SUMMARY

The authority of State and local governments to issue mortgage revenue bonds (MRBs) is scheduled to expire on December 31, 1991. Along with other expiring tax provisions, any extension of MRBs must confront the pay-as-you-go (PAYGO) provisions of the Omnibus Budget Reconciliation Act of 1990. If MRBs are to be extended, a source of funding must be identified -- new revenue must be raised, expenditure on some other activity must be reduced, or some other tax preference must be scaled back. Dan Rostenkowski, Chairman of the House Ways and Means Committee, has commented on the paucity of suggestions for financing an MRB extension that witnesses before the committee claim is critical to providing access to owner-occupied housing for lower income people.

MRBs are one of several income tax provisions intended to subsidize increased homeownership. The most important of the other provisions are the mortgage interest and property tax deductions, and the exclusion from taxable income of the homeowner's imputed net rental income from the house. This report explores the option of financing MRB extension with an adjustment of the existing ceiling on the deduction for mortgage interest. This analysis is based on the proposition that under such an option almost all households subject to a reduced mortgage interest ceiling have income and other characteristics that would cause them to be home owners rather than renters, anyway, whereas one-third of potential MRB recipients would remain renters rather than become home owners if MRB financing were allowed to expire. This suggests that, if the Federal Government's subsidy of the housing industry is to be reduced by the $1.6 billion currently spent on MRBs, more social benefits sought from increased homeownership would result if the $1.6 billion were taken from mortgage interest deductions rather than MRBs.

Rough calculations indicate that $1.6 billion could be raised, for example, by denying 8.5 percent of mortgage interest deductions of individuals with adjusted gross income in excess of $100,000 or 25 percent of deductions of individuals with AGI in excess of $200,000. Either of these results could be achieved simply by lowering the current ceiling on the size of the mortgage for which interest payments can be deducted. Of course, any percentage of denied deductions and associated income levels could be utilized to achieve $1.6 billion of revenue provided the denied deductions are targeted to homeowners whose tenure choice is not affected by the lost deductions.

                              CONTENTS

The authority of State and local governments to issue mortgage revenue bonds (MRBs) is scheduled to sunset on December 31, 1991. Along with other expiring tax provisions, any extension of MRBs must confront the pay-as-you-go (PAYGO) provisions of the Omnibus Budget Reconciliation Act of 1990. If MRBs are to be extended, a source of funding must be identified -- new revenue must be raised, expenditure on some other activity must be reduced, or some other tax preference must be scaled back. According to the 1992 Federal budget, MRBs will cost the Federal Government $1.6 billion in fiscal year 1992.

The PAYGO provisions are supposed to force the Federal budget process to do what the budget process has seemed incapable of accomplishing in recent years -- make marginal choices among alternative programs rather than deficit finance all or most of the alternative programs. Legislation that would make permanent the authority to issue MRBs has been introduced in the 102d Congress by Representative Barbara Kennelly (H.R. 1067) and Senator Donald Riegle (S. 167). The House proposal currently has 358 cosponsors and may acquire the 371 cosponsors for MRB extension as in the 101st Congress; the Senate proposal has 88 cosponsors. House Ways and Means Committee Chairman Dan Rostenkowski recently noted that despite such overwhelming congressional support for the expiring tax provisions, he still awaits suggestions on how to pay for them. ¹

The potential funding sources for an extension of MRBs are many in a Federal budget of almost $1.5 trillion and tax expenditures amounting to hundreds of billions. Inadequate information on the value of public services makes it difficult to compare the relative merits of $1.6 billion spent on MRBs, whose primary purpose is to stimulate homeownership, with $1.6 billion spent on such diverse goods as artillery shells or intangible drilling costs, whose primary purposes are national defense and stimulation of oil and gas exploration.

Such analytical difficulties can be reduced by choosing for comparison Federal budget and tax provisions whose primary purpose is to promote homeownership. The three primary Federal homeownership tax subsidies are the exclusion of the imputed rental income on owner- occupied housing and the deduction of mortgage interest and property tax payments. This report explores the option of financing MRB extension with an adjustment of the existing ceiling on the deduction of mortgage interest from Federal taxable income. ² The mortgage interest deduction is estimated to cost the Federal Government $40.5 billion in 1992.

The first section of the report provides a brief overview of the nontax factors that determine homeownership, what the housing literature calls tenure choice. The second section analyzes the effect of the three individual income tax homeownership subsidies on tenure choice and the equilibrium quantity of owner-occupied housing services. The third section describes the income tax treatment of MRBs and analyzes their impact on homeownership and owner-occupied housing services. The final section explores funding MRBs by lowering the ceiling on the size of the mortgage for which interest payments can be deducted, thereby increasing the incremental homeownership generated by Federal tax subsidy. This section makes a rough estimate of the percentage of interest deductions that would have to be denied to high-income taxpayers in order to finance an MRB extension.

DETERMINANTS OF TENURE CHOICE

Substantial effort has been devoted to investigating the factors that influence the choice between owning or renting housing. The most important factors seem to be income (particularly permanent or lifetime income), tastes for homeownership, demographic characteristics (e.g. age and marital status), credit market conditions, the distribution of income, the relative price of owning or renting, and expectations about inflation and the future path of housing prices. ³

The Federal income tax subsidies to homeownership influence this choice by lowering the cost of homeownership relative to the cost of renting. ⁴ For any subset of the population with all nonprice determinants held constant, the ownership choice can be characterized as in figure 1. The vertical axis measures the relative price of owning or renting a standardized housing unit. ⁵ The horizontal axis measures the number of owner occupied housing units. This is not the standard representation of the housing market where housing services are placed on the horizontal axis, but this characterization is a useful illustrative tool.

The demand schedule D traces out the desire of households to own (rather than rent) as the relative price of owning to renting declines. A lower relative price increases the number of households desiring to own their housing unit. In order to simplify the diagram, standardized owner-occupied housing units receiving no Federal tax subsidy are assumed to be supplied at constant cost set equal to 1 and represented as Su. The results would not change, other than magnitude, were this unsubsidized supply curve drawn as upward sloping.

Given these circumstances, Q(sub 0) households choose to own housing at a relative price of P(sub 0) = 1. They choose to own due to some combination of their permanent income (relatively high), demographic characteristics (older and not single), tastes, etc. Other households rent because their endowments of these factors make it more economic to rent. This simple example illustrates the obvious, that even without individual income tax subsidies to homeownership many households (Q(sub 0)) would choose to own rather than rent.

THE ECONOMIC EFFECTS OF HOMEOWNERSHIP SUBSIDIES

The availability of Federal income tax subsidies to homeownership causes the aftertax cost of owner-occupied housing units to decline relative to its pre-tax cost. This causes the subsidized supply curve facing potential home owners, S(sub s), to differ from the unsubsidized supply curve, S(sub u), and lowers the relative price of owner-occupied housing. The number of households choosing to own rather than rent increases from Q(sub 0) to Q(sub 1) as the relative price of owning decreases from P(sub 0) to P(sub 1). ⁶ Thus, as suggested by the literature on tenure choice, the mortgage interest and property tax deductions and the exclusion of imputed rental income do act to increase homeownership.

FIGURE 1. THE MARKET FOR OWNER-OCCUPIED HOUSING UNITS

[figure omitted]

The total subsidy being provided is equal to (P(sub 0) - P(sub i))Q(sub i). However, (P - P(sub i))Q(sub 0) of the subsidy is going to households that would have owned without the subsidy. Put another way, Q(sub 0)/Q(sub i) share of the subsidy is not effective in generating any additional homeownership. How large is this share? Rosen and Rosen suggest that about 25 percent of the post-World War II homeownership increase through 1974 was attributable to the combined effect of tax subsidies (exclusion of imputed rental income and the deductions for mortgage interest and property taxes). ⁷ This suggests Q(sub 0)/Q(sub i) = .75.

Figure 1 seems to imply that 75 percent of post-World War II home owners received an income transfer, and that redistribution of income from all Federal taxpayers to home owners is the primary economic effect. But a more realistic representation of the market for housing services indicates that resource allocation is distorted. Figure 2 changes the horizontal axis to represent the quantity of owner-occupied housing services (note that each home from Figure 1 provides multiple units of services in figure 2). The relative price variable represents the price of a unit of housing service rather than the price of a house, but is set equal to 1 as in figure 1. Let the demand curve represent any subset of households of given permanent income, demographic characteristics, tastes, etc. These households can purchase additional housing services, that is more housing space per household (bigger homes or second homes), at an increasing cost per unit as represented by the unsubsidized supply schedule, S(sub u). Given their demand schedule D, they choose to purchase Q(sub x) owner-occupied housing services at the unsubsidized relative price P(sub x) = 1.

The effect of tax subsidies for owner-occupied housing is to lower the supply curve to S(sub s), reduce the aftertax cost of housing services to P(sub y), and induce households to increase consumption of housing services to Q(sub y). The noteworthy point here is that the increased housing consumption is coming from all Q(sub 1) households in figure 1 -- both the Q(sub 1) - Q(sub 0) households who switch from renting to owning AND the Q(sub 0) households who own with or without Federal tax subsidy. In effect, tax subsidies intended as devices to increase homeownership have, for the majority of recipients, actually provided an incentive to increase the size or quality (including preferred locations) of the home purchased rather than the incidence of homeownership. Were the purpose of the subsidy to stimulate the construction and real estate industry, the subsidy would be considered quite successful.

The extra savings being attracted by the subsidy into purchase of larger homes or second homes are wasted if the social objective is to promote homeownership. What is being sacrificed? The savings could be used for alternative investments yielding higher pre-tax rates of return. One economist has estimated that preferential taxation of owner-occupied housing, post-1986 Tax Reform Act, imposes a reduction in national income or welfare loss (area "abc" in figure 2) equal to about 0.1 percent of GNP (about $5 billion). ⁸

FIGURE 2. THE MARKET FOR OWNER-OCCUPIED HOUSING SERVICES

[figure omitted]

THE ECONOMIC EFFECTS OF MORTGAGE REVENUE BONDS

The tax law classifies MRBs as private-activity bonds. Bonds so classified are generally not eligible for tax-exempt status because they are considered to provide benefits primarily to private individuals or businesses rather than to the taxpayers as a whole. Several types of private-activity bonds, however, are considered to provide a sufficient degree of public benefits in addition to their private benefits that they are designated tax-exempt private-activity bonds. MRBs fall into this latter category. Due to the mixed public/private nature of their benefits, all tax-exempt private- activity bonds are subject to a State volume cap equal to the greater of $50 per State resident or $150 million. ⁹ In 1989, new issues of MRBs subject to a volume cap amounted to $5.6 billion, about 37 percent of the total volume of new issue tax-exempt private-activity bonds. ¹⁰

The initial use of MRBs was essentially unconstrained, at least by the Federal Government. Unless the State or local government issuing the bonds imposed household eligibility criteria, the bonds were available to any present or potential owner occupier, just as are the more traditional homeownership subsidies discussed above. But just as Congress has limited MRB volume, so it has also made considerable effort beginning in 1980 to target this restricted bond volume to households whose economic and demographic characteristics suggest their tenure choice decision might be sensitive to a tax subsidy.

Eligibility for MRBS financing must meet several targeting criteria. ¹¹

1. Income limitation -- The income of mortgage recipients can not exceed 115 percent of the higher of (a) median family income for the area in which the residence is located, or (b) the statewide median family income.

2. First time purchasers -- 95 percent of the mortgages must finance homes purchased by individuals not having an ownership interest in a principal residence for the preceding three years.

3. Purchase price limitation -- The purchase price can not exceed 90 percent of the average area purchase price.

4. Principal residences -- All homes purchased must be principal residences financed with new mortgages.

5. Recapture (permanent income restriction) -- Owners of MRB- financed homes sold within 10 years must pay the lesser of 1.25 percent of the original loan balance for each year the loan is outstanding or 50 percent of the gain realized on sale. This provision was adopted because many young and single individuals whose lifetime income was substantially higher than their current income were found to be using the MRBs for a short time, selling the homes, and using capital gains from appreciation to "move up." Some analysts observed that the bonds were financing the "starter" homes of the children of the middle and upper middle classes.

These targeting criteria do produce MRB home and buyer characteristics that differ significantly from all home buyers. In 1989, the average home purchased through State housing finance MRB programs cost $56,788 and was bought by a household with a $24,750 income. The median existing home cost $93,000 in November 1989 and required a minimum qualifying income of $31,700, while the median new home sold for $127,000 in November 1989 (and required a higher qualifying income). ¹²

The targeting criteria are not sufficient, however, to guarantee that all the MRB users will fall within the (Q(sub 1) - Q(sub 0)) households of figure 1 whose homeownership was induced by tax subsidy. It is possible that some households meeting these targeting criteria would have been able to find alternative financing to substitute for MRB financing, and that they were using the MRB financing to purchase a slightly bigger or better house, such as in figure 2. The General Accounting Office (GAO), using a standard mortgage affordability test, found that two-thirds of households receiving MRB financing from 1983 to 1987 could have bought the same home with other available mortgage financing. ¹³ This suggests that one-third of MRB recipients fell in the Q(sub 1) - Q(sub 0) incremental home owner range of figure 1.

The GAO one-third estimate is not, however, derived from a multiple determinants model of homeownership as is the Rosen and Rosen 25 percent estimate for the influence of the other homeownership tax subsidies. The GAO estimate assumes that every one of the other two-thirds would have purchased a home even without MRB financing because the affordability test is passed. But the tenure choice literature (see footnote 2) indicates that other factors influence the choice. Thus, the one-third estimate may be a minimum estimate of the homeownership increase attributable to MRBs.

GAO does raise another interesting issue. It suggests that the demographic characteristics (income, age, marital status, and race) of the one-third of MRB recipients who could not qualify for a conventional mortgage provide strong indication that they would have become home owners in the near future even without MRB assistance. If this is true, the only effect of the MRB subsidy is to influence the timing of the homeownership decision. Of course, the same type of criticism can be leveled at the three individual income tax homeownership tax subsidies that supposedly account for 25 percent of the post-World War II homeownership increase. ¹⁴

THE OPTION OF FINANCING MRBs WITH A LOWER CEILING ON MORTGAGE INTEREST DEDUCTIONS

One financing option for MRBs would be to restrict the deduction for property taxes. Property tax deductions, however, are viewed both as a subsidy to homeownership and a subsidy to State and local spending. There is no policy consensus on the wisdom of removing the state and local subsidy. ¹⁵ Another option would be to tax imputed net rental income, but efforts to use imputations to more nearly approximate an economic definition of income have foundered on questions of administrative complexity and public skepticism. Thus, among the three homeownership tax expenditures, the mortgage interest deduction seems the most likely source of financing for MRBs -- neither corollary economic objectives nor administrative complexities complicate the issue.

The issue here is whether it is possible to finance MRBs with reduced mortgage interest deductions and avoid an offsetting reduction in homeownership. Restated in economic terminology, can a reduced mortgage interest deduction be focused on inframarginal home owners? Recall from the earlier discussion that somewhat more than one third of MRB financings generate increased homeownership, while about one fourth of household ownership decisions were influenced by the combined effect of the deductions for mortgage interest and property taxes and the exclusion of imputed net rental income from the tax base.

This one third versus one fourth difference in incremental homeownership does not seem to provide much potential for increasing social benefits by reallocating some of the mortgage interest tax expenditure to MRBs. But the difference may well be greatly understated. Limiting mortgage interest deductions would be very unlikely to cause one fourth of the affected households to change their ownership status for two reasons. First, the one fourth estimate is based upon the price differential determined by comparing the price of ownership with and without all three individual income tax subsidies. The option being discussed here would partially eliminate only one of the three subsidies. So the price increase implicit in a lowered ceiling would be considerably smaller than the price decrease that produced the one fourth estimate.

Second, the one fourth estimate assumes household tenure choice sensitivity to any given change in the price of homeownership is uniform across all income classes. This is probably not the case. At this point, it would be desirable to present information showing that the tenure choice of households becomes less sensitive to the relative price of owning to renting as income grows. Unfortunately, estimates of the price sensitivity of tenure choice by income class are not available.

Table 1, column 2 contains data on homeownership rates by Federal adjusted gross income class (AGI) for 1987. Households above $100,000 AGI have ownership rates of 90 percent or higher. What is being suggested here is that a small change in relative price to high income households such as those with AGI in excess of $100,000 would have minimal effect on their ownership rates. Put another way, these households' income, marital status, age, and tastes indicate that most would remain homeowners (or that others who acquire these characteristics would become homeowners) even if the mortgage interest deduction were curtailed.

                               TABLE 1.

The flip side of the preceding argument is that the bulk of the 25 percent homeownership increase induced by tax subsidies is focused in the middle and lower income classes because they are more sensitive to price changes. Table 1 indicates that households with the average income of MRB users in 1989 ($24,000) have ownership rates of 25 percent; the $25,000-$30,000 class has an ownership rate of about 35 percent. A substantial share of these households who are not elderly may have their tenure choice affected by the combination of MRB financing and the three individual income tax subsidies. Paying for MRBs with a ceiling on mortgage interest deductions set above the dollar value of these income classes' mortgage interest deductions would increase (or maintain) the Q(sub 1) - Q(sub 0) of figure 1.

Existing treatment of mortgage interest in the Internal Revenue Code is fairly straightforward. Interest incurred in acquiring, constructing, or improving a qualified residence is deductible in the calculation of a household's Federal taxable income. A qualified residence includes the taxpayer's principal residence and one other residence (such as a vacation home). The deductible interest expense may not exceed that incurred on $1 million of debt.

If the average home were 25 percent equity financed, the current ceiling allows any household with a home(s) purchased for less than $1.33 million to deduct all interest expenses. It would appear that this ceiling could be much more restrictive, that is, reduce the size of the debt eligible for interest deductions, without influencing household tenure choice decisions. It does not seem rational that a significant number of households capable of carrying a $1 million mortgage would choose to become renters (thereby denying the Nation the social benefits of their homeownership) because they lose some of their mortgage interest deductions. It is more likely that the overwhelming majority of these households would, if anything, reduce the quantity of their housing services.

Given this analysis, consider the effect of tightening the ceiling on mortgage interest deductions to finance MRB extension. The intention here is not to provide a revenue estimate, but rather to discuss a few of the issues that might arise were such an option chosen. First, how much revenue must be raised? If we ignore the transition years and focus on future years when the equilibrium stock of MRBs has been achieved, the easiest approach is to assume the outstanding stock of MRBs would remain unchanged. This suggests the required revenue need would be equal to the current revenue loss estimate of $1.6 billion. ¹⁶

This $1.6 billion might be achieved through several types of mortgage interest restrictions, all of which could be designed to decrease the deduction for households whose mortgages and incomes are so large that their homeownership choice is thought to be insensitive to the price and income changes produced by the restriction. Lowering the existing acquisition debt ceiling below $1 million until $1.6 billion were raised would concentrate the denied deductions among the fewest number of home owners. In effect, this approach would deny 100 percent of interest deductions for any portion of a mortgage above a particular amount. A second option would be to deny a fixed proportion of interest deductions on mortgages above some amount. Any share less than 100 percent would lower the mortgage debt ceiling (relative to the ceiling required by a 100 percent denial) and spread the deductions to households with somewhat lower incomes and interest deductions.

The question is how low would the debt ceiling and how high would the percent disallowance of interest deductions have to be in order to raise $1.6 billion. Data on mortgages by AGI class are not available. Column 3 of table 1 contains the share of all returns itemizing mortgage interest. Column 4 presents the average mortgage interest deduction per return itemizing mortgage interest. Each income class contains unknown but presumably wide distributions of interest deductions and mortgage values. About all we can do with these aggregate data is to ask what the revenue consequences would be if a fixed proportion of interest deductions on all mortgages (rather than on mortgages above some amount) was denied. Note that the calculations do not account for any interactions with the itemized deduction phaseout equal to 3 percent of the amount by which AGI exceeds $100,000.

Revenue of $1.6 billion could be raised by denying 8.65 percent of mortgage interest deductions taken by all households with AGI of $100,000 and above; or 25.0 percent for households with AGI in excess of $200,000. These estimates include households with relatively small mortgages who would escape the chosen debt ceiling, thereby overstating the revenue gain and understating the required percentage reduction. But they also include households with relatively large mortgages whose revenue contribution is underestimated, thereby overstating the required percentage reduction.

Thus, assuming these two groups balance each other, a mortgage debt ceiling could be chosen that would raise $1.6 billion with a percentage disallowance of interest deductions around 8.65 percent for income classes above $100,000 or 25.0 percent for income classes above $200,000. ¹⁷ The size of the mortgage ceiling is unknown until data on mortgages by AGI class are available. Of course, basing the restriction solely on a mortgage ceiling would imply two things: first, the restriction on the interest deduction would apply to those households below $100,000 (or $200,000) whose taste for housing led them to have large mortgages relative to their income; and second, the restriction would not apply to those households above $100,000 (or $200,000) who have small mortgages either because their taste for housing led them to below average housing consumption relative to their income or because their wealth (self financing) or age (mortgage paid down) allows large housing consumption with small mortgages. Arguably, these effects are desirable, or at least tolerable, for the purpose of the policy is not to readjust the distribution of the income tax burden but to increase homeownership and, coincidentally, reduce the overconsumption of housing services attributable to the mortgage interest portion of the Federal subsidy for owner-occupied housing. ¹⁸

² Some would argue against this proposal on the grounds that it is what economists call a second best solution -- from an economic perspective the best or optimal policy might be to reduce both mortgage interest deductions and MRBs. It is argued that reduction of both subsidies is optimal because, on the margin, the value of the social benefits provided to the Federal taxpayer by both programs falls short of the reduction in national income suffered by the Federal taxpayer from the reallocation of scarce savings away from more productive non-housing investments to less productive housing investments. The question here is whether a policy to enact a suboptimal reallocation of housing resources to increase the increment of homeownership is worth doing when another policy that redirected savings to private sector allocation could generate a larger increase in national income. In a practical sense, public policy acts as though second best policies are efficiency improving because it continually engages in them. In practice, public policy often would be paralyzed if no changes could be made until all marginal conditions were satisfied.

³ Patric H. Hendershott and James D. Shilling. The Economics of Tenure Choice, 1955-79. In C. Sirmans (ed.), Research in Real Estate. Greenwich, CT: JAI Press, Inc. 1982, 105-33.; Harvey S. Rosen and Kenneth T. Rosen. Federal Taxes and Homeownership: Evidence from Time Series. Journal of Political Economy. Vol. 88, February 1980. 59-75; Harvey S. Rosen, Kenneth T. Rosen, and Douglas Holtz-Eakin. Housing Tenure, Uncertainty, and Taxation. Review of Economics and Statistics. Vol. 66, August 1984. 405-416; and Raymond J. Struyk. Urban Homeownership: The Economic Determinants. Lexington, Mass.: Lexington Books. 1976.

⁴ These subsidies can also influence ownership choice indirectly through the expectations variable. If the future path of tax policy is thought to be uncertain, that is, if the probability of the value of the subsidies remaining constant in real terms is less than one, the ownership choice will be affected. Rosen, Rosen, and Holtz-Eakin (1984) discuss the effect of uncertainty on tenure choice.

⁵ Of course, the housing market as currently structured does not generally offer households the same kinds of units in similar locations for either renting or owning. Thus, the standard housing unit referred to here is a (useful) theoretical construction.

⁶ The tax-induced lower price for owner-occupied housing also increases the household's aftertax income. If D is drawn as an uncompensated demand curve, the positive influence of this higher income on tenure choice is incorporated in the move from Q(sub 0) to Q(sub 1). If D is a compensated demand curve, the income influence is not incorporated.

⁷ Rosen and Rosen (1980). Their results indicated a tax- induced ownership increase from 60 to 64 percent over the sample period, which amounts to 25 percent of the total postwar increase from 48 to 64 percent.

⁸ Jane G. Gravelle. Differential Taxation of Capital Income: Another Look at the 1986 Tax Reform Act. Norton Tax Journal. Vol. 42. December 1989, 441-463.

⁹ The policy reasons for capping the issuance of private- activity bonds included limiting the Federal revenue loss, reducing the adverse impact on resource allocation (welfare losses), and reducing the supply of tax shelter instruments. For a more complete discussion of these issues, see Dennis Zimmerman. Tax Private Use of Tax-Exempt Bonds: Controlling Public Subsidy of Private Activity. Washington, D.C.: The Urban Institute Press. 1991.

¹⁰ See Dennis Zimmerman. The Voluntary Cap for Tax-Exempt Private-Activity Bonds: State and Local Experience in 1989. Washington, D.C.: Advisory Commission on Intergovernmental Relations. July 1990.

¹¹ For a more complete discussion of MRB targeting, see Dennis Zimmerman (1991), chapter 11. In particular, the income limits are somewhat more liberal for homes purchased in targeted areas (economically disadvantaged areas).

¹² The data for MRB programs comes from National Council of State Housing Agencies. HFA Homeownership Survey: 1989. December 17, 1990; the data on all new and existing home sales is from National Association of Realtors. Home Sales. December 1989.

¹³ U.S. General Accounting Office. Homeownership: Mortgage Bonds Are Costly and Provide Little Assistance to Those in Need. GAO/RCED-88-111, March 1988.

¹⁴ This discussion seems to imply that the dollar value of the MRB subsidy is not large enough to lower the income threshold for homeownership sufficient to increase the long-term stock of owner- occupied housing units. Since the three individual income tax subsidies are much larger, they may lower the income threshold sufficient to increase this long-term stock.

¹⁵ For a detailed discussion of the State and local tax deductibility issue, see Nonna A. Noto and Dennis Zimmerman. Limiting State-Local Tax Deductibility: Effects Among the States. National Tax Journal. Vol. 37, December 1984. 539-549. The Tax Reform Act of 1986 did eliminate the deduction for sales taxes, restricting deduction to income and property (real and personal) taxes.

¹⁶ This ignores the transition years in which the revenue loss would be considerably lower. The revenue gain in 1992 from allowing MRB authority to sunset would not be $1.6 billion. That figure represents the loss for the existing stock of MRBs, bonds whose maturity varies and whose tax exemption would continue until they were retired. Sunset would affect only new issues of MRBs. Thus, the PAYGO requirement in 1992 for extending MRB authority would be considerably less than $1.6 billion because it would depend only upon the volume of 1992 new issues of MRBs. Suppose 1992 MRB new issues are $7 billion. If that $7 billion instead could have been invested in taxable debt issues yielding a 10 percent return, the taxable income generated (foregone) would be $700 million. If the average effective tax rate on that $700 million was 30 percent, the foregone tax revenue to be generated in 1992 from limiting mortgage interest deductions is $210 million. The PAYGO revenue requirement would grow as the stock of post-1991 MRBs grows. If we assume the current stock is a good indicator of the future stock of MRBs, then in equilibrium we must raise $1.6 billion from the mortgage interest deduction.

¹⁷ Some might criticize this analysis because it uses a partial equilibrium framework and ignores behavioral responses that might flow from changes in the tax treatment of homeownership. Given existing budget procedures and the status of MRBs as an expiring tax provision, Federal spending on owner-occupied housing at the time in the future when the outstanding stock of MRBs is retired will be lower by $1.6 billion dollars. The analysis has focused on funding MRB borrowing authority by curtailing mortgage interest deductions by an equivalent $1.6 billion. Thus, whether the option discussed here were chosen or not, the real estate and construction industry would have to adapt to a $1.6 billion budgetary reduction (unless revenues to support MRB extension were found in some other part of the Federal budget). An issue would be whether the real estate and construction industry adjustment would be likely to differ significantly if the $1.6 billion decrease were taken from the MRB segment of the housing market or the upper-income end of the housing market. These differences would probably not be very great.

¹⁸ An issue some consider important has been ignored in this analysis -- the possibility of imposing windfall losses on households. In effect, this concern suggests that some households already have paid, at least partially, for the tax benefit in a higher housing price. This question of the captitalization of tax benefits arises any time the tax treatment of an asset is changed.