POLICIES FOR LOW AND MODERATE INCOME HOUSING HOUSING ALLOWANCES AND DEMAND ORIENTED HOUSING SUBSIDIES John M. Quigley, University of California-Berkeley and Yale University ABSTRACT It is estimated that the Federal government will spend more than $170 million on the Experimental Housing Allowance Program (EHAP). So far more then $50 million has been spent on EHAP research alone. This paper indicates some of the analytical issues in evaluation and suggests that housing allowance policies have ample historical precedent. SUPPLY AND DEMAND SUBSIDIES. Public subsidies for housing can be classified as supply or demand oriented. Supply oriented policies subsidize directly the construction or rehabilitation of dwelling units for occupancy by deserving households. They directly produce additional dwelling units for occupancy by households who meet certain program criteria. Indirectly, by increasing the housing supply, they may reduce prices in the short run for those existing dwelling units which are substitutes for the newly constructed or rehabilitated units. Demand oriented policies subsidize the housing consumption of deserving households. Such policies stimulate directly the purchasing power of selected households. They thus provide only an indirect stimulation to the construction sector and, by increasing effective demand, they may bid up prices in the short run in those housing sub-markets in which the subsidized households compete. During the past half century, the principal federal housing subsidies have been demand oriented that is, they have stimulated demand directly by reducing the price of housing for certain classes of households. The demand subsidies the treatment of mortgage interest, property tax payments and capital gains from housing on personal income tax returns amounted to $9.59 billion in 1977.1 The subsidies are given only to a certain class of households, namely to homeowners who itemize deductions. The amount of the subsidy for each recipient household is proportional to its marginal tax rate and to the quantity of housing it selects. The tax subsidy is equivalent to a voucher (or a "housing allowance") given to each household in this class. The voucher entitles the recipient to a specified percentage reduction in the price of housing, with larger percentage reductions going to households with higher taxable incomes. The administrative cost of this program is negligible. The subsidy is calculated annually when tax returns are filed, and vouchers are issued monthly in the form of reduced withholding taxes. It is estimated that these vouchers are equivalent, on average, to a reduction of 15 percent in the price of owner-occupied housing.-^ Of the $9.6B in vouchers issued in 1977, however, about $4.OB (or 42 percent) were redeemed by households with annual incomes in excess of $30,000.^ 300
By comparison with these demand oriented subsidies which accrue principally to middle and upper income households, federal subsidies for lower income households have been almost completely supply oriented, that is, payments have been made to those who built or substantially rehabilitated dwelling units for occupancy by low income households at reduced rents. Before 1974 only a few federal programs subsidized the demanders of low income housing or were not tied to newly constructed units. ^ Since the passage of the Housing Act of 1974, the principal federal housing assistance programs to low income households have included public housing, homeowner assistance (Section 235), new construction/rehabilitation (Section 8), and existing housing (Section 8). With the exception of the latter program and a small fraction of public housing (leased under Section 23), all other subsidies for low income households are tied to increases in the overall supply of housing. Deserving households, in this case households with particularly low income levels, are subsidized by making payments to suppliers who contract to produce a new or substantially rehabilitated dwelling unit for occupancy at reduced rents. The Section 8 existing housing program also provides for payments to housing suppliers to permit occupancy by low income tenants at reduced rents. As the name suggests, however, the dwelling units provided by landlords may be from the current stock of standard housing and need not be rehabilitated or newly built units. Under this program a household receives a commitment from the federal government to pay part of its rent and then uses this entitlement to seek its own housing. Tenants have some financial incentives to seek housing that meets minimal quality standards and also rents for less than the so-called "fair market rent" (they are able to keep some of the difference). Recipients are limited in choice; they cannot spend more than 25 percent of their income on rent; they cannot occupy a unit which costs more than the "fair market rent"; and they can only choose units owned by landlords who participate in the program. This program is demand oriented, in the sense that it subsidizes rent paying ability without t3dng it to increases in housing supply. Nevertheless it is clearly more restrictive than the housing vouchers given to homeowners. As of June 1978, it is estimated that 2,120,000 low income housing units are provided by supply oriented programs and only about 328,000 low income households are assisted by demand oriented programs. It is anticipated, however, that the relative importance of demand oriented subsidies will increase under current programs. For example, it is estimated that the "processing pipeline" currently includes subsidies for an additional 356,000 low income households under the Section 8 existing housing program compared to 804,000 additional units under supply oriented programs. Nevertheless, for about 85 percent of the low income households currently assisted by federal housing programs, the subsidy is still tied to the production of a brand new or substantially rehabilitated dwelling unit, 5 When compared with existing programs of housing subsidy, the concept of housing allowances is thus not radical or even particularly novel. The principal difference between housing allowance
proposals and the current array of housing subsidies is the payment of allowances to poor households. This "housing allowance" concept represents the confluence of two distinct social programs. Through a variety of supply oriented programs (beginning with public housing in the early 193O's) the Federal government has sought to insure "safe and decent" housing for the poor. At the same time, and only tangentially related to housing concerns, a variety of income transfer entitlement programs has emerged. A "housing allowance" program melds the two. A major criticism of current housing programs is the glaring inequity caused by the limited availability of subsidized units, so that equally needy households receive differential treatment. Further, the limited supply of subsidized units has meant that the most needy usually defined in income terms are not necessarily the first served. This has led to the charge that the programs are "inefficient" in that they do not accurately target subsidies according to need. Welfare programs have been subject to a different sort of criticism. Because of the incentive structure built in through eligibility rules, there can be very large work disincentives. Any work disincentives are magnified by complicated interactions among the eligibility rules for different support programs, including those rules governing current housing subsidies and traditional welfare programs. These criticisms have led to reform movements both in the housing and the welfare area; "housing allowance" plans represent an appealing reform. The inequities of housing programs are attacked by targeting subsidies to needy families, through entitlement subsidies to potential housing consumers based upon need criteria (i.e., household income and the cost of "standard housing"). The work disincentives of welfare plans are reduced by a negative income tax scheme designed to eliminate "notches" in payment schedules and, in general, to reduce marginal tax rates on work income. THE EXPERIMENTS. During the period through June 1, 1979 the Federal government has spent more than $120 million under the Experimental Housing Allowance Program (EHAP)^ a large scale program which combines some features of a "feasibility study" or demonstration with several features of a true "social experiment." The EHAP program consists of three separate experimental programs the Housing Allowance Demand Experiment (HADE); the Housing Allowance Supply Experiment (HASE); and the Administrative Agency Experiment (AAE). Under the HADE program, which comes closest to being an experiment in the scientific sense, some 1200 households in each of two local housing markets (Pittsburgh and Phoenix) were offered various experimental payments under differing payment formulas and earmarking restrictions for a three year period. The behavior of these households and a sample of control households was monitored for a period of two years. Public expenditures for this experiment were $29 million $3.6M in transfer payments to low income households, $14.OM in program administation, and $10.4M in research.^ 302
Under the HASE program, au households in two local housing markets (South Bend and Green Bay) who met specified eligibility criteria were offered housing allowance subsidies under a single payment formula with a uniform earmarking restriction. Households were offered these subsidies for periods up to ten years. So far, public expenditures for this experiment have exceeded $72 million $25.6M in transfer payments, $10.OM in program administration, and $36.5M in research. Under the AAE program, four different types of agencies administered housing allowance subsidy programs for two years in eight local markets. These subsidy programs were available to all renter households who met specified eligibility criteria under a common payment formula, but with local variations in earmarking restrictions. The overall budget for this experiment was more than $19 million $7.1M in transfer payments, $2.8M in program administration, and $9.2M in research. WHY CONDUCT EXPERIMENTS? The immediate reaction of most economists, schooled in notions of consumer sovereignty, is "why not simply give money?" The answer to this question seems twofold, and interrelated. In the first place, there may be physical externalities in housing consumption that warrant overriding recipients' choices. These externalities, for example, may come from neighborhood effects if the social cost of slums is imposed on others. Despite the fact that many individuals seem to regard these externalities as self evident, they are not well supported by scientific evidence. There is little or no evidence, for example, that the incidence of bad health, safety, or social pathology is affected by housing conditions, as distinct from income conditions. ^ Secondly, and perhaps of more importance, a subsidy program tied to housing consumption has ample historical precedent, and may be more feasible politically than a pure cash transfer. Donor sovereignty is not inconsistent with paternalism. If considerations of housing consumption per se motivate the notion of housing allowances, it still does not follow that an expensive set of experiments is necessary. Indeed, if social scientists could provide policy makers with reliable estimates of the price and income elasticities of demand, it would be possible to forecast or simulate directly the effects of certain allowance programs. For example, the long run effects of an unconstrained transfer targeted to housing "need" and the long run effects of a "percent of rent" subsidy similar to that given to homeowners depend only upon these two parameters. For a variety of technical reasons, however, there is still considerable uncertainty about the values of these parameters. ^ Thus a significant portion of the research sponsored by the HADE experiment is devoted to providing better estimates of these key relationships by analyzing the behavior of those Pittsburgh and Phoenix households receiving unrestricted "percent of rent" payments (price subsidies) or unrestricted payments geared to the "gap" between current income and the cost of "standard housing" (income subsidies). 303
There is also considerable uncertainty about the responsiveness of housing supply in the short run to increases in demand, i.e., about the value of the supply elasticity of housing services. Despite this uncertainty, however, it does not follow that the $72M supply experiment is really necessary to the choice among allowance policies geared to "housing gap" or "percent of rent" payments. Clearly, in the short run, a large-scale commitment to demand oriented subsidies could drive up prices. Just as surely as supply is unresponsive in the short run, however, it is responsive over the long term. No serious scholar or policy maker would propose a "crash" program of housing allowances, regardless of the outcome of the supply experiment, nor could one be dissuaded from a phased program on the basis of any experimental results indicating price effects in two local markets. Indeed, the obvious way to phase in a program of demand oriented subsidies would be to begin with a national program for the elderly poor a group which has been the target of other federal assistance programs (including particular housing programs) and which represents only a small fraction of overall housing demand. The one clear-cut case where the results of these experiments are critical to policy design is the consideration of "tied" subsidies, that is, programs which require households to live in units with particular characteristics in order to receive subsidy payments. The crucial issues in experimental design involve the kinds of housing restrictions imposed. If the restrictions are very loose (or nonexistent), the program is simply a negative income tax program (with the feature that a "housing gap" transfer recognizes regional cost of living differences as reflected in local housing prices and a "percent of rent" transfer reflects differences in tastes for housing). In a situation where housing restrictions are initially binding for a proportion of the population, those households who already meet the restrictions who previously expressed a taste for "standard housing" again receive a simple income transfer. For those who do not initially meet the standards, the program is a combination of a pure transfer program and an in-kind benefits program; the pure income transfer equals the gross transfer payment less the cost of upgrading to meet the standards. As particular requirments are more binding, and as there is less upgrading by eligible households, a housing allowance program incorporating restrictive standards looks more like a general welfare, or income transfer, program with eligibility based upon household tastes for a given type of housing. The inequities of past housing programs, determined by supply constraints on available units, are replaced by inequalities determined by housing tastes tastes which are hardly immutable. The effectiveness of any particular form of a housing allowance, as both a welfare program and housing program, relates directly to these question. ISSUES IN EVALUATION. There are two very different ways of evaluating the overall effects of an unrestricted housing allowance program, either a "housing gap" or a "percent of rent" payment. On the one hand, the results can be evaluated in terms of their ['effectiveness" in stimulating increased housing consumption by low income households; on the other hand, the results can be summarized in terms of the high "rent burdens" faced by poor people. A program could be viewed as successful if it furthers either of these objectives. 304
Note, however, that a subsidy program will always indicate an improvement in one or the other of these dimensions, and that no program can further both objectives. A price subsidy will always be more "effective" in stimulating housing consumption than an income subsidy of the same amount, but will have a smaller effect on relative "rent burdens." In general, if the housing consumption of low income households is highly responsive to price, then a "percent of rent" subsidy would be viewed as relatively "effective," but would have little effect upon the "rent burdens" of low income households. Conversely, if housing consumption is insensitive to price, then "percent of rent" allowances are relatively "ineffective" in stimulating consumption, but they will greatly reduce the "rent burdens" of low income households. If, on the basis of the HADE research, estimates of the demand parameters (price and income elasticities) are available, the effect of housing allowance programs on housing consumption and rent burdens can be forecast. However, the desirability of any particular program will depend upon policy makers' views about the importance to the larger society of these different outcomes an increase in the housing consumption of poverty households who currently live in slum conditions, or in a decrease in the large fraction of their small incomes which must be devoted to shelter. Changes in rent burdens relate most directly to the "welfare" goals of a cash transfer, while increases in housing consumption relate most directly to "housing" goals of such a transfer. Undoubtedly, the most complex problems in evaluation relate to notions of standards and tied subsidies. The imposition of a requirement that households must live in "standard" housing to qualify for assistance presupposes some common and precise definiton of minimally adequate housing. An evaluation of the effects of tied subsidies, moreover, presupposes some conceptual model of household demand for the characteristics of housing. Unfortunately neither of these conditions is met. The appropriate physical or occupancy standard is not well defined. (Should fenestration requirements be included in the definition of "standard" housing? Should ceiling heights?) Second, for any definition, the standard is not easily measured, and verification is both expensive and intrusive. Third, even if measurement is possible, the relationship between household preferences and any specific standard is problematic. Little is known about the relationship between various types of incentives, on the one hand, and decisions to participate and to upgrade or move, on the other hand. Two basic types of information are needed from the experiment: (1) What is the response of households to different kinds of housing requirements and transfer payments? (2) What are the costs of particular entitlement programs keyed to these differences? For "standards" requirements that are more restrictive, larger subsidies are required to induce households to participate. For any gnlven set of standards and payments, some projections of long run program costs could be made on the basis of HADE and HASE experimental results. 305
Again, however, the normative choice among standards cannot be made on the basis of any experimental evidence. Policy makers will have to decide whether standards dealing with narrowly defined physical components are more important than standards dealing with the quality of the neighborhoods and schools servicing those dwelling units. NOTES %.S. Congressional Budget Office, Federal Housing Policy: Current Programs and Recurring Issues, Government Printing Office, June 1978, p. 44. S. Mills, "Economic Analysis of Urban Land Use Controls,' in P. Mieszkowski and M. Straszheim (eds.). Current Issues in Urban Economics, Johns Hopkins Press, 1979, p. 531. U.S. Congressional Budget Office, O. cit., pp. 40-43. ^These programs include: relocation payments to households displaced by renewal programs (cash payments to small numbers of low income households for periods of up to four years); Section 23 leased public housing (cash payments to housing authorities for leasing existing housing for occupancy by low income households); and rent supplements (cash payments to private suppliers to make up the difference between reduced monthly rents and the amortization of the cost of new construction. Neither of the latter two programs made payments directly to low income households, however, and under neither program could a recipient household transfer the stabsidy if it moved to another unit. ^Figures in this paragraph are derived from program sxmimaries in Congressional Budget Office, o. cit. ^Federal expenditures are currently projected to be about $170 million by the time the entire set of experiments is completed. U.S. Department of Housing and Urban Development, The Experimental Housing Allowance Program: A 1979 Report of Findings, April 1979 (mimeo), pp. 1-5. 'Expenditures for all parts of the experiment were obtained from Mr. Garland Allen, U.S. Department of Housing and Urban Development. Many studies have indicated that the incidence of bad health is correlated with bad housing. This is, of course, beside the point. uncertainty arises principally from two sources, first, the costs of relocation indicate that the appropriate notion of income is some "long run," or "permanent" or "normal concept; second, housing prices are never observed in the market (market rents are in the units of price-times-quantity). 306