Rethinking Federal Housing Policy

Abstract

Problem: Federal housing policy is made up of disparate programs that a) promote homeownership; b) assist low-income renters’ access to good-quality, affordable housing; and c) enforce the Fair Housing Act by combating residential discrimination. Some of these programs are ineffective, others have drifted from their initial purpose, and none are well coordinated with each other.

Purpose: We examine the trends, summarize the research evaluating the performance of these programs, and suggest steps to make them more effective and connected to each other.

Methods: We review the history of housing policy and programs and empirical studies of program effectiveness to identify a set of best principles and practices.

Results and conclusions: In the area of homeownership, we recommend that the federal government help the nation's housing markets quickly find bottom, privatize aspects of the secondary mortgage market, and move to eliminate the mortgage interest deduction and replace it with a 10-year homeownership tax credit. In the area of subsidized rental housing, we recommend that the current system of vouchers be regionalized (or alternatively, converted into an entitlement program that works through the income tax system), sell public housing projects to nonprofit sponsors where appropriate, and eliminate some of the rigidities in the Low Income Housing Tax Credit program. In the area of fair housing, we recommend that communities receiving Community Development Block Grants be required to implement inclusionary zoning programs.

Takeaway for practice: In general, we recommend that federal policy build on proven programs; focus on providing affordable housing for low- and moderate-income families and provide the funding to meet that goal; avoid grandiose and ideological ambitions and programs; use fewer and more coordinated programs; offer tax credits, not tax deductions; and promote residential filtering.

Research support: Partial funding support was provided by the National Science Foundation.

Keywords:

• federal housing policy
• rental assistance
• homeownership
• housing affordability
• fair housing

Acknowledgments

We would like to express our appreciation to numerous reviewers for their helpful criticism and suggestions, to our university faculty colleagues for key insights, and to the editors of JAPA for their willingness to take a chance on a long and sure-to-be-controversial article.

Notes

1. This estimate is drawn from data compiled by the Mortgage Bankers Association and published in a report to Congress by the U.S. Department of Housing and Urban Development, Office of Policy Development and Research (2010). It is important to note than many mortgage payment delinquencies do not result in default or foreclosure, and that many foreclosures do not result in families losing their homes. By the first quarter of 2009, the 90-day delinquency rate had risen to 4% of all residential mortgages.

2. This discrepancy does not reflect differences in housing need. Nationwide, homeowner median income is twice that of renters: $59,886 to $28,921 (U.S. Census Bureau & U.S. Department of Housing and Urban Development Office of Policy Development and Research, 2008, Tables 3–20 and 3–21). Whereas the typical homeowner paid 20% of their income for housing-related expenses (U.S. Census Bureau & U.S. Department of Housing and Urban Development Office of Policy Development and Research, 2008, Table 3–13), the typical renter paid 30% (U.S. Census Bureau & U.S. Department of Housing and Urban Development Office of Policy Development and Research, 2008, Table 4–13). On the quality side, the share of owners and renters reporting that their units had had major structural problems was the same: 6% (U.S. Census Bureau & U.S. Department of Housing and Urban Development Office of Policy Development and Research, 2008, Tables 3–2 and 4–2).

3. Numerous excellent housing policy histories are available, including Carliner (1998), DiPasquale and Keyes (1990), Hayes (1985, 1995),Orlebeke (2000), and especially, Schwartz (2007).

4. Following Cohen, March, and Olsen (1972) this has come to be known as the “garbage can” model of policy formulation

5. The 1974 National Housing Policy Review resulted in one of the most significant changes in urban policy history, substantially replacing existing production-oriented housing program with the demand-oriented Section 8 housing allowance program.

6. Husock (2003) is a similarly short, but far more ideological critique of federal housing programs.

7. To fund the increase in voucher expenditures, Glaeser and Gyourko (2008) proposed reducing the maximum home value against which homeowners could deduct mortgage interest on their income taxes to $300,000, but only in high-price low-construction-cost areas.

8. As an alternative, Glaeser and Gyourko (2008) propose that the federal government cut off all aid (not just housing assistance) to communities with high-priced housing and little housing construction. To avoid such an outcome, communities could agree to adopt more housing-friendly zoning and subdivision codes.

9. The “hard to house” are defined as those with incomes less than 30% of area median income for whom health, physical, or service needs drive their housing requirements (Brophy & Godsil, 2009, p. 74).

10. This interpretation is reinforced by the fact that Congress required the GSEs to increase their purchases of mortgage loans to underserved households in 1993, presumably in exchange for the continued granting of this guarantee and subsidy.

11. Some historians date the beginning of federal efforts to promote homeownership to the 1917 creation of the federal income tax system, but the deduction established for interest covered all interest payments, not just those on residential mortgages.

12. One of the great advertising lines of the “Own Your Own Home” campaign was, “To install your wife in a home of her own is a convincing demonstration of your affection and consideration for her comfort and happiness” (Vale, 2007, p. 27).

13. Most residential mortgages originated before 1934 were either simple-interest, balloon mortgages (in which monthly payments cover only interest and principal was repaid at the end of the loan in a balloon payment) or fully-amortizing loans of shorter terms like 10 years.

14. Data on bank lending activity is available to the public under the provisions of the Home Mortgage Disclosure Act (HMDA) of 1975. Initial HMDA data releases included information on loan approvals but not applications. In 1993, Congress amended HMDA to require banks to also list information on mortgage applications.

15. As of 2006, they included the following provisions: a) more than 50% of housing units financed by mortgages purchased by the GSEs must be for families with incomes not greater than the area median; b) at least 23% of housing units financed by GSE mortgage purchases must be for low- or very-low-income families; and c) at least 38% of housing units financed by GSE mortgage purchases must be for homes in central cities, rural areas, or other underserved areas, based on income and minority concentrations.

16. Whereas Fannie Mae and Freddie Mac securities issuances were more stringently underwritten and carried the implicit guarantee of the federal government, private issuances (such as those by Bear Stearns, Lehman Brothers, Morgan Stanley, and Goldman Sachs) included a much higher proportion of uninsured, high-risk loans. As housing prices started falling in 2006, it was the GSEs’ congressionally mandated ownership of these high-risk securities, and not their prior issuance activity, that most adversely affected their balance sheets, leading ultimately to a complete federal takeover in September 2008.

17. Implemented on a trial basis in 2005, the Bush proposals proved less attractive and more expensive than anticipated and were discontinued in 2007.

18. See Newman (2010) for a discussion of data on mortgage foreclosures.

19. As of the first quarter of 2009, the value of U.S. residential real estate had declined to $15.7 trillion (Federal Reserve Board, 2009).

20. To fill the funding gaps between these three programs as well as to give state and local governments a greater say in how rental housing subsidies are spent, Congress created the HOME program in 1990 as part of the Cranston-Gonzales National Housing Affordability Act. HOME provides formula grants to eligible states and localities to fund a wide range of activities that build, buy, or rehabilitate affordable housing for rent or homeownership; or provide direct rental assistance to low-income households. HOME funds are awarded annually as formula grants to eligible jurisdictions. Total HOME allocations in 2009 were just under $1.85 billion (HUD, Office of Policy Development and Research, 2010). Funded by Fannie Mae and Freddie Mac based on their volume of security issuances and administered by HUD, the NHTF distributes grants to states to provide housing for low and extremely low-income households, with a production goal of 1.5 million new affordable homes by 2018.

21. In Rucker v. Davis (2001), the Ninth Circuit Court of Appeals declared the one-strike law to be unconstitutional in cases where the criminal wrongdoing in question was actually committed by someone other than a legal tenant of the unit, such as the tenant's grown child or guest. This ruling was overturned in 2002 by the United States Supreme Court.

22. Estimates of the share of HOPE VI units offered at rents comparable to those tenants had paid previously in traditional public housing ranged from 39% to 50% (Popkin, Katz, Cunningham, Brown, Gustafson, & Turner, 2004)

23. As originally enacted, Section 8 was an umbrella program which, in addition to housing allowances, included two construction programs. Altogether, the Section 8 New Construction and Section 8 Substantial Rehabilitation programs produced 1.3 million units, a portfolio nearly as large as the entire stock of public housing. The funding for this program was canceled in the Reagan budget cuts of 1982 and was effectively replaced with the adoption of the LIHTC program in 1986. Currently, HUD also funds a project-based Section 8 program, which allocates vouchers to units in new projects rather than tenants.

24. HUD publishes FMRs annually for each housing unit size for all metropolitan counties. FMR is defined as the 40th percentile of the rent distribution for standard-quality rental housing units occupied by households who moved to their present residence within the past 15 months. FMR includes shelter rent plus the cost of all tenant-paid utilities except telephone, cable or satellite television, and Internet service.

25. There is some evidence from the Gautreaux, Holman, and Moving-to-Opportunity demonstration programs that combining vouchers with counseling, landlord recruitment, and other services results in better rental mobility and opportunity outcomes than rental vouchers alone (Schwartz, 2007, p. 172)

26. HUD no longer publishes average voucher utilization rates for the nation. Instead, utilization rates are reported for samples of public housing authorities.

27. The LIHTC program includes two types of annual credits: 9% credits for projects receiving no other federal funds or subsidies, and 4% credits for projects with other federal subsidies or that were financed using tax exempt bonds. Credits are prorated by the share of project units affordable to households with incomes that are 60% (or, at the developer's election, 50%) or less of the area median income. Credits are allocated on a competitive basis by state housing finance agencies and are capped at $2.25 per state resident per year. Credits can be taken for the first 10 years, although the units themselves must remain affordable for at least 15 years. In addition to new construction, credits may be used for substantial rehabilitation of older projects.

28. Although desirable for social policy reasons, the specifics of the LIHTC program mitigate against mixed-income housing.

29. Many state housing finance agencies stipulate longer periods of affordability than 15 years.

30. For the clearest analysis of the income incidence of affordable housing need, see Nelson (1994).

31. This comparison assumes that all those receiving housing assistance are low-income households with excess rent burdens and are eligible for assistance. While true in the main, there are a significant number of long-time residents of public housing and low-income housing credit units who have been allowed to remain in their units even after their incomes have risen above the ceiling levels.

32. Millions of households with incomes above $20,000 in 2000 were eligible for rental housing assistance, so, if anything, this method understates the level of rental housing need.

33. Yinger (1999) provides a concise summary of the enforcement provisions of the Fair Housing Act of 1968.

34. The 1989 and 2000 audits used similar methods, but the 1979 audit took a different approach, so direct comparisons of its results with those of the 1989 and 2001 audits are problematic.

35. Born out of a 1966 fair housing consent decree, the Gautreaux Program provided Section 8 vouchers to 7,000 families living in central Chicago to enable them to move to predominantly White neighborhoods in the suburbs. Follow-up evaluations showed significant improvements, particularly for children, compared to a group who did not move (Rosenbaum, 1995). Most notably, whereas only 4% of the children of non-movers eventually attended four-year colleges, 27% of the children of Gautreaux families went on to college. Inspired in large measure by the success of the Gautreaux program, in 1993 HUD commissioned the MTO experiment in which several thousand families living in public housing or project-based Section 8 housing were given supplemental vouchers and counseling to enable them to move to low-poverty neighborhoods of their choosing. As with Gautreaux, MTO compared mover outcomes to those of a non-mover control group; unlike Gautreaux, however, MTO focused on issues of poverty, not race.

36. A dissimilarity index measures how evenly two groups are distributed across subareas of a larger geographic area. It can be interpreted as the percentage of one of the groups that would have to move in order for both groups to be evenly distributed across all subareas. Dissimilarity index values vary between 0 (complete integration) and 1 (complete segregation).

37. FHA insures lenders against borrower default. Wrap insurance is additional insurance that covers pools of individually-insured mortgages. Because the underlying mortgages are individually insured, wrap premiums can be suitably reduced.

38. For households with incomes above $80,000, the mortgage interest deduction is most often a windfall rather than an incentive for saving to become homeowners. Concern over the U.S. budget deficit will increase the scrutiny of this expensive program.

39. For each state listed in Table 4, the number of needy households not currently being assisted was multiplied by the average monthly Section 8 federal assistance amount (as reported in HUD, Office of Policy Development and Research, 2006) for that state in 2000. Federal assistance amounts ranged from $263 per month in Arkansas to over $600 per month in New York. State totals were then summed to yield a national total of $17 billion.