«reNtal HousiNg Policy iN tHe uNited states Volume 13, Number 2 • 2011 U.S. Department of Housing and Urban Development | Office of Policy ...»
Finally, the international discussant, Priemus (“Renting in the USA: A Dutch Perspective”), offers a useful perspective on U.S. policy and reminds us that other housing assistance models might be examined and adopted. Priemus highlights some areas of similarity, calling for reforms to policies in the Netherlands, which are very similar to those recommended by the authors in this volume for the U.S. system (such as eliminating the mortgage interest deduction and adopting universal housing vouchers), but he also points to substantial areas of difference.
The need to learn from and avoid repeating the recent housing crisis provides an opportunity to rethink the role that rental housing should play in meeting the needs of the huge variety of American households during the different stages of their lives. We hope that the creative and thoughtful articles in this volume will provoke and enrich that conversation.
4 Rental Housing Policy in the United States Rethinking the Federal Bias Toward Homeownership Edward L. Glaeser Harvard University and the National Bureau of Economic Research
Introduction More than 85 percent of single-family dwellings are owner occupied; more than 85 percent of dwellings in homes with more than three units are rented. When the federal government subsidizes homeownership explicitly, through the home mortgage interest deduction, and implicitly, through the government-sponsored enterprises (GSEs) Fannie Mae and Freddie Mac, it is pushing Americans away from dense multiunit dwellings toward sprawling single-family detached homes. Local zoning rules that prohibit high-density and multiunit dwellings contribute to the governmental bias against compact development. Does it really make sense to stack the deck against energy-efficient, economically vibrant urban density?
Federal pro-homeownership interventions have long had three common justifications: (1) homeownership provides a path toward individual prosperity, (2) alleged market failures in the mortgage market require market-making federal action, and (3) homeownership establishes a connection with good citizenship. The great housing convulsion of the past decade challenges the first justification Cityscape Cityscape: A Journal of Policy Development and Research • Volume 13, Number 2 • 2011 U.S. Department of Housing and Urban Development • Office of Policy Development and Research Glaeser by reminding us that prices go down as well as up and that subsidized borrowing can lead to a “foreclosure” rather than an “ownership” society. Moreover, even if the secondary mortgage market needs public organizers to solve information-related market failures (a highly debatable proposition), those public organizers do not need to subsidize home borrowing as the GSEs have done. As the public is now beginning to see the bill for the GSEs’ largesse, the national appetite for further subsidy of this nature seems to have diminished substantially.
Thus, the post-crash case for subsidizing ownership comes down to the connection between homeowners and citizenship. One can reasonably debate whether the correlations between ownership and citizenship-related variables such as voting in local elections are causal or large or valuable enough to subsidize (DiPasquale and Glaeser, 1999).1 One can even question whether subsidies that scale up with the size of a mortgage are either well designed or fair. In this article, however, I focus on a different reason to rethink federal housing policy and shift the focus from owning toward renting: ownership is tightly tied to structure type, and it is a mistake to subsidize people to live in large single-family houses instead of dense apartments.
Economists have often emphasized that the mortgage deduction would be nondistortionary if we effectively taxed the implicit housing services provided to themselves by homeowners. This argument is correct, and would have some weight if a tax on implicit housing services were politically plausible in the United States. But, in the absence of that tax, the mortgage deduction makes investing in homes far more attractive than investing in other, often more productive investments that yield financial returns (which are taxed) rather than housing services (which are not).
The connection between structure type and ownership is strong and understandable. In the Evidence on Homeownership, Structure, and Policies section of this article, I explore the simple logic of property rights and investment (for example, Grossman and Hart, 1986; Klein, Crawford, and Alchian, 1978). In general, ownership should be lodged with the agent who is in the best position to make investments and, in the case of a single-family detached house, that agent is the resident. In a multifamily dwelling, occupants share infrastructure and often lack any comparative advantage in investing in a roof or a boiler. Moreover, multiple owners make coordination costly, so renting, rather than owning, comes naturally to residents in bigger structures.
The implication of this connection between structure type and ownership is that subsidizing ownership is inherently subsidizing single-family detached dwellings, which means that these policies are implicitly encouraging moving away from high-density living. Because the home mortgage interest deduction subsidy scales up with the size of the mortgage, it also functions as a subsidy for larger homes, and, under some conditions, larger lots as well. It would be problematic to distort decisions about density and home size if, in the absence of interest rate subsidies, the free market was making socially optimal decisions.
Even without the subsidy, however, there would be too little density relative to the social optimum. Private decisionmakers ignore the pollution- and congestion-related externalities that are Green and White (1997) presented evidence suggesting that homeownership improves children’s outcomes, which may present a more serious case for subsidizing homeownership. The correlation of omitted parental characteristics with homeownership, however, makes this connection difficult to tease out. Coulson and Li (2010) is an important paper that estimated homeownership spillovers.
6 Rental Housing Policy in the United States Rethinking the Federal Bias Toward Homeownership associated with larger homes away from the urban core. Larger homes use more energy and singlefamily detached homes, which are typically larger, use more energy than rental units.2 Local governments have consistently enacted rules that prevent dense development. From this perspective, federal pro-homeownership policies are distortions on top of other distortions and this fact implies that the social costs of these policies are likely to be even larger.
Although the positive externalities that could come from citizenship may suggest a case for subsidizing homeownership, the offsetting negative externalities will exist as long as the world fails to charge the social cost of carbon emissions and traffic congestion. Subsidizing homeownership is not just about subsidizing an asset class; it is also about subsidizing single-unit living, which tends to be away from the urban core. Encouraging people to flee cities to live in low-density homes hardly seems sensible. Although improving federal rental policy is certainly possible, perhaps by improving the portability of vouchers, the biggest way that the federal government affects the rental market is through its subsidy of homeownership, and it is surely time to rethink the size and extent of that subsidy.
Reducing the pro-homeownership bias would increase the demand for rental housing, and an unfortunate side effect of this increased demand might be that rental costs could increase. One way to mitigate a rental cost increase would be to reduce the barriers to multiunit building that exist in much of the United States. Although these barriers are enacted at the local level, they affect the nature of America’s built environment and even the distribution of population across the United States as a whole. In many metropolitan areas, suburban enclaves have made it extremely difficult to build the kind of dense developments that would typically be rented. For example, a large share of suburban Boston is completely off limits to multifamily dwellings (Glaeser, Schuetz, and Ward, 2006). In cities, significant barriers to building have also existed, including in large historic preservation districts. These barriers make housing more expensive, and, when they prevent large multiunit structures, they limit America’s stock of rental housing.
In the penultimate section of this article, I discuss these barriers and what can potentially be done to encourage localities to internalize the social costs of barriers to building up. Because the federal government cannot directly control local land use rules, the most practical approach would be to create financial incentives for more permissive construction of multifamily units. One approach would be to use funds taken from either a decreased home mortgage interest deduction or from the problematic low-income housing tax credit (LIHTC) to sponsor a density fund, based loosely on the U.S. Department of Education’s “Race to the Top” fund, that would reward a few localities with unrestricted aid in exchange for their meaningfully reforming local land use controls in a way that creates more high-density rental units.
In the next five sections, I (1) revisit three nonstructure-related reasons to question the home mortgage interest deduction; (2) focus on the link between ownership and structure type; (3) present facts about ownership, structure type, and energy usage; (4) provide a brief discussion of landuse-related issues; and (5) conclude with a modest policy proposal that involves reforming the home mortgage interest deduction and creating a fund to prod localities toward constructing highdensity, presumably rental, dwellings.
Three Nonstructure-Related Reasons To Rethink the Mortgage Deduction Before addressing the connection between ownership and structure type, I will first restate the three primary objections to the home mortgage interest deduction in its current form.3 (1) Although support for homeownership is often billed as providing a path toward wealth accumulation, the deduction is just as likely to encourage excessive risk-taking; (2) as Poterba and Sinai (2008) documented, the deduction is extremely regressive; and (3) the deduction distorts the purchase decision and encourages excessively large housing units.
Does the Interest Deduction Lead to Wealth Accumulation?
One core rationale often given for subsidizing ownership is that homeownership can be a path toward asset accumulation. Although one can ask why homes, of all forms of assets, should be so privileged, it is reasonable to revisit the question of whether the home mortgage interest deduction increases asset accumulation at all.
I start by assuming that the government has some interest in promoting wealth accumulation, perhaps because wealthier individuals are less likely to burden people around them or because the tax code reduces other forms of investment by taxing capital gains and dividends. If this is the primary motivation for encouraging asset accumulation, then it seems likely that the social value function is quite concave in private wealth because the goal is primarily avoiding extreme hardship. This challenges any strategy that leads toward excessive private risk-taking, such as encouraging people to borrow money with which to bet on a risky asset such as housing.
The best case for homeownership as asset accumulation is that (1) the individual saves more ex ante to get a down payment; (2) the person then buys a home, and the value of the home increases as home prices rise; and (3) the homeowner then pays down his or her mortgage. If all three events occur in this fashion, then homeownership will increase asset accumulation. Certainly, homes have typically been the largest asset for many American households.
Yet some of these events might not occur and might even be reversed because of home mortgage policies. The volatility of the housing market over the past 4 years has clearly illustrated that housing prices go down as well as up. In general, the Americans who bought houses during the boom have lost money and are now in far worse financial shape than if they had not bought. The asset accumulation approach will only operate effectively if we are confident that homes are likely to increase in value over time. In large parts of America, we have no reason to expect prices to rise regularly, and, in some cases, encouraging homeownership is a recipe for increased wealth loss.
As with any other properly functioning industry, the construction industry experiences technological improvements and increased efficiency. In the computer industry, technological progress has By current form, I mean without an accompanying tax on implicit housing services. If housing services were taxed, then the tax deduction would eliminate a bias favoring self-financing versus borrowing and could be quite justified. I do not believe that either a housing service tax or a consumption tax is on the table, so I will constrain my discussion to a comparison of deduction or not, holding everything else constant. Also notably, the discussion is pertinent primarily for the Americans for whom the deduction actually materially lowers their housing costs.