«reNtal HousiNg Policy iN tHe uNited states Volume 13, Number 2 • 2011 U.S. Department of Housing and Urban Development | Office of Policy ...»
8 Rental Housing Policy in the United States Rethinking the Federal Bias Toward Homeownership led to affordable laptops with computing power that would have been exorbitantly expensive in 1980. As a result, no one buys a computer assuming that it will rise in value. The real cost of building a home declined by about 3 percent in both the 1980s and the 1990s (Glaeser, Gyourko, and Saks, 2005). In principle, declining construction costs could also lead to homes becoming more affordable year after year.
Declining construction costs have not always led to declining home prices, because homes have two other crucial ingredients: land and permits. When those two ingredients are not scarce—and they are not in much of America—then construction prices dominate, and we should expect prices to stay flat or fall. Indeed, in many major metropolitan areas, real housing prices barely rose between 1980 and 2000. Even Census self-reported home values, which do not adjust for generally increasing housing quality, suggest that prices stayed flat between 1980 and 2000 for many of America’s fastest growing cities, such as Las Vegas and Miami. In Houston, which is expanding, prices dropped dramatically. The core economics of housing prices suggests that this price stagnation is perfectly normal, because America has abundant land and because we expect the construction industry to become more efficient over time.
Nonetheless, even if housing prices decline over time, people will still accumulate wealth through homeownership if they pay down their mortgage. This savings channel has operated for millions of American households throughout history, but mortgage innovation and the subsidization of home mortgages can actually reduce this savings channel. Subsidizing homeownership will lead to savings if people are not regularly drawing value out of their homes with second mortgages. The home mortgage interest deduction, however, creates an incentive for people to draw more wealth out of their home by subsidizing the interest rate at which they will borrow. At the extreme, the interest deduction encourages people to have the maximum possible loan-to-value ratio, which ensures that they will have little wealth accumulation and great risk regarding changes in home values.
A similar effect is at work when we look at the fact that needing a down payment can affect wealth accumulation. A pure homeownership incentive will encourage people to save to ensure that they have the down payment needed to borrow. Subsidizing leverage and working to lower down payment requirements can work in the opposite direction. Subsidizing borrowing encourages lenders and borrowers to work together to ensure that the borrower needs as small of a down payment as possible. This subsidy reduces the incentive to save. Altogether, it is hard to see the home mortgage interest deduction as a sensible way to encourage Americans to accumulate wealth.
Progressivity and the Deduction A second obvious problem with the home mortgage interest deduction is that its benefits go disproportionately to the wealthiest Americans. Poterba and Sinai (2008) estimate that the home mortgage interest deduction typically saves $523 per year in taxes for homeowning families earning between $40,000 and $75,000. The average annual tax savings for families earning more than $250,000 dollars is $5,459. This calculation treats current mortgage levels as fixed, and, if the deduction were eliminated, some people would pay down their mortgages and prices would also change. Nonetheless, the 10-fold gap in savings illustrates the extent to which the deduction’s benefits flow to more prosperous individuals.
Many people may object to the regressive nature of the deduction on social equity grounds, but that is more a matter for philosophers and politicians than for economists. From a purely economic perspective, the regressive nature of the tax deduction may be problematic because it suggests that the deduction is poorly targeted. From a narrow economic perspective, the fact that the benefit goes disproportionately toward more prosperous individuals raises doubts about whether it effectively encourages homeownership.
Jesse Shapiro and I used the 1998 Survey of Consumer Finance to examine homeownership and deductibility by income group (Glaeser and Shapiro, 2003). We found that more than 90 percent of households in the top quintile of the income distribution are homeowners. This group appears to be at a corner solution in which essentially all rich people own houses. This finding surely reflects a strong connection between income and wanting to live in a single-family detached house (Glaeser, Kahn, and Rappaport, 2008). The model in the appendix also suggests that homeownership should yield higher benefits in high-quality units that require a lot of regular maintenance.
Similarly, richer people may choose to modify their homes, which is easier to do in an owneroccupied house. In the model, the ease of modification in an owner-occupied home would come from the fact that any such investments in a rental unit would then lead to higher rent charges from the landlord.
Although it seems likely that most wealthy people would own houses with or without the tax deduction, poorer Americans are much more likely to be split between owning and renting. For example, people in the third, fourth, and fifth income deciles are almost evenly split between owning and renting. The Poterba and Sinai (2008) evidence, however, suggests that this group is benefiting far less from the deduction. Moreover, Glaeser and Shapiro (2003) found that these middle-income homeowners often do not even itemize when they file their annual income tax returns. As such, the deduction yields substantial benefits to people who are likely to own anyway and does much less for the middle-income group that seems more likely to be in the margin between owning and renting.
Glaeser and Shapiro (2003) also marshaled evidence suggesting that changes in the value of the deduction have only modest effects on homeownership levels. For example, as Poterba (1984) demonstrated, the value of the deduction is sharply tied to the inflation level, because nominal interest rate payments rise with inflation. Little evidence suggests, however, that homeownership rose substantially during periods when the deduction was more valuable. Glaeser and Shapiro also examined the effect of state-level heterogeneity in the value of the mortgage interest deduction and again found essentially no link between the size of the deduction and homeownership. This finding may reflect that the deduction is poorly targeted at increasing homeownership.
These criticisms are less telling toward the implicit mortgage subsidies that the GSEs created. The canard that suggested that these entities improved mortgage markets without subsidy has been proven false by the markets’ collapse and subsequent bailout, and, as such, they can be seen as providing—in part—a second mortgage subsidy to homebuyers. This subsidy, however, is more effectively targeted toward middle-income Americans. Of course, other significant problems exist with the GSE model, which creates a private for-profit entity that can borrow at extremely low rates because of an implicit government guarantee.
10 Rental Housing Policy in the United States Rethinking the Federal Bias Toward Homeownership Moreover, the GSE interest-rate subsidy, like the home mortgage interest deduction, also encourages Americans to borrow heavily to bet on the vicissitudes of the housing market. Before the housing crash in 2008, many believed that encouraging people to borrow money to buy housing was an attractive way to create an “ownership society” in which more Americans accumulated assets. After the crash, this homeownership incentive seems just as likely to have created a “foreclosure society,” filled with people who were encouraged to bet everything on the unlikely event that housing prices would only go up.
The Deduction and Home Size The home mortgage interest deduction also creates a subsidy that encourages people to buy bigger homes. Because the amount that can be deducted scales up with the size of the mortgage, and because the size of the mortgage is related to the size of the home, people have an incentive to buy bigger homes if they want to receive a larger deduction.
In the first half of the 20th century, it was possible to believe that Americans should be consuming more housing. In 1940, 35 percent of American homes lacked a sewage or septic connection, 20 percent had more than one person per room, and 9 percent had more than 1.5 people per room (Glaeser and Gyourko, 2008). In those years, legitimate public health concerns prompted an interest in larger, better quality homes. Moreover, in the wake of the Great Depression, it was certainly reasonable to consider in-kind and cash-based redistribution, and federal support of housing could be justified on those grounds as well.
By 2010, however, the case for encouraging Americans to own bigger houses had essentially disappeared and, today, a better case is to be made for discouraging living large. By 1990, only 1.1 percent of homes lacked a sewage or septic connection and, 10 years later, the Census stopped asking the question because decent sewage was so universal. In 2000, only 5.7 percent of American homes had more than one person per room. On average, each American has more than two rooms and 992 square feet of living space (Glaeser and Gyourko, 2008). This number is astonishingly large by both historical and world standards. Average living space per capita is less than 450 square feet in Great Britain, France, and Germany. America is an extraordinarily well-housed nation.
Even more astonishingly, poor Americans are, on average, living in very big homes. In the bottom quintile of the American income distribution, the average person has 855 square feet of living space, more than twice the overall average consumption in France and the United Kingdom (Glaeser and Gyourko, 2008). In fact, among homeowners, the connection between square footage and income is extremely weak, in part because poorer Americans often live in places where housing is extremely cheap.
Although little reason exists to encourage Americans to buy bigger homes, increasingly legitimate reasons encourage thinking that externalities are associated with more housing consumption.
Perhaps the most obvious externality is associated with home energy consumption. Bigger homes typically use more electricity and home heating. According to the Residential Energy Consumption Survey homes with between 2,500 and 3,000 square feet of heated living space use 41 percent more kilowatt hours of electricity, 8 percent more natural gas, and 19 percent more fuel oil than
homes with between 1,500 and 2,000 square feet of heated space.4 Greater energy use in larger homes leads to more carbon emissions, and, if those carbon emissions create a negative global externality associated with climate change, then the government should be pushing for smaller rather than larger homes.
Negative externalities are also associated with bigger homes if those homes occupy larger lots. A larger amount of ground space per capita implies more travel. Because this distortion is also linked to structure type, I revisit this issue in the section on homeownership, structure, and policies.
The Economic Logic of Ownership and Structure Type The regressivity of the home mortgage interest deduction, or its effect on structure size, would occur even if everyone lived in a single-family detached home or if everyone lived in a multifamily dwelling. Because of the strong connection between homeownership and structure type, however, subsidizing homeownership implicitly subsidizes single-family dwellings. Although many condominium owners do take advantage of the deduction, a natural connection exists between ownership and structure type, and that means that subsidizing homeownership affects the physical structure of America. If all owners lived in single-family detached homes and all renters lived in multifamily dwellings, then subsidizing homeownership would be exactly the same as subsidizing single-family detached dwellings. Reality is not far from that extreme. In this section, I discuss the theory behind the structure-ownership connection.
I have already argued that the home mortgage interest deduction may not play that much of a role in encouraging homeownership; in that case, it may not play that much of a role in encouraging structural change either. Nonetheless, the deduction serves as a catalyst for homeownership and other federal policies, such as the implicit subsidies for Freddie Mac and Fannie Mae, which encourage people to own rather than rent. This section examines the effect that any such push for homeownership is likely to have on structure type.
Ultimately, homeownership is simply about the decision to assign property rights over an asset to one person or another. Therefore, making sense of homeownership requires investigating the core insights of research on property rights that have blossomed in recent decades (for example, Grossman and Hart, 1986). One hallmark of this literature is that the assignment of ownership can serve to mitigate the social losses associated with different contracting problems. For example, Klein, Crawford, and Alchian (1978) argued that the General Motors Corporation (G.M.) bought Fisher Body to minimize hold-up problems associated with relationship-specific investments.
G.M. needed Fisher Body to invest in very specific products to meet its needs as a downstream customer, but once Fisher Body made those investments, G.M. could hold up Fisher Body and threaten to pay less. Klein (2006) shows that the difficulties of this relationship were solved only when G.M. bought Fisher Body.
That same intuition helps us understand homeownership. Two critical forms of investment relate to the homeownership decision. The first, emphasized by Henderson and Ioannides (1989), relates http://www.eia.doe.gov/emeu/recs/recs2005/c&e/spaceheating/pdf/allTables1-13.pdf.