«reNtal HousiNg Policy iN tHe uNited states Volume 13, Number 2 • 2011 U.S. Department of Housing and Urban Development | Office of Policy ...»
If encouraging larger homes also encourages larger lots, then this situation also creates negative— not positive—externalities. Of course, as I discuss in the next section, local policies that limit the development of high-density housing similarly lead to more driving and more carbon usage.
Rental Housing, Structure Type, and Local Barriers to Building One unfortunate side effect of reducing the home mortgage interest deduction could be an increase in rental rates, at least in the short run. If the demand for rental units rises, then prices will rise.
One natural means of counteracting that force is to increase the supply of rental housing by loosening the barriers to building multiunit housing.
The federal pro-homeownership policies are not the only public policies that stack the deck against higher density rental housing. At the local level, communities have long made it difficult to build, with land use restrictions that are often specifically targeted at multiunit dwellings. Although very small governmental units frequently enact these restrictions, they can have national implications. A reasonable case can be made that the extraordinary post-1990 growth of Atlanta, Dallas, Houston, and Phoenix, and the far more limited expansion of the regions around New York, Boston, and San
Francisco, owes much to the differences in land use regulations between these two sets of places (Glaeser and Tobio, 2008). Restricting high-density dwelling can lead to higher carbon emissions and movement away from more productive areas, which will have an adverse effect on national income. These restricting factors make local land use policy a national concern.
In this section, I review two types of land use policies and their impact on multiunit dwellings and prices. First, I consider suburban land use restrictions. I then consider land use policies within the large cities that are the natural places for tall, multiunit structures. I then discuss the pattern of land use restrictions across space.
Land Use Restrictions in the Suburbs A long and distinguished literature exists on land use restrictions in suburban areas. For example, Katz and Rosen (1987) examined growth controls in the San Francisco region and found less development and higher prices accompanied these limits on building. Much of this literature has focused on a narrow region of the country, because institutional structures differ from state to state and rules often change at very low levels of geography. No one section of the country can be considered typical, and, at best, a particular region can be illustrative of only a particular style of land use policy.
With those caveats, I will briefly describe the fairly intensive investigation of land use rules in greater Boston undertaken by Amy Dain, Jenny Schuetz, Bryce Ward, and me, which is found in Glaeser, Schuetz, and Ward (2006) and Schuetz (2008). This work gives a sense of the extent to which some areas have stacked the deck against multifamily dwelling. Both papers focused on 187 cities and towns that were within 50 miles of Boston. Boston itself was not included because the city’s planning process—like that of many larger cities—is quite sui generis. We used a combination of census data, Banker & Tradesman sales price data, zoning data from the MassGIS system, and the results of an extensive questionnaire filled out by representatives of each of the 187 cities and towns.
Although a great maze of regulations affects various forms of development, two rules are particularly relevant for the development of multifamily dwellings: (1) much land is zoned to allow only single-family detached dwellings, and (2) In areas that do allow multifamily dwellings, restrictions on minimum lot size per unit make multifamily dwellings more difficult to build.
For example, Schuetz (2008) found that 127 out of 186 areas have literally no land where by-right development (that is, no special permits or procedures are needed) of multifamily dwellings could occur. Another 47 have less than 10 percent of their area where by-right multifamily housing could occur. Sixty communities have no land area that allows multifamily dwellings even with a special permit, while an additional 65 allow multifamily construction by special permit on less than 10 percent of their land area. These data are fairly remarkable, because the survey includes places such as Brookline and Cambridge, which are quite close to Boston and quite urban in character.
Moreover, those places that do allow multifamily dwellings often have fairly large minimum lot sizes associated with the development. For example, out of the 59 communities that allowed some multifamily building by right, 26 required a minimum lot size of 20,000 square feet. Of the 126 communities that allowed multifamily dwelling somewhere, by special permit, 79 required a lot size of more than 20,000 square feet and 28 required a lot size of more than 80,000 square feet. In other words, these places allow multifamily dwellings but only if each unit is associated with about 2 acres of land.
24 Rental Housing Policy in the United States Rethinking the Federal Bias Toward Homeownership We have every reason to think that these rules are effective at limiting the production of multifamily dwellings. Schuetz (2008) found a strong, positive effect of the number of by-right lots on the amount of multifamily permitting. Glaeser, Schuetz, and Ward (2006) found that more multifamily permitting occurs as the average lot size for multifamily units decline. These results control for distance to Boston and historic density levels, which suggests that these rules are not merely reflecting the low state of demand for multifamily dwellings in outlying areas. The rules preventing the construction of multifamily units are, in fact, limiting the supply of new multifamily units.6 The increase in land use controls may explain the significant decline in permitting within the Boston region and, in particular, multifamily permitting. In the 1960s, 172,459 permits were issued in the Boston area, but, even though prices rose dramatically, only 84,105 permits were issued in the 1990s. In the 1960s, more than one-half of all permits were for multifamily construction. In the 1990s, more than 80 percent of permits were for single-family housing. Some of this change surely reflects the decline in publicly subsidized multiunit projects, but it is hard not to suspect that limitations on multifamily building played some role in the massive change.
The work on greater Boston illustrates a region that has made it quite difficult to build multiunit dwellings. Because multiunit dwellings are disproportionately rental, this difficulty for people to build multifamily units means that the supply of rental units is likely to have also been restricted by these rules. In some cases, these communities are sufficiently far from the metropolitan area’s core so that it would be surprising to find much multiunit building, but these rules also apply to many inner ring suburbs, where multiunit dwellings are quite plausible. Moreover, some of these outlying areas might also have developed higher density neighborhoods if zoning had allowed it, perhaps built around a major employer or a rail stop.
Land Use Restrictions in the City Restrictions on multiunit construction tend to be far more severe in suburbs than in central cities, but the barriers to building in the cities may be more important because they affect the area that is the most natural place to build multifamily units. Suburban building is difficult, but urban building is far more complicated. Few green fields for new construction exist in older cities, and most large buildings get produced through a complicated, ad hoc procedure. Glaeser, Gyourko, and Saks (2005) used condominium sales prices and construction costs to estimate that perhaps one-half of the cost of units in New York City can be seen as the cost of restrictions on building up, but their approach has little ability to determine which specific rules and regulations are most important in driving down construction and driving up costs.
It is certainly clear that permitting activity has decreased in many older cities and that, in many places, the construction of rental units has particularly slowed. Glaeser, Gyourko, and Saks (2005) documented the decline in new construction in New York City between the 1960s and the 1990s, which is all the more remarkable because prices rose dramatically over this time period. In the Schuetz (2008) found little clear connection, however, between these policies and rental costs, perhaps because neighboring areas are close substitutes for one another, or because her work was not able to control for many unit characteristics.
country as a whole, the number of new rental units being produced declined dramatically from the 1980s, when 358,000 new multiunit rental properties were produced during an average year,7 to the last decade, when 179,000 new multiunit rental properties were produced annually.8 Cities constrain the development of new buildings in many ways. The permitting process can take years, zoning codes often restrict building heights, and older areas can be landmarked in ways that restrict new development. It is often easy to see the increasing scope of regulations that restrict redevelopment of older areas, even though it may be hard to ascertain the effect of any particular policy. For example, Glaeser (2010) described the growth of historic preservation districts in Manhattan, and interest continues in expanding the size of Manhattan’s historic districts, which would make even more areas of the city essentially off limits to new, larger scale development.
During the past 45 years, large areas of New York City have come to be part of historic districts.
The estimate in Glaeser (2010) is that nearly 16 percent of Manhattan south of 96th street (excluding parks) is in a historic district. Moreover, those districts occupy much of the city’s best real estate, such as the land abutting Central Park. More than one-half of the land in historic districts is made up of three very large districts: Greenwich Village (added in 1969), the Upper East Side District (added in 1981), and the Upper West Side District (added in 1991). Since the Mayor Dinkins era, relatively limited new districting activity has occurred in Manhattan.
The Landmarks Preservation Commission, which reviews all proposals to change exteriors within historic districts, is charged with maintaining the traditional character of these areas. It should be no surprise, therefore, that they have not typically approved large-scale demolitions that would create taller buildings. For example, the Landmarks Commission denied permission to the developer Aby Rosen, who proposed erecting a Norman Foster-designed tower on 980 Madison Avenue that would have kept the façade of the Sotheby-Park-Bernet building. Glaeser (2010) reported evidence suggesting that significantly fewer tall buildings have been constructed, and less growth in the number of housing units has occurred, in historic districts.
Glaeser (2010) also reported the demographic changes that happened in historic districts relative to other areas in Manhattan south of 96th street. Over time, the demographics of the areas that became historic districts became relatively richer, better educated, and more White. The prices in these areas also rose faster than in other comparable areas. Perhaps because they are not building new rental housing, these areas are becoming ever more elite. Jane Jacobs postulated that preserving older housing would make cities more diverse, but historic districts appear to be having the opposite effect.
I do not mean to suggest that preservation districts are the most important means by which cities restrict the development of highrise units. They surely are not, at least in the United States. They simply illustrate one of the many ways in which new residential construction is limited in older cities. By restricting new construction, cities are not providing as much new housing as they could, which, in turn, increases congestion by pushing development out toward the urban fringe and increases pollution by pushing people into larger suburban homes.
Patterns Across Metropolitan Areas Although much of the research on land use restrictions has focused on particular policies within metropolitan areas, a smaller literature looks across areas. Some papers (for example, Glaeser and Gyourko, 2002) have drawn inferences based on prices, construction levels, and land density alone. That work certainly seems to suggest that important supply differences exist across space that cannot be accounted for by land availability, and that the value of land is itself far too low to explain differences in home prices across metropolitan areas. Other papers (for example, Gyourko, Saiz, and Summers, 2007; Saiz, 2010) have used the Wharton Land Use Restrictiveness Index to try to explain the differences in home prices across metropolitan statistical areas.9 The Wharton Index was based on a questionnaire sent out to land use professionals throughout the country. Gyourko, Saiz, and Summers (2007) then averaged the responses to relevant questions to form a place-specific index. The place-specific indices were then averaged to form a metropolitan-areawide index of land use regulation. Typically, the index is normalized to have a mean of 0 and a standard deviation of 1 across places. Several cross-metropolitan area facts have been established using the index.
One fact is that a robust positive correlation exists between the Wharton Index and price across metropolitan areas (Gyourko, 2009). For example, a one-standard-deviation increase in the index is associated with a 23-percent increase in housing prices. Naturally, a causal interpretation of this relationship is difficult because the Wharton Index is associated with many other factors, such as the education level of the metropolitan area. Saiz (2010) also showed that the index is positively correlated with topographic limits to supply, such as hilliness and coastline. Saiz (2010) also demonstrated that demand shocks have had more of a positive effect on price in areas with more land use regulation.