Zoning and the Tax Treatment of Housing

Decentralization—or suburbanization—is probably the single most important fact of recent urban development. Peter Mieszkowski and Edwin Mills have convincingly argued that suburbanization has been occurring in most developed countries throughout most of the twentieth century.¹ Decentralization in the United States has been particularly rapid, and it has been associated with considerable geographic sorting by income. Lower-income households have become increasingly concentrated in American central cities, while more wealthy households have opted for suburban communities.² The rapid pace of U.S. decentralization is often seen as the reflection of the intrinsic tastes of the American public, and hence some argue that it should not be an issue for public policy.³ Persistent concerns about central-city decline and suburban sprawl, however, suggest some dissatisfaction with current patterns of metropolitan development.

Observed patterns of decentralization are broadly consistent with the implications of the basic monocentric urban model developed by Edwin Mills, Richard F. Muth, and William Alonso more than thirty years ago.⁴ In the basic model and the generalizations that followed, land prices adjust to compensate for differential commuting costs across residential locations. This model implies that, as rising incomes increase the demand for residential land and as improved transportation systems lower the cost to commute to distant locations, metropolitan areas should become more decentralized and have lower population densities. The model further suggests that higher-income households will outbid lower-income households for the most desirable locations, leading to communities that are stratified by income. If high-income households strongly prefer large residential lots, they will tend to locate in more distant suburban locations where land is relatively less expensive, while lower-income households reside in more central locations.

To the extent that decentralization is driven by the forces described in the basic urban model, the process of decentralization is efficient and there should be few public policy concerns regarding the trend toward less dense residential patterns or the geographic sorting by income. There are at least three reasons, however, to reexamine the role of public policy in the process of decentralization. First, the pace of decentralization appears to be more rapid in the United States than in other developed countries, and it has frequently been associated with severe central-city decline.⁵ Severe decline in urban centers is not necessarily an implication of the monocentric model. Second, there is little evidence regarding the relative importance of preferences versus prices in the process of decentralization. Both commuting costs and house prices are significantly affected by public policies, which makes it less clear whether the observed pattern is solely a result of preferences, or whether the outcome is altered by policy-related effects on the prices of land and transportation. Third, the basic urban model does not address the actions of local jurisdictions, such as those related to zoning, that may affect patterns of development. Although more sophisticated Tiebout-type models suggest that efficient sorting by income across communities can occur without zoning, large-lot zoning, which has the effect of making some suburban communities inaccessible to lower-income households, remains common.

One national policy that can potentially affect both individuals’ choices regarding residence and land consumption and communities’ choices regarding zoning rules is the federal tax treatment of housing. Two aspects of the tax treatment of housing may affect patterns of metropolitan development. First, the U.S. tax code effectively reduces the price of housing, including residential land, relative to other goods.⁶ This increases the demand for housing and, because housing is a composite good consisting of structures and land, it also increases the demand for residential land. Because land is often a fixed factor in cities, the suburbs, with an elastic supply of land, have a productive advantage in housing. Thus the tax treatment of housing, by increasing the demand for housing, tends to favor suburban communities with elastic supplies of land. Second, the value of the tax-related housing subsidy is greater for higher-income individuals. In an earlier paper, Richard Voith and Joseph Gyourko show that in the presence of exclusionary zoning, these factors result...