How large lot zoning and other town regulations are driving up home prices

function bigPopUp(url, width) { newWindow = window.open(url, ‘new’, ‘width=’+ width +’, height=650, resizable=1, scrollbars=1′); }

Large lots and other requirements drive up the cost of Bay State homes

Over the past 25 years, housing prices in Boston have exploded. According to the National Association of Realtors, in the third quarter of 2005 the median sale price of single-family homes in the Boston metropolitan statistical area (which also includes southern New Hampshire) was $430,900, higher than any other region of the continental US except for portions of California, the New York metropolitan area, and Washington, DC, and its suburbs. Between 1980 and 2004, housing prices in three of the Census Bureau’s four divisions of the Boston metropolitan area (excluding the two counties in New Hampshire) grew between 179 and 210 percent, which made these areas—Boston-Quincy, Cambridge-Newton-Framingham, and Essex County—second through fourth in the nation, behind only the New York area’s Nassau-Suffolk division (Long Island). This housing price explosion is partially a reflection of Boston’s remarkable success in reinventing itself, but it is also a reflection of policy choices that have led almost inevitably to an affordable housing crisis.

The starting point for Boston’s housing price boom is the city’s economic reinvention as a capital of the information age, a development that boosts the economy of the entire region. To have high housing prices in any given area, demand must be high—people must want to live there. But in a well-functioning housing market, high demand doesn’t necessarily lead to high prices. In Phoenix and Atlanta, both cities whose booms are even more striking than Boston’s turnaround, rising demand has been met with flexible supply, and price increases have been muted. In eastern Massachusetts, the housing supply response to the region’s reinvention has been anemic, and the growth of the region in the long run is surely threatened by a failure to build.

Indeed, over the past 40 years, supply has fallen, even as prices have climbed. In the 1960s, permits were issued for 172,459 units in the Boston area, and in the 1980s, 141,347 units. But in the 1990s, only 84,105 units were permitted, a drop of 40 percent from the decade before. The decline in the construction of multi-family buildings at the end of the 20th century was particularly sharp. In the 1960s, less than half of all permits in the Boston area were for single-family homes. In the 1990s, over 80 percent were for single-family homes, although multi-family buildings have made a partial comeback recently, accounting for 39 percent of the 61,800 units permitted in the region since 2000.

In other regions of the country that have boomed without dramatic increases in house prices, the supply of new housing has been much greater. In 2004, local governments in the Houston area issued 45,103 permits for single-family homes; comparable numbers were 57,360 in the Phoenix area and 31,741 in the Las Vegas region. That same year, cities and towns in Greater Boston issued permits for 8,204 new single-family homes. Not surprisingly, the prices in boom towns that have been building more housing have remained reasonable: As recently as the first quarter of last year, median home prices in Las Vegas were less than $300,000, and in Phoenix less than $200,000; in Houston, the median sales price last fall was $145,000.

There are two theories about why so little new housing is being built in Greater Boston. One is that the region has run out of land. The other is that restrictive regulations are impeding the development of new housing.

Even though the Boston metropolitan area is one of the country’s densest, there is little evidence that the region lacks the land to build new homes. Within the urban core, it would be quite feasible to build taller buildings. In fact, with strong support from the city, a host of new high-rise residential housing is being built in the heart of Boston. Outside of the city, densities are high relative to the US as a whole, but there is still a lot of land. For example, in Middlesex County, there are 0.37 acres per person, or 0.9 acres per housing unit. When we consider 187 cities and towns in eastern Massachusetts, excluding the city of Boston, the average density is 1.4 acres per home. This does not suggest that there is no land available. Towns like Weston, Lincoln, and Concord have densities of less than one home for every two acres.

Another way to test the theory of land scarcity is to look at the price of land for extending an existing lot, rather than building a new home. If land is scarce, it should be expensive whether there’s a house on it or not. But on average, an extra acre of land under an existing house adds only $16,600 to the value of the home. An additional acre under a new home is, by itself, worth $450,000.

UNTANGLING THE REGULATORY WEB

That leaves regulation as the reason for low levels of new construction and high housing prices in Greater Boston. To test this hypothesis, the Pioneer Institute for Public Policy Research and Harvard’s Rappaport Institute for Greater Boston spent the past two years compiling a unique dataset on land-use regulation in 187 cities and towns in eastern Massachusetts. Working under the direction of Pioneer’s Amy Dain, researchers answered more than 100 questions about each of these localities by reviewing documents from each community and interviewing local officials, who were subsequently given the opportunity to review the data about their community.

The most striking fact that emerges from this research is that, far from doing business in a unified market for housing construction, developers face an unbelievably heterogeneous set of conditions, varying town by town. The variation begins with minimum lot size, which remains the most powerful and widespread form of land-use control, and results in very different types of community. The 22 municipalities in the region with minimum lot sizes of less than a quarter of an acre contain more than one-quarter of the region’s population. In contrast, the 14 municipalities where minimum lot size is greater than 70,000 square feet —nearly two acres—contain only 4 percent of the region’s population but more than 10 percent of its land.

While minimum lot size is one way of managing development, communities have adopted a wide range of other controls that limit growth. The most direct approach is a growth cap, which limits the number of new units that can be built during a given year; a variant is a phasing schedule that limits the pace of construction within a single subdivision. Such regulations have become more common in the last decade. While 10 of the communities studied adopted these measures prior to 1994, 28 did so between 1995 and 2000, as did nine more since 2000. (We were unable to determine adoption dates for growth caps in another seven communities.) Another form of regulation prohibits irregularly shaped lots. Slightly more than half the communities with the largest minimum lot size requirements use this approach.

Almost two-thirds of the communities in the database also have wetlands bylaws or ordinances stricter than state regulations. Only a handful of these local environmental codes were adopted before 1980. More than 50 communities adopted the wetlands measures in the 1980s, and more than 50 others have imposed them since 1990. Similarly, 109 communities in our sample had septic-system regulations stricter than the state’s. All but six communities have rules governing subdivisions. Some adopted the regulations before 1950 and most did so by 1980. However, more than 70 amended their bylaws after 2000, almost always in ways that made it harder to build new subdivisions.

Not all local housing regulations impede high-density development. Some communities have adopted measures that allow construction on lots smaller than their zoning allows. Cluster provisions, for example, let developers build at higher densities if they set aside some amount of land for open space. In low-density communities, cluster zoning typically allows almost a two-thirds reduction in the minimum lot size required for each home. Still, many communities limit the units in a cluster development to the number that could be built on the entire parcel under conventional zoning, without the open space preserved.

Some communities also have inclusionary zoning provisions, which sometimes allow higher densities if builders keep some housing units affordable for low- or moderate-income households. While 99 municipalities—just over half of our sample—have some form of inclusionary zoning, with nearly half of them adopting the provision since 2000, in at least 43 of those communities the measures have never been used.

Many communities also offer a reduced lot size for developments open only to older adults. Almost 60 percent of communities with minimum lot sizes greater than 20,000 square feet have some provision for such age-restricted housing, including more than 40 percent of communities with minimum lot sizes of more than 35,000 square feet.

HOLDING DOWN PRODUCTION

While these restrictions are dizzying in their number and variety, do any of them make a difference in housing production? To answer that question, we added to the regulation information data on permits by locality going back to 1980, Banker and Tradesman data on house sales in eastern Massachusetts, Census data going back to 1910, and data on lot sizes for all the state’s localities, from the Massachusetts GIS system. Given that minimum lot sizes were imposed early in the century, we controlled for characteristics of the locality at the dawn of the zoning era (1915 or 1940). We then looked at differences in subsequent development.

Of the various forms of housing regulation imposed by municipal governments, minimum lot size has the clearest impact on pace of development. Larger minimum lot size is strongly associated with lesser amounts of housing stock and lower rates of permits issued for new homes. On average, for each increase of one-quarter acre in minimum lot size, there were approximately 10 percent fewer houses in place as of 1970, 9 percent fewer in 2000, and 10 percent fewer housing permits issued between 1980 and 2002. Conversely, if the average acres per lot fell one-tenth of an acre, then permitting would increase by 4 percent and total housing stock by as much as 5 percent in the long run. These results are, perhaps, unsurprising, but they do confirm the impact zoning has on new development. Perhaps more surprisingly, the connection between minimum lot size and development is declining over time, as even places that permit smaller lots are also radically reducing the amount of new construction they allow.

We then turned to rules that make new construction more difficult. Wet-lands regulations, septic rules, and subdivision restrictions do appear to have negative impacts on new construction, although the results do not meet the highest standards for statistical significance. When localities impose wetlands regulations stricter than those imposed by the state, new construction appears to be reduced by 5 percent to 10 percent. Stricter septic rules reduce new construction by about 4 percent and restrictive subdivision rules appear to decrease production by about 10 percent. Putting all three together, it seems that each additional regulation reduces annual permits by about 10 percent.

What about measures intended to alleviate the burdens of zoning? Cluster zoning does appear to increase new development, but we were not able to discern any impact from inclusionary zoning. We were also unable to assess the impact of Chapter 40B, the state’s anti-“snob zoning” law, which allows the state to overrule local land-use decisions for affordable-housing projects, because it impacts most municipalities. Nonetheless, considering the number of units constructed, it is clear that, in many areas, 40B is a significant part of new development.

Click here to view graphs

DRIVING UP PRICES

While regulations clearly affect permitting in each community, their impact on price is more diffuse, because housing markets are regional, not just local. Stringent land use restrictions in Wellesley, for example, would push up housing prices in Needham, even if Needham has less stringent land-use restrictions. Think of Needham and Wellesley as being two members of a cartel, like OPEC, that sells homes rather than oil. In OPEC, if Saudi Arabia restricts its crude production, this would raise the price of crude globally, but it wouldn’t raise the price of Saudi crude relative to Kuwaiti crude. In the same way, if Wellesley restricts production, it wouldn’t necessarily raise its prices relative to Newton (or at least not much), but it would likely push up prices in the entire region.

That makes it all the more remarkable that we do see a correlation between local land-use regulations and housing prices in individual communities. After controlling for community characteristics prior to recent changes in land-use regulations, each additional acre required per lot raised the median sales prices of homes by 15.8 percent in 1987, 11.3 percent in 1995, and 19.5 percent in 2001. The fact that the impacts were higher in 1987 and 2001 than in 1995 suggests that more restrictive land-use regulations are more potent at the high point of the real estate cycle.

Median sales price, however, does not control for differences in housing characteristics. The housing in areas with larger lots could be more expensive because the homes are larger and have more land. But there continues to be an effect on price even holding constant such factors as number of rooms, square footage, age of home, and total acreage of lot. Further controlling for other important characteristics (including distance to downtown Boston and whether the locality is home to a major college or university—a major distinguishing characteristic associated with a number of amenities), a one-acre increase in minimum lot size is associated with a 13 percent increase in housing prices, even for comparable house types. These results do not hold, however, when you control for housing density in the locality as of 1940, because the high correlation between housing density then and minimum lot size today makes it difficult to disentangle the impact of these two variables.

Remember, though, that price effects spill over into neighboring communities and, as such, raise prices throughout the region. With a rough, back-of-the-envelope calculation we also can provide a basic estimate of the effects of regulation on regional housing prices. In the 1990s, for example, the housing stock in Greater Boston increased by only 9 percent. If it had increased by 27 percent, as it did from 1960 to 1975, the median price of a home in the region might be 23 to 36 percent lower than it is now. That is, instead of costing $431,900, the median house price would have been as low as $276,100.

o, what have we learned from our analysis of land-use regulation, housing production, and home prices? First, while stringent land-use regulations may have advantages for an individual community, they clearly impose costs on the rest of the state. When a Boston suburb restricts production and pushes up prices, other people are prevented from buying homes in the restricted suburbs, and employers throughout the region end up paying higher wages. Indeed, there is little reason to expect that land-use regulations imposed on the local level will lead to the most socially desirable outcome without some form of state-level intervention, in the form of both carrots and sticks. As a carrot, the state could change local aid formulas to reward communities that allow their fair share of new housing growth. As a stick, the state could use 40B-type measures to override local zoning when necessary to meet housing production goals. The state could also set maximum standards for local environmental regulations that impose stricter limits than statewide environmental rules.

Second, property rights currently are diffuse and ill-defined, which results in a great deal of uncertainty in the development of new housing. The current permitting process frequently takes years from start to finish and is fraught with potential litigation. As a result, it imposes vast costs on developers. Policy actions that clarify rights and limit the potential for litigation, while simultaneously providing protection to current homeowners in a clear and predictable fashion, could potentially help both homeowners and developers.

Finally, it must be recognized that new development involves a trade-off between the interests of developers and those of existing homeowners. Indeed, if more housing development would limit the rise of housing prices, current homeowners would be better off with strong limits on development. The most natural way to resolve this conflict is through a system of cash payments between developers and existing homeowners. While some Massachusetts localities do, in fact, charge impact fees on new developments, the state’s courts have set stringent limits on such fees. Courts in other states, such as California, have given localities more leeway on the use and extent of such fees. In practice, local officials and community groups often work with developers during the approval process to devise a package of amenities that developers “voluntarily” agree to as part of their project. In theory, such amenities could include direct payments to abutting homeowners. While appealing from an economic point of view, such an approach would have to be carefully crafted because federal and state courts have often struck down such policies.

Each of these approaches would be controversial. But they would address the fact that the supply of housing is limited by regulation across the cities and towns of Greater Boston, and that this regulation is associated with more expensive housing. If policy-makers, leaders, and concerned citizens of the region and the state are serious about making housing more affordable, they must work together to lower the barriers to development. Doing so will be difficult, but the only way to effectively reduce the price of something is to produce more of it.

Edward Glaeser is director of Harvard’s Rappaport Institute for Greater Boston and is the Fred and Eleanor Glimp Professor of Economics in Harvard’s Faculty of Arts and Sciences (FAS). Jenny Schuetz is a doctoral student at Harvard’s Kennedy School of Government and Bryce Ward is a doctoral student at Harvard. They are the authors of Regulation and the Rise in Housing Prices in Greater Boston, a white paper published jointly by the Rappaport Institute and the Pioneer Institute for Public Policy, available at www.ksg.harvard.edu/rappaport/.