portland or maryland? Sticks or carrots? That’s the choice that the new administration needs to make as it translates candidate Deval Patrick’s stated support for smart growth into Gov. Patrick’s smart-growth policies.
Specifically, should Massachusetts follow Oregon’s “big stick” approach of forcing localities to follow state guidelines when deciding where growth should and should not occur? Or should the Commonwealth follow former Maryland Gov. Parris Glendening’s approach of using carrots, notably state capital spending, to encourage desired development patterns—while rarely, if ever, using the stick of forcing localities to comply with state mandates concerning growth?
Facing the same choice four years ago, the Romney administration chose a modified form of Maryland’s approach, creating a super-cabinet agency, the Office for Commonwealth Development, to oversee the actions of state agencies responsible for transportation, housing, environment, and energy in a way that would encourage compact development in existing city and town centers and steer development away from pristine countryside. From the outset, Romney, a pro-business Republican who took office during an economic downturn, made it clear that he wanted to use Maryland’s carrots, not Oregon’s sticks, to achieve his goals.
During the fall campaign, Democratic nominee Deval Patrick lauded smart growth as “the best idea to come out of the Romney administration,” but, like many smart-growth advocates, he also asserted it was an idea that his predecessor “never executed very much.”
Now that’s he’s governor, Patrick can improve on Romney’s record. But before he does so, he might want to review the extensive literature on the smart-growth policies of Oregon, Maryland, and several other locales. Such a review would, I believe, raise questions not only about the choice of Oregon versus Maryland but also about what smart-growth policies can actually accomplish. That’s because the literature suggests that while the Portland approach is likely to achieve more of smart growth’s core goals, it requires policies that are so politically controversial that they are not likely to be feasible in Massachusetts. In contrast, while the Maryland approach is more feasible politically, it’s not clear that it would significantly affect land-use patterns here—nor does an initial assessment of Romney’s Maryland-like smart-growth policies suggest anything different.
Such conclusions don’t mean that smart growth isn’t a worthy goal. Rather, they suggest that we should be realistic about what plausible, politically acceptable smart-growth policies are likely to achieve. And they suggest that smart growth may represent not the development nirvana its proponents purport it to be (meeting housing, transportation, and land conservation needs all at the same time) but rather a set of policies whose unintended consequences, particularly in housing costs, could make the cure as bad as the disease.
DRAWING THE LINE IN OREGON
While exact definitions of smart growth vary, virtually all start from the assertion that, for a variety of reasons, we should stop, or at least slow the pace of, sprawling development on the suburban/ rural edges of major metropolitan areas. To do so without stopping economic growth, we should at the same time encourage higher-density development in existing, built-up areas—particularly urban areas but also historic town centers and other places well served by public transit.
The state of Oregon has been trying to accomplish many of these goals since 1973, when it passed a landmark and still largely unique state law to guide and control growth. Under the law, local governments set a firm boundary dividing the areas where they will accommodate expected growth over the next two decades from largely agricultural areas where new development will not be allowed. In Greater Portland, these boundaries govern growth in a 400-square-mile area that encompasses 26 cities and towns and is home to 1.3 million people. Moreover, state agencies are charged with reinforcing these land-use plans by making investments that support desired growth patterns, such as downtown-oriented light-rail lines rather than new roads into undeveloped areas.
In many ways, this approach has worked. Compared with the years before these rules went into effect and with other metropolitan areas in the American west, much less farmland near Portland has been developed. Moreover, in keeping with policies requiring that the large amounts of undeveloped land inside the boundary be rezoned to require six to 10 units of housing per acre, housing densities seem to be increasing.
These trends are the opposite of what is happening in Massachusetts. As MIT’s Henry Pollakowski has shown, the median lot size for new single-family homes in Greater Boston grew from 0.76 acres in the early 1990s to 0.91 acres for houses built between 1998 and 2002. Moreover, as groups such as the Massachusetts Audubon Society have noted, farmland and open space in Massachusetts have been converted to residential uses at rates that far exceed the minuscule growth in the state’s population.
There are, however, important limits to what Portland has accomplished. According to a 2004 article in the Journal of the American Planning Association (JAPA) by Yan Song, a professor of planning at UNC/Chapel Hill, and Gerrit Knaap, director of the University of Maryland’s National Center for Smart Growth Research and Education, while most recent new developments in Greater Portland are more pedestrian-friendly than previous developments, the neighborhoods generally do not have the mixed uses favored by many smart-growth advocates.
Despite its touted accomplishments, the city of Portland has only about 1,766 housing units per square mile, or less than three units per acre; the densest community in the region is Johnson City, which has almost 5,000 units per square mile (about eight units per acre); and only one other community, King City, has more than 3,000 housing units per square mile. In contrast, Somerville has almost 8,000 units per square mile (more than 12 per acre); Cambridge almost 7,000; and Chelsea and Boston more than 5,000. Nine other communities in Greater Boston have more than 3,000 units per square mile.
In a similar vein, despite the investments in several new light-rail lines, there is no community in Oregon where more than 20 percent of residents take transit to work. Only in Portland do more than 10 percent of workers commute by public transit. Here in the Bay State, more than 30 percent of Boston’s residents use transit to get to work, as do more than 20 percent of residents in Brookline, Cambridge, Chelsea, Malden, Quincy, Somerville, Quincy, Revere, and Winthrop, along with more than 10 percent of residents in another 18 Greater Boston cities and towns.
Even more serious are questions about the effect of Portland’s growth controls on the cost of housing, which rose dramatically in Portland in the 1990s and the early part of this decade. Many economists believe that much of the rise was due to the fact that Portland’s growth policies reduced the supply of developable land, particularly inexpensive outlying land but also land for single-family homes within the growth boundary. In a 2005 JAPA article titled “Smart Growth, Why We Discuss It More Than We Do It,” Anthony Downs, a well-known urban and transportation economist at the Washington, DC–based Brookings Institution, observed that “this is why some analysts have concluded that smart growth and affordable housing are inconsistent goals.”
Others, however, contend that Portland offset most of the problem by requiring denser development within the urban growth boundary. Increases in Portland housing prices beyond those in other metropolitan areas, defenders say, were due to the fact that smart growth made Portland a more attractive place to live. But even if this were true, Downs responded, it means that affordable housing is possible under smart-growth schemes “only under unusual measures,” such as state policies mandating that communities accept denser development than they might desire and that those developments include at least some subsidized housing. The history of Chapter 40B, Massachusetts’s landmark affordable housing act, shows that these policies are often bitterly controversial, and that opponents may never stop trying to repeal them.
Finally, Portland–style systems have political consequences as well. Sharply demarcating where growth can and cannot occur creates economic winners and losers, and that sets the stage for political backlash. In This Land: The Battle Over Sprawl and the Future of America, Anthony Flint, a former Boston Globe reporter who later worked for the Office for Commonwealth Development, reported that some land just outside Portland’s urban growth boundary was one-tenth the value of land on the other side of the road. Owners of such land became the public face of a 2004 campaign that saw Oregon’s voters overwhelmingly approve a referendum requiring localities to provide compensation for lost value due to land regulations, or else allow owners to develop their land as they pleased. Given the cost of compensation, most localities have chosen to let development proceed, leaving Oregon’s growth-control system badly wounded. Explaining why the measure passed even in liberal strongholds that strongly voted against George W. Bush in the same election, Flint noted that “most Oregonians didn’t think [the measure] had anything to do with sprawl or farmland being destroyed. They thought it was about government bureaucrats screwing people over.”
The lesson for the Patrick administration is that “sticks,” like those used in Oregon, can be powerful tools. But they can also be so powerful that they ultimately lead voters to rebel against the entire smart-growth system.
OFFERING CARROTS IN MARYLAND
Portland, of course, has always been unique. Over the past three decades, several other states have flirted with systems that required localities to comply with statewide land-use goals. But none implemented a system as strict as Oregon’s. Rather, most states followed a course most fully implemented by Maryland in the 1990s.
In 1991, concerns about sprawl and the pollution of Chesapeake Bay led the state to consider an Oregon–style system. But those concerns were not sufficient to overcome the many forces that supported the existing system, which gave local governments and counties control over land uses in the state. Recognizing this political reality, Parris Glendening, who was elected governor in 1994, chose to rely on incentives rather than mandates.
“Regulations, Glendening believed, created more enemies than friends,” wrote John Frece, who was the Maryland program’s spokesman for six years, in “20 Lessons from Maryland’s Smart Growth Initiative,” a 2005 article in The Vermont Journal of Environmental Law. “Intuitively and politically, he knew that an incentive-based program would have a softer landing in the legislature and with the public-at-large.”
Under Glendening, Maryland sought to influence development decisions by “where the state spent its money,” explained Frece. “To this end, the Governor and his staff restricted almost every growth-related financial or technical assistance program the state offered to [targeted growth areas]: housing assistance programs, job creation tax credits, brownfield cleanup assistance, historic preservation tax credits, business expansion loans, park improvement funds, highway improvement funds, and the location and placement of state offices.” In addition, the state provided grants to enable local governments and private land trusts to buy development rights for land in rural areas where the state wanted to discourage growth. Such policies brought Glendening accolades, including an award from the Innovations in American Government Program, a program at Harvard’s Kennedy School of Government funded by the Ford Foundation.
But Maryland’s policies seem to have had only modest impacts. A report published by the University of Maryland’s National Smart Growth Center (authored by Frece, Knaap, Jungyul Sohn, and Elisabeth Holler) cites state data showing that while 76 percent of all new housing units built in Maryland in 2000 were within areas targeted for growth, that figure was only one percentage point higher than in 1997, the year before the law passed, and was much less than the almost 90 percent figure recorded in the 1950s and 1960s. Moreover, they added, low-density development outside of targeted areas was consuming about 75 percent of all the land being used for new development in the state. They added, “though this study does not prove that housing markets and development trends in Maryland have been adversely affected by [state] land use policies, there is evidence to suggest that state and local constraints on development are contributing to problems of housing affordability and deflecting growth to outlying areas. The result could be more, not less, urban sprawl.”
The findings from Maryland are particularly noteworthy because the Romney administration followed an approach here that might be described as “Maryland light.” As in Maryland, the state’s Commonwealth Capital program tried to direct capital spending to targeted areas. The Office for Commonwealth Development also tried to use its regulatory powers to ensure that 40B projects met smart-growth standards. And while it did not initiate the effort, OCD supported the Commonwealth Housing Task Force’s successful efforts to pass state laws (known as 40R and 40S) encouraging communities to develop denser, mixed-income housing in historic centers by offering financial incentives to create smart-growth zoning districts; the legislation even included subsidies to cover the marginal school costs incurred for the additional children who would be living in those districts.
Taken as a whole, however, these measures amounted to less than what Maryland did under Glendening. Although it’s too soon to assess the impact of the newer smart-growth incentives, sprawl is continuing apace. In the Greater Boston Housing Report Card, 2005-2006, Northeastern University’s Bonnie Heudorfer and Barry Bluestone report that between 2000 and 2004 (the last year that data was available), the population of Boston’s outer suburbs grew by about 5 percent while the population of the region’s inner core fell by almost 1 percent. They also found that while employment fell throughout the region between 2001 and 2006, the declines were particularly steep in the urban core.
RIDING THE RAILS
Oregon also offers important lessons about transit and transit-oriented development, two important parts of the smart-growth agenda and ones extolled by both Patrick and Romney before him. The idea of transit allowing people to move about without traveling in air-polluting cars and allowing development to be dense without resulting in traffic gridlock is an attractive one. But the reality is less comforting.
In his JAPA article, Downs observed that “building additional public transit facilities almost never reduces traffic congestion in a region once that congestion has reached the point of serious slowdowns during major rush hours.” As evidence, he cited, of all places, Portland, which doubled its light-rail system in the 1990s but saw traffic congestion get worse than ever before. “Why? First, a high percentage of new light-rail riders shifted from buses rather than private vehicles. Second, population growth in the region overcame any slight improvements in traffic congestion caused by the added light-rail facilities.”
Even Worcester, which Lt. Gov. Tim Murray, the city’s former mayor, touted in the campaign as a smart-growth success story, offers a sobering tale. According to Murray, the return of commuter rail in the 1990s helped reinvigorate Worcester. But according to the US Census, the city’s population fell from 165,759 residents in 1990 (before rail) to 154,389 residents in 2000 (after the line had opened). Moreover, the share of the city’s residents using transit to get to work fell from 5.8 percent in 1990 to 4.5 percent in 2000.
The census data do suggest that rail may have had some small, but concentrated, impacts in the three-square-mile census tract that includes the city’s downtown railroad station. Between 1990 and 2000, the population of that area increased (from 2,300 to 2,474 people), and the area became home to more working-aged adults, fewer elderly people, and about the same number of children. However, the share of the tract’s residents using transit to get to work fell from 5.8 percent to 4.5 percent in the same time period. Moreover, MBTA data show that the number of people taking the train from Worcester hasn’t really changed in the past six years and instead has fluctuated between 800 and 950 weekday morning boardings.
The problem may be that the combination of local land-use policies and transit agency inertia has stymied efforts to build developments that take full advantage of rail lines. Twenty years after the San Francisco’s Bay Area Rapid Transit system went into operation, for example, there was virtually no increase in the density near BART stations built to serve existing residential areas and very little new residential development near BART stations built in less developed areas, according to a study done by Robert Cervero and John Landis, both professors of city and regional planning at University of California–Berkeley and both supporters of transit-oriented development.
What more could we expect from aggressive efforts to spur development near transit stations? In a detailed study for the National Transit Cooperative Research Program in 2004, Cervero and his colleagues found that the “doubling of mean residential densities from 10 to 20 units per gross acre” in developments within walking distance of transit stops increases transit’s commute share from 20.4 percent to 24.1 percent for a typical Bay Area rail station. They added that when the areas also feature pedestrian-friendly designs, this figure rises to 27.6 percent. Elsewhere, however, the report notes that while increased densities produce more transit riders, the vast majority of people living and working near rail stations still commute by car. In other words, denser, transit-oriented development is likely to worsen traffic congestion, even as it increases transit usage.
While this study suggests that smart-growth advocates may have overstated the potential impacts of transit and ignored the trade-offs that come with transit-oriented development, other analyses suggest they have missed two other factors that shape land use: education and crime. As Edward Glaeser, a professor of economics at Harvard who also directs the Rappaport Institute, has noted, “aside from climate, education may be the most powerful predictor of urban growth.” While Glaeser focused on college graduates, he also asserted that improving K-12 education is important because “individual cities and towns with high dropout rates benefit less when their broader metropolitan area grows.” Similarly, scholars such as Julie Berren Cullen and Steven Levitt have found strong causal links between violent crime and city depopulation.
SMART SMART GROWTH
Taken as a whole, the findings reviewed here suggest four important lessons for the new Patrick administration.
First, while state policies can modify growth patterns, they probably cannot dramatically alter them. Indeed, the more ambitious the attempts to control growth, the more likely they will have unexpected impacts and invite a political backlash that could undermine the entire program, and even the governor himself. Therefore, Patrick should seriously consider whether it’s worth the political effort to launch a Maryland-like initiative at a time when so many other problems are vying for his attention.
Second, despite advocates’ claims to the contrary, serious smart-growth policies can impinge on other agendas, most notably efforts to make housing more affordable. Given the state’s still-high housing prices, the Patrick administration has to ensure that its smart-growth policies do not exacerbate the housing affordability problems that have contributed to the state’s recent economic woes.
Third, not all policies associated with the smart-growth ideal, particularly extensive investments in rail transit, are likely to be effective, and even some that appear to produce results, such as true transit-oriented development, may do so at higher-than-expected costs. That’s particularly important because the new administration has to decide how to allocate a limited pot of money among an almost unlimited set of demands for new spending, such as education, the state’s new health care law, and community policing.
Fourth, traditional efforts to encourage smart growth are likely to fail unless they include both sound capital investments and effective programs to improve education and reduce crime in older urban areas. The administration, therefore, should consider a broad-based approach that funds programs only if careful analyses show they are actually likely to achieve desired goals.These approaches might be politically unappealing, particularly because some are likely to pit the administration against some of its strongest supporters. But the new governor says he is committed to a new way of governing. Imagine the message Gov. Patrick could send by making it clear that his administration’s policies are going to be based on a prudent combination of important core values, political savvy, and critical analyses showing that the planned policies really are likely to address the problem at hand. Now that would be something new.
David Luberoff is executive director of Harvard University’s Rappaport Institute for Greater Boston.