In late March, I'll fly to Berkeley to participate in an OECD Roundtable on sprawl. Below, I report a draft of the paper. I apologize that I don't show you the tables or figures. I can't figure out how to upload a .pdf file so I simply pasted this in. I'm trying to signal to you that not only can bloggers blog but we can also write "real" papers. I will let the market judge the value of this piece but I enjoyed writing it.
The Quality of Life in Sprawled versus Compact Cities
Matthew E. Kahn
Today, most Americans who live in metropolitan areas live in single detached homes and commute to work by automobile. New York City is America’s sole urban center where a significant fraction of the population lives in apartment buildings, works downtown and commutes by public transit. As transportation costs continue to decline and household incomes rise, we are choosing sprawl as we live and work in the suburbs.
The conventional wisdom is that this trend imposes major social costs relative to its benefits. An advanced Google search reveals that there are 39,500 entries for the exact phrase “costs of sprawl” while there are only 455 entries for the exact phrase “benefits of sprawl”. The beneficiaries of sprawl may be a “silent majority” who are not as politically active as center city boosters, environmentalists and the urban poor’s advocates in voicing their views on the merits of the ongoing decentralization of jobs and people taking place across cities in the United States.
This paper seeks to address this intellectual imbalance by presenting original empirical work documenting some of the benefits of living in a sprawled metropolitan area. This paper uses a number of U.S data sets to explore how sprawl improves quality of life. I focus on how sprawl affects firms, workers and consumers.
Opponents of sprawl often argue that suburbanization may offer private benefits but that it imposes social costs. The “cost of sprawl” literature posits that there are many unintended consequences of the pursuit of the “American Dream” that range from increased traffic congestion, urban air pollution, greenhouse gas production, farmland paving, to reducing center city tax revenues, and denying the urban poor access to employment opportunities. The last section of the paper argues that environmental regulation, new markets, and technological advance have helped to mitigate several of the social costs of sprawl.
Measuring Sprawl in the United States
The first step for comparing quality of life indicators in compact versus sprawl cities is objective data that allows major cities to be classified by “sprawl category”. A 2005 study by Reid Ewing, Rolf Pendall and Don Chen titled Smart Growth America creates such data for 83 major U.S metropolitan areas in the year 2000 (see www.smartgrowthamerica.org). These areas represent nearly half of the nation’s population. Table One lists these areas and reports their “compactness” ranking. Major metropolitan areas are listed from most sprawled to least sprawled. These authors base their sprawl index on four factors; residential density, neighborhood mix of homes, jobs, and services, strength of activity centers and downtowns, and accessibility of the street network.
As discussed by Ewing, Pendall and Chen (2005), “The most sprawling metro area of the 83 surveyed is Riverside, California, with an Index value of 14.22. It received especially low marks because it has few areas that serve as town centers or focal points for the community: for example, more than 66 percent of the population lives over ten miles from a central business district;• it has little neighborhood mixing of homes with other uses: one measure shows that just 28 percent of residents in Riverside live within one-half block of any business or institution; its residential density is below average: less than one percent of Riverside’s population lives in communities with enough density to be effectively served by transit; its street network is poorly connected: over 70 percent of its blocks are larger than traditional urban size.”
It is important to note that even in compact metropolitan areas such as New York City, there is significant suburban growth at the fringe. A broader definition of the New York City metropolitan area would include large pieces of New Jersey and Connecticut.
In previous research, I have used the share of employment within a certain radius of the CBD as my prime measure of sprawl (see Kahn 2001, Glaeser and Kahn 2004). The Ewing, Pendall, Chen (2005) measure is more comprehensive and offers an independent measure of sprawl.
Throughout this paper, I use their compactness index (see Table One) to partition metropolitan areas into four groups (i.e high sprawl, sprawl, low sprawl, very low sprawl). The most sprawled metro areas are those whose compact index lies between the 0 and 25th percentiles of the empirical distribution as listed in Table One. The least sprawled metro areas are those in the top 25th percentile of the empirical distribution. This simple classification system allows me to compare outcome indicators in low sprawl versus high sprawl areas.
Ideally, I would like to observe how people who currently live in sprawled cities would have lived their lives had they lived in a compact city. This counter-factual would allow me to measure how sprawl affects household wellbeing. If this information could be combined with preference information on how much people are willing to pay for such amenities as a short commute or a nice house then it would be straightforward to estimate the benefits of sprawl. In reality, this counter-factual can only be approximated by examining the outcomes for observationally similar people who live in high sprawl and compact cities.
The 2003 American Housing Survey (AHS) micro data set is a representative national sample for examining housing consumption in high sprawl and low sprawl cities. Over 20,000 people are sampled. Using the geographical identifiers in this data base, I merge the metropolitan area sprawl measures to this micro data. For 77 major metropolitan areas, I examine housing consumption in compact versus sprawled cities.
In Table Two, I focus on home ownership propensities and land consumption as a function of urban form. As shown in the top row of Table One, home ownership rates are 8.5 percentage points higher in the most sprawled cities relative to the most compact cities. In compact cities, the median household lives on a lot that is 40% smaller than the median household who lives in a sprawled city (i.e the 0 to 25th percentile of the compact distribution). The differential with respect to interior square footage is smaller. The median household in a compact city lives in a unit with 158 fewer square feet than the median household in a sprawl city.
While there are clear housing consumption gains for households in sprawled metropolitan areas, these observable differentials do not reveal how much households value such gains. The population differs with respect to its housing preferences. Those people with the greatest taste for large single detached housing will migrate to cities and areas where they can cheaply achieve their housing goals.
Some cities such as New York City remain compact due to maintaining a large share of employment downtown. Other cities have increased their compactness by fighting sprawl through Smart Growth policies of land use controls. A political economy literature has examined the distributional effects of who gains and who loses when cities battle sprawl (Katz and Rosen 1987, Portney 2002, Glaeser and Gyourko and Saks 2006, Quigley and Raphael 2005). Incumbent homeowners gain twice from such from anti-growth policies. By limiting increases in housing supply, these policies raise the value of existing homes. If these policies increase the quality of life of the city, then this will increase the demand for the existing homes.
Who loses from “Smart Growth” policies? It is well known that minority homeownership rates have lagged behind whites (see Colllins and Margo 2001). Part of this gap is due to differentials in wealth accumulation. In previous research, I have documented that blacks who live in sprawl cities “catch up” on some housing consumption dimensions to whites relative to the black/white housing consumption differential in compact cities (Kahn 2001). In Table Three, I present some new evidence on this question. I use the 2003 AHS data and focus on one measure of housing consumption, the number of rooms in the housing unit. I use multivariate regression techniques (i.e ordinary least squares) to control for such important demographic features as household income, the household’s size, presence of children. Controlling for these factors, I examine how urban form affects housing consumption.
As shown in Table Three, an increase in the metropolitan area compactness index reduces minority household housing consumption. This estimate is statistically significant. For white households, the compactness index has a negative but small statistically insignificant coefficient. Moving the average minority household from a high sprawl city (Atlanta) to a low sprawl city (Portland) would reduce its rooms consumption by -.52 = -.6658*log(126.1/57.7). These results support the hypothesis that sprawl encourages housing convergence. Why could this be? Housing is more affordable in high sprawl areas. Such areas are not erecting entry barriers and developers are building homes. Future work might study whether immigrant housing consumption in European cities is more comparable to natives in less compact cities.
Are commute times higher or lower in compact cities? In compact cities, people are likely to live closer to their downtown jobs but people are more likely to commute by relatively slow public transit. In a monocentric city, workers who commute by private vehicle are likely to slow each other down as they each impose congestion externalities on each other. In contrast, in sprawled metropolitan areas featuring multiple employment centers, workers commute by private vehicle at faster speeds (Gordon, Kumar and Richardson 1991, Crane 2000).
To begin to examine these issues, I use commute data from the 2003 American Household Survey. This data set reports the distance to work, and commute time for heads of households. In Table Two, I report summary statistics for workers in compact versus sprawled cities. Relative to workers in compact cities, workers in sprawled cities commute an extra 1.8 miles further each way but their commute is 4.3 minutes shorter. Over the course of a year (400 trips), they save 29 hours. While the workers living in sprawled cities have a longer commute measured in miles, they are commuting at higher speeds. Table Two shows that workers in sprawled cities commute at a speed 9.5 miles per hour faster than workers in compact cities.
The Neighborhood Change Database reports the share of census tract commuters who have a less than 25 minute commute by year. In Figure One, I graph this with respect to the census tract’s distance from the Central Business District (CBD). The figure shows that in both 1980 and 2000, the share of commuters with a short commute declines over the distance 0 to 10 miles from the CBD. Starting at the 11th mile from the CBD, the share of commuters with a short commute actually stops declining. This is strong evidence of the effect of sprawl. A large share of residents at such locations are not commuting downtown. Note the differential between the 1980 and the 2000 graphs. Over these twenty years, suburban households (i.e those living more than ten miles from the CBD) have experienced a large percentage increase in short commutes. For example, ten miles from the CBD between 1980 and 2000 there has been over a fifteen percentage point increase in the share of commutes with a commute of 25 minutes or less. This is strongly suggestive evidence of the commuting gains brought about by employment suburbanization. Employment sprawl has shortened commute times for suburban residents as such workers can commute faster over a shorter distance relative to if they worked downtown.
Additional Benefits of Sprawl
This section briefly highlights a variety of potentially important benefits of sprawl. Data limitations preclude presenting original data analysis measuring the size of each of these effects but I believe that each contributes to household well being in sprawled cities.
The Location of Employment Within the Metro Area
In the year 2000, only 21% of Atlanta’s jobs were located in zip codes within 10 kilometers of the CBD. In Boston, 52% of this area’s jobs were located within 10 kilometers of the CBD (Baum-Snow and Kahn 2005).
Firms gain by having the option of locating some of their employment further from the high land priced CBD. The key reasons for why firms choose particular locations include 1. land costs, 2, access to ideas, 3. access to workers and 4. transport cost savings for inputs and output. For example, manufacturing industries which are more land intensive are more likely to decentralize while skill intensive industries are less likely to decentralize (Glaeser and Kahn 2001). Those firms that gain from “Jane Jacobs” learning from other types of firms have an incentive to locate in diverse high density downtowns.
Within firms, non-management occupations are increasingly being sited at the edge of major cities (Rossi-Hansberg, Sarte and Owens 2005). This cost savings increases firm profits. Firms that are able to split their activities between headquarters and production plants are likely to gain greatly from sprawl. Standard agglomeration forces encourage firms to only keep those workers at the center city headquarters who benefit from interactions in the denser downtown (Rossi-Hansberg, Sarte and Owens 2005).
Other firms may gain by being able to construct large campuses where members of the firm can interact across divisions. Microsoft’s Richmond, Washington campus will be ten million square feet after it completes its expansion and there will be 12,000 workers there. Google now has 5,680 employees and is adding 1 million square feet to the 500,000 it now occupies in Mountain View, California.
There are at least two quality of life benefits from employment suburbanization. The previous section documented the reduction in commute times in suburban communities as more suburbanites now live closer to their jobs rather than commuting downtown. A second quality of life benefit from suburbanized employment is that this creates a type of separation of land uses. In the past, when cities where much more compact, millions of people lived too close to dirty, noisy manufacturing and slaughterhouse activity (Melosi 2001). Declining transportation costs have allowed a separation of where goods are produced and where people live.
Suburban Consumer Prices and the “Walmart” Effect
Walmart and other “superstores” could not exist in an urban world of compact cities with binding zoning laws. “Wal-Mart has sometimes had difficulty in receiving planning approval for its stores. Currently, Wal-Mart has either no presence or an extremely limited presence in New England, the New York metro area, California, and the Pacific Northwest. However, its expansion into new areas has proceeded over the past few years (Hausman and Leibtag 2005).”
These stores require large physical spaces and large parking lots to accommodate their inventory and to attract shoppers. Such stores offer one stop shopping and prices that can be 25% lower than regular supermarkets (see Hausman and Leibtag 2005). The diffusion of these stores may mean that the U.S consumer price index over-states inflation because this index does not properly reflect the prices that people face for core goods. These stores are disproportionately located in suburban and rural areas where land is cheap. Center city residents often drive to suburban locations to shop at such stores. While the popular media often reports stories critiquing Walmart’s employee compensation and its effects on driving out of business smaller “mom and pop” stores, it cannot be denied that consumers gain from having access to such stores. The key counter-factual here is what prices would residents face in a compact monocentric city without Walmart and other superstores?
Local Government Competition and Services and Taxes
Relative to a compact city, a sprawled metropolitan area is likely to have more political jurisdictions allowing households to have greater choice (Dye and McGuire 2000). Such political competition forces local jurisdictions to provide services such as garbage collection more efficiently per tax dollar spent relative to how they would provide services if they knew they were a monopolist. A central tenet of local public finance is that diverse households will gain if they can “vote with their feet” and seek out communities that offer the local services and taxes that meet their needs. Households willing to pay high taxes for good local schools will move to certain communities that households with no children would not consider. Many rich people are seeking out suburban communities both due to the housing stock, local public goods offered and the types of neighbors such communities attract.
A defining characteristic of cities in the United States is diversity. Such cities feature diversity with respect to ethnic groups and income inequality. The social capital literature has argued that an unintended consequence of the rise of ethnically diverse cities featuring significant income inequality is that people are less civically engaged (Costa and Kahn 2003, Alesina and LeFarina 2005). In such cities, part of the attraction of living in the suburbs may be the opportunity to self segregate into more homogenous communities within the greater metropolitan area.
It is important to note that spatial separation of different groups within the same metropolitan area reduces the likelihood of social interactions and this can have perverse consequences. In a sprawled city, if the heterogeneous population migrates and forms more homogenous communities with the poor in the center city and the wealthy in the suburbs, then bridging social capital across ethnic and income groups is less likely to take place. In this case, stereotypes can persist and collective action may be more challenging to achieve. In the past, when rich and poor clustered together in center cities, wealthy urbanites could not so easily escape the problems of their less fortunate neighbors. Pollution or disease spread easily and quickly from the tenements of the poor to the mansions of the rich. As a result, upper-bracket taxpayers were more likely to support policies that improved the living conditions of the worst off. For example, as Troesken (2004) points out, “In a world where blacks and whites lived in close proximity ‘sewers for everyone’ was an aesthetically sound strategy. Failing to install water and sewer mains in black neighborhoods increased the risk of diseases spreading from black neighborhoods to white ones.”
Today suburbanization has greatly increased the distance between the middle and upper middle class and the poor. Alesina and Glaeser (2004) investigate why Europe has more generous redistribution for the poor then the United States. They argue that part of the explanation is the fact that the United States has a more diverse population. Another possible cause is sprawl. If more U.S cities were more compact, would U.S tax payers be willing to redistribute more because they would have greater contact with the poor?
Does sprawl protect the suburban rich from crime? If criminals have less access to cars, then physical distance from the urban poor is likely to reduce the risk that the relatively wealthy face.
It is true that over the last decade center city crime has sharply decreased (Levitt 2005). While the causes of these quality of life gains continue to be debated, the consequences of this trend are clearly visible. Center cities will be better able to compete for the skilled (especially those with few children living in the household) against suburbs if the city is perceived to be safe. The reduction in urban crime will differentially increase quality of life in more compact cities such as San Francisco and New York City.
Compact cities do face greater risks from terrorist attacks. While only a small share of any city’s population is killed in even very large attacks such as 9/11/2001, people do tend to over-estimate the probability of unlikely events (Rabin 2002). Sprawled cities are also less attractive targets for terrorists (Glaeser and Shapiro 2002, Savitch 2005). It is no accident that the major terrorist attacks have taken place in dense cities such as at the World Trade Center, and the London bus bombs. A sprawled city offers the terrorists fewer causalities and thus less media coverage.
Sprawl and Urban Quality of Life
The previous sections have focused on individual subcomponents of urban quality of life. I have made no attempt to prioritize which dimensions of urban quality of life are most important to people. The economics literature on compensating differentials has attempted to answer this question. The theory of compensating differentials says that it will be more costly to live in “nicer” cities (Rosen 2002). This theory is really a “no arbitrage” result. If migration costs are low across urban areas and if potential buyers are fully informed about the differences in non-market attributes bundles then home prices and wages will adjust such that in nicer cities wages are lower and home prices are higher.
An enormous empirical literature has estimated cross-city hedonic price functions to estimate the implicit compensating differentials for non-market goods. In these studies, the dependent variable is the price of home I in city j in community m in year t. Define Xit as home I’s physical attributes in year t. Ajt represents city j’s attributes in year t. Given this notation, a standard real estate hedonic regression will take the form:
Priceijmt = 0 + 1*Xit + 2*Ajt + ijmt (1)
Multivariate regression estimates of this regression yield estimates of the compensating differentials for city level local public goods (based on 2). Intuitively, such estimates reveal how much higher are home prices for observationally identical homes in nice climate areas (i.e San Francisco) versus bad climate areas (i.e Houston).
In one prominent cross-city quality of life study, Gyourko and Tracy (1991) estimate equation (1) using 1980 data for 130 center cities. They use ordinary least squares estimates to construct a city quality of life index equal to 2*Ajt . In their empirical application, this A vector includes city attributes such as rainfall, cooling degree days, heating degree days, humidity, sunshine, wind speed, air pollution levels (measured by particulates), coastal access, cost of living, crime, student teacher ratio, insurance company ratings of the local fire department, hospital beds per-capita, taxes and population size. By estimating equation (1), Gyourko and Tracy provide index weights for the revealed relative importance of each of these factors in local quality of life. Intuitively, if a specific city attribute such as clean air is highly valued by people then cities with clean air should feature higher home prices and pay lower wages.
Their city quality of life rankings are useful for me because they allow me to study whether more compact cities have higher quality of life. The Gyourko and Tracy index can be used to rank center cities with respect to their quality of life from best to worst. I am able to merge the Gyourko and Tracy data and the Ewing, Pendall and Chen (2005) data (see Table One) for 47 of the metropolitan areas. In Figure Two, I graph their rankings of each city ranked from best (#1) to worst (#130) as a function of the metropolitan area’s sprawl. This figure allows me to examine whether there is objective evidence that quality of life is higher in more compact cities. Figure Two shows that there is no relationship between quality of life and city compactness for this subset of major metropolitan areas. Put differently, center city quality of life is not lower in more sprawled metropolitan areas. It is important to note that Gyourko and Tracy do not explicitly measure housing consumption or commute times by metropolitan area. Instead, they are focusing on non-market local public goods such as climate, street safety, and public services. Figure Two shows that these non-market services are neither better nor worse in more sprawled cities. An interesting extension of this research would examine cities over time. In sprawling cities, do we see urban quality of life declining? As shown in Figure Two, at a point in time across cities there is little evidence supporting this hypothesis. In the next section, I will present some evidence that air pollution has not grown worse in growing cities.
Some of the Local Environmental Costs of Sprawl are Declining
Sprawl’s opponents are likely to concede that the “American Dream” offers private benefits. They would counter that suburbanization imposes important social costs that no one household has an incentive to internalize. This section seeks to examine some of these environmental costs.
Environmentalists often argue that sprawl contributes to a large ecological footprint because people consume more resources when they live at low density. Table Two presents some evidence supporting this claim. The 2001 National Household Transportation Survey reports for each household how much gasoline they consume each year. Merging the city compactness index (see Table One) to this data, I examine gasoline consumption in compact and sprawled cities. As shown in Table Two, the average resident in compact cities consumes 335 gallons less per year of gasoline than the average resident of sprawled metropolitan areas. Within metropolitan areas, suburban drivers drive over 30% more miles than center city residents and are more likely to drive low fuel economy SUVs (Kahn 2000, 2006). The average Atlanta household would drive 25 percent fewer miles if it relocated to relatively compact Boston (Bento et. al. 2005). As a result, there are significant differences in average gasoline consumption across the country. Cross-national studies suggest that gasoline consumption could be 20 percent to 30 percent lower in sprawling cities like Houston and Phoenix if their urban structure more closely resembled that of Boston or Washington, DC.
People are less likely to use public transit when they live in sprawled cities. This has environmental implications because public transit is a “greener” transport technology than private vehicles. To document this fact, I use census tract data from the Urban Institute and Census Geolytics’ Neighborhood Change Database. This is a set of repeated cross sections from the 1970, 1980, 1990 and 2000 decennial censuses at the census tract level normalized to 2000 tract geography. Census tracts are areas of roughly 4,000 people. Using GIS software, I calculate each census tract’s distance to the CBD and focus on those census tracts within 25 miles of the CBD for the metropolitan areas listed in Table One. As shown in the bottom two rows of Table Two, in 1970 6.8% of workers in sprawled metropolitan areas and 24.6% of workers in compact metropolitan areas commuted using public transit. In both areas, these shares shrank between 1970 and 2000. In the year 2000, 2.8% of workers in sprawled metropolitan areas and 17.1% of workers in compact metropolitan areas commuted using public transit.
Income growth plays some role in explaining this trend. As household incomes increase, people are less likely to use public transit, which is typically slower than commuting by car. Car travel takes about two minutes per mile for commutes under five miles. In contrast, bus commuting takes more than three minutes per mile for commutes under five miles. In addition, the average bus commuter waits 19 minutes to board the bus. Using data from the 2000 Census of Population and Housing, I find that the probability of using public transit is 2.5 percentage points lower for a household at the 75th percentile of the income distribution ($65,339) than for a household at the 25th percentile ($41,159).
However, sprawl also helps explain the decline in public transit ridership. Based on the same data, I find that simulating sprawl by moving a person from the 75th percentile of the population density distribution (2,528 people per square mile) to the 25th percentile (142 people per square mile) reduces public transit use by 8.6 percentage points.
Baum-Snow and Kahn (2005) examine public transit use trends in sixteen major United States cities that have spent billions of dollars constructing new light rail and heavy rail lines between 1970 and 2000. They study whether the share of workers who commute using public transit increases in communities that have increased access to rail transit because they now live close to a new rail line. While they find some evidence of increased usage (especially in more compact cities such as Washington D.C and Boston), the observed “treatment” effects are small. New rail transit expansions are unlikely to encourage mode switching from vehicles to public transit. To reduce the ecological footprint impacts of private vehicle use, induced innovation is needed to encourage producers to develop high fuel efficient vehicles and for consumers to demand such vehicles. Expectations of high future gas prices would play a key role in providing incentives for such products to be demanded.
A standard argument that environmentalists make about sprawl is that this trend contributes to urban air pollution. But, new vehicle emissions regulation has offset increased vehicle mileage. The Los Angeles Basin suffers from the highest levels of air pollution in the United States, with the pollution caused mainly by vehicle emissions. But Los Angeles has made dramatic progress on air pollution over the last 25 years. For ambient ozone, a leading indicator of smog, the average of the top 30 daily peak one-hour readings across the county’s 9 continuously operated monitoring stations declined 55% from 0.21 to 0.095 parts per million between 1980 and 2002. The number of days per year exceeding the federal one-hour ozone standard declined by an even larger amount—from about 150 days per year at the worst locations during the early 1980s, down to 20 to 30 days per year today.
Recent pollution gains are especially notable because Los Angeles County’s population grew by 29 percent between 1980 and 2000, while total automobile mileage grew by 70 percent (Census of Population and Housing 1980 and 2000; California Department of Transportation 2003). For air quality to improve as total vehicle mileage increases indicates that emissions per mile of driving must be declining sharply over time.
To document this fact, I use two waves of the California Random Roadside Emissions tests spanning the years 1997 to 2002 to estimate vehicle level emissions production functions (see Kahn and Schwartz 2006). Intuitively, I control for a number of vehicle characteristics such as the vehicle’s mileage, and the zip code of the vehicle owner. Holding these factors constant, I estimate how vehicle emissions vary as a function of vehicle model year. How much cleaner are 1990 makes relative to 1980 and 1975 makes?
In Figure Three, I present predicted vehicle emissions by model year holding all vehicle attributes at their sample means. For each of the three pollutant measures I normalize the predictions by dividing through by the predicted value for 1966 model year vehicles. The Figure shows sharp improvement with respect to model year and documents emissions progress even during years when new vehicle regulation did not tighten.
The vehicle emissions progress by model year means that the average vehicle on the road in any calendar year is becoming greener over time. In each subsequent calendar year, there are fewer high emitting pre-1975 model year makes on the roads. This greening of the average vehicles has greatly contributed to the reduction in ambient pollution despite ongoing city growth and increased vehicle mileage. To document this, I use ambient air pollution data from California Ambient Air Quality Data CD, 1980-2002 (California Air Resources Board). This CD-ROM provides all air quality readings taken in the state during this time period. In Figure Four, I graph the percent change in ambient ozone smog for 29 major California counties over the years 1980 to 2000 with respect to county percent population growth. I include data for the 29 California counties that had population levels greater than 200,000. Ambient ozone by county/year is measured by the maximum one hour reading at each monitoring station within the county and then I average these maximum readings by county in each year.
Anti-sprawl advocates would argue that counties experiencing greater population growth should experience rising ambient air pollution. As shown in this figure, there is no correlation between county growth and ambient air pollution. The correlation equals -.08. These major counties, even those such as Riverside that have experienced the greatest growth, have enjoyed large pollution reductions over this time period. The vehicle pollution progress documented in Figure Three has helped to offset the scale effects of California’s population growth.
In addition to greenhouse gases and ambient air pollution, a third environmental concern often voiced by sprawl opponents is the conversion of farm land. Farmers provide green space. Such green space is privatized when farmers sell their land to suburban developers. If nearby households value the open space, then farmers impose a negative externality on existing urban and suburban residents when they sell to a developer. Fortunately, new markets in land development rights have helped mitigate this problem.
Throughout the United States, municipalities are purchasing open space around their borders to guarantee that the land is not developed. For example, the city of Boulder, Colorado, has earmarked a 0.73 percent sales tax to fund the purchase of 25,000 acres to establish a greenbelt around the city. It has also set aside 8,000 acres in the Boulder foothills to be used as parks. Some of the Boulder open space is leased to farmers and remains in agricultural use. Other parcels are maintained as natural areas. This allows residents to enjoy recreational activities such as walking, bicycling, and horseback riding. In the Seattle metropolitan area, King County has adopted a different strategy with a similar goal. Drawing upon a $50 million bond issue, the county is purchasing development rights from farmers. Farmers gain an increase in their income and in return they promise not to convert their “green space” into suburbia (see Kahn 2006).
Such government initiatives solve a free rider problem. In the absence of government intervention, environmental organizations such as land trusts might go door to door, asking people to contribute money to help preserve open space. But few people are likely to give under these conditions. The “win-win” for any one household is to contribute nothing to such programs and let everyone else underwrite their cost. As a result, too little money is invested in protecting local public goods. Government’s unique ability to collect taxes and allocate revenue solves this problem. However, not all governments can take this approach: like many green policies, “open space” initiatives are more likely to succeed as local incomes rise. After studying voting patterns for all open space referenda in the United States between 1998 and 2003, Kotchen and Powers (2005) found that richer jurisdictions and jurisdictions with more homeowners were more likely to vote to hold such ballot initiatives and to enact them. Nearly 1,000 jurisdictions had open space referenda and nearly 80 percent passed. From an ecological perspective, the key issue here is whether jurisdictions hire competent ecologists who can prioritize what are the most valuable pieces of open space to purchase and protect.
Compact cities featuring all employment located in the Central Business District limit economic opportunities. There is significant diversity of types of people and types of firms. Firms that need large parcels of land to operate and people who have a strong preference for their own large private plots of land face significant tradeoffs if they must locate in compact cities. Sprawled cities offer both firms and households more choices. How much such economic actors gain from these additional choices is a complex function that merits future research. This paper has presented original evidence on some of the relevant margins. Similar to the United States, Europe faces an increase in the diversity of its workforce as immigration changes urban demographic patterns. Would sprawled cities offer greater opportunities for these newcomers?
This paper has attempted to present a balanced analysis of the private benefits and social costs both of compact and sprawled cities. Compact cities feature greater congestion and higher commute times while in sprawled cities certain global environmental externalities such as greenhouse gas production are likely to be exacerbated. Technological advance has mitigated many of the environmental problems associated with sprawl.
Today the diversity of major cities within the United States offers households a wide menu to choose from. People with a taste for “new urban” living can move to a New York City while those that want their own private space can move to Houston, Texas.
Do European cities feature “too little” sprawl? As documented in this paper, an unintended consequence of urban compactness is that the diversity of choices for consumers and firms is shrunk. How much would these economic agents value increased choice?