The High Cost of Smart Growth

Over the past several years, many skeptics of smart growth policies and other restrictions on development and land use have argued that such regulations tend to harm those with moderate incomes by raising house prices and apartment rents. Such growth control regulations in common use today include impact fees, oversized lots, growth boundaries, and mandated amenities--all of which raise costs and restrict supply.¹

In response to such criticism, many smart growth advocates either deny that their policies are costly or contend that the cost increases are minor and are offset by other benefits. At the same time, however, a small but growing faction of growth control advocates endorse the prospect of higher prices as a necessary cost of achieving elitist goals, as did the California planner who recently argued:

If a community or region refuses to grow, the result may be higher prices, economic displacement and hardship, and dangerous crowded housing in exchange for keeping a desirable quality of life for the "already landed" middle and upper income groups. If the local voters are willing to pay this price, why should planners try to prevent it?²

Despite denials by smart growth advocates, a growing body of academic evidence demonstrates that lower-income households seeking home ownership bear a disproportionately high amount of smart growth costs. Over the past few years, analysts at The Heritage Foundation, the Brookings Institution, the National Center for Public Policy Research, Harvard University, Tufts University, and the National Bureau of Economic Research have published reports concluding or acknowledging that the more commonly applied smart growth regulations imperil home ownership opportunities or that the absence of such restrictions improves the prospects for minority home ownership.³

Environmentalists Counterattack

As the evidence of disproportionate harm to the less well-off builds, advocates of growth controls and their allies in the environmental movement and the transit industry have been searching for ways to challenge the claim that growth controls increase housing costs and hurt the poor. Typical of the defensive effort getting underway is a July 2003 report by the STPP that, notwithstanding all evidence to the contrary, attempts to link the existence of sprawl to burdensome automobile costs that in turn absorb a share of a family's income that could otherwise be devoted to buying a house.⁴

In other words, the STPP claims that sprawl reduces home ownership. This hypothesized relationship between cars and houses leads the experts at the STPP to urge the "working poor" to reject auto ownership and turn to public transportation or bicycles so that they can afford to become home owners. The STPP also recommends that government spend more on public transportation and bicycles to give the poor more transportation choices.

Cars are apparently acceptable for the better off, and the STPP acknowledges that "For middle- and upper-income families, the cost of transportation is taken for granted."⁵ The poor, however, are different: "But for the poorest American families the high cost of owning and maintaining a car may put homeownership out of reach."⁶

The STPP report offers no meaningful evidence to support this claim. Instead, it relies on the reader's intuition to accept that a cost connection exists between cars and homes. Indeed, what little data the STPP does provide prove nothing about the postulated relationship between homes and cars, and the report completely ignores the more firmly established positive relationships between car ownership, employment, and economic opportunity.

Instead, the report focuses only on costs, not outcomes or benefits. For example, the STPP notes that the working poor who drive spend 21 percent of their income on commuting, while "the working poor who were able to take public transportation, bicycle, car pool, or walk to work spent far less."⁷ But did more of the latter own their own homes than the former? On this, the STPP is silent. The STPP also cites a study that claims an "inverse relationship between increasing car and truck ownership and diminishing family savings."⁸ But if the two events are connected, what does this mean for home ownership? Again, the STPP fails to make the connection, most likely because there is none. Since World War II, both home ownership and vehicle ownership have soared--an accomplishment the STPP analysis implies could not have happened.

But it did happen, thanks to rising prosperity. For the most part, the growth in ownership of both types of valuable assets reflects the extra mobility and freedom that automobile ownership allows individuals, often giving them access to better jobs and the means to buy a house and a car. In other words, the extra benefits of car ownership offset the extra costs. Although the STPP denies the existence of such a relationship, it is largely alone in this belief. Most other social welfare advocates on both sides of the political spectrum believe that access to an automobile is essential to economic advancement.

The Job Benefit in Auto Mobility

For example, President Bill Clinton proposed relaxing auto ownership restrictions in food stamp eligibility requirements "so that people can access reliable transportation to get to work without sacrificing their food stamp benefits."⁹ An October 2000 press release from the Clinton Administration in support of the eligibility change noted: "One national study found that twice as many welfare recipients with cars were working than those without cars, and 25 percent more low-income families with cars were working than those without cars."¹⁰

More recently, a May 2003 Brookings Institution survey found:

Most welfare recipients do not have access to a dependable automobile, and research indicates that lack of access to an automobile is one of the most prevalent barriers to employment. Research further indicates that car ownership improves the likelihood that low-income people will get and keep work, and improves access to better jobs.¹¹

According to another recent Brookings study on transportation policies for the working poor:

In recent years, new sources of federal funds have helped agencies initiate transit services aimed at moving low-income adults into the labor market. By contrast, policymakers have paid far less attention to increasing automobile access among the poor. Given the strong connection between cars and employment outcomes, auto ownership programs may be one of the more promising options and one worthy of expansion.¹²

An earlier study on a similar subject by the Progressive Policy Institute, a think tank affiliated with the Democratic Leadership Council, noted:

In most cases, the shortest distance between a poor person and a job is along a line driven in a car. Prosperity in America has always been strongly related to mobility and poor people work hard for access to opportunities. For both the rural and inner-city poor, access means being able to reach the prosperous suburbs of our booming metropolitan economies, and mobility means having the private automobile necessary for the trip. The most important response to the policy challenge of job access for those leaving welfare is the continued and expanded use of cars by low-income workers. Across the country, state and local decision-makers are inventing new programs to do just that and devising new ways that public funds can help.¹³

Since transit service is so much slower than cars and is focused principally in the core and central business districts of major metropolitan areas, people who use transit because they do not have a car face limited mobility and diminished job prospects. Much of a given metropolitan area simply cannot be accessed by transit. This imposes a significant economic burden according to Steven Raphael and Michael Stoll of the University of California, whose research indicates that raising the rate of automobile ownership among African-Americans to the rate found among non-Hispanic Whites would eliminate 45 percent of the unemployment gap between the two ethnic groups. Similarly, if the Hispanic automobile ownership rate were increased to that of non-Hispanic Whites, 17 percent of the differential would be erased.¹⁴

It is easy to see why cars can help close the unemployment gap. At average transit operating speeds of 15 miles per hour, a maximum "job shed" of 175 square miles can be accessed in 30 minutes. In reality, however, the actual job shed would be much less because of the necessity of transfers and limited service areas. On the other hand, a person with an automobile can expect to average 30 miles per hour and reach a job shed of 700 square miles in 30 minutes. This vastly increases employment and other opportunities, offering a better quality of life.

In spite of the long-standing bipartisan support for improving low-income access to automobiles and the growing body of evidence indicating that an automobile offers the working poor much greater access to better jobs, the STPP has maintained that such access should be discouraged. In June 2002, the STPP took its case to Congress when STPP board member and former director Hank Dittmar expressed his opposition to greater auto mobility for the poor in testimony before a Senate committee:

It is indeed ironic that many social scientists believe that the best way to help former welfare recipients secure jobs is to give them automobile purchase assistance, thereby trapping them into the poverty cycle even more profoundly, as the poor typically end up with less reliable cars which are more expensive to operate and maintain.¹⁵

Recognizing that effective advocacy in Washington requires that one offer a counter-solution as an alternative to the one you oppose, the STPP revealed what it thought government should be doing to accommodate the mobility needs of low-income families. In a letter to Senator James Jeffords (I-VT), the STPP proposed that government create and fund "Bike Purchase Programs and Car Sharing Programs to help low wage earners secure reliable and affordable transportation options."¹⁶ One can only imagine the ridicule that some Senators would have heaped upon such a "let them eat cake" proposal if it had come from a fiscally conservative organization rather than one supported by environmentalists, unions, and transit system contractors.

End Sprawl, Use Transit?

Although the STPP's postulated relationship between cars and homes has no basis in fact and is at variance with the findings of most social scientists regardless of where they stand on the ideological spectrum, the STPP has persisted in its efforts to keep the poor on the bus. In addition to advocating bike and bus subsidies, the STPP recommends changes in land use and residential development patterns in order to push people onto transit as a way to resolve the nonexistent relationship between transportation costs and home ownership.

As the STPP sees the world and its deficiencies, the suburbanization of America is the heart of the problem, which it believes has led to increased transportation costs as people living in "sprawling" suburbs make greater use of their cars than they would have if they had remained in inner-city neighborhoods with greater access to public transportation. In a letter to Congress's Millennial Housing Commission, the STPP argued that its earlier "report, Driven to Spend, found that residents of more sprawling metro areas tend to spend a much higher proportion of their family budget than residents of more compact, traditional metro areas with good public transportation service."¹⁷

That statement, however, has no basis in fact. In fact, data presented later in the current report demonstrate that the relationship between degrees of sprawl and household transportation spending is exactly the opposite of what the STPP claims--using the same data on the same set of metropolitan areas as the STPP used in its report.

In its most recent report of July 2003, the STPP hedges its bets by noting that "the sample size is too small to allow a rigorous statistical analysis," but then adds that "a quick glance at the list of metro areas shows that in many sprawling metro areas, families spend a much larger portion of their household budget on transportation than in more compact, transit- or pedestrian-oriented areas."¹⁸ Yet a lingering glance over the STPP's data would reveal precisely the opposite, as presented in Table 1, which categorizes the STPP cities by degrees of sprawl.

Measuring Sprawl

There is no precise definition of "sprawl." In this analysis, sprawl is defined by population density, or the number of inhabitants per square mile of land in the area. The more sprawling a community is (i.e., the greater the share of single-family detached houses on large lots, few or no apartments and town homes, and shopping centers surrounded by vast parking lots), the lower will be its population density. Conversely, traditional cities and older suburbs with their smaller lots and preponderance of multi-family housing would have much higher population densities. For example, Manhattan's population density is 70,000 per square mile,¹⁹ and Hong Kong's is 83,000, while Atlanta's urbanized area has only 1,800 people per square mile.

This measure of sprawl would also seem to conform with the STPP's definition: The STPP report contrasts the differences between "sprawling" cities and those that are "compact," meaning more dense. Using people per square mile as the measure of sprawl, the 26 metropolitan areas cited by the STPP can be arranged in quintiles by the degree of sprawl.

The first column of Table 1 shows these measures for each of the five degrees of sprawl, with the "most sprawled" metropolitan areas²⁰ averaging a density of 2,146 per square mile (Atlanta, Boston) and the "least sprawled" regions having a density of 5,379 people per square mile (San Francisco and New York City). Table A in the Appendix ranks these 26 metropolitan areas by density and by the five degrees of sprawl: least sprawling, less sprawling, middle, more sprawling, and most sprawling. Each quintile includes five metropolitan areas, except for the middle quintile, which contains six.

The Advantages of Sprawl

Ranking areas by their degree of sprawl allows more precise inferences about the influence of sprawl on transportation costs and allows the validity of the STPP's assertion about the relationship between the two to be assessed. This sprawl index can also be used to determine the influence of sprawl on other important measures of well-being that are believed to be adversely affected by sprawl. For example, if sprawl raises transportation costs--as the STPP contends--and higher transportation costs limit home ownership, then the data should reveal a negative relationship between sprawl and home ownership. But the data presented in this report show that no such relationship exists in the cities listed by the STPP. Indeed, transportation costs, housing prices, home ownership, and traffic congestion are all more favorable in communities with more sprawl than in communities with less sprawl.

Household expenditures are lower where sprawl is greater. The STPP argues that sprawl is associated with higher consumer expenditures for transportation and housing. In fact, the opposite is true. When "food at home" is added to the equation, the disparity becomes even greater. Indeed, the lowest transportation expenditures per household were in the "most sprawling" areas ($7,249) and were $915 less than households spend in the "least sprawling" areas.²¹ (See Table 1.)

The least sprawling areas had transportation expenditures of $8,164, or 13 percent above the most sprawling areas. Similarly, the lowest transportation and housing expenditures were in the most sprawling areas, at $20,334. The highest transportation and housing expenditures were in the least sprawling areas, at $25,552, and were 26 percent above the most sprawling areas. Finally, when food-at-home expenditures are added, the lowest figure is also in the most sprawling areas: $23,363 per capita. The highest figure is in the least sprawling areas, at $29,045, or 24 percent above the most sprawling areas. (See Figure 1.)

Housing affordability is greater where sprawl is greater
Another measure of the American dream is the affordability of home ownership, as measured by the relationship between housing prices in a region and the income level of that region. As the data reveal, housing is by far the least affordable in the least sprawling areas. Median house values are highest in the least sprawling areas, at $213,100, or 82 percent above the $116,840 of more sprawling areas.²² (See Table 2.)

The less sprawling areas have the highest median household incomes, although the least sprawling areas rank nearly as high, at $49,688 per household. This is 11 percent above the lowest value, $44,899, in the more sprawling areas. The median house value in the least sprawling areas is 4.29 times the corresponding median household income. Similarly, in the less sprawling areas, median home values are also high, at 3.44 times income. The most affordable housing relative to incomes is in the more sprawling areas, where median house values are 2.60 times median incomes. The middle and most sprawling areas have similar house affordability ratings, of 2.68 and 2.72, respectively. (See Figure 2.) In other words, the greater the sprawl, the higher the share of the local population that can afford to become home owners.

Home ownership is higher where sprawl is greater
If housing prices are more affordable in sprawling areas, then one might expect home ownership rates to be higher in more sprawled areas--precisely the opposite of STPP's hypothesized relationship between home ownership, cars, and urban sprawl. Contrary to the STPP's assertions, however, the lowest home ownership rate is in the least sprawling areas, while the highest rates are in the more and most sprawling categories.²³ (See Table 3 and Figure 3.)

As the fourth and fifth columns of Table 3 show, home ownership rates tend to be lower for minorities due to their generally lower incomes. Minority home ownership is thus an indicator of lower-income home ownership rates. The least sprawling areas have slightly lower than average African-American and Hispanic home ownership rates. The highest African-American home ownership rates are in the middle sprawl areas, while the highest Hispanic home ownership rates are in the more sprawl areas. The lowest minority home ownership rates are in the most sprawling areas. However, if transit-rich Boston is excluded from the most sprawl category, the African-American home ownership rate rises to the second rank, while the Hispanic home ownership rate becomes third.²⁴

For example, a breakdown of the home ownership patterns by income class in Washington, D.C. (a sprawling metropolitan area), reveals a striking relationship between sprawl, income, and home ownership. As noted earlier, one attraction of outlying communities is lower land prices than are found in communities closer to the metropolitan center. Lower land prices, in turn, contribute to lower housing costs. As a result, moderate-income households and entry-level buyers typically seek their home ownership opportunities in outlying communities because comparable housing in the closer-in communities is often unaffordable.

Table 4 illustrates this tendency in the Northern Virginia suburbs of Washington, D.C. In the inner suburbs, where land and housing costs are high, households with incomes below $35,000 represent only a tiny fraction of the home owners, whereas as many as 50 percent to 70 percent of the home owners in the outer suburbs--where land and homes are cheaper--have incomes of $35,000 or less.²⁵

Transportation is better where sprawl is greater
Residents of the most sprawling areas also face fewer transportation problems, according to data from the U.S. Census and the Texas Transportation Institute. The shortest average one-way work trips (24.2 minutes) occur in the more sprawling areas, while the longest trips (29.5 minutes) occur in the least sprawling areas.²⁶ (See Table 5 and Figure 4.) Commuters in the least sprawling areas spend nearly one workweek (39.6 hours) more traveling to work each year than is spent by commuters who live in the more sprawling areas.²⁷ Similarly, the roads are most congested in the least sprawling areas and least congested in the most sprawling areas.

The Texas Transportation Institute's Roadway Congestion Index averages 1.34 for the least sprawling areas, indicating that roadways operate at 34 percent above their rated capacity (as measured by trucks and autos per lane mile) during peak periods. In the most sprawling urban areas, roadways operate at an average congestion index of 1.07, or 7 percent above capacity.

These relationships between sprawl and congestion also undermine another important STPP contention: the value of transit service in relieving congestion. As the next section demonstrates, transit service levels are highest in the least sprawling but more congested areas, with 1,866 place miles per capita annually.²⁸ In contrast, the more sprawling areas have the lowest transit service levels, at an average of 685 place miles per capita annually.

The Cost of More Transit

Higher Consumer Expenditures, Less Affordable Housing, and Worse Traffic
The 26 metropolitan areas were also ranked in quintiles based on place miles of transit service per capita--a measure of transit intensity or service levels. Higher levels of transit service tend to be associated with the opposite phenomena that the STPP implies. While transportation expenditures tend to be higher in regions where transit service is least, housing and food-at-home expenditures tend to be much lower, offsetting whatever cost benefit consumers derive from subsidized transit systems. (See Table 6.)

The result is that the lowest overall household expenditures are in the least intense transit urban areas ($24,451). The most intense transit urban areas have overall expenditures of $28,652--over 17 percent higher than the areas with the least transit service. Although total transportation expenditures are lower in transit-intensive places, this may be more a function of fewer people owning cars, or households owning fewer cars, than a reflection of transit representing a low-cost option. Indeed, transit fares are nearly as high as the full cost of driving. In 1999, according to data collected by the federal government, total expenditures on personal vehicles (purchase, financing, and operating costs) were approximately $0.20 per passenger mile.²⁹ The same year, average transit fares per passenger mile were nearly $0.18, and total rider and taxpayer expenditures were $0.66 per passenger mile--more than three times the cost for automobiles.³⁰

It is a sad commentary on the evolution of transit costs that it is now much less expensive to move people by car than it is to move them by transit. Low-income households that rely on transit would be far better served if their fares and part of their transit subsidies were used instead to purchase the same type of mobility that 90 percent of American households already have--a car.

Home ownership becomes higher as transit service levels decrease, with a greater difference among African-American and Hispanic households. (See Table 7.) African-American households in areas with the least transit service are 17 percent more likely to be home owners than are African-American households in areas with the most transit service. (See Figure 5.)

Similarly, Hispanic households are 40 percent more likely to be home owners in areas with the least transit service than are Hispanic households in areas with the most transit service. (See Figure 6.)

Generally, housing affordability tends to be greater where there is less transit service. (See Table 7.) Indeed, New York City, with by far the most extensive transit service in the nation, ranks second worst in both African-American and Hispanic home ownership. Only Boston, also one of the leading transit urban areas, ranks worse. Traffic congestion is also less of a factor where there is less transit service. (See Table 8.) Average one-way work trip travel times are the lowest in the least intense transit areas, where commuters spend an average of 45.6 hours less traveling to work each year than is spent by commuters in the most intense transit service areas.

These findings are contrary to the relationship between transit and home ownership that is postulated by the STPP and therefore cast doubt on the wisdom of its recommendation that government should subsidize expanded transit service to assist lower-income and minority households to become home owners.

Consumer Expenditures, Costs, and Quality of Life
Although none of the STPP's implied relationships is consistent with the available data, the basic presumption that a certain level of transportation expenditures is somehow "forced" on helpless families by regional land use patterns and lack of transportation choices is itself a highly dubious proposition. Equally dubious is the implication that land use patterns and commuting preferences have a significant influence on household transportation costs.

Individuals in a modern, prosperous economy buy products and services for reasons that often have less to with the specific utility of a product or service and more to do with the pleasure and value their use provides. A person can healthfully and inexpensively sustain life by eating only green pea soup flavored with a little chicken, carrots, and onions for a cost of around 69 cents a serving. Yet most individuals spend much more than this and eat less healthfully (e.g., steak, onion rings, and mashed potatoes with butter) for reasons that have little do with the food's nutritional value or the need to sustain life.

In the same way, households could dramatically reduce their transportation costs by limiting car purchase and operation to pre-owned, fuel-efficient, compact cars. Some do, but many want something new and sporty, big and gas guzzling, or imported and expensive because those factors, which have little to do with cost-effective mobility, give them pleasure.

More practical, however, are the many suburban households that choose to spend more on transportation and save much more in housing costs, and thereby buy even more housing services than would otherwise be possible, in contrast to the STPP's contention. In most metropolitan areas, land and housing prices on the urban fringe tend to be much lower than those closer to the center of the metropolitan area. These houses on the distant fringe tend to be the only ones that individuals of moderate incomes can afford, as Table 4 illustrates for the Virginia suburbs of Washington, D.C. Thus, rather than being the burdensome expense that the STPP asserts, auto ownership opens up vast opportunities for prospective home owners of modest means to reduce housing costs.

Finally, as the STPP's own numbers reveal, only a fraction of the typical household's annual transportation costs is a result of commuting or driving. Three-fourths of the cost of car ownership is essentially fixed overhead (e.g., purchase price, insurance, license, and interest on the car loan), while the other 25 percent (e.g., repairs, gas, oil, and maintenance) is accounted for by actual operations. Thus, if someone limits his driving to non-work purposes, which may account for half of all driving, transportation costs shrink by only 12 percent, a portion of which would be offset by money that must then be spent on transit fare.

Indeed, the STPP notes that the annual savings to commuters using transit versus driving is only $515 per year. With the average household in the 26 metropolitan areas spending an estimated $7,633 per year on total transportation services, the transit saving amounts to less than 7 percent of this. It is for this modest sum that the STPP is asking the working poor to give up automobiles altogether.

Conclusion

The lessons for Congress from the STPP's misplaced effort and revealing data are that automobiles continue to gain market share, are the preferred and most cost-effective form of mobility, and have an important role to play in helping low-income families gain financial independence. Despite these obvious advantages, the federal transportation program has suffered from an advanced case of mission creep over the past several decades and now spends tens of billions of dollars each year on marginal projects that benefit only a small fraction of the traveling public.

In a class by itself, however, are the STPP's recommendations that government fund bicycle purchases, car-sharing programs, and more transit to help the poor. As Congress crafts a new federal transportation program, there is a risk that some Members may actually believe this advice and divert scarce federal financial resources to counterproductive programs that both waste money and hurt the poor. Hopefully, those believers will be in the minority.

Whether it is money for the renovation of covered bridges, thousands of pork-barrel projects, buses and trolleys, historic preservation, hiking and biking trails, maglev research, or highway beautification programs, all of these diversions are funded by the fuel taxes paid by motorists who suffer from worsening congestion because of these diversions. This year's reauthorization of the transportation program is a good place to start restoring balance to the program and begin funding projects that people actually want to use, such as more road capacity.

Wendell Cox, Principal of the Wendell Cox Consultancy in St. Louis, Missouri, is a Visiting Fellow at The Heritage Foundation, and Ronald D. Utt, Ph.D., is Herbert and Joyce Morgan Senior Research Fellow in the Thomas A. Roe Institute for Economic Policy Studies at The Heritage Foundation.

Appendix