Transit alone cannot solve the systemic problems behind job inaccessibility

cleveland job sprawl 2000-2012
bus stop at route 237 & Eastland Road

An actual stop for the #86 bus alongside Route 237 in Cleveland (courtesy of Google Maps).

In my last post, I dug into the debate over which approach is better – bringing people to jobs or jobs to people – by examining case studies from San Francisco and Minneapolis-St. Paul. But, as I noted, the studies I examined assume that the relationship between transit and job accessibility occurs in a vacuum. Things are not that clean in the real world.

Furthermore, one can argue – as Peter Truog and others have – that, while it may make sense to invest in transit-oriented development (TOD) and the “jobs to people” approach on paper, it is impractical. Planning and building fixed transit assets can take a decade or more. Encouraging businesses to relocate along existing transit lines and form new, transit-rich clusters also takes time.

In the meantime, people of color in high-poverty neighborhoods will continue to struggle to access jobs in far-flung suburbs. It must be better then to connect these neighborhoods with existing low-skill, entry level jobs with buses in order to give them a leg up and not keep them locked out of the job market.

There is merit to this argument. Shifting transit systems from hub-and-spoke (radial) networks focused on the central business district (CBD) to a grid system focused on multiple destinations is one effective alternative. The success of Houston’s redesign bears this out.

Research supports the benefits of moving away transit systems away from a radial service orientation to a multi-destination one, as well. Jeffrey Brown and Gregory Thompson, researchers at Florida State, have published several studies illustrating this point. In a 2008 paper (paywall), they found that metropolitan statistical areas (MSAs) with multi-destination transit systems outperformed those with radial systems for ridership, service productivity, and cost-effectiveness.

With this in mind, I want to examine the economics – both financial and political – of connecting people to jobs with transit in Northeast Ohio. Ultimately, when we explore these broader factors, it becomes clear that, while busing a low-income worker from Central to a manufacturer in Solon may help that individual, it remains insufficient to the problem at hand.

Are we bringing people to jobs or just chasing jobs with transit?

While improving system performance is an absolute good that all transit agencies should pursue, it does not necessarily mean that a city will see improvements in job accessibility for low-income people of color.

Moreover, system redesigns are useful, but transit systems still require minimum levels of population and job density in order to deliver high-quality service and remain cost-effective. According to Robert Cervero and Erick Guerra, job density is more important than residential density. These minimum thresholds of job density simply do not exist in most metros any more, particularly Cleveland.

In a 2009 paper, Elizabeth Kneebone from Brookings traced the outward shift of jobs in American cities. She examined the location of jobs in the 98 largest MSAs from 1998-2006, focusing on three bands around the CBD: a 3-mile radius for the central city, a 10-mile radius for the “beltway,” and a 35-mile radius to define the metro’s sprawl area.

Her analysis showed that just 21% of employees work within 3 miles of downtown, while a plurality (more than 45%) work between 10 and 35 miles away from the CBD. Additionally, she found that jobs were growing fastest in these outer areas; jobs at least 10 miles from the CBD grew 2.5 times faster than those in the beltway ring and 57 times faster than those in the central city.

We may have heard much since the end of the Great Recession about the revival of cities and the shift away from suburbia, but this shift has not occurred for jobs.

Another 2015 report from Kneebone and colleague Natalie Holmes makes clear that the suburbanization of people and jobs quickened during the 2000s. As a result, “by 2010 in the nation’s largest metro areas, the majority of every major ethnic and racial group and the majority of the poor lived in suburbs for the first time.”

The rate of outward growth was faster for jobs than people, though. Kneebone and Holmes concluded that the number of jobs within a typical commute for a resident of 96 major MSAs fell by 7% from 2000 to 2012. Nowhere was this change greater than in Cleveland, which saw a 26.5% decrease in the number of jobs available within a normal commute.

Given these trends and the realities facing GCRTA, trying to bring low-income workers from neighborhoods like Central to employment centers like Solon would be less about bringing “people to jobs” and more about chasing jobs with transit. What’s the likelihood that a new bus route to a suburban job campus will still be useful in 3 to 5 years?

cleveland job sprawl 2000-2012

No city saw a larger decrease in the number of accessible jobs from 2000 to 2012 than Cleveland (courtesy of The Brookings Institution).

Whither the job hubs?

And even if you know the jobs would remain stationary in the short-term, which locations should GCRTA prioritize?

Jobs have sprawled to such an extent in Northeast Ohio that few actual job hubs remain. NOAcA has found that fewer than one-quarter of all jobs in Northeast Ohio are concentrated in the region’s 6 largest job hubs, and only one such hub – Cleveland’s CBD – contains more than 10% of jobs. Most parts of the region simply lack the job density needed for new transit investments to secure funding and be cost-effective.

Given GCRTA’s well-documented and ongoing budgetary crisis, shouldn’t it be on the employers in places like Solon to step up and address their own workforce recruitment challenges?

There are several options that they could pursue, on their own or in concert with GCRTA. They could join GCRTA’s new vanpool program, which can help up to 15 people get to work in one vehicle. They could run their own shuttle service, like many companies already do in the Bay Area. They could invest in first-mile/last-mile initiatives like carshare or bikeshare programs to make it easier from employees to get from a bus stop to the worksite. (Because, yes, the 41 and 41F buses already run to Solon)

Each of these options provides some of the benefits of car ownership without placing the steep financial burden upon the low-income worker. Research suggests that owning a car doubles the odds of low-income person finding a job and quadruples the odds that s/he keeps it. Transit, on the other hand, only seems to help people keep jobs they already have.

And if these employers are, in fact, committed to bringing new transit investments to their doorsteps, why don’t they approach GCRTA with public-private partnership options? Perhaps employers in a hub like Solon could use tax increment financing (TIF) and take advantage of the increased property values that fixed transit investments would generate to cover the capital costs.

We can’t fix systemic issues with “imagination”

The problem is not just a lack of imagination and creative solutions, as Peter Truog suggests. In reality, it is much larger than that.

As I’ve illustrated, employers have known for at least 20 years that their outward migration was creating a spatial mismatch between them and potential employees, yet they continued to sprawl regardless. Hell, public officials were discussing this in detail during the 1990s, and GCRTA even created a vanpool program to address it at the time.

Some observers have argued that employers may be relocating to suburbia and exurbia in response to residential sprawl. If that was the case, perhaps facilitating the suburbanization of low-income residents and people of color could actually address spatial mismatch.

Unfortunately, that argument does not hold up to even the slightest scrutiny. In a 2007 paper, Michael Stoll and Harry Holzer of Brookings examined population and job growth rates in central cities, low-income suburbs, and high-income suburbs.

Residential populations grew at a faster clip in low-income suburbs than in high-income – 36% and 16%, respectively – reflecting the suburbanization of poverty in recent years. In contrast, jobs grew 30% faster in high-income suburbs, suggesting that jobs are shifting towards these more exclusive locales, where few low-income workers live. These location decisions are likely exacerbating spatial mismatch.

Stoll also tackled this topic in regards to race in an earlier study. In this analysis, he examined the relationship between job sprawl and spatial mismatch in 300 MSAs in the year 2000. He found that, while there is no statistically significant relationship between job sprawl and spatial mismatch for whites, one exists for blacks and Latinos.

Each 10% increase in the job sprawl index score was associated with a 3.1% and 1.7% increase the spatial mismatch index scores for blacks and Latinos, respectively. The link between job sprawl and spatial mismatch for blacks was especially significant in large Midwestern metros, like Cleveland and Detroit.

Transit alone won’t solve civil rights issues

But, perhaps all of that is in the past. What if you had a particularly woke set of employers who wanted to address that mismatch by encouraging low-income black workers to move closer to available jobs?

Well, you’d likely just run head first into systemic questions of housing discrimination and exclusionary zoning. Dozens of studies and reports from several cities – including Cleveland – have documented our job accessibility crisis for years now. And yet, not only has the problem persisted, it has gotten worse.

Susan Turner, a researcher at Wayne State University, provided a valuable insight into this question with her 1997 study (PDF) in Detroit.

Turner interviewed a number of employers in suburban Detroit who faced comparable workforce challenges. What she found was revealing. Instead of relocating to the suburbs to follow workers, Turner’s interviews suggested that many employers were intentionally moving to such locations in order to avoid black workers and keep their workforces white.

She noted that, while every predominately white suburban firm cited poor transportation as the reason why they couldn’t hire black workers, the firms that hired black workers not only faced these same challenges, they were often located in more remote sites. But they and their employees still managed to overcome these barriers.

Turner argued that her study “challenges the assumption that merely removing physical barriers to places of employment would improve the labor market outcomes of blacks.”


Negative racial attitudes aided the creation of the spatial mismatch. Thus, spatial mismatch, while partly due to economic factors related to finding cheaper land for new production processes, also results from persistent racial animosities and discrimination that find their expression through spatial variables.

One anecdote in the study was particularly eye-opening. Turner recounted a story that an urban planning official in Detroit told her. The planner’s agency attempted to set up a vanpool program for low-income Detroit workers in the mid-1990s, but they could not find any interested in employers.

According to Turner, when the planner asked the employers why they would not take part, “most indicated that they did not want to hire black workers, stating ‘Why do you think we moved out here in the first place?’”

Transportation (and the provision of public transit) may be a civil rights issue, but it is not sufficient to overcome these sort of systemic civil rights abuses.

Truog and those who support his view have not engaged with this massive topic. The fact of the matter is, Solon has pursued and continues to pursue exclusionary zoning policies that harm low-income people of color. When developers proposed a multi-family housing development in 2014, residents created the committee against rezoning Solon (CARS). (Side note: that acronym is a little too on the nose.)

These opponents tried to couch their opposition to the apartments under the guise of zoning, but the group’s leader, Don Gallo, was more explicit. He told reporters that the apartments “will not attract the wealthier young people…They will attract riff-raff.”

The ever-present specter of racism, which has only grown more ominous in recent months, continues to plague these discussions.


Ultimately, in my view, a problem this large requires a multifaceted approach.

  1. In the immediate term, employers with workforce challenges in suburbs like Solon should step up and tackle their own problems by implementing the sorts of programs I outlined earlier. Spatial mismatch is a problem that warrants the attention of the public sector, but that does not mean the private sector can wash its hand of it. We’ve spent more than enough public money to subsidize sprawl and then clean up its consequences.
  2. In the short- to medium-term, GCRTA should pursue a system redesign, moving from a radial to a multi-destination service orientation, like Houston and Columbus. This approach would be beneficial for the agency and the region, even if it fails to address job accessibility.
  3. And, in the long-term, the region’s elected officials and planning agencies should devote their attention and resources on TOD in order to take advantage of the benefits that it creates. Not only will it reduce the financial and environmental burdens of sprawl and chip away at job inaccessibility, it can also help firms take advantage of the benefits of industry agglomeration, which would increase their productivity and output.

Fortunately, there is no inherent conflict between 2 and 3. On the contrary, Brown and Thompson, in a 2012 paper analyzing transit in Broward County, Florida, wrote,

Contrary to what many planners believe, developing a multi-destination transit system does not necessarily work at cross-purposes with the planning goal of clustering urban development. One consequence of a multi-destination grid is the emergence of a number of strong routes. Planners develop transfer centers along such routes, and these centers can develop into strong development nodes.

We can begin to address Northeast Ohio’s overarching job accessibility issues. But we cannot do it if we ignore the systemic challenges behind them, and we cannot do it if we fail to take a regional approach. Both of these will require much more than just running another bus to Solon. It will require us to acknowledge and begin to ameliorate the 800-pound gorilla in the room.

Which approach is better: ‘jobs to people’ or ‘people to jobs’?

healthline health tech corridor
healthline health tech corridor

A GCRTA Healthline bus rolls down Euclid Avenue past the MidTown Tech Park campus in Cleveland’s Health Tech Corridor (courtesy of the Anchor District Council).

Back in September 2015, I was fortunate enough to take part in a focus group discussion at the Fund for Our Economic Future on a report they were finalizing that examined access to jobs in Northeast Ohio.

On the whole, I found the report, The Geography of Jobs, to be a valuable first stab at a conversation that desperately needs to occur in this region. While the comments from some of the business community representatives gave me pause, as they did not seem to fully grasp the gravity of sprawl and weak transit for our economy, it was nevertheless a useful discussion.

Jobs to people or people to jobs?

It was with that in the back of my mind that I eagerly read an op-ed on this topic from Peter Truog, the director of civic innovation and insight at the Fund. In this piece, Truog responded to a study out of Cleveland State by Richey Piiparinen and James Russell, which examined whether or not transportation policy is inhibiting the region’s movement towards a “knowledge-based economy.” The authors reached a fairly clear conclusion on job access and transportation in Cleveland, writing:

Specifically, it may not (sic) prudent to advocate for limited transportation funding in the creation of transit connections to disparate areas governed by maturing labor markets. Put simply, will bringing a bus to the likes of Solon fix what is largely structural? Probably not.”

It was this conclusion that Truog took exception to in his piece. He argued that the CSU study “risks encouraging ‘zero-sum’ transit funding discussions where urban centers vie to win a larger share of a fixed transit investment ‘pie.'” Instead, relying on the Fund’s report, he argued that we would be doing a disservice to low-income, low-skilled workers in struggling neighborhoods if we fail to improve transit service to manufacturing job hubs, like Solon.

So who is correct here? Should Northeast Ohio, forever plagued by enormous challenges and insufficient resources, choose to invest in the types of supply-side approaches to addressing job accessibility advocated by Truog? Or should policymakers instead focus on Piiparinen and Russell’s demand-side approaches, which call for concentrating economic development in areas with existing infrastructure and labor pools? To use Truog’s nomenclature – should be focus on bringing people to jobs or jobs to people?

Focus on Central

First, I want to examine the validity of Truog’s claims, particularly as they relate to Central, the Cleveland neighborhood that the Fund focused on in its report. According to Truog, residents of neighborhoods like Central have not benefited from new investments in the City, like the Health Tech Corridor along Euclid Avenue.

Instead, these residents depend upon suburban locations, like Solon, which “have a high concentration of manufacturing jobs” and “will remain important economic hubs with jobs for multiple skill levels.” As a result, Truog asserts that it is ” an economic imperative to include these communities in the transit discussion.”

But what do the data say? To answer that question, I decided to check the 2015 5-year data on commuting habits from the American Community Survey (ACS) for the 5 census tracts located in Central, as illustrated in the map below.

census tracts central

Census tracts from the 2010 Census located in Cleveland’s Central neighborhood (courtesy of NEO CANDO).

At first glance, the data seem to support some of Truog’s claims. Residents of Central are far more likely to take transit to work (35.5%) than Cleveland residents as a whole (10.2%). The share of Central residents who do not own a car (42.7%) is also vastly higher than the City’s (10.2%). But that is more or less where Truog’s claims start to lose merit.

Manufacturing is not as important for Central residents as Truog asserts. Just 11.2% of workers in the neighborhood are employed in the manufacturing sector, compared to 13% for all city residents. In contrast, a plurality (40.6%) of workers in Central are employed in educational services and health care/social assistance – more commonly known as eds and meds – nearly double the citywide rate (26%). And, contrary to what Truog may claim, manufacturing sector employees are not more transit-dependent than other workers. In fact, a larger percentage of eds and meds workers take transit both within Central (28.6%) and citywide (8.35%) than manufacturing workers (24% and 7.5%, respectively).

Interestingly, not owning a car does not necessarily seem to make one totally transit-dependent. Nearly 12% of Cleveland residents who live in a zero car household drove alone to work, while another 10.8% carpool. Just over half (54%) take transit.

Now, none of this discussion is to say that the status quo represents the ideal situation for every worker or would-be worker. In a perfect world, I imagine that we would see people travel differently, but the end result would likely be a higher drive alone rate among low-income Clevelanders, regardless of the location of their jobs. Car ownership remains a major aspiration for many low-income workers, though one that is largely too expensive to obtain.

When it comes to length of commutes, 15.2% of workers in Central spend more than 60 minutes traveling to work each way, more than twice the city average of 6.2%.  Of those Central residents with commutes shorter than 30 minutes, roughly one-third drive to work, while just 10.8% take transit. For those with hour-long commutes, these numbers are 7.4% and 85.9%, respectively.

These numbers may suggest that those who already have lengthy commutes – perhaps to suburban job hubs like Solon – are already able to utilize transit. More likely, it reflects how much longer it takes transit riders to get work, limiting the number of jobs they can access. According to the report from the Cleveland Fed, just one-third of all jobs in Northeast Ohio are accessible via transit.

Examining the evidence from the Bay Area

While the latter explanation seems far more likely, the ACS data do not tell us the direction of the relationship. To get a better sense of whether the “jobs to people” or “people to jobs” approach is better, we need to dig into the literature.

In a 2003 study (paywall), Harry Holzer from Georgetown and John Quigley and Steven Raphael from Berkeley looked at the results of a natural experiment in the Oakland area to parse this issue.

Voters approved a sales tax increase in 1986, allowing the Bay Area Rapid Transit (BART) to expand its network into the Oakland suburbs. Two new stops – the Castro Valley in the Oakland urban core and Dublin/Pleasanton in its eastern suburban ring – opened in May 1997.


To examine how the transit expansion may have affected job access and employment opportunities among blacks and Latinos, the authors conducted two surveys of employers in the area. The first occurred in April-May 1997, while the follow-up survey was done in April and July 1998, after the lines had been fully operational for a year.

They broke employers into two groups: those who were located within 6 miles of one of the new transit stations were part of the treatment group, while those outside of this radius were part of the control. This allowed the authors to determine whether proximity to the new stations influenced an employer’s propensity to hire people of color.

Interestingly, the authors found a strong, statistically significant relationship between an employer’s distance from a station and its odds of hiring a Latino worker. The likelihood that an employer would hire a Latino declined by 1.5-2.% for each mile of distance that an employer is located from the station, increasing the demand for Latino labor by roughly 8%.

But no such relationship existed for black workers. On the contrary, there was a slightly higher chance – though not statistically significant – that employers located more than 6 miles from a station would hire black workers.

The authors suggest this may stem from the fact that, while Latinos were more evenly dispersed throughout the Oakland area, black workers were heavily concentrated in urban core neighborhoods. Latinos were also employed at nearly double the rate of black workers to begin with. These conditions may have made it easier for them to locate and take advantage of jobs opened up by the new transit service.

Ultimately, Holzer, Quigley, and Raphael concluded,

Given some of the extreme distances between urban neighborhoods and suburban employment centers in modern metropolitan areas, along with the low-density sprawl development that characterizes many suburban employment centers, these patterns indicate that the potential of transit policy to foster large increases in reverse commuting is limited.

What about the Midwest though?

But that’s just the Bay Area, so perhaps the study has limited applicability to Northeast Ohio. Fortunately, we can draw from another study by Nebiyou Tilahun of the University of Illinois at Chicago and Yingling Fan of the University of Minnesota, focused on Minneapolis-St. Paul.

In their article, Tilahun and Fan consider the effects of long-range transit plans (PDF) from the region’s metropolitan planning organization (MPO), the Metropolitan Council. In their plans, the Council calls for adding 14 new transitways by 2030, which would ramp up transit connectivity in the metro area.

By modeling various scenarios, Tilahun and Fan are able to examine the “jobs to people” or “people to jobs” debate in a Midwestern context. They model the effects of four separate scenarios:

  1. Transitway-focused centralization: people and jobs cluster along transit lines; closest to transit-oriented development (TOD)
  2. Decentralization scenario: people and jobs continue to sprawl outwards
  3. Reference scenario 1: jobs cluster along transit lines, while people continue residential sprawl
  4. Reference scenario 2: people cluster along transit lines, while jobs continue to sprawl

To examine the impacts of these scenarios, the authors consider the accessibility of jobs in the Twin Cities via transit, including walking time, waiting time, and transfers. If a person can reach a job via transit within 30 minutes, it is considered accessible to her/him.

In 2010, a randomly chosen person in the Twin Cities could access 117,611 jobs within 30 minutes via transit. Under the transit system envisioned by the Metropolitan Council, this number would grow by 7.5% through 2030. Considering the alternatives, the authors note,

Results from the scenario analysis show that the highest gains in accessibility result from a policy of concentrating both jobs and population along transitways. Given the current (and anticipated base) patterns of population and jobs, if one had to choose between centralizing population or jobs, the accessibility gains suggest that one should focus on centralizing jobs along transitways.

Reference scenario 1, in which jobs but not people locate along transit, increases accessibility by 4.5%. This is more than double what can be achieved by centralizing people only. Even if residential sprawl continues, focusing on TOD for jobs significantly improves accessibility.

Ultimately, Tilahun and Fan conclude,

Under these scenario analyses, we show that centralizing housing and jobs along transitway corridors is the best strategy to follow if increasing regional accessibility is the goal. Particularly a strategy that focuses on targeted jobs centralization along transitway corridors would have significant payoffs…

By bringing jobs closer to public transportation corridors, higher accessibility gains can be achieved than can be by the provision of the transportation service alone. This can move forward broader access equity questions among the region’s population by enhancing access to car-less or other transportation disadvantaged groups.

Taken together, these results throw cold water on Truog’s advocacy for the “people to jobs” approach to Northeast Ohio’s serious accessibility challenges. They also lend support to Piiparinen and Russell’s findings, which call for more coordinated regional investment in TOD projects like the Healthline.

Nevertheless, each of the articles and studies I have considered here treats transit as an exogenous variable. In other words, they act as though the provision of transit and its effects on job accessibility occur in a vacuum, free from larger institutional questions. That is simply not the case (though it makes sense in an academic sense, as you want to isolate your independent variable).

But, given that this post is already absurdly long, I will punt that issue to a second piece. There, I will examine the broader trends of job sprawl in Northeast Ohio and how larger, systemic issues like racism and exclusionary zoning influence the impact of transit on job access for minorities in Cleveland.

Employers play a major role in shaping commuting behavior

miami rush hour traffic
miami rush hour traffic

Rush hour traffic along I-95 in Miami (courtesy of Wikimedia Commons).

For sustainable transportation advocates, changing people’s commuting behaviors can seem like our white whale.

While commutes account for just 19% of total personal trips in the US, they play an outsized role in our transportation system, accounting for 27.8% of total vehicle miles traveled (VMT).

Their timing is also critical. The concept of rush hour revolves around our commute patterns. In cities like Washington, DC and Los Angeles, rush hour congestion can make life hell commuters, costing them time, money, and sanity. But in cities that are not growing and have no real congestion issues normally, these rush hour periods are particularly important.

For a city like Cleveland, commuting patterns directly influence the transportation infrastructure we end up with. The influx of drivers heading to and from work each day provides justification to expand our already overbuilt road system, which has serious impacts on development patterns, travel choices, and mobile emissions. If we could smooth these demand spikes by reducing the number of single-occupant vehicles (SOVs) on the road, we could potentially upend this vicious cycle, which justifies the continued addition of freeway lane miles to the system.

Unfortunately, the evidence suggests that we’ve made little, if any ground in this area. In 1960, the Census Bureau reported that 64% of Americans drove to work; they did not differentiate between driving alone and carpooling at the time. Another 12.1% of commuters used public transit, while just under 10% walked. By 1980, thanks to OPEC oil embargo, 19.7% of Americans carpooled, and the drive alone number stayed at 64.4%, even as public transit use fell.

But by 2014, these trends had reversed; 76.3% of Americans drive alone to work, while just 5% take transit, and 3.4% walk or bike. To date, our efforts to get people to stop driving alone to work have failed spectacularly.

Commuting: What is the role of employers?

Part of the problem with these efforts is that we have focused far too much on the individual. Commute mode decisions are a two-way street (pun intended). They depend not only on the whims of the individual, but also on employers’ decisions. People don’t just decide to drive to work in a vacuum. Their universe of choices are shaped by a number of endogenous and exogenous factors, including things entirely in the control of their employers.

This thought really crossed my mind recently while I was reading an article on electric vehicles (EVs) by Christopher Mims in the Wall Street Journal. Mims noted the vital role that employers can play in normalizing EVs for their employees:

Placing charging stations at workplaces, where cars spend much of their time, will be uniquely powerful. When a workplace installs a charging station, employees are 20 times as likely to buy a vehicle with a plug, according to a survey from the U.S. Department of Energy.

In light of this fact, it’s important to consider what, precisely, employers can do to influence the commute patterns of their employees.

Earlier this year, the City of Cleveland Office of Sustainability and NOACA partnered together to launch what they called the Commuter Choice Challenge. The goal of the program is “encouraging Northeast Ohio organizations of all shapes and sizes to take action in sustainable transportation.”

While some people may scoff at the idea that we should reward organizations that provide pre-tax transit passes to their employees, there really are a number of steps employers can take to foster mode shift. Collectively, this effort to provide alternatives and enhance the efficiency of our transportation system is known as transportation demand management (TDM).

Changing jobs and the importance of signalling

One of the simplest things that an employer can do is to act as a knowledge broker and paragon for their employees. For most people, commuting is a habit – once people start driving to work everyday, it becomes very difficult to shake them out of it.

Because commuting is a habit, there are only so many potential points at which an intervention is likely to succeed. But habits become weaker when your personal circumstances change. This context tends to shift most abruptly after major life events, such as moving or changing jobs.

In a study published earlier this year, Ben Clark, Kiron Chatterjee, and Steve Melia from the University of the West of England explored how these life events affect people’s commutes. They found that, while one-fifth of all British commuters change their mode from one year to the next, car commuting is far stickier. Just 8.6% of car commuters changed away from driving, and the mean duration of their commute mode was 6.3 years, twice as long as those for public transit (3.0 years) or active transportation (3.2 years).

Targeting new hires can be a highly effective way to disrupt the stability of car commuting. The odds that a person will switch from driving to alternative modes increases 2.5-fold when people change jobs. This highlights the importance of providing new hires with comprehensive TDM options and information, not just a parking pass. Demonstrating from day one that your workplace acknowledges and supports non-SOV modes helps to normalize them for employees. Signalling is an important part of behavioral change.

Pull factors matter…

If employers wish to reduce their SOV share, they need to provide a suite of incentives to get them out of their cars. These pull factors can come in a variety of forms, from reduced health insurance premiums for people who use active transportation to subsidized transit passes.

Considerable evidence suggests that these sorts of TDM packages can go a long way. In a 2005 study, researchers from the U.S. EPA examined the impact of the Agency’s Best Workplaces for Commuters (BWC) program, which recognizes employers that encourage, educate, and incentivize their employees to try alternative commute modes.

The researchers compared the commute patterns among employees using the BWC programs to the average commuter in these same Census blocks. They then modeled the impacts of these commute patterns to see the associated reductions in gasoline use and mobile emissions. According to the authors:

The results of this survey indicate that where employers provide employees with incentives to commute by means other than driving alone, significant percentages of them take advantage of these benefits. Comprehensive benefits packages such as those enjoyed by commuters in the BWC group, with financial incentives, services (such as guaranteed ride home, carpool matching, etc.) and informational campaigns, appear to produce reductions of trips, VMT, pollutants, and fuel consumption of around 15 percent even under conservative assumptions.

Another 2012 paper from Virginia Tech professor Ralph Buehler found that providing bike parking, showers, and locker rooms increases the odds that employees will bike to work nearly 5-fold. Clearly pull factors, such as financial incentives and facilities investments, play a central role in this equation.

But push factors – especially parking – matter more

But, as with anything else, changing commute behaviors requires both push and pull factors. And the latter are particularly key, as the single most effective strategy that an employer can use to reduce SOV share is to remove parking subsidies.

In the US, some 95% of US commuters receive free parking at work. The provision of this benefit can increase the SOV rate for commutes by up to 50%.

UCLA professor Donald Shoup, the godfather of parking research, has explored the effects of curbing this parking subsidy. In a 2005 report, he outlined the benefits of implementing a parking cash out program, by which employers provide commuters with the option of receiving a cash incentive equal to their parking subsidy if they don’t drive alone to work. Such programs allow employees who really want to drive to work to continue getting discounted parking, but it also incentivizes alternatives for those who would rather try them.

Shoup’s research in California found that cash out programs can cut SOV share by 17%. A separate study from Daniel Hess, also of UCLA, concluded that by charging $6 per day for parking, Portland was able to cut its SOV share by 16%.

Getting parking right is even more important than these numbers show, however. The lure of free parking so strong that if an employer rolls out a TDM program but fails to price parking, the latter will simply crowd out the former. As Dr. Shoup put it, “Advocating ridesharing while offering free parking is like denouncing smoking while offering free cigarettes.”

A separate paper Buehler and his colleague Andrea Hamre explored this issue. Their research showed that providing free parking increases the share of commuters who drive alone, regardless of what other incentives the employer may provide. Without free parking, 75.9% of Washington, DC area commuters would drive alone. Free parking increases that share to an astounding 96.6%. Providing subsidized transit and incentives for active transportation, while also supplying free parking, only takes that SOV share down to 86.8%.

As the authors concluded, their research “suggests that benefit combinations that include free parking either overwhelm or render insignificant the positive effects of benefits for public transportation, walking, and cycling.”

Location, location, location

But even the most comprehensive TDM packages will struggle to overcome another factor that employers can control – their location.

Often times, sustainable transportation advocates focus on the negative effects of residential sprawl, but neglect workplace sprawl. Just as people in the US have spread farther and farther outward, so too have employers.

Consider Northeast Ohio. The region boasts five major employment hubs, like downtown Cleveland and University Circle. Yet, combined, these hubs only account for less than one-quarter of all jobs in the region. The rest are distributed broadly across the five counties.

This outcome poses a major challenge to TDM. Transit ceases to be viable when households and destinations are sprawled out. The same holds true for active transportation. No one is going to choose to walk 10 miles to work in an exurb without sidewalks.

The Clark, Chatterjee, and Melia study illustrates this clearly. If a worker’s commute increases from less than to more than two miles, the odds that s/he will switch from active commuting to driving increases 30-fold. The research seems to indicate that two miles is a key threshold; most people simply will not bike to work if their commute is longer than that.

According to the National Center for Transit Research, location may be the most important variable in the commute equation. No matter how strong the TDM package or how much the organization supports alternative modes, locating your office in the middle of an exurban office park locks in your employees’ commute options.

Ultimately, I think we have focused our attention too narrowly on the individual commuter for too long. Research has shown time and time again that the most effective TDM strategies target the employer first, as that is the critical leverage point.

Large institutions that claim to support sustainability need to back up their words through their actions when it comes to commute options. It’s not enough to simply post an annual sustainability report or get your buildings LEED certified if you subsidize parking and locate your office in exurbia. The transportation sector is now the largest source of carbon pollution in the US. It’s time for employers to act like it.

Does sprawl make the urban heat island effect worse?

urban heat island effect by city
urban heat island effect by city

Cleveland experiences the fourth strongest urban heat island effect in the United States. Could our sprawling development patterns be to blame? (courtesy of Debbage and Shepherd, 2015).

A few weeks ago, NASA officially announced that the record-breaking, “Godzilla” El Niño event that dominated much of our weather over the past year plus had finally come to an end.

But while the monster has returned to its hibernation deep below the surface of the Pacific Ocean, its impacts have already been and will continue to be felt across the United States. Around the same time that it made this announcement, NASA also revealed that April and May were the warmest such months on record in the US, meaning that every month since October 2015 has broken the existing record for that month. This eight-month streak of heat is, obviously, unprecedented. To date, the average temperature in 2016 is 1.9°F (1.08°C) above the average for the 20th century, making it a full 0.43°F (0.24°C) above the mark for the first five months of 2015.

You remember 2015, right? The warmest year on record? Well, not for long. NASA scientists are already more than 99% certain that 2016 will break that record, just as 2015 had claimed the mantle from 2014.

The impacts of 2016’s extreme heat

The extreme heat is having clear effects. It is contributing to wildfires consuming wide swathes of the West. Ozone levels are higher than normal across the country, as high temperatures foster the development of harmful, ground-level ozone more readily. So far, Greater Cleveland has already experienced six days when ozone levels exceed 70 parts per billion (ppb), the most at this point since 2012.

But the most acute impact of high temperatures heat-related mortality, a subject that I’ve written about considerably. Extreme heat is the deadliest type of disaster in the US, killing more people than hurricanes, floods, tornadoes, and lightning strikes combined each year. As I’ve discussed in the past, climate change is only exacerbating this issue; the World Meteorological Organization (WMO) noted that the global death toll from extreme heat rose by around 2,300% from 2000-2010, compared to the previous decade.

change in disaster deaths by decade

The change in the number of deaths, by disaster, from 1991-2000 to 2001-2010 (courtesy of WMO).

Nearly all regions have seen a spike in dangerous heat, but the risk of heat-related mortality is not distributed evenly. While an individual’s vulnerability to extreme heat is the function of a number of factors, one of these is where s/he lives. Generally speaking, those of us living in cities are at greater risk due to the so-called urban heat island (UHI) effect. I won’t go too far into the science behind the UHI effect; suffice it to say that the combination of dark surfaces, a lack of urban trees, and the production of waste heat from various sources like air conditioners increases the temperature of cities, relative to rural areas. According to the U.S. EPA, the temperature of a large city can be more than 20°F higher than surrounding rural areas under the right (or wrong?) conditions.

Last September, Forbes published an article examining the scale of the UHI in various cities throughout the US. Strikingly, it included a map (see above) stating that Cleveland has the fourth strongest UHI effect in the country. Now, if you’re one of the literally tens of people who has inexplicably read something I’ve posted on this site, you may be familiar with my general dislike of sprawl. I’ve discussed research linking it to population decline, limited social mobility, climate change, and poor air quality, among other things.

So, I wondered, could Cleveland’s strong UHI effect be the result of our development pattern? Given that sprawl affects so many important phenomena, it seems reasonable to assume it would have an effect on UHI, right? To the peer-reviewed literature! [Cue 1970s Batman transition music].

Is suburban sprawl actually linked to the urban heat island effect?

At first glance, it may seem odd to posit that suburban sprawl would play a role here; the phenomenon is called the urban heat island effect, after all. But a handful of studies strongly suggest that sprawling development patterns do, in fact, exacerbate the UHI effect.

Two of the most convincing papers come from Brian Stone, Jr., a professor at the Georgia Tech School of City and Regional Planning and an expert on urban environmental planning and climate change.

In a 2006 study (paywalled) that he coauthored with Jon Norman from the University of Wisconsin-Madison, Stone examined the link between land use patterns and the UHI effect in Atlanta. The researchers broke properties into groups based on four variables: extent of impervious surfaces, lawn and landscaping, tree canopy, and the number of bedrooms per residential structure. This categorization enabled them to study the magnitude of surface warming produced by property type.

Stone and Norman concluded that the size of residential lots – in other words, residential density – was closely tied to black body flux, a measure of surface warming. As one moves from the highest density lot type to the lowest density, the amount of surface heat released increased 6-fold. Other land use features closed tied to suburban and exurban development – namely large lawns – also exacerbate the UHI effect. A one unit increase in the area of a plot covered by lawn and landscaping increases the net black body flux by 0.51 units.

As the authors conclude,

The results of this analysis provide compelling evidence that the size and material composition of single-family residential parcels is significantly related to the magnitude of surface warming in the Atlanta study region. Specifically, smaller, higher density parcels were found to be associated with a lower net black body flux than larger, lower density parcels…

[The] results of this study support the hypothesis that lower density, dispersed patterns of urban residential development contribute more surface energy to regional heat island formation than do higher density, compact forms.

Connecting sprawl and the UHI across cities

On its own, one study does not prove the relationship. Fortunately, Stone followed up with a 2010 paper that he co-wrote with Jeremy Hess and Howard Frumkin of the University of Washington, which studied the connection between urban sprawl and the number of extreme heat events (EHEs) in 53 cities from 1956-2005.

To measure the relationship, they took the correlation between the mean annual change in the number of EHEs from 1956-2005 and the sprawl ranking for each of the cities in 2000. Whereas the most compact cities experienced 5.6 more extreme heat days in 2005 than in 1956, that number was 14.8 for the most sprawling cities. In other words,

The most sprawling cities experienced a rate of increase in EHEs that was more than double that of the most the most compact cities…These findings are consistent with the hypothesis that urban sprawl contributes to EHE frequency.

Exploring some competing research

Now, I should note that there is other research that does not jibe with Stone’s work. Last year, Neil Debbage and Marshall Shepherd of the University of Georgia took another look at urban form and the urban heat island effect. Using a different measure for UHI (the difference in average rural and urban temperatures) and a different measure for urban form (an index measuring various variables of city shape, contiguity, and land uses), Debbage and Shepherd studied the degree to which city configuration affected urban heat in the 50 largest US metro areas from 2001-2010.

Contrary to Stone, Debbage and Marshall found that both more compact and more sprawling cities experience a stronger UHI effect, provided they are highly contiguous. That is, the contiguity of urban form may matter more than its composition; designing cities so that they are made up of either cul-de-sacs or skyscrapers as far as the eye can see makes them more vulnerable to extreme heat. According to the authors,

A ten percentage point increase in the spatial contiguity of high-intensity urban development, the equivalent of shifting roughly from Orlando to Seattle, was predicted to enhance a city’s average UHI intensity by 0.4°C…[In turn] a ten percentage point increase [in low-intensity urban development] was predicted to enhance a city’s annual average UHI intensity by 0.3°C. Therefore, as suggested by the bivariate analysis, both low and high-density urban land uses appear to amplify the UHI effect if they are high contiguous.

Importantly, Debbage and Marshall note that, while compact development may not solve the UHI on its own, it does provide a litany of other benefits, from improved air quality to better public health. Accordingly, urban planners may want to promote less contiguous, higher density urban development by designing networks of smaller green spaces, expanding the urban tree canopy, and installing white and green roofs throughout cities. A 2014 study by Stone and colleagues found that implementing these types of policies can offset projected increases in heat-related mortality due to climate change by anywhere from 40-99%.

Wait, so does sprawl make the UHI effect worse?

So what can we take away from all of this?

First, yes – the evidence does suggest that sprawl exacerbates urban heat islands. Low-density, suburban-style development increases the amount of impervious surfaces, which raises lowers the surface albedo of urban areas. It also increases the amount of excess waste heat that cities produce, as larger houses require more energy. And sprawl typically leads to forest clearance for development, reducing the extent of the urban tree canopy. All told, these factors increase the amount of heat cities generate, and they prevent this additional heat from dissipating rapidly at the urban fringe.

Second, the fact that sprawling development patterns are not the only type urban form that increases the UHI effect may not be as relevant as it may seem. While dense urban areas may also promote UHIs, they also make it easier to address both the causes and effects of heat-related mortality risks. Residents of dense cities produce fewer carbon emissions per capita, mitigating climate change. And the economies of scale in these dense neighborhoods increases the efficacy of mitigating extreme heat; opening a cooling station or installing shade trees are more effective in these areas, for instance.

All told, we can add the urban heat island effect to the list of social problems that sprawl makes worse. Maybe we should rename it the (sub)urban heat island effect?

Raising the sales tax is not the answer to GCRTA’s funding woes

joe calabrese town hall
joe calabrese town hall

GCRTA CEO Joe Calabrese prepares to speak at the Clevelands for Public Transit town hall meeting on June 23 at Antioch Baptist Church (courtesy of Clevelanders for Public Transit).

Pragmatists have long invoked the phrase “Don’t make the perfect the enemy of the good.” Generally speaking, that’s solid advice. And, as a card carrying incrementalist sellout™, it’s something I can get behind. Most of the time, that is.

Some issues are so substantial, so systemic in nature, that tinkering on the margins is unlikely to remedy the problem. And that’s one of our major pathologies here in Cleveland. We seem to try tackling these big, hairy problems with the same tired toolkit of solutions, despite the fact that they haven’t worked yet. There’s only so many times you can run headlong into a brick wall.

The problem isn’t that we make the perfect the enemy of the good. It’s that, in Cleveland, we tend to make the facile the enemy of the good.

In other words,  I mean that we almost always fall back on old ideas, regardless of whether or not they have worked or are the best tool for the job.

Our ongoing transit funding dilemma

Take public transportation. For years, transit boosters here have offered up exactly one solution for our system’s funding woes: make the State of Ohio pay its fair share. And, for the love of God, of course it should! But it’s not going to happen anytime soon.

Yet, even when we finally start talking about local funding options, we fall into familiar trap. At a town hall meeting last week, Newburgh Heights Mayor Trevor Elkins – the only member of the GCRTA Board of Trustees to vote against the new suite of fare hikes and service cuts – suggested that we raise the county sales tax to fund the agency.

Now, I’m not saying that there’s anything prima facie wrong with funding transit with sales tax revenues. Cuyahoga County voters approved a 1% sales tax to find GCRTA when it was established back in 1975, and this tax – still the most generous in Ohio – has been a resounding success. The fact that it continues to provide a majority of GCRTA’s revenue 41 years later, despite not being altered, is a testament to the fact that sales taxes can be useful policy tools.

The problem is that, while the transit tax itself hasn’t changed since 1975, the overarching conditions in the region have. Cuyahoga County’s ceaseless population loss has slashed the revenue generating potential of this tax by some $68 million.

Why is the answer always “raise the sales tax”?

Moreover, over the past two decades, Cuyahoga County leaders have turned consumption taxes into a veritable slush fund for bad ideas, many of which have degenerated into little more than thinly veiled corporate welfare. Want to build an oversized convention center and Medical Mart Global Center for Health Innovation? Let’s raise the sales tax! The Browns want a free, new scoreboard? Sin tax! The Republicans need hotel rooms? Bed tax!

Eventually, when you keep throwing tens of millions of good tax dollars after bad into these sorts of projects, people get wary. And, to be fair to Mayor Elkins, he said as much at the town hall meeting. But he still fell back on sales tax hike, perhaps because we lack the policymaking creativity to try something new. Cleveland’s leaders have no incentive to come up with good ideas when they can get away with the same old facile ones.

The problem is, just because something is easy to understand doesn’t make it an effective policy. Cuyahoga County already has the highest county sales tax rate (2.25%) and combined sales tax rate (8%) in the state. Franklin County is second at 7.5%. While we are no longer at the statutory sales tax cap (the state raised that to 8.75%), continually raising consumption taxes is not a long term recipe for success in a county that has been treading water, at best, for nearly 50 years.

Wait, you may say, didn’t you just say that sales taxes could be useful for funding transit? Yes, I did; but I don’t support this idea for a number of reasons, a few of which I will expound upon here.

The problem with raising the sales tax — again

First, as I already mentioned, Cuyahoga County residents already bear the single highest sales tax rate in the state. Raising it further will simply stress a shrinking tax base further.

Economists typically prefer consumption taxes to income taxes, because the latter tend to generate larger distortionary effects. But that’s not to say that consumption taxes don’t do this as well. By raising the relative costs of goods, sales taxes alter the equilibrium between supply and demand by decreasing the amount that people consume. This creates a market inefficiency, known as a “deadweight loss,” which reduces social welfare. It’s wonky as hell and seems weird, but it’s at the heart of economic theory.

Such deadweight losses have clear effects. According to a recent study by researchers at the Institute for the Study of Labor (IZA) in Germany, sales tax increases in the US harm localities. The study, which considers the effects of changes in combined sales tax rates on business activity from 2002-2011, found clear, negative effects.

Each 1% increase in the combined sales tax rate reduces total payrolls within a county, particularly among retail businesses and smaller firms. According to the authors, “These results show a clear negative and significant association between the combined state and county sales tax rate and total annual payroll.”

These deleterious effects are particularly apparent when a county’s sales tax rate is high in relation to neighboring counties. As the map below illustrates, the four counties bordering Cuyahoga have combined see taxes ranging from 6.5% to 7%.

ohio county sales tax rates

Combined sales tax rates in Ohio by county (courtesy of Ohio Department of Taxation).

Multiple studies have demonstrated that these types of cross-border differences can negatively affect the county with the higher rate. In a 2012 study (PDF), Jeffrey Thompson of the Federal Reserve and Shawn Rohlin of Kent State University examined how sales tax rate differences across state borders affected economic activity from 2004-2009. They found that when border counties raise their sales taxes by 1%, total employment falls by 3.8-5.8%. The bulk of this effect occurs in the retail industry, which sees a 7.6% decline in employment.

Second, it is a well-established fact that consumption taxes are regressive, as lower income residents end up paying a larger portion of their income than wealthier residents. This effect is concerning in Cuyahoga County, where we see extremely high rates of poverty and economic inequality in many areas. But this issue is even greater, in this case, as low-income residents are largely concentrated in impoverished neighborhoods within the City of Cleveland and inner-ring suburbs. Wealthier residents, in contrast, have moved out into the outer suburbs and exurbs.

Why does this matter? Because these suburbs are closer to the neighboring counties, making it easier for middle- and higher-income families to take advantage of lower sales tax rates. Consumers are far more likely to cross county borders to make purchases at lower sales tax rates if they live near those adjacent counties. According to Gary Cornia and colleagues (paywall), consumers are more likely to travel 5 kilometers (3.1 miles) a border to take advantage of lower sales tax rates, but that effect essentially disappears when you increase the distance to 30 kilometers (18.6 miles).

Moreover, the low-income residents residents living in Cleveland and the inner-ring suburbs are also far less likely to have ready access to a car. Compare Cleveland, where more than 10% of households lack a car, to Solon or Gates Mills, where that number is just 1.1% and 0.4%, respectively.

As Thompson and Rohlin demonstrate, “cross-border shopping is more prevalent when transportation costs are low.” This essentially guarantees that a sales tax hike will be uber regressive for low-income, transit-dependent Cuyahoga County residents, as they lack the means to avoid it readily. So they would bear an even more disproportionate burden of funding GCRTA, in addition to already shouldering the fare increases.

Third, raising the transit sales tax seems neither seems imprudent from a political perspective. Under Ohio law, county leaders can increase the sales tax by 0.25% to raise general fund revenues without triggering a vote. Our former County Commissioners took advantage of this back in 2007 for the convention center.

That option is not available for transit. Any effort to increase GCRTA’s sales tax rate would have to go before voters for approval. Now, it’s entirely possible that such a proposal would pass. Cuyahoga County voters have approved similar tax increases for far less beneficial purposes in the past. Yet, it’s not guaranteed. Summit County voters rejected this same sort of proposal in 2014 by a convincing 54-46 margin.

Are Cuyahoga County residents that much more inclined to support transit, particularly the type of voters who show up in off-year elections? Does it make sense for GCRTA and local officials to spend considerable political and financial capital trying to get approval for a proposal that does nothing to staunch the winds that have buffeted the transit system for decades?

Let’s choose good policy over simple policy

We know that GCRTA is a fighting a heroic, sisyphean battle against sprawl, population loss, and overbuilt vehicular infrastructure each day. So why don’t we pull policy levers that raise transit funds while simultaneously helping to remedy the underlying challenges?

There are two good options available that serve this dual purpose: a parking tax, which I have already proposed, and scaling up GCRTA into a real regional transit agency by incorporating the collar counties.

Both of these are possible under current law. They will require legislative action and voter approval at the local level, yes. But if we have to launch a campaign to gin up support for local transit funding, we might as well do it right.

How America’s anti-urban bias distorts infrastructure spending

portsmouth bypass construction
portsmouth bypass construction

Crews prepare embankments for the Portsmouth Bypass project (courtesy of Midwest Energy News).

The relative struggle for power between urban and rural areas is a defining feature of the American political system, one that dates back to the founding of the country. In the immediate aftermath of the Revolution, the two dominant political theories were the urban republicanism of Alexander Hamilton and the agrarian democracy of Thomas Jefferson.

Hamilton’s ideas and biography are en vogue again, but the Jeffersonian push for agrarian power was enshrined in the Constitution. During the Constitutional Convention, small states successfully imposed the Connecticut Compromise, which created a bicameral legislature that included an upper house where every state would have equal standing. Thanks to this compromise, Wyoming – a state with 586,000 people – has the same power in the Senate as California – a state with five cities larger than 500,000 people.

Legislative proportionment and Baker v. Carr

Although rural areas unquestionably have disproportionate power on the federal level, it’s even starker at the state level. For a variety of reasons, rural voters tend to hold sway in state legislatures. One driver behind this outcome is gerrymandering, as typically Republican-dominated statehouses can draw districts in such a way as to enhance the relative power of rural and suburban residents.

The situation was actually worse prior to 1962. In that year, the Supreme Court issued its decision in Baker v. Carr, ruling that states need to continually update their legislative districts based on decadal Census data. This ruling effectively enshrined the principle of “one person, one vote,” which helped to somewhat level the playing field between rural and urban voters. But the situation has not been remedied. As I’ve said before, life is still difficult for blue cities located in red states.

Anti-urban bias in infrastructure spending

Recently I came across a great quote on this issue:

Ohio has practiced a rural and suburban philosophy that ignores big central city problems because those who run the state win their positions by soliciting the solid backing of farmers and small town residents.

I think this sentence encapsulates the current relationship between Cleveland and the State of Ohio quite well. Except it’s not a contemporary quote. This is actually from a speech (start at 17:00 mark) that former Mayor Carl Stokes delivered at the City Club of Cleveland on July 24, 1970.

Stokes spoke at a time when the State of Ohio showed no inclination to support the City of Cleveland. One year before the Cuyahoga River infamously caught on fire (for the 13th time) in June 1969, Cleveland residents passed a $100 million bond issue to upgrade the city’s water and wastewater infrastructure. The bond issue was necessary, because – six years after Baker – the State of Ohio refused to finance the improvements.

This imbalance remains today. Ohio actually spends more state money on its rural transit program than its urban program, despite the fact that 78% of Ohioans live in urban areas.

ohio transit funding 2000-2014

Transit funding, by program, from the Ohio Department of Transportation from 2000-2014 (courtesy of ODOT).

The bias towards rural areas also remains when examining spending on social issues at the state level. How many red states refused to expand Medicaid under the Affordable Care Act, for instance? Here in Ohio, while Governor John Kasich loved to brag on the campaign trail how he did expand Medicaid, he never seemed to mention how he only expanded food stamp allocations for rural areas, not cities.

But it’s possible to alter these types of social spending decisions in the short term. Governors and legislatures can pull the policy levers to increase welfare or education financing much more readily than they can build a new transit line in cities.

So what is it about infrastructure that makes it harder to influence, over the short-term, than social spending? What accounts for this inertia?

The ‘persistence of highways’

In a recent working paper (PDF), Stanford political science professor Clayton Nall and PhD candidates Simon Ejdemyr and Zachary O’Keeffe examine this issue. The study examines how the anti-urban bias in legislative apportionment affected the distribution of public goods before Baker. It also questions to what extent the case remedied this issue. As a proxy for public goods, the study considers the provision of federal-aid highway funding.

The authors identify two key mechanisms that explain the “persistence of highways” and that make highway spending less responsive to political change than social spending.

First, infrastructure clearly carries a sense of permanence. While the food that SNAP benefits help families buy may be gone a day later, a new bridge or highway will be fixed in steel and concrete for decades. Politicians are suckers for groundbreakings and ribbon cutting ceremonies.

Once these structures are built, there is typically tremendous pressure to maintain them. This is true even if the road or bridge outlives its useful life or if the cost of maintaining it outweighs the benefits. Like everyone else, lawmakers are subject to the sunk cost fallacy, which leads us to falsely believe that we need to keep investing time and money into projects or tasks, simply so that we don’t lose the time and money we have already spent.

Secondly, building infrastrastructure creates policy feedback. In other words, the development of the federal highway system gave rise to new interest groups that benefited from their construction. Construction and engineering firms directly benefited by winning state and federal contracts.

New infrastructure also creates what the authors call spatial policy feedback. New highways alter development patterns, facilitating the growth of suburban and exurban areas. In turn, these new suburban populations form powerful interest groups. Their support for and dependence on these highways becomes a self-perpetuating political force.

The unique persistence of highways is an important issue, as the framework for the federal-aid highway system was laid before Baker and was, thus, inherently anti-urban. In the Federal Highway Act of 1944, Congress explicitly barred the use of federal funds to build roads in communities with more than 2,500 residents.

While Congress altered its funding formula in 1956, it allowed rural-dominated state legislatures to influence the allocation of highway funding for decades. According to the authors, “state legislative malapportionment would have compounded already biased federal policies that limited states’ freedom of action to develop their urban areas, while promoting the biases within state legislatures.”

So, did the Baker ruling make a difference?

To examine how anti-urban bias within the political system may have affected the allocation of highway spending, the authors developed two models.

First, they compare each county’s portion of total highway mileage to its share of the state’s population. They then account for each county’s representation within the state legislature using the Relative Representation Index (RRI); a county with more political representation than its population should justify will have a score above 1, and vice versa.

To study the impact of the Baker ruling, they compare results from 1934-1960 to those from 1970-present. Per their results,

We find that legislative representation had approximately the same effect on a county’s share of state highway construction, regardless of region and urbanism. Second, and more importantly, we find that pre-Baker malapportionment had a persistently significant effect on highway-mileage bias in the decades after Baker. We find that representation mattered, but that it was the timing of the representation—prior to the construction of most of the American highway network—that dictated the distribution of highway infrastructure for decades to come.

According to this model, from 1934-1960, a one standard deviation increase in a county’s RRI score increased its share of highway-mileage spending by 0.3 standard deviations. While this number has been halved since the Baker ruling, the effect remains.

Second, the researchers compared the relative impact of Baker on both highway and social spending from 1972-20002; to measure the latter, they used state spending on public welfare and education. In this model, counties with higher RRI scores continued to see higher levels of state highway funding after Baker, even as they received less funding for welfare and education.

The implications of this study are significant. We live in an era of budget cuts and austerity. But, in a lot of ways, this crisis is political. The Ohio legislature could not be bothered to allocate even $1 million more for public transit, even as the largest transit system in state, GCRTA, muddles through a $7 million budget deficit.

Yet, while Cleveland residents brace themselves for fare increases and service cuts, Ohio has no qualms about spending $1.2 billion on the Portsmouth Bypass, which serves no purpose other than to let drivers avoid a few traffic lights.

As Ejdemyr, Nall & O’Keeffe conclude,

The current plight of American infrastructure – widely described as a “national infrastructure deficit” – is not a universal phenomenon but represents a long-term legacy of legislative malapportionment and decisions about infrastructure made before cities had equal representation in state legislatures. The poor state of American infrastructure is not merely a result of overall underinvestment, but stems from a historical legacy of unequal treatment that left some areas (notably cities) with a host of social and economic problems, including underfunded road infrastructure.

In an era in which people seem to be rediscovering the value of our center cities, we cannot afford to keep recreating the mistakes of our predecessors. Remedying the anti-urban bias in infrastructure spending will not happen overnight, but it’s well past time that we start.

Ohio won’t save GCRTA, so let’s tax parking to fund transit instead

rta healthline buses
rta healthline buses

RTA HealthLine buses in downtown Cleveland (courtesy of

One of the biggest stories in Northeast Ohio right now is the Greater Cleveland RTA’s budget shortfall. It’s probably because of the company I keep, but my Facebook and Twitter feeds have been inundated with posts, comments, and tweets about every new update and public meeting for the past several weeks.

It’s a big story. GCRTA has reported that, in order to balance its books, it needs to cut expenses by $7 million this year. CEO Joe Calabrese and his staff have proposed a suite of route cuts and fare increases to plug this hole. Options include raising the base fare from to $2.50 per ride from $2.25 currently, increasing paratransit fares to $3.50 from $2.25, and curtailing or eliminating bus service along 18 routes. Alternatively, the agency could maintain existing service and increase the normal fare to $2.75 per trip.

Rather than just approving some combination of these options, the GCRTA Board of Directors tabled this discussion at its December 2015 meeting, opting to hold a series of 15 public meetings around the county. The last of these hearings occurred on Wednesday, and the ball is now back in the Board’s hands.

If I had to wager, I would guess they’ll raise fares by $0.50 to minimize the service cuts. Keep in mind that Cleveland has already cut annual bus revenue miles by nearly 40% since 2006, the single largest decrease in the country, according to Jake Anbinder of the Transit Center. Given the nearly overwhelming opposition to some of these service cuts, making it that much harder to get around town seems pretty untenable.

In addition to hosting this litany of hearings, Mr. Calabrese testified before the Cleveland City Council Transportation on Wednesday. He came to distill the agency’s challenges, justify its plans, and hear feedback from the Committee members. For the most part, the tenor from the Council members was pretty standard – they all agreed GCRTA is in a tight spot, they opposed service cuts in their wards, and they pilloried the State of Ohio for not doing its part.

Enough ink has been spilled – including by me – on the sad state of public transit funding in this state, so I won’t belabor the issue. Suffice it to say, as I once did, that I’m not sure it would be possible for Ohio’s elected officials to care less about public transit if they tried. Hell, even if Ohio devoted every one of the $7.3 million it kicks in for public transit to GCRTA, that would barely be enough to paper over its budget hole. The state needs to fund transit, full stop.

We can’t depend on Ohio to fund transit

But, while I don’t like cutting service on the 81 to the Lakeview Terraces or raising paratransit fares, I found myself agreeing most with Councilman Zach Reed. It was strange. I rarely see eye-to-eye with Councilman Reed on transportation issues, but his comments were dead on. He called on local officials to disabuse themselves of the notion that Ohio is suddenly going to find religion on transit funding.

Instead, Councilman Reed broached a subject that most local officials have sidestepped – we need to increase local funding for transit. Currently, GCRTA gets around 60% of its funding from a 1% county sales tax assessed in 1970. But this tax generates far less revenue today than it did in 1970; population loss costs the agency nearly $68 million in funding each year. That could close this budget hole nearly 10 times over.

gcrta sales tax revenue

Sales tax revenue by year (courtesy of GCRTA).

Additionally, Councilman Reed was the only person to note another key detail – federal and state transit dollars come with strings attached, including the local match requirement. Local governments need to cough up 20% of the cost of a project in order to spend federal transit dollars. This issue has increasingly become a hurdle. According to ODOT’s transit needs study (see page 46 of PDF), the state is sitting on more than $21 million in transit funding that it cannot disperse due to a lack of local matching funds. We need more transit spending for Cuyahoga County from Cuyahoga County; there’s no way around it.

Granted, continually increasing taxes on a shrinking population to fill budget gaps is a recipe for disaster. But not all taxes are created equal. There are certain levers that officials can pull to help rectify social harms and raise funds at the same time. And since sprawl is among Cleveland’s most pressing issues, taxing land uses that promote it can be beneficial. So let’s tax parking to fund transit.

Cleveland already has a parking tax, but…

First, I’ll note that Cleveland already taxes parking.* In 1995, City Council approved an 8% sales tax on commercial parking transactions in the city. This tax raises roughly $10-11 million per year for the City’s coffers. Or it, would if Council hadn’t passed this tax as part of its plan to finance a new Browns stadium. Cleveland doesn’t actually see a dime of this money, as it just goes to pay off debt from bonds issued for FirstEnergy Stadium. Argle bargle.

The easy way to raise funds for transit would be to simply raise this existing tax. Compared to other cities with this sort of tax, Cleveland’s is relatively low. New York City, Miami, and Los Angeles impose taxes of 18.5%, 20%, and 25%, respectively. Pittsburgh, which has impressively progressive transportation policies, imposes a 50% tax. Cleveland could, say, double its tax to 16% – raising $10 million per year for transit – and remain on par with other cities.

Unfortunately, despite its ubiquity, this tax is flawed. Because it’s assessed on reported transactions, parking lot operators have an incentive to underreport their sales, something that has occurred in Cleveland. Additionally, it can have the unintended consequence of reducing the supply of paid parking and increasing the supply of free parking in city centers.

According to a study (PDF) from the Victoria Transport Policy Institute (VTPI), “it makes urban centers relatively less competitive compared with suburban locations where parking is unpriced. In this way, commercial parking taxes can increase total parking subsidies and sprawl.” Not only are we wasting the revenues from our parking tax to subsidize a football stadium, the tax itself may be contributing to urban sprawl. Argle bargle.

Taxing parking lots by surface area instead

What other options exist? A number of cities outside of the US – chiefly in Australia and Canada – take a different approach. They levy a tax on the total area of surface parking lots or on the total number of parking spaces. In this way, the tax generates a double dividend; it produces tax revenue while also driving down the demand for parking and reducing congestion.

Following the advice of a study from Eran Feitelson and Orit Rotem, I propose that we implement a tax on the surface area of private parking lots. But this tax would be imposed based on the square footage of each lot at ground level; this would ensure that a surface parking lot like those blighting the Warehouse District would have the same tax burden as a 4-level parking deck with a comparable surface area on each level. Not only would this create revenue and cut into parking demand, it would also push developers towards parking decks, as the effective tax per parking spot falls with each additional level added.

Moreover, the supply of free parking in the region would decrease, as developers would now have a greater motivation to recoup tax expenditures by charging. There is a legitimate risk that hiking up parking taxes could push people away from downtown or other districts within Cleveland. Accordingly, this policy should be implemented at the county level. Doing so would increase the cost of developing in suburban and exurban areas, relative to the city center, because the latter has an existing supply of parking decks and underground lots.

Because this would be applied countywide, it would be essentially a commuter tax imposed on both county residents and those people who live outside but enter the county to work, go to school, shop, see a sporting event, etc. This would help to address the types of equity concerns we face when dealing with similar taxes, like the Sin Tax.

Coincidentally, ODOT actually explored the idea of levying an annual tax on parking spaces to fund transit all the way back in 1993. The agency estimated (see page 20 of PDF) that this sort of tax could generate $187.3 million per year. If we adjust for inflation, that would be equal to $307 million in 2016 dollars. If we conservatively assume that a tax in Cuyahoga County could generate 10% of this revenue, that would still equal roughly $30 million per year.

Excess parking is a scourge for urban areas. It consumes valuable land, encourages driving and sprawl, contributes to air pollution and climate change, increases surface runoff, and harms water quality. On a good day, GCRTA struggles to compete and keep its budget balanced. Parking makes this challenge that much harder. So let’s try to remedy our incentives and tax parking to fund transit.


*Credit to Cleveland real estate lawyer and part-time blogger Christian Carson, whose 2014 post helped put me onto this idea.

Actually, fuel economy standards are a great way to tackle carbon emissions

plug-in hybrid prius
plug-in hybrid prius

A Toyota Prius plug-in hybrid vehicle (courtesy of Wikimedia Commons).

It feels like it’s been ages since I wrote a post taking down something that someone else has written. I get the impression that is what people enjoy on the World Wide Web these days, plus it’s pretty fun to rip apart a person’s specious argument – using peer-reviewed literature and well-sourced facts, of course.

With that in mind, I feel somewhat obligated to address an op-ed I read in the Los Angeles Times on Monday from Salim Furth, a research fellow at the conservative think tank The Heritage Foundation. In the piece, Furth argues that state and federal fuel economy standards are a poor policy tool for limiting mobile greenhouse gas (GHG) emissions and that they unfairly harm low-income families. Instead, he calls for California state officials to focus their attention on land use reforms that would “allow denser, environmentally conscious construction” to “make residents less dependent on car.”

On the surface, this seems reasonable. I’ve written in the past about how promoting denser, infill development patterns in sprawling metro areas like Cleveland could go a long way towards improving air quality and limiting GHG emissions. To the extent that Furth is calling for these sorts of policies, we are probably on the same page.

Except, when you dig into his argument, it collapses like the proverbial house of cards. As the folks at Climate Nexus argued, “Instead of suggesting policy that would preserve more land to act as a carbon sink, Furth writes that California should instead relax the permitting process so that development is even easier.” And this is exactly what he argues. While it’s true that NIMBY-ism can inhibit the development of denser, multifamily housing (see: Washington, DC), it takes a certain amount of rhetorical gymnastics to assert that the fault lies with environmental regulations. I guess that’s what you get when dealing with stuff from Heritage.

Comparing fuel economy to land use planning

But none of this gets to the central thesis of Furth’s argument – that fuel economy standards are less effective tools for curbing GHG emissions than “streamlined” permitting and “more permissive zoning laws.” Why enforce regulations that cost the average family roughly $4,000 to only mitigate global climate change by a fraction of a percent?

Leaving aside the fact that Furth demands California repeal state fuel economy rules that even he admits were superseded by President Obama’s 2011 CAFE standards, does his main point hold water? Well, he never actually provides a shred of evidence to support his argument, for one. How can we know if the CAFE standards will cut GHG emissions less than land use reform if we don’t have numbers for the latter?

Fortunately, there exist a number of studies and reports that dig into the potential for land use reform to mitigate climate change. At the local level, several of these analyses have come from metropolitan planning organizations (MPOs), which are federally-mandated agencies that conduct transportation and environmental planning activities for urban areas.

Back in 2008, California lawmakers passed SB 375, the Sustainable Communities and Climate Protection Act, which requires every MPO in the state to develop a sustainable communities strategy (SCS) that outlines its approach to meeting its GHG emission reduction target. These targets are established by the California Air Resources Board (CARB). To what extent can land use planning by MPOs contribute to these these goals? And do the projected reductions in GHG emissions exceed those from fuel economy standards?

In a word, no. CARB estimated (PDF) in 2010 that regional transportation and land use policies can only account for one-sixth of the GHG reductions generated by federal CAFE and low-carbon fuel standards through 2020. That proportion will likely increase after 2020, as the full effects of those long-term policies are realized, but they still pale in comparison. Given that Furth is writing about California, you think he’d be aware of these data.

Evidence from outside California

Findings from MPOs in other states back up CARB. Washington, DC’s MPO found similar results (PDF). The region’s leaders set a goal of reducing GHGs 80% versus a 2005 business as usual (BAU) scenario by 2050. According to their analyses, enacting new land use and transportation policies at the metro level can only make up 3.3% of this 80%. Increasing CAFE standards to 99 could account for 30% of the reduction, however, making this approach 10 times more effective. While raising CAFE standards would likely lead to something of a rebound effect by making driving cheaper, the results are still impressive.

The Puget Sound Regional Council (PDF) – Seattle’s MPO – has also modeled the potential GHG savings from various policies. They found that more compact development and better pricing transportation could cut GHGs by 6% and 9% compared to BAU, respectively. Emissions control strategies, like stricter fuel economy standards and the electrification of the vehicle fleet, have the potential to cut GHG emissions by 25-43%, depending on how aggressive they are. Even in the more conservative scenario, these standards outperform land use controls. The benefits of land use policies take an outsized role in Seattle, as transportation accounts for two-thirds of the city’s total emissions, because it’s electric grid is so much cleaner than the national average. Accordingly, Seattle is the best case scenario for Furth’s argument, but it still falls short.

ghg savings from different scenarios

Potential GHG reductions from various policy instruments under different scenarios (courtesy of Puget Sound Regional Council).


And just to hammer my point home further, the Transportation Research Board (TRB) published a comprehensive report on this topic back in 2009. The authors modeled the impacts on GHG emissions from 2000 to 2030 and 2050 under two scenarios, which assumed that 25% and 75% of all new housing would be be built in compact development areas, respectively.

While scenario 1 only sees GHG emissions fall 1.3-1.7% by 2050, while scenario 2 bumps this number up to 8-11%. But, again, tightening CAFE standards wins the day. The study finds that adding aggressive fuel economy requirements to scenario 2 can increase the GHG reduction potential up to 39-51%. The report states, “In short, over the longer time frame (i.e. to 2050), the impacts of continuing improvements in fuel economy beyond 2020 on energy use and CO2 emissions significantly outstrip those from more compact development.”

As is so often the case, an op-ed emerging from The Heritage Foundation is tripped up by the think tank’s old nemesis – math.

In the effort to cut GHG emissions and battle climate change, we don’t need to privilege better land use planning at the expense of tighter fuel economy standards. We need to harness every policy tool at our disposal, and these are two great tastes that taste great together. While it’s true that better fuel economy can undermine some of the GHG benefits of compact land use, we should clearly pursue these approaches in tandem. For, in the long-run, more compact, mixed-use development and more efficient vehicles are both important tools for improving air quality, reducing transportation costs, revitalizing our neighborhoods, enhancing public health, and battling climate change.

There’s no congestion in Cleveland, so why are we adding capacity?

480-271 junction
480-271 junction

The junction of Interstates 480 and 271 near Warrensville Heights (courtesy of Google Maps).

Most people probably recognize TomTom as a company that manufactures and sells GPS units and fitness trackers. But developing these tools has allowed the company to develop considerable information regarding road networks and traffic conditions worldwide. TomTom uses these data to produce its annual Traffic Index, and the firm released its much awaited 2015 edition of the report earlier this week.

The report includes a wide array of information on congestion across the globe, from the amount of time that drivers spend in traffic to the individual day in 2015 when congestion was worst in each city. Many of the cities notorious for their soul crushing congestion topped the rankings, with Mexico City taking over first place from Istanbul, which fell to third behind new entrant Bangkok. Surprising no one, Los Angeles has the worst traffic in the United States, coming in 10th. Brazil, which has some of the worst gridlock in the world, had more cities in the top 10 than any other country. Rio, Salvador, and Recife ranked fourth, seventh, and eighth respectively.

The pros and cons of congestion for cities

TomTom ranks cities using a metric it terms “congestion level.” It defines this variable as the “increase in overall travel times when compared to a Free Flow situation.” In other words, it indicates how much longer a given trip will take in a city, due to congestion, compared to a situation in which the flow of traffic was completely unobstructed. In Mexico City, for example, traffic ensures that the average trip takes 59% longer than it would otherwise. Capitalinos spend 219 hours – more than 9 full days – per year waiting in traffic.

Sitting in traffic sucks. There’s no question about that. It consumes time people could spend doing something more productive or fulfilling, it wastes substantial amounts of fuel,  exacerbates air pollution, interrupts economic output, and discourages people from entering congested areas. As Yogi Berra famously quipped, “Nobody goes there anymore. It’s too crowded.”

But, congestion also hints that a city is succeeding. If a lot of people are trying to get into and around your city, that probably indicates that the city is doing well – at least up to a certain threshold. It’s when your roads are totally empty and you can get around town without tapping the brake that you should be worried.

cleveland congestion level history

Cleveland’s congestion level history from 2008 to 2016 (courtesy of TomTom).

Congestion? What congestion?

If you ask the average commuter in Greater Cleveland, s/he is likely to complain that we have issues with traffic. After all, many of our major interstates, particularly the Innerbelt, have been under construction for years now. This constant construction, combined with the ballyhooed revival of downtown must be making the streets more crowded, right?

LOL. The TomTom index ranked 174 cities based upon their respective levels of congestion. Cleveland came in 168th with a congestion level of just 11%. In other words, of the 174 cities for which TomTom had traffic data, just 7 of them have less traffic than Cleveland. What’s more, traffic actually eased somewhat from 2014 to 2015, falling from 12%. In the eight years that TomTom has been publishing its index, Cleveland has had congestion levels of 11% six times and 12% twice. That’s it.

According to the report, travel times for the average Cleveland commuter increase by just 20% and 25% during our morning and evening “rush hours,” respectively. The single worst hour of the week for travelers is 5:00-6:00pm Wednesdays, when the congestion index tops out at 27%. For the sake of comparison, there are 78 cities that have an average congestion level of 27% or more.

Expanding capacity to solve a nonexistent congestion problem

Given the fact that Cleveland has little, if any, traffic problems, one would expect our local and state transportation planners to avoid adding capacity. We clearly have enough roads as it is.

And this is exactly NOACA has done. The NOACA Board of Directors officially adopted a “fix it first” policy last September, indicating that it would prioritize funding to repairing our existing road network, rather than adding to it. Every lane mile that we add simply increases our financial liabilities, because it’s extra pavement to lay, repair, clear from snow and ice, etc.

There have been some rumors floating around the interwebs that ODOT would adopt a similar policy. This would be a huge deal, albeit one reportedly driven solely by financial realities, as opposed to enlightened self-interest. And yet, when the agency announced $2.1 billion in road and bridge projects it plans to complete this year, the list included plans to widen three separate freeways in Greater Cleveland:  I-271, I-77, and I-76/77. Granted, these projects have been in the works for a long time, and it was highly unlikely they’d be scrapped, regardless of whether or not we really need them. It is what it is.

Yet, we know that widening urban highways has directly contributed to the sprawl and population loss that has plagued Cleveland for 50 years. In a 2007 study, Brown University economist Nathaniel Baum-Snow concluded that every new highway that passes through a center city cut its population by 18%. According to Baum-Snow, “had the interstate highway system not been built, instead of declining by 17 percent, aggregate central city population would have grown by 8 percent” from 1950-1990.

We have entirely too much pavement in this region. NOACA (PDF) has indicated that there are 670,795,906 square feet of federal aid roadways in Northeast Ohio. That’s equal to just over 24 square miles of pavement, roughly the size of the City of Strongsville. But here we are again, adding capacity, even as this pavement falls into disrepair. NOACA estimates that 46% of the region’s pavement falls short of being in a state of good repair. The cost of just bringing that roughly 308 million square feet of pavement into good repair is somewhere on the order of $1 billion.

But hey, that’s for another day! Why fix what we already have when we can widen highways so that “motorists [who] are putting on make-up or even shaving” don’t have to worry about keeping their eyes on the road. That’s just good public policy.

If you want to improve air quality, end the sprawl

interchange los angeles
interchange los angeles

The I-10/I-110 Interchange in Los Angeles (courtesy of Politico).

For centuries, people have fled the supposed squalor of cities in pursuit of the fresh air that is so vital for our health and well-being. Before Louis Pasteur’s development of germ theory, most scientists and physicians subscribed to the belief that miasmas – essentially the foul smells associated with rotting organic matter – were the source of major diseases. The cure for illness, they argued, was for people to escape cities to get fresh country air.

Doctors prescribed fresh air as a treatment for various illnesses into the 20th century. American physicians encouraged their patients suffering from tuberculosis to head West in pursuit of the restorative benefits of the clean air. This movement helped foster the growth of many prominent Western cities, including Denver and Phoenix.

The clean air premium

Today, we tend to refer to the deleterious emissions that plague many cities by a different term: air pollution. But that same underlying precept – that we need to leave cities behind in the pursuit of fresh air and better health – remains.

Multiple studies demonstrate that people view air quality as an amenity and are willing to pay for it. According to economists Kennethy Chay and Michael Greenstone, reductions in particulate matter during the 1970s were “associated with a $45 billion aggregate increase in housing values,” while a separate study found that Americans were willing to pay $149–$185 for a one unit reduction in particulate matter levels.

Clearly, we place a premium on the concept of “fresh air.” But could our pursuit of this good actually be making air quality worse for others?

How might sprawl affect air quality?

The link between our sprawl-based development patterns and air pollution seems pretty obvious on the surface. The more we spread out, the more we have to drive to reach workplaces, schools, stores, entertainment venues, etc. All of this adds to vehicle miles traveled (VMT), and each additional mile we drive increases our mobile emissions.

But perhaps it’s not as simple as it seems. Moving people away from the worst polluters was clearly beneficial for public health. And pollution levels are typically higher in denser areas with high traffic volumes.

Additionally, one of the main targets of the landmark 1970 Clean Air Act Amendments (CAAA) was vehicle emissions. As a result, cars purchased today run more than 90% cleaner than those manufactured 46 years ago. This trend has helped to offset the rise in VMT. Thanks to these regulations, emissions of carbon monoxide and ground-level ozone have fallen by 85% and 33%, respectively, since 1980, while particulate matter emissions have declined by more than one-third since 1990. These improvements have occurred even as VMT has nearly trebled from 1.12 trillion when the 1970 CAAA was signed to just under 3.14 trillion last year.

annual vmt

Rolling 12-month change in vehicle miles traveled in the US (courtesy of St. Louis Federal Reserve).

With all of this in mind, I decided to comb through the literature to see what the best available science says on the relationship between sprawl and air pollution, and what I found may surprise you.

Just kidding, no it won’t.

Sprawl and air quality: The evidence

One of the first scholars to explore this relationship, in depth, was Brian Stone, Jr., who published his findings back in 2008. He probed the relationship between the number of days that ground-level ozone concentrations exceeded the EPA’s National Ambient Air Quality Standard (NAAQS) and sprawl index scores for 45 large cities from 1990 to 2002.

Stone, Jr. argued there are three major reasons that urban form could affect ozone levels. First, it can influence emissions of nitrogen oxides (NOx) and volatile organic compounds (VOCs), the two pollutants that combine to form ground-level ozone. Secondly, sprawling cities are more prone to the urban heat island effect, which can increase the ambient temperatures in urban areas by more than 20°F; higher temperatures facilitate the formation of ozone. Thirdly, large, sprawling cities are more likely to have a broad ozone monitoring network, which may increase the odds that high ozone levels are observed.

The study found a strong, statistically significant link between urban form and air quality, even when controlling for weather conditions. According to Stone, Jr.’s results, a one standard deviation increase a city’s sprawl index score produced 5.6 more ozone exceedance days per year. In turn, a one standard deviation increase in population density – one of the four components of the sprawl index – was associated with 8 fewer exceedance days. Based on his findings, he notes that “urban form is significantly associated with both ozone precursor emissions and ozone exceedances…Overall, the most sprawling cities were found to experience over 60% more high ozone days than the most compact cities.”

To put that into perspective, Cleveland has a composite sprawl score of 85.62, meaning it is just over 14 units less compact than the average metro area. If it was as dense as Madison, Wisconsin (136.69) or Detroit (137.17), we would have had 11.2 fewer ozone exceedance days per year through 2002. Given that we averaged 20.5 exceedances per year from 1997-2002, this would represent a 55% reduction.

Strengthening the connection

Multiple subsequent studies support these findings. A 2013 article from Bradley Bereitschaft and Keith Debbage examined the connections between ozone and fine particulate matter (PM2.5) emissions and five separate sprawl indices. Each of the indices computed sprawl in a slightly different way, accounting for various measures of urban form. The authors folded these into two key measures – urban continuity (the degree to which the urban landscape is fragmented) and urban complexity (the degree of the “jaggedness” of the urban boundary).

The authors note that each of the sprawl indices produced a significant connection between sprawl and pollution levels. A one standard deviation increase in the most prominent sprawl index score was associated with 3.4% and 7.8% increases in ozone and PM2.5, respectively. The density of residential properties appears to be a particularly important correlate for air quality. Increasing residential density by one standard deviation lowers ozone and PM2.5 levels by 8% and 16%, respectively. This change in density can also lower on-road CO2 emissions by an average of 1.9 million tons. Using a social cost of carbon of $37 per ton, this produces social benefits of $70.3 million.

Bereitschaft and Debbage argue that “an increase in residential density might improve air quality and contribute to a reduction in per capita CO2 emissions at the metropolitan scale primarily by decreasing automotive dependency and tailpipe emissions.”

Additionally, the authors examined the connection between urban form and the direct emissions of NOx, VOCs, and PM2.5. Their results are striking. A one standard deviation rise in urban shape complexity increases PM2.5 emissions by 3,055 tons (12.4%) per year. Using EPA damage factors for on-road emissions, this additional pollution would cause 128 to 287 premature deaths and carry public health costs of up to $2.47 billion per year.

If more sprawl leads to worse air quality, will reversing that trend in Rust Belt cities lead to cleaner air? Stone, Jr. looked at this issue in a 2007 piece for the Joumal of the American Planning Association. He studied how shifting from a business as usual scenario to a more compact growth approach modeled on Portland could affect VMT and air quality in 11 Midwestern cities through 2050. By shifting from our current sprawl-heavy trajectory, Cleveland could lower household VMT by 9% and reduce emissions of CO, NOx, VOCs, and PM2.5 by anywhere from 7-9.2% each.

All the evidence points to one conclusion – sprawl is exacerbating air pollution. In our haste to find fresh air, we’ve simply made things worse for those who do not have the means to keep moving farther and farther out.

As Bereitschaft and Debbage put it, “Planning for density therefore becomes an issue of environmental justice, particularly at the metropolitan level. Simulations suggest that by relocating to peripheral suburban areas, residents might reduce their exposure to certain air pollutants…[while] simultaneously contributing to a decline in regional air quality by increasing the total volume of automotive traffic.”

I’ve argued before that, despite our poor air quality, almost no one in Greater Cleveland seems interested in talking about the issue and how we might solve it. Perhaps that’s because, unlike in the past, the real source of the problem isn’t just some large coal-fired power plant or steel mill. No, the problem is our individual driving habits. We’ve met the enemy, and it is us. Unless we face up to that fact, we aren’t going to change things for the better.