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