"Gretchen, Stop Trying to Make Megaregions Happen."

For someone who likes geeking out about public transit, I discovered Alon Levy’s blog embarrassingly late in the game (May of this year). The most fun posts he writes are those along the lines of American Transportation Is in the Hands of a Bunch of Parochial, Self-Satisfied Idiots—see, for example, his reaction to the GAO report on construction costs, which he describes as “a total miss,” and that’s before he’s even really gotten started—but the most recent post hit closer to home. Levy writes about the rise and fall of the idea of the “megaregion,” and where the term might still be applicable:

The key takeaway is that rich cities do not have to be in megaregions. The Northeast Corridor is a rich megaregion, and San Francisco, Los Angeles, and Chicago anchor smaller megaregions of their own; but in Europe, among the richest cities only Frankfurt and Amsterdam are in megaregions, while London, Paris, Hamburg, and Munich are not. Megaregions are areas of high population density and interlinked social networks. Their size may give them economic advantage, but it doesn’t have to; urbanists and urban geographers must avoid overselling their importance.

“Closer to home” because for a good chunk of my grad-school career, 2005 to 2009, I worked at the Center for Quality Growth and Regional Development (CQGRD) at Georgia Tech, and CQGRD was one of the loci of the megaregion idea that Levy’s discussing. (The other major center of megaregional research, as Levy notes, was the Regional Plan Association, through its America 2050 initiative; and there were other groups contributing, such as SCAG and the planning programs at UT-Austin, Penn, and Virginia Tech.) Catherine Ross, who edited the book Megaregions: Planning for Global Competitiveness, was my doctoral adviser, and we co-authored one of the book’s chapters together. My dissertation research ended up being at an entirely different scale, so I haven’t had a chance to discuss megaregions with Dr. Ross in a while; but the concept is still part of her research agenda.

Now, if you read Levy’s post, you may be raising an eyebrow right now, because Levy very specifically states that Atlanta is not in a megaregion; and yet we at CQGRD were pretty consistent on insisting that there was such a thing as the Piedmont Atlantic Megaregion, and that it ran from Birmingham to Raleigh, including parts or the entirety of six different states. And at this point you’ve probably got both eyebrows up, because: states? If megaregions consist of interlinked metropolitan areas, why bring states into it?

The original idea of the “megalopolis,” first put forward by Jean Gottmann in the early 1960s, was meant to describe the Boston-New York-Philadelphia-Washington corridor. The idea was that connectivity—economic, social, in terms of trip choices—was so dense between those cities that it was hard to define any one of their areas as isolated from the others. Not surprisingly, that’s the area where Levy finds the description works. But a similar number of trips doesn’t exist between Birmingham and Raleigh, or even Birmingham and Atlanta. (For those who don’t spend a lot of time in the American South: Birmingham and Raleigh are almost 900 kilometers apart by highway.) To claim that Birmingham-Atlanta-Raleigh is similar to Boston-New York-Philly means straying from Gottmann’s definition quite a bit.

Which is what we did, when we talked about megaregions. Instead of relying solely on existing density, we talked about cultural and environmental similarities, expected population growth trends, and freight flows. (This is part of the reason why we talked about states rather than metropolitan areas: a Piedmont Atlantic Megaregion defined solely by metropolitan-area social connections wouldn’t include the port of Savannah.) But that flowed from the bigger difference between Gottmann’s megalopolis and the megaregions of the 2000s. Gottmann started with the existing connectivity and asked what changes that connectivity implied in how people thought about where they lived, and what the connectivity might portend. But the megaregion was a framework to get to the connectivity. It was more prescriptive than descriptive.

You can still find, for example, the original 2005 briefing by Robert Lang and Dawn Dhavale, then both at Virginia Tech (Lang is now at Brookings), in which the hopeful language is hard to miss:

When the Census Bureau does formalize a geographic concept, it gains power. Consider a recent example…. In 2003, the US Offi ce of Management and Budget (OMB), which oversees the Census Bureau, responded with the designation “Micropolitan Area.” Now micropolitans are literally on the map. Businesses, government agencies, and planners have new geography to work with. Publications took notice—Site Selection Magazine, for example, started a list of “Top Micropolitans” in which to locate businesses…. Megapolitan Areas (or “Megas”) have a similar potential. Once they are officially recognized, private industries and government agencies would embrace this new geography.

The argument all along was that if the planners and policy-makers of the various metropolitan areas and states could just be talked into putting their heads together, the resulting growth could be shaped to reap more of the potential benefits of density and connectivity. A megaregional approach could be used to help policy-makers brainstorm across state and metropolitan-area lines and mitigate the kind of at-each-other’s-throats infighting that led to the likes of the Tri-State Water Wars. Whereas the Bos-Wash corridor had had geography and history contributing to its connectivity (pre-car urban settlement + smaller amounts of land to settle + a high concentration of academic and research activity), PAM was going to need some top-down direction.

Hence one of our megaregion-focused initiatives was a 2009 one-day symposium at which elected officials and policy-makers from throughout PAM showed up and talked infrastructure, with the idea of fostering megaregional coordination and pressing for national-level goals (not least a commitment to high-speed rail). By the way, if you check the America 2050 archives, you’ll see a picture of one of the speakers, Pat McCrory, who was then mayor of Charlotte. Then-Mayor McCrory was very affable and pro-cooperation and pro-transit and charming, and so when he got elected governor of North Carolina later on, I assured a worried friend that he would be pro-cooperation and pro-transit and generally not too awful. And if you take away nothing else from this blog post, remember: do not under any circumstances ask me for stock tips.

I’m joking, but you can start to see the problem. Megaregion-as-policy-driver was a hard sell that got harder after that symposium: neither the recession nor increased political polarization at the national level helped its cause. (The one place where it did get some purchase was the FHWA, thinking about freight routes.) And megaregion-as-theoretical-construct didn’t travel well: could the same word be applied to growing dense population centers in Asia and growing (but considerably less dense) population centers in the United States? There was a lot of discussion of potential European megaregions, as Levy notes, but the following ten years didn’t bear out the megaregional predictions there either.

This is all, admittedly, the insiderest of baseball, possibly not even of interest to people doing megaregional research right now. Having not kept up with that research, I’m not in a good position to disagree with Levy’s finding the whole concept weak. (His invocation of fans in the Bos-Wash corridor crossing state lines to attend conventions does make me wonder: would we have had a stronger case for PAM if we’d showed how people travel for SEC football games?) I did want to add to the history of the megaregion idea in the 2000s while I could, though—some of the work we shared and relied on is already dead links. And the whole effort does illustrate the tensions in planning for the spatial forms you want, as opposed to the spatial forms you’ve got.

Why Didn't a More Diverse Gwinnett County Vote For MARTA?

The New York Times’s new 1619 Project, reflecting on the 400th anniversary of the bringing of slaves to American soil, includes an article on how two centuries of segregation influenced metropolitan Atlanta’s urban form, leading to, among other things, traffic jams. A large part of the article discusses how racism fueled the two Gwinnett County votes, in 1970 and 1990, to turn down (partially federally funded) MARTA expansion into the county. Which is pretty well established. If you’re not familiar with the history of MARTA being shunned by the white-flight counties, Doug Monroe’s 2012 piece for Atlanta Magazine describes it in more detail, and more passionately. (During our public-transit unit of the class this summer one of my students found that piece for me, unasked, unrewarded by extra credit.)

I will add here that fear of crime also played into both the 1970 and 1990 votes, as violent crime was more prevalent both generally in the United States and in the Atlanta metropolitan area than it is now; it was particularly salient for the 1990 vote, because IIRC there was an extensively-covered stabbing at a MARTA station in the months leading up to the vote. But generally in the United States and in the Atlanta metropolitan area, it is hard to untangle “fear of crime” and “fear of people of color,” and at the time of the 1990 vote Gwinnett residents were going on the record calling Atlanta residents “the wrong people,” so the point stands.

But if white racism -> isolationist settling in far-flung counties -> spurning of public transit, and then the far-flung counties get less white and still reject public transit, what’s going on? That’s the question alluded to but not answered in the last part of the article:

Earlier this year, Gwinnett County voted MARTA down for a third time. Proponents had hoped that changes in the county’s racial composition, which was becoming less white, might make a difference. But the March initiative still failed by an eight-point margin. Officials discovered that some nonwhite suburbanites shared the isolationist instincts of earlier white suburbanites. One white property manager in her late 50s told a reporter that she voted against mass transit because it was used by poorer residents and immigrants, whom she called “illegals.” “Why should we pay for it?” she asked. “Why subsidize people who can’t manage their money and save up a dime to buy a car?”

The quote is an odd veering-off—one would have expected that, following an assertion about nonwhite suburbanites, the reporter would then quote a nonwhite suburbanite. So the reader is left with the overall impression that little has changed and Gwinnett is still full of racists—and I won’t be the foolhardy soul to say there aren’t white racists in Gwinnett, but I think the nonwhite population of the county is being dismissed too easily.

First, a quick recap of how the demographic shift happened:

Data for 2018 is from  Census Population Estimates ; I extracted the 1970 and 1990 data from  the NHGIS data set from IPUMS .

Data for 2018 is from Census Population Estimates; I extracted the 1970 and 1990 data from the NHGIS data set from IPUMS.

Apologies for making that a picture and not a real embedded table; if your screen reader is baffled, the quick summary is that Gwinnett went from 95% white and 5% black in 1970, to roughly the same in 1990, to 62% white and 25% black by 2018.

(Obviously when we say “people of color” or “nonwhite” we generally include people of Hispanic/Latino/Latinx descent, but that’s not how the Census works; it treats race and ethnicity as two separate categories. The 2017 American Community Survey, which is a slightly different data set than the one I used above, has a total population for Gwinnett of 889,954, of whom 184,621, or 20.7%, identified as Hispanic or Latino. Of that population, 3.6% selected “black or African-American” as their race, so we can safely say that the black and Hispanic/Latino/Latinx populations of Gwinnett don’t overlap very much.)

So what happened? First of all, turnout was relatively low: approximately 90,000 votes were cast, which meant 89% of Gwinnettians didn’t participate. Second, the Times isn’t quite right in implying that suburbanites of color generally voted against the measure. If you look at the Atlanta Journal-Constitution’s map of the vote, you can see pretty clearly that the “Yes” corridor clusters roughly around I-85, an area that has large populations of Latino (closer to DeKalb) and Asian (in Duluth / Suwanee) residents.

The “No” vote, meanwhile, gets stronger the further out from I-285 you get, the Daculas and the Loganvilles. By the white-racist hypothesis, this makes sense. If you reject the white-racist hypothesis, it actually still makes sense: Dacula to Doraville is about 25 miles, which, even if you hire Alon Levy and somehow give him a magic wand so he can make rail construction happen at closer to European costs, is still a good $4 billion. (Officials were guessing $200-300 million per mile before the vote, which, reading Levy, seems optimistic.) If you know the rail project is never going to get close enough to you to let you ditch your car, you either need to be selfless indeed or see other benefits, and the Daculans may simply have never seen those other benefits.

Third, George Chidi, who writes for the GeorgiaPol blog (and who supported MARTA expansion), argued soon after the vote that MARTA supporters didn’t make their case well enough even to black voters:

The MARTA vote lost in large part because of a revolt among black taxpayers in southeastern Gwinnett around Snellville. These voters — who are somewhat more conservative than black voters generally — perceived little personal benefit for paying $100 a year in increased taxes for the services of a bus line, maybe.

…let’s be clear: this was an old-fashioned tax revolt, of exactly the same flavor that killed the statewide T-SPLOST a few years ago. The rebranding of the MARTA governance body as The ATL didn’t help; state government and GDOT haven’t built enough goodwill with the electorate of either party to be trusted with new projects.

I’ll add a fourth possible reason: a lot of potential voters might not have been all that excited by the hub-and-spoke approach MARTA rail is locked into. That goes even for black voters in south Gwinnett. To get from there to black communities in the likes of Stone Mountain or Lithonia, you drive. If the Gold Line gets extended from Doraville to Gwinnett Place Mall, you’d still drive. (Or take the bus, but MARTA, like a lot of transit authorities in the United States, has a hard time selling its buses to the public.) The same calculation applies for residents going from Dacula and Buford to Johns Creek or Roswell. Gwinnett is more diverse, both in the casual way we use the term and in terms of actual demographics; and so is DeKalb, and parts of Fulton north of the city, and parts of Cobb. Because of that, it may be that Gwinnett residents may not value the rail link to downtown and yet still not be “isolationist.”

In short I think the white-racism-led-to-highways-led-to-traffic analysis, while more fairly applied to the 1970 and 1990 votes, loses steam when applied to the March 2019 vote. You can be pessimistic about both race relations generally and about metropolitan Atlanta’s ability to mitigate its congestion by expanding transit options; but the latter doesn’t map as neatly onto the former as it used to.