We have recently moved the
Creative Class Exchange.

Please update your bookmarks with our new address at www.creativeclass.com

We look forward to your comments and discussion.

Thank you.

Posts by Author

  • Global Trends
  • Ask Rana: Advice on Work, Life and Play
  • Urban Digs, Creative Class Communities
  • Workplace
  • Entrepreneurship, Creative Class Strategies
  • Creative Class Research and Indicators
  • Architecture + Design

Video Interview

Watch a Speech

Hear a Speech




October 31, 2007

Richard Florida


« Buff-Tor-Chester? | Main | How Not to Fix Your City »

Interesting back and forth over at Tory Gattis's blog on Hou-Orleans, the mega-region running fro Houston to New Orleans and megas in general.

Caught this Richard Florida blog post on the rise of global mega-regions ... Unfortunately, something in their methodology led them to split Texas into two mega-regions, the Dallas-San Antonio corridor, and the Houston-New Orleans corridor, extending all the way the Florida panhandle (see map on p.27). Of course, anybody familiar with Houston's economy knows it has far more connections to the rest of the Texas Triangle cities than it does to NOLA and points east, and I let him know in the comments.  The true mega-region is the Texas Triangle Megalopolis, as identified by the Federal Reserve Bank.

But some disagree:

I wouldn't be so quick to dismiss the Houston-New Orleans corridor. That's the petrochemical belt, and though Houston is admittedly the dominant pole in that belt, we do have a very strong connection to the points in between, such as Beaumont and Lake Charles and Lafayette and the famed cancer alley between Baton Rouge and NOLA. There are a lot of strong connections tieing Houston in that direction.  I don't see a similar connection between Houston and San Antonio, nor the points in between.

Another points out:

Probably a much better and more thorough examination of this phenomenon in the US is happening at america2050.org ... It's very interesting stuff, and really just getting started, having had an official organization for just three years. But it's got some top demographers and planners working on it.

Hmmmmm ... These are empirical issues. We use a comprehensive methodology to identify mega-regions, using light-emissions to pinpoint contiguous geographic areas.  It's Hou-Orleans that's a distinct mega that way: There are significant light gaps in the Texas Triangle.  I'm a big fan of what America 2050 is doing, but our method for identifying megas is different and we believe to be a significant advance in two respects. One, it takes guess work out of the process, allowing mega boundaries to be determined purely on observed light data.  Two, it enables us for the first time to systematically identity and compare megas not just in the US but worldwide.


TrackBack URL for this entry:

Listed below are links to weblogs that reference Hou-Orleans:



I've also struggled with the light data versus the actual connections between cities in the Mega Region (as I've said before, in Cascadia it seems companies in Vancouver are more tied to San Jose than Seattle, and those based in Seattle similarly more connected to California than Portland or Vancouver.)

Perhaps we need yet one more study of "mega regions" that starts with Richard's groups findings based on light data combined with Gross Local (?) Product.

Next we need a methodology to trace the economic connections between the mega regions or sub-regions within them. On a map, perhaps these would be lines of varying thickness showing the weight of the connection in dollar value. This way you could see how strongly Houston-NOLA is connected to Dallas versus other points in the US.

I'm not sure if the data exists, however. It might have to be compiled by a survey of large employers or large companies.

Michael Wells

I think we need to define what we're measuring. If it's industry sectors, then there may be more in common between software firms in Seattle and Silicon Valley than with the more hardware-oriented high tech manufacturers like Intel in Portland, who also have more connections with California than Seattle. But Seattle software firms also have more in common with California software firms than they do with Tacoma lumber companies or food distributors -- and we wouldn't argue that Seattle/Tacoma's not a coherent region.

If we're talking about job markets, supply chains, social networks it may be that the light-based data shows distinct regions. So for the creative class, what do we want to study?

Michael Wells

On the West Coast we also have to realize that California is huge and is going to dominate any economic relationship. With a population of 36 million and an economy of $1.5 trillion, it accounts for about 13% of the US in both areas and is larger than Canada (unless the strong Loonie has made Canada catch up). Much larger than Oregon, Washington and B.C. put together.


Interesting links related to megapolitan regions- with discussion related to methodology and definitions:




Good comments all. Wendy, I think the most efficient way to do what you suggest is with air traffic data, which incidentally is on our list.

The hardest part by far was developing these geographic boundaries and creating techniques to estimate economic activity based on light emission. The next steps are easier.

An alternative, and also very interesting method, for identifying regional locations of economic activity, based on "gross cell product" has been proposed by William Nordhaus and his team at Yale. So progress is being made.

Mega-regions are a big project of the Prosperity Institute, so expect more to come. And expect that some of the issues everyone is raising here to make it into the pages of Who's Your City.

Charles Rostkowski

I would like to put in a plug for another region in North America: the Salt Lake-Boise corridor. Its population is approaching Denver metro and Glasgow-Edinburg areas and the Salt Lake connections to southern Idaho are substantial. The Morman Church is one of the key connecting factors. Fast Company magazine included both Salt Lake and Boise on their lists of hot cities. Both ends of this region have booming economies. Richard, this region may be worth a look.

Michael Wells

Air traffic data works between megas and sometimes within. It certainly makes sense to fly from Portland to Vancouver. But it's actually as fast or faster to drive (3 hours or so) to Seattle -- by the time you drive to PDX an hour early, fly for 25 minutes, get transportation at Sea-Tac and get into the city it's taken the 3 hours. Suburb to suburb is even worse.

Of course the best would be AmTrac if it had a reasonable schedule, didn't get sidelined or break down regularly. A Cascadia high-speed rail would probably be the most productive thing for creating a coherent region here.


In this case, Richard, I would suggest that your light-emissions approach is getting distorted by peoples' inherent desire to live close to the coasts, especially waterfront, even if they have no economic or personal ties to the next metro down the coast. It makes it look like Corpus Christi is linked all the way to Mobile and Pensacola, which is simply not the case.

Even with that, I'm actually surprised there is continuous light between Houston and NOLA, because there's mostly a giant swamp in southern LA, including a 35+ mile continuous bridge ("Atchafalaya Swamp Freeway") over the empty swamp on I10 between Lafayette and Baton Rouge.

On the other hand, there is little incentive for people to live on the nondescript land between Houston an the other Texas Triangle cities, but there is certainly a tremendous amount of daily traffic between those cities - car, truck, and plane. Check out the mapped flows on p.8-9 of this report on the Texas Triangle:

The reason the light looks more continuous between San Antonio and DFW is that that's the edge of the Texas Hill Country, which is a much more attractive area to homestead. Most retirees or others who want to "move to the country" in Texas go there, supporting many small towns along I35.

There are certainly energy ties between NOLA and Houston, and maybe that justifies extending the Texas Triangle mega-region east to include NOLA, but it doesn't make sense to cut off Houston from the rest of the Triangle just because few people choose to live in-between.


My personal experience with the whole concept of "mega-regions" involves people - often young people, going to college or leaving college - choosing the best college or career opportunity they can within a roughly one-half to one-day drive of their friends and family. Social networks are powerful, and people like the idea of being able to make easy weekend trips to visit it. It's a tougher decision to completely leave that network and move to one of the coasts, for instance - although plenty of people make that choice too. As people get older, and their local social network gets larger and "denser", the resistance to abandoning it goes up. It's really a huge "asset" to give up, and it takes a heck of a long time to build one from scratch in a new place. This regional draw, IMHO, is a big part of the success of cities like Atlanta, DFW, and Chicago (and, to a lesser extent, Seattle and Denver) - who each have very large and populated "draw zones" .

Of course, those regional social connections often lead to business connections, thus interlinking that regional economy over time and validating the mega-region concept.

One issue with this phenomenon is that it often stops at a national border for obvious reasons (with maybe the European Union being a recent exception, because of the new ease crossing borders and gaining employment there), thus the comments questioning regions crossing the Canadian border.


Tory - I like all of that too. I am a huge RPA fan and have worked a great deal with them. But their data is for the US only. Not the world. The key to our research was to map the megas for the world. Nordhaus has one technique; we have another. These mega-regions are the world's key economic centers. We have given them real identification and definition. And we will be revising, adding and tweaking.


Richard and I worked together in coming up with a light based approximation of the megas and I wanted to weigh in briefly on the RPA's mega work and the Texas Triangle issue.

Underscoring Richard's point in the comment above -- we think highly of the RPA's work and our light-based approach has its basis in the mega definitions that came out of their multi-factor approach (along with Lang's, etc.). We saw that we could identify essentially the same set of US megas using contiguous development as a proxy for integration and found that it worked remarkably well. Because the approach was so simple and objective, we could then extend the same approach to the rest of the globe. It would be really, really hard to extend the RPA's approach to the the globe -- yet this seems to be a scale where mega-regions are particularly useful abstractions.

With regard to Texas, it is worth noting that the RPA places Houston in _two_ megas -- the Texas Triangle and the Gulf Coast. It is the only city in the US that the RPA classifies as part of two megas. They have some very pretty maps here:


Check out the US map that spans pages 10 and 11.

Our emphasis on contiguity grouped Houston unambiguously with the Gulf, but the overall picture is not radically different. A Houston city planner would be foolish to start ignoring Dallas based on our maps, but nobody thinks they should forget that they are part of the Gulf Coast.


I'm familiar with the very nice RPA map and that they chose to put Houston in two megas, a compromise that I'm sure was hard for them to make. I respect your methodology, I just believe it's misleading in this case to split Houston off from the Texas Triangle. I think there are two more reasonable alternatives:

1) Merge the Texas Triangle and Gulf Coast into a single mega-region, which is not unreasonable at all if Minneapolis and Pittsburgh can be considered part of the same mega-region. Houston is really the linchpin between the two regions in the same way Chicago is between those two, or NYC is between Philly and Boston - any split just doesn't make sense.

2) Since Katrina and the depopulating of New Orleans, I think there's a fair argument for dropping the concept of a Gulf Coast mega altogether, since, without Houston, they are a string of relatively smaller metros that would not even come close to ranking on your list. If places like St. Louis, Kansas City, Memphis, Nashville, Las Vegas, and Salt Lake City/"Wasatch Front" (comparable to Denver) can be left out of regions, then I don't think Mobile, Pensacola, Gulfport, Baton Rouge, or NOLA have much of a case, relatively speaking.


Tory - Lots of folks around the world would have us re-jigger the methodology to jibe with their region's idiosyncrasies. But then it wouldn't be a systematic, replicable method, would it. And lots of places have ties to other places. But in this case, the Hou-orleans mega is just what the light pattern says it is. The Texas Triangle surely exists. It's just not what comes up in our light emission methodology.

The comments to this entry are closed.