Wednesday, March 26, 2008

Geek Toys

I love numbers. I love the Census Bureau's Website. I love Adobe Readers' Snapshot Tool. I'm a fan of Google Applications.

I put the following together to tell a story about the unique characteristics of the 12th Congressional District, but if you're interested in any part of Ohio, you might find this interesting (darker always means higher/more, range is calibrated for Ohio).

The full size slides are available here.
This data was pulled from the Geographic Comparison Data in the 2006 American Community Survey as accessed through


Paul said...


This is very interesting stuff. I'm a visual learner, so this is a very helpful way to represent the data for me.

The most obvious thing is how often central Ohio shows up at the top of the scale for the good things (e.g. education). But I think there's something nefarious hiding underneath.

I've had the opportunity to participate in a group called Greater Common Good that is seeking to make recommendations to government policymakers regarding the economic health of Central Ohio.

My observation to them is that we seem better than we actually are. We talk about growing when the rest of Ohio is declining. We talk about the high fraction of advanced degrees and professional jobs. All those numbers sound pretty positive.

The economic health of a region is largely defined by how successful it is in 'importing wealth.' By that I meant the ability to export products and services to places outside the economic zone, which of course brings money in the opposite direction. 'Export' does not have to mean internationally, just outside the economic region (e.g. exporting to Cleveland and to Moscow create the same economic benefit for central Ohio).

Some regions of the country have become powerful exporters in this way. New York has been that for 100 years (with their primary product being financial services, but also manufactured goods). It used to be cities like Buffalo, Cleveland, Pittsburgh and Detroit. Today, it's cities like Seattle, San Diego, San Francisco, Atlanta and Charlotte.

Where does Columbus fit in that landscape? For one thing, notice that none of those cities I list as today's exporters are the capital of their state.

But Columbus is. I'd argue that, while we do have some powerful exporters here like Nationwide Insurance, The Limited, and AEP, a primary cause of the cash flow into central Ohio is our position as the state capital. In support of that statement, let's observe that on any list of major employers in central Ohio, the State of Ohio is at the top of the list. Add to that the various trade and professional associations which are headquartered here for access to legislators.

We have lots of people with advanced degrees and professional jobs partly because of OSU and Battelle, but perhaps most of those advanced degrees are from law schools and schools of public administration, and the professional jobs we're talking about are lawyers, legislators, and policymakers. Could be that our profile is just a miniature version of Washington DC.

So what?

As a dedicated free-market capitalist, what that means to me is that we don't so much earn our money in central Ohio as it is we take it in the form of taxes collected all over the state and dispersed as payroll here in central Ohio. The association dues and management fees collected by organizations such as the Ohio Education Association and the Public Employees Retirement System (who have a couple of the larger non-government buildings downtown) are much the same. I doubt that either would be here if we weren't the capital.

You don't have to be very good to take money - only powerful.

The other things the new exporters have in common? They're actually interesting places to live. No matter how nice of a dress we put on the old girl, central Ohio is still just a flat piece of farmland with not much of interest. I'm sure others have different perspectives, but let's look at the objective measure - how much tourism money gets spent here? Besides 6-7 OSU home football games each year, the Arnold Classic, and the Quarterhorse Congress (has that relocated already?), what draws people here? How many people tell their kids - "hey kids, we're going to Columbus for Spring Break!!" The line I tell people is that Columbus is a great place to live, but you wouldn't want to visit here.

So we can be fooled into thinking we're a dynamic region that people flock to. I think the truth is more that the money thrown off by government operations is the primary driver of economic stability here, and otherwise we're not very good at nurturing private sector growth.

And that's a shame, Ohio State should be an exceedingly powerful engine. They're trying, with things like the Business Technology Center. If they get the right amount of investment capital in the mix, like the $billions of VC money that floats around San Francisco, San Diego and Seattle, it could pay off.

Meanwhile, a major growth industry in central Ohio for 20 years has been homebuilding. Unfortunately, the macro effect of this industry is to transfer money within our community, not bring in new money. It fact, it probably causes a net money flow out of the region for building materials. It has also drawn in thousands of immigrants, primarily Hispanic, as construction workers. (by the way, my feeling is that by itself, immigration to fill available jobs is a good thing)

But notice that the homebuilding industry was created with government help as well. The result of the late 70's busing-for-desegregation ruling and the Win-Win Agreement was to instigate the White Flight to the suburbs that created the boom for the homebuilding industry. The consequence is a hodgepodge of school districts with unsustainable economics. But that's another story.

With the sharp downturn in new home construction, many of those immigrants are now unemployed, adding to the social services burden of a region that already has a substantial population living in poverty. Just as they came here seeking jobs, many are leaving to find work elsewhere. But some are staying to see if they can weather the downturn, again increasing the burden on us all.

If it were my decision how to best distribute tax dollars to improve Ohio's economics, I'm not sure central Ohio would get much of that money. We'll be alright as long as we can keep grabbing those tax dollars from our fellow Ohioans.

But if you live on tax revenues, you'd better be sure there are taxpayers. I think I would help other regions crank up their economies, and I start with Cleveland and Cincinnati because they are - or have promise of being - pretty interesting places to live if you need to attract the best and brightest.


bonobo said...


I know I have ben really lousy at responding to comments recently, and this isn't going to be much of an improvement...

I think you're right to some extent about the role of public institutions in keeping Central Ohio from sliding as fast. I also think you're right when you say that Central Ohio can't sustain an isolated economy.

On the other hand, I tend to think that what attracts people to a community is the net attributes of the people already there. 30 years ago Seattle was that dreary place where it rained all the time, San Jose was the opposite of Urban Hipness ("Do you know the way..."). It only takes a couple of seeds sprouting for a forest to erupt.

And those seeds may or may not be the ones you plant. Seattle wouldn't be the same without college dropouts...

Anyway, you would probably be interested in a slide I didn't include, because it's difficult for me to analyze. This is percent privately employed (private for-profit, private non-profit, and incorporated self-employment)

Paul said...

Yeah, that's an interesting chart. Are farmers included in this set?

The range of data is only six percentage points, but I trust that you selected this range because it's were the significance lies. But I wonder what this looks like with raw counts, since 1% in NE Ohio might be around 30,000 individuals, while 1% in central Ohio might be more like 15,000.

You're also right that a number of factors play into making an area attractive for economic development. San Jose may have been the northern California version of Bakersfield, but the Silicon Valley can serve as a place for business, education and research (although tons of people live here) while the region surrounding it includes some of the most spectacular areas of the country (Lake Tahoe, Redwoods, Pacific, Monterey, San Francisco itself).

Gates founded Microsoft in Seattle because that's his hometown (not a bad place to live either - more days of sunshine than Columbus I've heard?). We got some of that local-entrepreneur-does-good here: The Limited, Wendy's, White Castle, NetJets, Cheryl's Cookies are some of the best known names, but there are others. But we've also lost some significant ones: Bank One, CompuServe, to name two.

Columbus isn't a bad place - in fact it's a nice town to call our home and to raise a family. But let's not fool ourselves into thinking we know how to do economic development effectively just because our region is growing in population and the demographics look pretty good.

We know how to build houses (and create urban sprawl) and schools (but not how to fund them) and shopping centers (but tear them down as fast we build them).

But what are we doing to create middle-class jobs that can be had without a college degree - jobs that employ the young people trapped in the cycle of poverty in our inner city?

Not enough. In Good to Great Collins observes: "The enemy of Great is Good." I think that's us.

bonobo said...


We're, as usual, mostly on the same page.

Because you bring it up, I will expand a bit on the technical elements of the charts:

1) The maps are actually created by the Census Bureau's GIS in response to a user query. The user doesn't get to choose the range, it appears to be calibrated such that the minimum observed value is the lower bound, the maximum observed value is the upper bound, and the categories are derived by creating equal sized intervals in between those two points. If some values are so different from the rest that they would result in a distribution that could only use the highest and lowest categories, those categories are split off, and the intervals divide up the rest of the districts.

2) The unit of analysis is the Congressional District, and CDs have been gerrymandered specifically to heighten some of these differences and to smooth over others.

3) None of these differences can be assumed to be significant. One runs into some sticky decisions when trying to apply a significance criterion in the context of multiple simultaneous comparisons. What you can do, if you go to the census website yourself, is add a statistical significance layer to the maps. You then click on the district of interest, and the map will distinguish between the districts that are not significantly different from the CD you clicked on, and those that are significantly different based on a pairwise comparison.

I decided against using that feature as I realized pretty quickly that these maps would be of interest outside the 12th CD, and that I'd rather not give an indication of statistical significance than to fret over and defend the arbitrary method available. If I could link directly to the dynamic maps, I would, so that everybody could click and play, but all I can do as of this moment in time is down load static PDFs of the current map state (The Census Bureau continually upgrades their functionality while I continually upgrade my competence in using the site... eventually I imagine I'll be able to do anything I want with any data I could desire, effortlessly. Not yet, though.)

Paul said...

Thanks for the additional explanation. I too am a data geek, especially data well presented visually. Hope you find time to do more of this kind of stuff.