Monday, February 01, 2010

Natural Experiments Of Rust Belt History

As a geography graduate student, I took data analysis with budding psychologists. Mastering research methods in another discipline is quite useful for crafting a clever line of inquiry. Furthermore, the professors leading the course trotted out one peer reviewed article after another and highlighted tragic flaws in the design. Mind you, this was published work. Burned into my psyche is the careful crafting of research questions.

Such a scholarly approach is central to Jared Diamond's new book, "Natural Experiments in History". Appropriately, a history professor interviews Diamond about the strength of the comparative method (i.e. natural experiments). For historians, the technique is blasphemy. For Diamond, it is wonderfully illuminating (see "Guns, Germs & Steel"). The conversation is an hour long and offers a number of fascinating insights.

I learned about the podcast from blogger David Campbell, who writes about economic development in Atlantic Canada. Campbell tries to employ Diamond's methodology and thinks about a natural experiment that could help explain why his region is chronically the poorest:

If you used his model you might find that Atlantic Canada might have been far more prosperous as a separate country. If you read Donald Savoie’s work, you see clearly that many of the leaders at the time in Atlantic Canada predicted this region would wither and be entrenched as the poor region of the country due to Upper Canada’s political domination. And it’s hard to deny that we have become the entrenched poor region of Canada in the intervening years.

To many folks this is an inevitable consequence of history. The Maritime region is in a bad geography. It has relatively little oil and gas (at least until now). It is physically far from the centres of power and control. Too bad. Every country has poor areas. Accept your destiny.

But I think that is too simplistic. If I had the cash, I’d get a guy like Diamond to look at it. Why didn’t Halifax become Boston? Why did this region (and I guess we can include northern Maine) stagnate while other areas boomed?

I'll start with the secession question. Ideally, you'd have a region very similar to Atlantic Canada that chose not to join the Confederation. Controlling all the other variables, you could trace the effect of joining Canada. Or, you could dig up a similar situation somewhere else in the world and see what happened.

I rather like the Boston-Halifax comparison. That got my wheels turning concerning an effective line of inquiry. The problem is the international border. The dominant effect would be national geography. That's less than ideal, but could suggest a few policy avenues.

A better design would be economic variance within Atlantic Canada. Such a geographic scope makes it easier to control more variables. The places compared should have the same advantages and constraints.

This can be applied to the Rust Belt. Sean Safford's "Why the Garden Club Couldn't Save Youngstown" is Diamond-esque. Youngstown (Ohio) and Allentown (Pennsylvania) are two similar cities with diverging economic fortunes. Given the research design, Safford is able to draw some useful conclusions about the effect of social network on economic development. To offer a critique, I wonder about the effect of Allentown's proximity to NYC. Might two similar cities in NE PA provide a more rigorous comparison?

Settling on a common geography isn't easy, at least as far as academic standards are concerned. I think blogging about research design is a good solution. I've floated the following idea a few times concerning the Sun Belt advantage. How do Rust Belt cities located in the Sun Belt compare to Rust Belt cities located in the Frost Belt? My hypothesis is that the diverging economic fortunes of the two regions is mostly a result of brownfield versus greenfield development opportunities. In other words, the difference is in the legacy costs.

I remain unconvinced that policy in Sun Belt states explains much of anything. Using Diamond's methodology allows for easy dismissal of the promoters of tax reform. Why else would Pittsburgh be outperforming Sun Belt industrial cities such as Birmingham? In this regard, the Wendell Cox production "Empire State Exodus" is just awful. The conclusions are not to be taken seriously.

When I discuss brain drain, I tend to use the comparative method. Relatively speaking, just how bad is the out-migration of talent in Northeast Ohio? Using net-migration data torpedoes the analysis of retention policies. What you are measuring doesn't match your research question. Joe Cortright made this mistake when addressing his Akron audience. Doing so undermined his recommendations. Such oversights are common and I suspect most of them are unintentional. As Diamond contends, and I would confirm, graduate student training is hit-or-miss. You needn't be an expert in research design to get tenure or make a lot of money as an economic development consultant.

No comments: