When the definition of “city” – at least on an administrative level – varies so widely from country to country, how can you draw any meaningful comparisons on topics like density, health, and wealth between cities in different parts of the world?
A new paper posted on LSE Cities takes a stab at at addressing that problem in a straightforward but ingenious way. If you’ve got a number of incompatible ways to measure the extent of urban areas, why not just throw them out and create a new, consistent standard from scratch?
That’s pretty much what Antoine Paccoud, the author of the paper, did. He writes that his objective is to define
a new type of geographical unit, which I will now refer to as the extended metropolitan region (EMR), which can be defined as the administrative unit or combination of administrative units which contain(s) the largest spatial extent of the city yet do(es) not exceed the population of the urban agglomeration by more than a factor of 2.
The “urban agglomeration” he refers to is actually a UN-calculated figure that – you know what, just go and read the paper itself if you really have to know. For those of you who just want the executive summary: from what I can tell, the population for American “urban agglomerations” usually seems to be in line with that of the corresponding metropolitan statistical area (MSA). It’s certainly not exact, but it’s close enough that you can safely use that comparison as a rough guide for understanding that term.
I like what Paccoud has done here. He intentionally takes an aggressively expansive approach to defining EMR, and as a result, population for any given EMR is going to be significantly higher than the corresponding MSA. For example, according to the Census Bureau, Atlanta’s MSA has about 5.3 million people in it. But according to Paccoud, the Atlanta EMR has 7.5 million people. That’s a difference of 2.2 million, which is not insignificant – in fact, it’s roughly the same population as Hong Kong’s EMR. But Atlanta’s people are spread out in an area that is about 25 times the size of Hong Kong. Of course, Atlanta dosn’t face the same geographic constraints Hong Kong does, so there’s no reason to expect it to be as compact.
Paccoud uses the EMR concept to compare the health and well-being of residents in 129 EMRs, using the UN’s Human Development Index (HDI). I think this application of the EMR concept shows a lot of potential. However, I did find one flaw in his methodology that bothers me. Specifically, Paccoud bases the true extent of the built-up areas of American cities on congressional districts. I don’t think that’s the appropriate geographical measurement for a project like this because of the way in which district boundaries are drawn. As you may know, congressional district boundaries are set every ten years by state legislatures. The party that holds a majority in the state legislature draws congressional district boundaries in such a way that their party’s chances of winning the majority of congressional races in that state will be maximized.
The result is that Congressional district boundaries have basically no relation to patterns of urbanization. Just look at some of these ridiculously-shaped districts in this Slate slide show. Hell, one of those slides is a map of Florida’s 23rd district, which just so happens to be included in Paccoud’s formulation for the Miami EMR. I’ve been to the west side of FL-23. I can assure you that it is neither urban, nor does it have much at all to do with the city of Miami.
Of course, this is a perfectly reasonable error for a European researcher to make. There’s no reason to expect Paccoud to necessarily be familiar with the arcane tendencies of our archaic and ridiculous electoral system. I would just suggest that he rethink how he defines American EMRs – perhaps Census tracts would be a better choice.