Mega Debate

Paul Krugman has more to say on megas:

My basic view is that at a point in time the economic geography we see is more or less an equilibrium — that is, the individual costs and benefits of moving from one place to another are roughly balanced. Suppose that we see two regions, M (the mega) and N (a not so mega). M will have higher productivity and wages, offset by higher commuting costs and/or higher housing prices. An individual considering a move from N to M will take these into account, so that roughly speaking the two locations will be equally attractive.

Now we know that’s not the whole story. Adding to M’s work force will reinforce the positive spillovers from a large concentration of talent etc.. That’s the argument for encouraging more people to move to M.

But there are other considerations. For one thing, there are negative spillovers too, such as traffic congestion. For another, there are spillovers in N as well. For example, reducing the smaller city’s work force may deprive it of the critical mass needed to support some amenities, or maintain a cost-competitive position in some industries.

Do we know enough about these cross-cutting effects to be sure that moving people from Des Moines to Acelaland makes America better off? I don’t think we do. And as anyone who’s tried to study positive externalities knows, it’s very hard to pin them down.

Hmmmmm. Seems to me a good deal of recent research in economic geography suggests that regions are diverging. Berry and Glaeser find powerful evidence of human capital divergence over time. It is also increasingly clear that urbanization, in general, is an important component of productivity.  Careful studies of US-Canadian regional productivity and competitiveness by my colleague Roger Martin and the Institute for Competitiveness and Prosperity show that urbanization is a key component of the difference. Anyway you slice it urbanization is important to economic growth. (BTW, no one is suggesting "moving people from Des Moines to Aceleland," that's the way China and Russia did it: just that existing conditions are moving people worldwide to mega-regions and we need a public policy regime that takes that into account in various ways and across various dimensions).

Furthermore, the Jacobs-Lucas theory of human capital externalities provides the basic microfoundations as to why.  In his book with Fujita and Venables, The Spatial Economy the authors essentially admit that conventional economic models are unable generate the Zipf rank-size distribution of cities and regions. This is exactly the problem that Axtell and Florida, Emergent Cities takes on and generates using a basic Jacobs-Lucas framework to undergird agent-based models which show how clusters, then cities, then metros and then mega-regions form based on these human capital externalities. The Santa Fe Institute PNAS paper on urban scaling also shows how urban metabolic rates essentially speed up to offset "negative spillovers" like traffic congestion which Krugman mentions.

But let me get to the nub of the matter. Nations are political jurisdictions. Metros are constructed from data on commuting patterns and the like.  Mega-regions, as we define them, reflect an observable physical reality.  If an alien were to descend from space into the earth's atmosphere what would he see?  He would not be able to discern the world's 191 or so nation-states, nor would he be able to easily pick up metros (which blur into megas). He would see mountains and oceans in the daytime, and mega-regions at night. Mega-regions of all man-made geographic phenomena are a readily observable unit.

Furthermore, there is no logical reason to believe that the basic Jacobs-Lucas microfoundations for city formation and economic growth stop at the (arbitrary borders) of a cluster or metro, especially when we observe mega units in real space.  In fact, there is good evidence that suggests that those micro-foundational processes can and do scale from Axtell-Florida and the Santa Fe Institute research.

I'm the first to admit that there is a lot more to do on this issue. But the fact remains that mega regions produce the lion's share of the world's economic output and innovation.  Developing policy for a spiky world seems like a pretty big issue, and not just from a competitiveness point of view. Krugman is very concerned about inequality. As Who's Your City points out, the divergence between mega-regions and other places makes geography an increasingly important component of economic and social inequality.

Comments

[Michael Wells]

It seems to me that this cycles back to the original creative class theory. The Creative Class cities/regions/megas are going to have not just higher "productivity and wages" but more innovation, business creation, venture capital investment, human capital, etc. Krugman's argument starts with "all things being equal" the regional advantages will cancel out, but the point of Rise was that all things are not equal...and are getting more unequal.

This was borne out in your "where the brains are" map a year or two ago, of the increasing concentration of college graduates in the cities/megas. And further in Who's Your City? with the maps of higher levels of concentration going from population to economic production to patents to researchers.

I'm a real admirer of Krugman for his focus on inequality and the real-life effects of policy on real people, as opposed to the Milton Friedman theory-based economics that are unconcerned with people's lives. But I think he's being short sighted here, and it may be that behind the mega disagreement he doesn't buy the creative class theory and so describes the "old economy" world of mass manufacturing and people moving only for jobs.