This post was inspired by Nolan Gray’s “Jane Jacobs’ Hayekian Critique of Urban Planning” and the discussion it recently sparked over at Strong Towns.
In Jane Jacob’s Hayekian Critique of Urban Planning, Nolan Gray argues for the futility of trying to master plan something as complex as an entire city. And he’s right. The last century’s Corbusian fantasies overwhelmingly ended in failure. And, in what’s a very even-handed article, he goes on to make room for some amount of centralization where decentralized planning just seems to break down. He’s right on that point as well.
But after reading Mr. Gray’s article and the discussion it sparked over at Strong Towns, I think we can take the conversation a little farther. Instead of a binary choice, we should be speaking in terms of a spectrum with centralization and decentralization on opposite, theoretical ends. Once we think in those terms, we can approach questions of planning as questions of determining what issues are best addressed at what scale (individual, neighborhood, district, municipal, regional, etc).
Looking at India’s Gurgaon or China’s Wenzhou, we can see how hard it is to produce certain kinds of city-wide infrastructure through decentralized market coordination. In the specific areas where relatively decentralized coordination produces sub-par results, it makes sense solve problems via a single entity with responsibility for an entire urban area.
The point here is not that municipal or regional planners are ever better at confronting knowledge problems than market participants; it’s that, in some places, transaction costs render decentralized coordination nearly impossible, so the potential benefits of better-leveraged local knowledge never even have a chance to appear. In these situations, any plan is better than no plan, and below a certain scale, no plan is what we end up with. And where transaction costs aren’t too high we should eschew with the Fordist style micro-management and let order organically emerge.
Ultimately, what we want for any given issue is that sweet spot where social coordination is small enough in scale to cope with knowledge problems, but large enough in scale to overcome transaction costs. And we want to recognize that this sweet spot will be very different for different things. And that instead of thinking in terms of central planning vs. decentralized coordination, we should recognize a scale of relative values and frame our thinking and our debates accordingly.