I recently discovered Jane Jacobs’s legendary book, The Death and Life of Great American Cities. I think there are a lot of interesting parallels between the arguments she makes about urban planning and the issues we argue about in technology policy. (one example that I’ll save for another post: her prescriptions for healthy cities are almost all oriented toward increasing the power of network effects generated by peoples’ proximity to one another. My previous post about non-money-mediated markets reminded me of the chapter in Jacobs’s book about congestion.
Jacobs lived in New York when she wrote the book in 1961, and she described a political battle over whether to widen a road that ran through a large park in the city. The city planners, citing congestion problems, wanted to widen it and turn it into an expressway. Jacobs, in contrast, says that local activists, after beating back this proposal (which would have split the park in half and generally made life miserable for the locals) succeeded in getting the road closed entirely. The officials predicted mayhem in adjacent streets, as all of those extra cars re-routed into the side streets. Yet, according to Jacobs, nothing of the sort occurred. If anything, traffic on nearby streets actually declined.
This seems startling at first glance, but it becomes obvious what’s going on once we apply a bit of economic reasoning. Congestion is essentially a rationing mechanism for over-used city streets. In an ideal world, we’d charge each driver for using the roads, and charge the market-clearing price, but because of administrative hurdles, that’s not possible. So instead, people “pay” for the resource with their time. Waiting in traffic is effectively the “price” of using the road.
Viewed from this perspective, it’s obvious why adding new roads won’t necessarily decrease congestion, and why closing roads won’t necessarily increase it. Adding new capacity reduces congestion in the short run, but that has the effect of reducing the “price” of those roads. If the demand for roads use is highly elastic with respect to waiting times, commuters will respond to this price cut by consuming more of the resource. Similarly, closing a road increases congestion in the short run, which raises the “price” and (assuming high elasticity) will reduce consumption. (This logic only applies in a city like New York where there are good substitutes for driving. In a more car-oriented city like Indianapolis, demand for roads might be inelastic, and adding more capacity might solve the problem)
OK, so what does this have to do with my previous post? Transportation departments and Internet content creators face a similar problem. Both would like to find some way for the consumers of their product to pay the producers. But in both cases, high transaction costs make collecting cash payments impractical. So each is forced to find alternative payment mechanisms.
But roads and web sites have an even more important difference: the “pricing” mechanism for roads is wasteful: when a motorist “pays” with his time, no one receives that payment–it’s simply wasted. Congestion is a “negative sum” pricing system, in the sense that the value of what’s paid by the buyer exceeds the value of what’s received by the seller (namely, nothing). In contrast, the non-monetary Internet-based markets I described in my last post use “positive sum” pricing schemes. The cost to the payer of providing the “payment” (bits, attention, code) is small relative to the value of that payment to the recipient. Money is a zero-sum pricing scheme: everyone values a dollar about the same amount, so wealth is neither created or destroyed when dollars change hands.
It’s a good idea to convert markets with negative-sum transactions into markets with zero-sum transactions. Hence, for example, charging congestion tolls on highways converts a negative-sum transaction into a zero-sum transaction: drivers have to pay with their wallets rather than their time, while the state gets revenue it wouldn’t have gotten previously. By the same token, it’s a good thing when zero-sum transactions are converted into positive-sum transaction, as when AOL replaces zero-sum subscription fees with positive-sum advertising.
To be clear, I’m not saying it’s always possible to do this. There probably isn’t any positive-sum pricing mechanism that can keep Ford’s workers on the assembly line. And so far, at least, no one has figured out a way to charge tolls efficiently enough to eliminate all congestion. But given a choice between a positive-sum pricing mechanism and a zero-sum pricing mechanism, it’s nuts to prefer the zero-sum one.