Saturday, October 22, 2011

Felix Salmon is wrong about congestion pricing

On the way to school the other day, fellow 4:17 A.M. contributor Charles and I were talking about congestion pricing; specifically, this Felix Salmon post about why it will never be popular. Here's the heart of his argument:
When congestion pricing is first introduced, people recoil against it — they expend quite a lot of effort to avoid the charge, and traffic goes down. Over time, however, it becomes just another part of the cost of driving, along with gas and insurance and parking tickets. As that happens, traffic goes back up again. Congestion-charge revenues go up too, of course, and those can be reinvested into public transport....

In Singapore, they set the amount of traffic they want, and then dial up the congestion charge until they get it. It’s much the same idea as the one behind SFPark: you set the number of empty parking spaces you want, and then dial up the parking-meter pricing until you get there....

[T]he charge has to be variable over time — specifically, it has to increase over time. Without those steady increases, drivers become inured to the congestion charge, and traffic will go back up to its former level.

As a result, drivers are pretty much never happy with congestion pricing. Either it’s painfully expensive and going up in price — expensive enough to keep them from driving — or else it doesn’t have much effect.

That doesn’t mean that congestion pricing isn’t good public policy. It is. But it’s always going to be unpopular with a powerful constituency.
Like Felix, I'm in support of congestion pricing. But I don't really buy his analysis here. If you unpack what he's saying, he's really making three assertions: one I'm sure he's right about, one I'm sure he's wrong about, and one I'm unsure about either way.

So let's look at it a little more closely.

His first argument is that the correct way to use congestion pricing is determine the optimal level of traffic and set prices so that level is achieved. This is clearly correct. There wouldn't be much point in implementing it otherwise! The whole idea of congestion pricing is that road space is a scarce good, and when it's given away for free, is overconsumed, creating a shortage (and gridlock).

There is, of course, the libertarian objection -- that the government shouldn't be arbitrarily setting the price of goods, but instead, goods should be priced by achieving an equilibrium of supply and demand. But that's a framing issue more than a real difference of opinion. Congestion pricing isn't actually any different than any other market mechanism -- instead, the problem is that the supply is virtually fixed in most dense urban areas, and therefore the only way to achieve the most efficient use of available road space is to adjust prices.

(The one real question here is what "the most efficient use of available road space" actually means -- high-traffic, near-gridlock conditions, or roads that are well-trafficked but quickly navigable. The answer depends on exactly what sort of good the city wants to provide when it provides roads. Once again, I'm sure this will raise accusations of improper government meddling in the marketplace -- but it's worth noting that the government makes exactly the same decision anytime it builds a road, congestion pricing or no, and metering simply allows it to more finely target particular goods.)

Then there's his second assertion:
But the point in all of these cases is that the charge has to be variable over time — specifically, it has to increase over time. Without those steady increases, drivers become inured to the congestion charge, and traffic will go back up to its former level.
In economic terms, this is sort of bonkers.

First, a caveat. It's totally conceivable that, in the short term, the introduction of congestion pricing would have an irrationally large impact on driving habits. People are excessively fee-averse, and I'd fully expect a widespread overreaction the introduction of driving fees. In other words, the initial reduction in city mileage would be the product of both a changed market equilibrium and the cognitive biases of potential drivers.

But as the fees became more widely accepted (as drivers became inured to them, if you will), the amount of traffic in the city would also more closely reflect the new equilibrium. If the traffic increases past an acceptable level, the fees were too low to begin with (possibly because the initial overreaction deceived the fee-setters).

I am very confident, however, that drivers becoming inured to congestion charges would not result in a perpetual increase in the amount of urban traffic, up until the point where roads were at maximum capacity and traffic literally couldn't get worse. That's not how people work.

I suspect Salmon has confused himself because it's hard to think of road space as a normal good. Traditionally, after all, it hasn't been treated as one. There was no easy way to exclude someone from a city street and as a result, we think of people's demand for access to city streets as insatiable.

So let's use an analogy instead -- something that is incontestably a normal good --let's say, bananas. The process that Salmon fears has already played out with bananas. You and I have become "inured" to the price of bananas in the sense that we've accepted that bananas cost what they cost. We don't go out of our way to avoid buying bananas because we feel they're unfairly priced. So are producers forced to continually raise prices to keep us from buying up all the world's bananas? Well, uh, no. None of the negative consequences Salmon predicts have manifested themselves here. We don't regularly go to the grocery store and clean out its banana stock. Our daily banana consumption doesn't perpetually increase. There are not widespread banana shortages. Somehow, society has managed to simply stave off mass banana consumption without penalizing consumers with constant banana price hikes.

Described in terms of bananas, the idea that people will become unaccountably "inured" to prices seems ridiculous and obviously wrong. But that's what Salmon is saying would happen to road use. Why? There's simply no reason to think that road use would be treated any different in the marketplace. People might not take undue caution to avoid driving, but they're certainly not going to completely ignore the cost of driving either. Some number of drivers will opt to stay home, walk, or take the bus, and if congestion prices are higher, that number will be higher as well. That's what a market equilibrium is.

Really, only three things should drive up adequately priced congestion charges over the long run: a reduction in supply (i.e., less road space is available for some reason), an increase in demand (i.e., more people want to be in cities, or want to drive), or a decrease in the value of money (i.e., inflation). If congestion continually increases even in the face of congestion, you can safely assume one of these three things is the cause. (Perhaps somewhat ironically, successful congestion pricing should result in a one-time increase in the demand for road space, since driving would be a significantly more pleasant experience.) On the other hand, if people simply stop caring about paying the fees, that means the fees were too low to begin with, any initial objections to the fees were essentially irrational, and any fee increase would just represent a move towards a stable price equilibrium.

Finally, there's his third assertion:
That doesn’t mean that congestion pricing isn’t good public policy. It is. But it’s always going to be unpopular with a powerful constituency.
I can definitely see the argument for this. There's no realistic way to create congestion pricing that isn't going to be subject to public pressure. And the public, in general, is always going to favor lower prices. (The same would be true of any good whose price was set by the public at large, rather than private entities.) As Salmon suggests, "I'll lower the cost of driving!" is a pretty obvious promise for aspiring elected officials to make. It's impossible to create a congestion pricing scheme without also creating the opportunity for indirect vote-buying.

With that said, there's a difference between a possibility and a certainty. There are already any number of small, annoying fees associated with the administration of local government, and they're all subject to the same dynamic. Even if sometimes they result in the fees being arguably lower than they should be, you don't see municipal governments falling over themselves to abolish, say, utility fees. Nor do you see widespread discontent with the concept of having to pay for utilities. Besides, governments have gotten pretty good at segregating fee-supported public entities from tax-supported public entities, and in the process have done a mostly acceptable job of protecting fee-supported entities from the vicissitudes of the electoral process. In the end, it's very hard to know how people will respond to something like congestion pricing over the long-term, and given the straightforward economic logic behind the idea, any number of reasons to pursue it.

No comments:

Post a Comment