Posted by Daniel Hall on January 16, 2008
Evan Sparks is excited over at his Aviation Policy Blog about the FAA proposal to allow congestion fees at airports. And well he should be. Congestion fees are a sensible way to help efficiently allocate runway space at high-traffic airports. Evan’s post has thorough coverage of the proposed amendments. Mike Giberson over at Knowledge Problem also has a great post discussing the economics of the proposal.
Despite his general enthusiasm, Evan sounds ambivalent about a section of the amendments that allows a congested airport to include in its fees a portion of the landing costs from some nearby (uncongested) airport. He ends up rationalizing it, however, by invoking one of my previous posts:
I’m curious about this proposal, which means that at least some of the additional fees paid by airlines flying to JFK (run by the Port Authority of New York and New Jersey) at busy times can go to support another PANYNJ airport, say, sleepy Stewart International at Newburgh, sixty miles north of Manhattan. I thought a good side benefit of congestion charging would be plowing those resources back into improving capacity at the congested airport. This is especially important because of the hub-and-spoke system: the airlines with the most flights at, say, JFK–Delta, JetBlue, and American–can’t move flights to Stewart because that goes against the whole idea of a hub system, which remains a very efficient air traffic model. But then again, my friend Daniel Hall at Common Tragedies has urged using roadway congestion charge revenue to cross-subsidize mass transit, so perhaps the FAA’s second point is valid after all.
I think Evan’s initial reaction may have been right here. It sounds like there are a couple differences between these two situations that weaken the case for a cross-airport subsidy. The key is to think about why it makes sense to use roadway charges to cross-subsidize transit.
First, it is economically rational to subsidize mass transit systems in general. I make this argument qualitatively in the post referenced above:
Mass transit systems are a network good with positive externalities — beyond reducing congestion, buses are a far less polluting alternative than cars — and hence should be subsidized to achieve the socially optimal level of provision.
A strong analytical case for this proposition has also be made: in an earlier post I looked at the results of a recent discussion paper from Ian Parry and Kenneth Small. They show that due to economies of scale, declining marginal costs of provisions, and the benefits of reduced congestion, mass transit systems should be heavily subsidized in order to maximize social welfare. As the authors conclude:
The results support the efficiency case for the large fare subsidies currently applied across mode, period, and city. In almost all cases, fare subsidies of 50 percent or more of operating costs are welfare improving at the margin, and this finding is robust to alternative assumptions and parameters.
(Lots of fascinating data and analysis in that paper for the interested.)
In contrast, it isn’t clear to me that there is a strong economic rationale for subsidizing uncongested airports. If the congestion fees at the busy airports push a few flights at the margin over to other airports, great. But unless there are other externalities involved there shouldn’t be a need to actively subsidize landings at these less used airfields.
The second reason I proposed the cross-subsidy is that driving and mass transit are roughly substitutes. (This is particularly true for many urban rush-hour commuters.) Since mass transit needs to be subsidized from somewhere, and since roads become over-utilized when given away for free, you solve two problems at once with the cross-subsidy. Further, you won’t have to raise congestion charges as high if you simultaneously make mass transit easier to use.
In contrast, Evan’s discussion makes it sound as if these uncongested airports are typically very poor substitutes for the busy ones. I don’t know nearly as much about the air transit system as he does, so I’m taking his word for it here. But if he’s right, and the hub-and-spoke model is really far more efficient than using smaller airfields further out, then the cross-subsidy may not help much.
I actually wonder of there are opportunities here to subsidize other transit options that might be closer substitutes, such as regional rail. It sounds as if one of the major problems with congestion at airports is small jets — current fees are based on aircraft size, but since small jets take essentially just as much landing space and time as large jets, they are getting an effective subsidy that is worsening the congestion headaches of the passengers in much larger planes. What are these small jets? Are these private corporate jets, or are there a large number of short-haul commercial flights which could be replaced by an efficient regional train network?
I’ll actually be headed up to New York from DC tomorrow on the train. This is currently one of the few routes in the U.S. where the train is a truly competitive option to a flight. (Just under three hours from station to station, which is not too bad in comparison to a flight once you start adding up the time to arrive, board, taxi, take-off, fly, land, taxi, and deplane.) But on many other routes train service is either too infrequent or too unreliable to serve as an effective alternative. It is not hard to imagine that a well-managed high-speed train network could reduce some of the pressures on airports from shorter flights (say, 500 miles or less).
The discussion above is intended to highlight the reasons behind my support for a cross-subsidy from roads to mass transit . Turning to the more immediate question of the proposed FAA amendments, while I agree with Evan’s first intuition that using these congestion fees to subsidize landing fees at other airports may not do a lot of good, I suspect that in reality they won’t much get used this way. This is because the congestion fees can also be used to construct new capacity, and even if they are used to subsidize other airports there are restrictions. As Mike Giberson explains in his post:
One change is a proposed ability to add the costs of facilities under construction into current rates (at present airports are allowed only to charge for facilities in use). The second change would allow airport authorities operating multiple airports to shift some costs from uncongested regional airports into the fees charged by the authorities congestion airports. This second proposed change is subject to several limits generally intended to insure that users of the congested airport can benefit from the shift in traffic expected to follow a shifting in costs.
Finally, before I close, one lexical issue: there is a difference between a ‘congestion charge’ and ‘congestion pricing’. Here’s Evan, with a quote in the middle from the FAA:
There’s a section on pricing that strikes me as disingenuous:
The proposed actions do not represent true congestion pricing because they do not authorize airport proprietors to set fees to balance demand with capacity without regard to allowable costs of airfield facilities and services.
Putting it differently, this rule authorizes a change in pricing to ameliorate congestion. Sounds like “congestion pricing” to me.
Ah, but if you read that exact statement from the FAA carefully you’ll see that it captures the difference. As Mike Giberson explains:
As explained in the notice issued by the FAA, the proposed policy change will not allow true congestion pricing, because they will not allow airport authorities to charge prices sufficient to balance demand will capacity without regard to “allowable costs of airfield facilities and services.” Instead, the FAA is proposing to allow airports to re-shuffle currently allowed costs in ways which reflect congestion at airports.
I suspect the FAA fears that the true congestion prices that would be necessary to actually balance airline demand with airport capacity at peak travel times would be so high as to cause public outrage. They’ve therefore settled for a congestion charge (capped by regulation, i.e. “allowable costs”) that should ameliorate congestion but which won’t necessarily eliminate it.