The difficulty of unintended consequences – airlines, HSR, and deregulation

Pittsburgh International Airport - CC image from Fred

Philip Longman and Lina Khan make the case for re-regulating America’s airlines, claiming that deregulation is killing air travel and taking de-hubbed cities like St. Louis with it (hat tip to Matt Yglesias).  The authors do indeed present compelling evidence that airline deregulation has indeed shifted the economic geography of many cities in the US – but as Matt Yglesias notes (channeling the aerotropolis thesis), in many cases this is merely an example of the air travel network’s ability to emphasize agglomeration economies:

They observe that… once the imposition of market competition caused some medium-sized midwestern cities to lose flights, the per flight cost of the remaining ones went up. That tends to produce a death spiral. Eventually the market reaches a new equilibrium with fewer, but more expensive flights. Except that equilibrium tends to drive businesses out of town. And once Chiquita leaves town, Cincinnati will have even fewer aviation opportunities which will further impair the business climate for the remaining large companies in the city.

This is a great concrete and usefully non-mystical illustration of agglomeration externalities.

Yglesias argues that fighting these agglomeration economies is counter-productive, but that’s not the only flaw in Longman and Khan’s thinking. Using the example of Pittsburgh, where the America West-US Air merger meant PIT losing hub status, they cite examples of the problems this represents for business travel:

K&L Gates, one of the country’s largest law firms, used to hold its firm-wide management meeting near its Pittsburgh headquarters, but after flying in and out of the city became too much trouble, the firm began hosting its meetings outside of New York City and Washington, D.C. The University of Pittsburgh Medical Center, the biggest employer in the region, reports that its researchers and physicians are increasingly choosing to drive to professional conferences whenever they can. Flying between Pittsburgh and New York or Washington can now easily take a whole day, since most flights have to route through Philadelphia or Charlotte. A recent check on Travelocity showed just two direct flights from Pittsburgh to D.C., each leaving shortly before six in the morning and costing (one week in advance) $498 each way, or approximately $2.62 per mile.

The problem is that Pittsburgh to New York and Pittsburgh to DC aren’t all that long as the crow flies.  Longman and Khan explain why that’s problematic, thanks to those pesky laws of physics:

One reason this business model doesn’t work is that it’s at odds with the basic physics of flying. It requires a tremendous amount of energy just to get a plane in the air. If the plane lands just a short time later, it’s hard to earn the fares necessary to cover the cost. This means the per-mile cost to the airlines of short-haul service is always going to be much higher than that of long-haul service, regardless of how the industry is organized.

Indeed, part of the economic logic of the airline hub was to ferry passengers to the hub via loss leader (or, hopefully, less profitable) short-haul routes so that they can then use the more profitable long-haul services – transcontinental and international flights, and the like.  The problem is that Longman and Khan can’t see beyond the end of the runway.  We have a transportation technology that has a different economic calculus, one that works well for those shorter trips up to about 500 miles – high speed rail.

This isn’t to counteract Matt’s first point – just because HSR can make travel time competitive with air travel over such distances does not mean building it will be cost-effective, but the broader point is about the need to think beyond the modal silos.  Current rail service from Pittsburgh to DC and New York isn’t time-competitive with flying, even with those connecting flights.  But HSR could be. Indeed, given the current economics of the aviation industry, HSR ought to have a larger role in key corridors.

Indeed, Longman and Khan do consider rail in their article, but they pick out the history of railroad regulation instead:

 By the 1880s, the fortunes of such major cities as Philadelphia, Baltimore, St. Louis, and Cincinnati rose and fell according to how various railroad financiers or “robber barons” combined and conspired to fix rates. Just as Americans scream today about the high cost of flying to a city like Cincinnati, where service is dominated by a single carrier, Americans of yesteryear faced impossible price discrimination when traveling or shipping to places dominated by a single railroad “trust” or “pool.”

This, more than any other factor, is what led previous generations of Americans to let go of the idea that government should have no role in regulating railroads and other emerging networked industries that were essential to the working of the economy as whole.

The problem with applying this logic to the current airline situation is that the railroads of the turn of the century didn’t just have a monopoly over a given town as the sole operator of service along the line, but they had a monopoly on the very technology that could offer such increases in mobility.

That technological mobility is no longer the case.  The excellent Mark Reutter article The Lost Promise of the American Railroad (now behind a paywall) documents the many reasons for the decline of American rail, including new competing technologies (both air travel and cars taking away long distance travelers as well as commuters), outdated regulations (such as WWII era taxes meant to reduce unnecessary travel during the war – and were quite successful at doing so – that remained in place until the mid 1960s), direct subsidization of competitors by the government (see taxpayer funded highways and airports, in the face of largely privately financed and taxed rail assets), and differing regulatory regimes.

The regulations present a compelling story.  The original regulations, as noted by Longman and Khan, were devised in an era before heavier-than-air human flight had even occurred – yet alone before the rise of commercial aviation.  Yet, the regulations devised by the Interstate Commerce Commission (formed in 1887) were the basis for a portion of the blame for the decline of American rail less than a century later.  Longman and Khan defend the need to regulate, despite these shortcomings:

To be sure, any regulatory regime can degenerate and wind up stifling competition, and the CAB of the late 1970s did become too procedure bound, ruled, as it came to be, by contending private lawyers rather than technocrats. It would have helped, too, if the country had not largely abandoned antitrust action after the Reagan administration. But even strong antitrust enforcement wouldn’t have helped that much, because airlines— just like railroads, waterworks, electrical utilities, and most other networked systems—require concentration both to achieve economies of scale and to enable the cross-subsidization between low- and high-cost service necessary to preserve their value as networks. And when it comes to such natural monopolies that are essential to the public, there is no equitable or efficient alternative to having the government regulate or coordinate entry, prices, and service levels—no matter how messy the process may be.

While this can be a compelling case for the need for regulation in the abstract, it doesn’t present a compelling case for the content of those regulations.  How can these regulations possibly change to reflect changing economic realities, such as the rise of new technology?

Chris Bradford put forth an interesting idea regarding land use regulation: give all zoning codes an expiration date (a similar idea to the zoning budget).  If the anti-trust and equity concerns are so great as to require this kind of regulation, requiring some sort of periodic review is an interesting idea for simulating some of the innovation and competition that a freer market might provide.

The extreme positions aren’t that illuminating.  Likewise, merely promoting the idea of regulation in the abstract (without speaking to the content and effects of those regulations) isn’t helpful, either.  The specifics matter. Regulation for the sake of regulation is pointless, and we must have mechanisms for continual re-evaluation of the regulations we do have to ensure they actually work towards our stated policy goals.  All too often, this re-evaluation falls short.

This isn’t meant to be a broadside against regulation – far from it.  There’s clearly a role for it.  Instead, I ask for periodic review to ensure the regulations are helping achieve our objectives rather than hindering them. Likewise, the inevitable reality is that whatever regulations we impose now will have unforseen, unintended consequences.