hub and spoke

This is Part III in my series on reforming the commercial air travel system in the United States. Part I, primarily about seat sizes and basic comfort for passengers, is here; Part II, about the abolition of food service on U.S. airlines, is here. You’ll note these topics, while important, are customer service issues. What I want to talk about tonight is more fundamental: the basic way in which airlines function.

If there’s one feature of our broken air travel system that, if fixed, would have a greater impact than any other, it’s the basic pattern on which air routes are organized. American air travel is based on the “hub and spoke” model of transit systems–that is, airline operations are based in a central point, the hub, and everywhere the airline flies is a spoke radiating out from that central hub. This form of organization seems so basic as to be an axiom. United Airlines’s main hub is Houston; Delta’s is Atlanta; Charlotte, North Carolina is the “fortress hub” of US Airways. The larger airlines have multiple hubs and sub-hubs. Small, no-frills airlines (a misnomer; there are no “frills” left even in the major carriers) also use hubs, typically smaller airports like Seattle, Las Vegas or Minneapolis-St. Paul.  If you want to fly from San Francisco to New York City on United, you will fly from SFO to either Houston or Chicago, and catch your connector to New York there. This seems only logical, right?

There is, however, a different logical model of transit organization, and that’s called the point-to-point model. Instead of flying through Houston, you board a single plane in San Francisco and get off that plane in New York. Or Wichita. Or Buffalo, New York. A point-to-point transit system is faster and more efficient. It takes less fuel, less time and generates fewer carbon emissions to fly one plane directly from the Golden Gate to the Big Apple than it does to fly two planes, each going out of their way either to and from Houston, which in any event is not on the way between San Francisco and New York.

airport bad weather

A “hub and spoke” model may make sense in a corporate boardroom, but its drawbacks are painfully apparent when bad weather hits a hub airport city.

Why do airlines like the hub and spoke model as opposed to the point-to-point model? Why do airlines like anything? It’s cheaper. Instead of 10 flights a day from San Francisco to 10 different cities, which will carry fewer passengers, you can schedule 10 flights a day from San Francisco to Houston, and let the folks in Houston worry about where to send eastbound passengers. Theoretically every flight you operate is either to the hub or from it. There’s less infrastructure to pay for–fewer terminals, gate agents, baggage handlers and kitchens (if airlines still cooked food anymore). If you want to cut costs in a large transit system, this is how to do it.

But there are serious problems with the hub and spoke system. Chief among them is that it only works under “best case scenario” conditions. I’ll get from San Francisco to New York just fine if both of my planes are on time, weather conditions are good for flying, nothing’s mechanically wrong with the planes, etc. But if a wing falls off the plane just before leaving San Francisco and it takes three hours to fix, and I miss my connecting flight from Houston to JFK which is on time, the airline  now has a problem. It’s not a big problem so long as there’s another flight from Houston to JFK they can put me on. But what if there isn’t? Or what if that flight is overbooked and they can’t get me a seat? The more that goes wrong, the greater the consequences–you have what’s known in computer jargon as a cascading failure. And by taking two flights, from SFO to Houston and Houston to JFK, United is now gambling on “best case scenario” conditions involving two planes, not just one. Fixing a failure on the SFO-to-Houston leg involves yet another plane, because a second plane from Houston to JFK must now be found for me. If it was simply a matter of waiting for a single plane to get fixed in San Francisco, and then it takes off for New York three hours late, I’m still delayed three hours but nothing else is affected.

A “point to point” air transit model would make situations like this, the January 2014 blizzard, much less severe in their impact on air travel.

The fragility of hub and spoke airline systems is graphically demonstrated whenever there’s a blizzard or other weather condition that knocks an airline’s major hub out of commission. This happens frequently, but the most catastrophic recent examples include the Christmas 2006 blizzard in Denver which grounded United Airlines for three days, and the New Year’s Eve 2013/14 debacle. If you’re hubbed in Denver and you can’t get a single plane in the air, or land one, for three days, your entire system will collapse, even in places that have never seen a single snowflake. Cancellations cascade outward from a weather-stricken hub and often paralyze an airline’s entire system. The effect of hub failures is made much worse by the fact that airlines deliberately try to operate at 100% capacity all the time, which means, at least in theory, there are no empty seats anywhere in the system at any time. So, if I get stranded by a blizzard in Denver and miss my flight to New York, there simply are no other empty seats on other flights to give me. As airlines have eliminated excess capacity for financial reasons, this is why flying “standby” at peak travel times like holidays has become, in the past 10 years, a virtually impossible task. Yet bizarrely this flawed system is exactly what airlines have opted for.

Thus, a major step toward root-and-branch reform of the U.S. airline system is to abolish the hub and spoke model. While a pure point-to-point model is not feasible, airlines could be prohibited from flying more than, say, 25 or 30% of their total daily flights from a single airport, thus making it impossible to establish major hubs and diffusing air traffic through more geographically balanced locations. Granted, a lot of people will be going to New York, Chicago or L.A. every day, but if less than 30% of them are getting there through Houston, thunderstorms in Houston will affect a much smaller percentage of that day’s travelers, as opposed to virtually all of them in a Houston-hub model. Furthermore, airlines should be required to maintain a certain number of planes in reserve and flight crews on call, at least at high-traffic times of the year. This could be combined with a requirement that airlines not operate flights at 100% capacity, but perhaps 98% or with a certain raw number of seats vacant to serve as excess capacity to ameliorate delays. This would make “standby” flying feasible again for passengers as well as provide airlines with the ability to clean up quickly after large-delay events like blizzards or closure of an important airport.

Okay, I admit this Western Airlines commercial from 1985 has nothing to do with the subject of this article, but I found it and just had to post it.

Changing from a hub and spoke model to a point-to-point model will also have a number of excellent fringe benefits. A point-to-point airline system will serve a larger number of smaller airports. This means, among other things, that cargo can be moved more efficiently. It also has economic benefits: with more airports to staff, airlines will be able to create more jobs for local communities outside traditional hub cities. Environmentally a point-to-point system is far preferable to the hub and spoke model. Burning less fuel per passenger and mile flown means fewer carbon emissions and lower fuel costs for the airlines.

These suggestions are not fanciful or fantastic. They’ve already been tried–and they worked. Before the disaster of airline deregulation in 1978, the United States air system was largely on a point-to-point model. More cities were served by air, passengers were moved more efficiently and service was better. Smaller carriers still use point-to-point systems. Before it became one of the majors, Southwest Airlines’ system was predominantly point-to-point. Many airlines outside of the U.S. use point-to-point systems. There is no reason in the world why this model won’t work for large national carriers–that is, no reason than the airlines’ go-to complaint, “But that will cost more money!

I do have an answer for the “But it will cost more!” argument, but that’s for a future article. Thanks for reading, and for thinking.