As a physicist, I do a lot of flying. Many physicists do a hell of a lot more, and many do less. I get to conferences and workshops, or join my wife at her experiment, four to five times a year. Throwing in family visits, unexpected life events, and vacations, the number winds up somewhere between four and seven flights (round trips) each year. When I first flew, during a CDF collaboration meeting at Fermilab in 1997, I was excited and terrified. I have a fear of heights, but I found the soaring heights of air travel, the acceleration, and the claustrophobia all fun and fascinating. Over time, this charm has worn off. In fact, I’m working on a collection of short stories written over many years on the theme of air travel and hotel stays. They are not good tales.
This past week, we watched the first airline meltdown since about 1997. Passengers were stranded on planes for hours, some up to half a day, within sight of the terminals but unable to leave. So many people, many healthy but some sick, crammed into a small space with limited water, food, and waste facilities is a recipe for misery, or even health disaster. This renewed an interest in an industry-wide “passenger bill of rights”, a document added to the FAA regulatory power that gives certain rights to passengers. Airlines in violation would be penalized. This doesn’t exist now, except where an individual airline chooses to have such a document. Enforcement is handled by the airline, a system which has repeatedly failed to work (in many industries).
I watched last night’s “Lehrer News Hour”:http://www.pbs.org/newshour/bb/transportation/jan-june07/planerights_02-23.html and was greatly disturbed by the airline industry’s arguments against such a bill. It was the usual stuff: government regulation is bad because we don’t want a bureaucracy to micromanage passenger comfort (which is NOT what a bill of rights represents). However, another argument was thrown in there: capitalism. Self-policing works because the only thing that differentiates airlines these days is customer service. Since there are so many airlines, you have a choice. If you don’t like how one airline serves you – food, baggage handling, flight attendants – you can choose another one to get to your destination. That choice is the self-policing mechanism that makes each airline want to serve you the best.
A choice? Really. Hmmm. Let’s conduct an experiment. First, let’s enter some typical passenger variables into the equation. I live in the Bay Area. SFO Airport is a hub for United Airlines, and San Jose is the next closest airport and is a major Southwest Airlines airport. So, like many passengers I live near a small airport and/or a major hub airport. I also refuse to fly Delta Airlines (here enters passenger choice), because they were so TERRIBLE during my trip to Moscow that I am boycotting them for five years (c.f. “TAOMPH297” and “TAOMPH298“). OK, so I have a few common initial conditions for passengers. Let’s construct two models.
* Model A: A passenger with a flight choice constraint (will not fly Delta) living near a major airport (in this case, SFO).
* Model B: A passenger with a flight choice constraint living near a mid-size airport (SJC, in this case).
Let us now pursue searches for airlines for a passenger who needs to get to Jacksonville, FL, for model A and B. Let’s add a few more initial conditions to the two models:
* A passenger will always want the cheapest flight possible
* A passenger will choose flights that fit their schedule. For instance, needing to leave late in the day to keep work intact, or leave early in the morning to make a timely arrival at the destination (west to east travel means getting in late).
The latter is hard to include randomly. Let’s modify the two models as follows:
* Model A: passenger flying from SFO, won’t fly Delta, wants cheapest possible flight, and needs to leave in the evening or late afternoon
* Model B: passenger flying from SJC, won’t fly Delta, wants cheapest flight possible, and needs to leave in the morning.
I use kayak.com to do my search. I will use Southwest.com to search for SJC as well, since it’s not included in kayak’s site search (also, ATA.com). The trip will be 7 days, from April 13-April 20.
Here are the results. In bins of price that are $50 wide, I look at flights up to $500 in cost from kayak.com. I take any flights I can get from Sourthwest, ATA, and JetBlue. For Model A, I find the following:
* $300-$400: US Air is the primary carrier for this price range from SFO. Most of the options pair US Airways with another carrier, including Continental, Delta, American Airlines, and United. As you can see, of the lowest cost choices, all are US Air and one fifth involve a second carrier the passenger refuses to use.
* $400-$450: This region is dominated by “Multiple+Carrier” choices, where the second carrier is one of Delta, Continental, and American Airlines. Most were Delta. In addition there are flights exclusively operated by AA and Continental. Of the choices, most of them includes a carrier that is excluded from consideration.
* $450-$500: This is the single-carrier-dominated region. United, Delta, Continental, and US Air are the choices. Of the flights offered, 3/8 involve Delta, excluding them from consideration.
* There are NO flights on Southwest, ATA, and JetBlue.
What have we learned in Model A, where a major airport (hub to United) is considered? Well, among the cheapest flights (which start at $300), US Air is the lead carrier on all of them. Of those they partner with, Delta is one of them, excluding that flight from consideration. In the next price bracket, multiple carriers are paired with single carriers. Of those single carriers, Delta is one and is excluded. The most expensive flight available (up to $500) is on Delta, which is excluded. Passengers in Model A would have to use US Air if they want the cheapest flight, and if they want to take something in the next price bracket but avoid Delta they’d get to choose from United, AA, or Continental. Clearly, big carriers dominate here.
For the San Jose (Model B) case, the analysis is easier. There are only TWO flights available for the morning, one on Multiple/AA and the other on United. Neither is cheap ($800 and $500, respectively). Southwest offers a flight for $400, and JetBlue for $728! For the passenger seeking the cheapest flight possible, they have one choice ($400 on Southwest).
These two examples are not unique. Jodi and I experience this kind of frustrated choice every time we travel. She has strong dislikes for Continental, and I loathe Delta. Northwest is a common carrier for us to the Midwest, and United or AA to the East Coast. US Air is decent for trips to the Southeast. But carriers dominate regions, and for cheap flights you have few choices. Constraints based on previous bad experience limit those options pretty quickly. I don’t buy the capitalism argument for self-policing – the big guys who can afford to cut costs, drive down prices, and maintain their regional dominance have done so very effectively. Based on the data, would you trust choice as the basis of self-policing?
