5 of 10 Most Expensive Airports Are UA/CO Hubs

There's a piece in today's NY Times about this subject, which shows how much more expensive hubs "controlled" by UA/CO are than other airports. Yes, there are some flaws in the survey, which are also discussed, but....

IAH is #1, EWR is #2, CLE is #6, IAD is #7, & ORD is #8.

http://fivethirtyeight.blogs.nytimes.com/2011/04/06/which-airports-have-the-most-unfair-fares/

 

It does not surprise me in the least that IAH is at the top of that list, having grew up around IAH. CO knows it has a fortress hub and prices accordingly. Odd, then, that IAH is usually cheaper than AUS but such is life.

 

The article might have me checking out the fare difference more often between EWR & JFK given the merger, even though EWR is more convenient for me:
" But perhaps the most basic strategy is to know which airports tend to be relatively cheap or relatively expensive to fly to or from. For instance, while individual fares vary significantly, the average passenger flying out of Newark Liberty Airport pays about 25 percent more than someone flying out of John F. Kennedy International for an equivalent seat on an equivalent flight."

 

Ummm...yeah. Not so much, methinks. Suggesting that airfare should be based primarily on distance traveled would skew fares all sorts of awful.

They've ignored schedule and non-stop versus connection so saying that there is an "equivalent" flight that is 25% cheaper seems highly suspect to me. How many domestic destinations can one get to non-stop from JFK versus EWR?

When you start with flawed assumptions of market forces you end up with ridiculous results like this "study" produced.

 

It's interesting, because Nate Silver (who wrote this blog) is a very sharp guy and is well known for high-quality analysis of baseball statistics (his day job) and political polling data.

I think the problem is his assumptions:

Regarding the least expensive airports:

Um, no, the unifying theme is the presence of WN and/or B6, which is correlated with vacation destinations.

Similarly, his list of the least expensive mid-sized airports is filled with airports whose dominant carrier is either WN, B6, or NK.

Essentially, his conclusion is that airports dominated by legacy airlines where WN/B6/NK/G4 have a minimal presence are expensive. That thump you heard was me falling out of my chair.

 

Actually, WN-dominated airports aren't always so cheap. Ask ALB.

 

Agreed...it is NK that makes the prices (base fares) drop. Add in the carry on, delta in checked bag fees between NK and OAL, drink fees, seat assignment fees, and then you will be comparing apples to apples.

 

Is there a glossary of these abbreviations available here on MP?

 

These are the IATA airline codes for Southwest, jetBlue, Spirit, and Allegiant, respectively. (I had to look up Allegiant. )

You can find the full list of IATA codes here, and a searchable database here.

 

Thank you!

 

Add in F9 (DEN, MKE) and FL (MKE, BWI) and we haaaaave a winnnnner. Basically, the airports where the LCC's ain't are the ones with stupid high prices.

 

in ter es ting ... SFO isn't on the list

so why can't I find good ex-SFO MRs yet

 

including US with DCA and CLT, that would make * having 7 of the 10

 

DCA is expensive because folks are paying to be closer to the District. Add in that it is slot-constrained, that one carrier dominates those slots and that the long-haul flights are artificially restricted, further reducing competition and you've got a recipe for monopolistic pricing, aka higher fares.

EWR is not more expensive, per se. But there are definitely routes from EWR which are more expensive, and it all comes down to competition. EWR-ORD/DCA/IAD/BOS have no competition today and they are more expensive than LGA/JFK-ORD/DCA/IAD/BOS. JetBlue is starting EWR-BOS and all of a sudden there is competition and the fares drop. Is this really something that requires advanced statistical regression analysis to document? I can figure it out with one pretty picture:


I give the guy credit for running a bunch of stats against the DoT data, but when you start with flawed assumptions about the market you almost always end up with crap results. GIGO. Ignoring connections v. non-stop, ignoring the value of speed for daily commuters on shorter trips and ignoring the presence or lack of competitive forces on individual routes basically damned this "report" to being a heaping pile of poo.
http://boardingarea.com/blogs/thewa...-ny-times-got-it-so-wrong-on-airline-pricing/

 

No surprise at all....especially lately.

 

VX and WN.

 

Lately I have found it cheaper to fly out of AUS thru EWR to east coast cities, than flying directly out of IAH, eg, just booked AUS-EWR-CLT for $238 a/i in July. IAH-CLT was $616 for same days. Added benefit is more miles.

 

Am I the only one who liked the article? The author establishes what the average fare would be at a given airport controlling for distance travelled, passenger demand, and airline competition. He then compares what the industry wide average fare should be for each airport and ranks them based on the difference between the average and the actual fare.

The controversial assumption is that there is no difference between non-stop and connecting flights. This assumption is justified if you define the airlines role as getting you from point A to point B and back. Of course when people book flights arrival and departure times make a difference, and also flight duration. While people may be willing to pay more for a more convenient flight, that does not change the "fair" price. If the data sample includes flight duration information it would be useful to get that data into the regression since extra time in the air is effectively charging a higher fare. If that information is not available then this assumption is the best that can be done. I think some useful information can still be gleened from the analysis even with this somewhat flawed assumption.

The next assumption is that demand will go up at an airport if prices goes down, and vice versa. A reference is provided to back up the claim that a 20% increase in fares results in a 20% reduction in travel. I have some concerns over the linearity of this relationship. While I am sure it is accurate at some price point, at some point the linear relationship will break down. I'm assuming it is at least in the ballpark when dealing with fares actually paid, so it is probably useful for the regression. I read on a blog somewhere that US is experimenting with this out of Rochester. They've dropped the prices and are watching to see if they get more business.

The last control is to find the average fare for the hypothetical situation where there is competition between the legacies, Southwest, and LCCs. This allows us to establish what a fair market price is in a competitive setting. I have to assume the author did this correctly...while I took a few sadisticsstatistics courses back in college regression analysis is not my expertise.

In the end, I think the author went to a lot of work to show what most of us already knew. Airports that have one dominate carrier tend to have higher fares. Small airports dominated by business travellers with little competition have high fares. Airports with a LCC or two competing with the legacies have lower fares. Leisure destinations that rely on fare conscious travellers have low fares.

A lot of the complaints about the article focus on how airlines actually price their fares due to market forces and not to distance traveled. Of course that is the point of the article; to show how the airlines are able to gouge customers in some markets and customers in other markets get to pay below the "fair" fare. I can drive up to PHL and catch a $200ish RT flight to SEA, but IAD-SEA is $450+ on UA. Sure there may be tens of dollars difference in taxes/landing fees, but essentially PHL travellers are paying half the price I am to fly the same distance and incur similar crew costs. I'm getting gouged at IAD because UA can get away with it. Heck, IAD even has competition from VX and B6 to SEA and it is still sky high.

Are leisure destinations cheap because they are leisure destinations or because of the presence of WN, B6, and NK? Chicken or egg? I tend to be of the opinion that the leisure markets are cheap because leisure travelers tend to be fare conscious. The LCCs in turn target these markets. This then helps to keep fares low. No argument that "Essentially, his conclusion is that airports dominated by legacy airlines where WN/B6/NK/G4 have a minimal presence are expensive." That pretty much sums up the whole article, but with numbers to back it up.

I have two problems with the article. There is no controlling for flight duration (basically non-stop versus connecting) or bag fees. Sure, Spirit reported that nice fare $25 less than the WN fare, but then charged the passenger $35 in fees. Since LCCs and legacies both have plenty of fees it probably doesn't make a huge difference in the relative fairness of airports.

 

I've thought about this a few times, as the two airports are an identical distance from my home. The key difference is that it costs me at least 4 times more (in both tolls and travel time) to get to and from JFK. This alone makes it worth the extra ticket cost to use EWR.

 

Indeed, if you start with a grossly flawed assumption then you can have a grossly flawed set of results.

Exactly.

I very much disagree that this is the best that can be done. Any analysis that begins with distance between the O/D pair as the basis for what is a fair fare is so horribly flawed that it loses credibility IMO. Certainly there is some information that is represented here, but it isn't what the author claims it to be.

This relationship certainly exists but it definitely is not linear. At some point folks will travel just because the price is right. The destination matters much less. Still, offering 1000 seats/day at $49 each from EWR to Fargo isn't going to fill up the planes. Sure, more folks will fly, but probably not enough for the demand to be linear.

This bit of his analysis I agree with. Competition is what breeds lower fares. Not distance traveled. How else can you explain the average transcon fares being so close to the average "commuter" fares in the WAS/NYC/BOS region?

But why is $200ish fair from PHL and $450ish not fair from IAD? Because they are a similar distance? You're not getting a non-stop UA flight on PHL-SEA, you're getting a connection; the IAD flight is likely a non-stop. The "cost" of that connection in terms of total travel time decreases as a percentage as the trip gets longer which partially explains why EWR-LAX is closer to JFK-LAX despite minimal competition on that route. Put another way, if you have a 20oish mile flight (e.g. LGA-DCA) and you need to make a day trip out of it you can fly non-stop and get there, have your meeting and get home or you can connect in PHL, CVG, ORD or somewhere else. Those might be cheaper, but you're going to spend 3-4 hours in transit, not 45 minutes. If you're making a day trip for a meeting then saving those hours of time is real value, something that you pay for.

It isn't only about the distance.

Those are two of many.

I don't doubt that the statistics are dead-on accurate. But the flawed assumptions to start the study ruin it.

And if you don't mind the 2-3x total travel time then you can save the cash.

 

Well, you can make arguments both ways. I was trying to think of a leisure destination that didn't have significant LCC penetration and the best I could do is MIA, which is on his list (though fares from my home airport to MIA are significantly higher than to FLL).

However, the presence of BUF and BWI and MKE on the list of big airports suggests strongly it is the LCCs, and not leisure destinations per se, that is driving pricing, unless you can convince me that there is a lot of wing or crab tourism out there (or people going to Niagara Falls by plane or doing the "Laverne and Shirley" reality tour).

I do agree that there may be some value in demonstrating that the obvious is true -- not all good analyses need to be contrarian -- but I still think he made some fundamentally flawed assumptions about how airline pricing works.

 

Try to MKE, ORD-MKE is 67 miles but you earn 500 EQM and 1000 RDM each way.

 

Both the PHL-SEA and IAD-SEA flight were connecting flights. It was pretty much an apples-to-apples comparison. To do a non-stop same-day turn (same parameters used for the above example) is $937 out of IAD. Of course I picked PHL because UA/CO are messing with US's hub at the moment. Most examples wouldn't show such an extreme difference.

My point-of-view on fairness keeps switching between fair relative to the national airfare market and fair relative to airline costs. I will stick with the national airfare market sense of fairness since that is what the article was about.

It seems to me this is the real problem with the article from a market POV. Passengers are definitely willing to pay more for shorter flights, thereby establishing a higher market rate for non-stop flights. I would like to see flight duration controlled for in the market price regression, but that data is not available. What is available is the number of segments and the length of those segments. So we can't account for abnormally long layovers, but we can account for connecting flights. The author should have used this in order to come up with his "equivalent flights."

Maybe I will grab the data and play around with it a little bit. There are 44 million flight segments in the database. It would be interesting to see the percentage of domestic connecting versus non-stop flights out of EWR if nothing else.