Sure, there are rules to follow and safety regulations if commercial service is going to exist at an airport, but when it comes to tracking competition and airfares the Feds take a slightly different approach. Not every market matters. And while there are certainly good reasons for that approach it does raise some further questions when reports are released claiming that consolidation has not harmed competition in the domestic market.
In June the GAO released a report which suggests that consolidation has not had a significant impact on service or fares. The conclusion of the report is relatively positive:
In recent years, the average number of competitors has not substantially changed in markets traveled by the majority of passengers, despite several major airline mergers. From 2007 through 2012, the average number of effective competitors (defined as airlines with more than a 5 percent market share) ranged from 4.3 to 4.5 in the markets with the most passengers. During this period, the average number of effective competitors in markets with the fewest passengers decreased slightly from 3.3 to 3 airlines. While these results reflect market changes that have occurred since several airlines merged, the American-US Airways merger occurred after GAO’s analysis. The mergers created larger networks and new connections in some markets. Also, low-cost airlines have expanded since 2007, thereby adding new competitors into some larger markets. The structure of the market will continue to evolve as economic conditions change and the recent airline mergers are fully implemented.
In recent years, consumers have experienced higher airfares, additional fees, and fewer flights in certain markets, but also new services and expanded networks. Consumers paid about 4 percent more in real terms, on average, for air travel in 2012 than in 2007, without considering additional fees. The airline industry has reduced flights, especially to smaller airports, and consolidated service at large airports. Airlines have also invested in new aircraft and introduced new services, such as early boarding and entertainment options, in an attempt to differentiate products and increase revenue.
And while there is a reference in the summary to reduction of service at smaller airports the report never really details which airports are included in the analysis. A FOIA request also failed to turn up the list, though it did result in instructions for how to recreate it myself. And so I did. What resulted was a list of just over 48,000 city pairs served in the USA in 2013 with at least one passenger buying a ticket to fly between those airports. The GAO only cares about markets with a significant number of passengers flying (just over 1,000 total in a year in either direction), which makes sense from a statistical analysis perspective. That means roughly 90% of the city pairs are dropped from the calculations. Of the 463 total airports in the list of routes flown a full 185 are dropped when the GAO passenger number filter is applied. In other words the GAO is ignoring the impact of competition on 185 airports when coming up with statements like, “[T]he average number of competitors has not substantially changed in markets traveled by the majority of passengers, despite several major airline mergers.” And while it is true that these are smaller markets, those are precisely the markets which are expected to suffer as a result of the mergers. It seems they should be the ones which the government cares about, at least if they are truly committed to protecting those consumers. So, which airports are dropped? All of these:
Map generated by the Great Circle Mapper - copyright © Karl L. Swartz.
Doesn’t look so bad, right?
A whole bunch of the airports which are excluded from the analysis are in Alaska.
Map generated by the Great Circle Mapper - copyright © Karl L. Swartz.
But there are another ~150 on the mainland including Essential Air Service airports (where competition is not expected). Back out the EAS airports – places where the operating costs are subsidized (though fares are not necessarily) and there is only expected to be a single carrier – and you’re down to just over 50 airports still on the list operating as commercial facilities but where the level of competition doesn’t matter, at least not to the GAO.
Map generated by the Great Circle Mapper - copyright © Karl L. Swartz.
I’m having trouble getting too worked up over the exclusion of these airports, but that’s mostly because I live in NYC where we have plenty of competition and plenty of airlines with non-stop flights to a ton of places. Still, I do wonder about the GAO’s choice to exclude these smallest markets – the ones most likely to be affected by reductions in service and competition – when claiming that the mergers have been OK for consumers. Then again, at least one of the listed airports has seen a significant up-tick in service, though that has more to do with MassPort subsidizing service into Worcester than demonstrated commercial demand.
If you’re interested in looking at the data showing all the city pairs and passenger counts the csv file is zipped here.
Never miss another post: Sign up for email alerts and get only the content you want direct to your inbox.
Fascinating stuff. Thanks for grabbing this data.
After reading what you’ve written it sure doesn’t appear that the inclusion / exclusion of these routes alters the lede, “In recent years, the average number of competitors has not substantially changed in markets traveled by the majority of passengers,” which remains true either way.
Consolidation may or may not reduce competition in these cities, it can also increase effective competition by creating viable second-largest airlines as competitors out of what were otherwise small/dispersed competitors at an airport.
It’s not correct to say that passengers wouldn’t enjoy the benefits of competition even if these airports dropped to a single carrier. Don’t passengers at FLG have access to a competitive hub at PHX? And should the federal government really care about competitive service at MOD when it’s (by air) 57 miles from San Jose, 78 from San Francisco, 90 from Fresno?
If the study said “changes in competition have not been bad for any consumers” it would be completely fair to say “but what about consumers x, y, and z?”
But it doesn’t say that. The study, while having normative implications, does not say changes have been good or bad. It suggests that for the vast majority of consumers the number of competitors has not changed in a material way. And it says that analysis does not incorporate the effects of the AA/US merger.
We also know the story on airfares — recent small uptick when including average fees paid, although on an inflation-adjusted basis still far lower than prior to deregulation even incorporating fees.
The relevant message coming out of this study isn’t a bold claim about “consolidation is good for consumers” or any such.
But rather there are a whole lot of effects out there, such as the advent of low cost carriers and increased effective competition, and when airlines cede markets others come in because there are very very few barriers to entry outside of slot controlled airports (or airlines with limited gate space where gates are already fully leased, but even there they can sometimes be subleased).
And so “it’s complicated, and certainly more complicated than the standard narrative would suggest” seems like a fair statement to make.
Indeed, Gary, most passengers are not facing fewer airline competition options in most markets. Even in these smallest markets – the ones the GAO ignores – most are likely still only going to have 1 or 2 airlines serving them.
Unfortunately, however, I take some issue with the claim that new services and expanded networks have offset the cuts in frequencies and some markets which the consolidation brought about. Especially in the smaller markets. The data is not all bad, and for you or me living near a major city with multiple hub operations the outlook is quite positive. But there are a lot of people in smaller markets feeling a squeeze. And this GAO report won’t help consumers demonstrate that very well.