A couple of weeks ago, American put out its list of cities that it was going to drop in October if the CARES Act wasn’t extended. This cliff was created by the CARES Act itself, which restricted an airline’s ability to actually drop cities until October 1. If more money is thrown at the airlines, then this is all moot. But if not, American is making it clear that things will be changing. Did American choose its cities based on political expediency? Or was it solely based on performance? Let’s take a look.
American announced that it would suspend service to 15 airports beginning October 7 — the day it begins operating its October schedule — though it has already had to back off three of them. I’ll get into that later. The suspension was announced as only lasting through November 3 since the airline just hasn’t made any decisions about its November schedule yet, but let’s be honest. If these are being suspended in October, they aren’t coming back a month later.
Overarching Themes
The cuts are all at small airports where American had been operating four flights per day or fewer even before the pandemic. These airports also generally tended to be weak performers, but that didn’t stop speculation that these cuts were political. It seems highly likely there was a political component here. After all, the cities were spread out into different states, and that could help put local pressure on Congress to give more money to the airlines. But it’s highly unlikely that this was the main driving factor in choosing these particular markets… or at least most of them.
Looking for broader trends, I found that with two exceptions, these routes were all very close to an alternate airport served by American. Judging from the city’s downtown to the next nearest airport, those 13 markets were all less than 90 miles away by car. I’ll talk about that in more detail with exceptions below.
I went into Cirium to look over whatever data I could find. The problem is that government revenue data hasn’t been released beyond the first quarter of 2020 yet, so it’s hard to see just how hard hit things have been. But I do have T100 load factor data that goes into May, so, I’ve cobbled this together the best I could. Let’s look at this city by city, ranked by stage-length adjust (SLA) unit revenue for the fourth quarter of 2019.
Williamsport, PA – 58 miles from State College
Williamsport pretty clearly seems to be the worst performer of the group. Even before the pandemic, American was filling just over half the seats in the market. That plus low fares meant that Williamsport had the lowest stage-length adjusted unit revenue of any city not propped up by Essential Air Service funding. It’s so close to State College — an airport that could use a boost — that this must have been a no-brainer, especially with not even a Little League World Series to help it this year.
Del Rio, TX – 158 miles from San Antonio
Del Rio is an outlier in that it is pretty far away from another airport, but it’s also an underperformer. Notably, it appears that travel really fell off after COVID. American’s system loads fell off about 33 points down to about 48 percent in May. But in Del Rio, it fell from an already light 66 percent all the way down to 28 percent. On top of that, Del Rio sits on the border, and land crossings have been restricted. So I imagine any Mexican customers are having trouble getting there to fly.
Florence, SC – 68 miles from Myrtle Beach
Florence is what you might call… geographically undesirable. It sits inland, not far from Myrtle Beach but also only a 2 hour drive from the massive Charlotte hub. It saw loads drop more than 40 points between January and May. You might try to compare it with Greenville/Spartanburg or Columbia, but both of those have higher SLA unit revenues and kept loads more than double Florence in May. American is the only airline in the market, so it’s not like it is losing a competitive edge.
Greenville, NC – 46 miles from New Bern
I’ve always been amazed at just how many random little towns in North Carolina are served out of the Charlotte hub, and this is one of them. Greenville is out in the eastern part of the state, and it’s close to several other small airports. It’s also only 100 miles away from Raleigh/Durham. If you compare it to nearby New Bern, it’s no contest. Greenville SLA unit revenue is about a point less than New Bern, and while New Bern saw loads drop 25 points from January to May, Greenville was cut in half.
New Haven, CT – 52 miles to Westchester or Hartford
New Haven is one of those airports that you’d think would do better. It is situated to attract from northern New York suburbs as well as other coastal towns in New England. Despite that, it just doesn’t work. Most people go to other airports in the region, and New Haven is left with very few travelers. It filled 31 percent of seats in May, something that wouldn’t be a death knell if the yields weren’t so low.
Joplin, MO – 69 miles to Northwest Arkansas
This one isn’t actually going away just yet. Joplin, it turns out, is an Essential Air Service market — albeit unsubsidized — and somehow American never realized this. For that reason, it can’t just stop flying to the city. Instead, it is now looking to go through the government process so it can pull out. That request was filed with the feds last week for a December 1 stop date, and it was accompanied by this chart:
This looks bad. Things started to recover in June and then fell of a cliff. But be careful. In July, American ramped up quickly to offering 3,570 departing seats, up from 1,482 in June. Also, August isn’t for a full month and September and October only include bookings made before August 25. So, while this does show that American gambled by expanding too quickly and lost in July, it’s still not as bad as American wants you to believe. It is, apparently, bad enough for American to walk away.
Huntington, WV – 57 miles to Charleston
West Virginia is a tough place to succeed as an airline in general, but with Charleston so close to Huntington, it’s exceedingly hard to make Huntington work. Huntington wasn’t the worst performer, but it did see its loads drop to a mere 18 percent in May. This one has dried up quickly and appears to have stayed depressed. It was the fourth worst load factor for American in May.
Kalamazoo, MI – 52 miles to Grand Rapids
Michigan has a lot of small airports, and Kalamazoo hasn’t done that poorly. It has decent fares and loads are about half full. So what gives? I think this is likely the airport that gets sacrificed to try to boost the others. Grand Rapids is north, South Bend is south, and Lansing is northeast. Getting rid of this could help boost the rest. Or this could just be political. Lansing and Kalamazoo both perform similarly, but Kalamazoo lost out for one reason or another.
Stewart/Newburgh, NY – 56 miles to Westchester, 69 to LaGuardia
Westchester has a lot of rich neighbors, and Newark, LaGuardia, and JFK are the big guys. But Stewart? Not so much. In fact, American pulled out on April 23, something that was allowed by the CARES Act since Stewart was a New York City metro area airport. Service didn’t return until August 18. Presumably forward bookings weren’t looking very good, and now it’s gone again.
Stillwater, OK – 68 miles to Tulsa, 73 to Oklahoma City
North of Oklahoma City and west of Tulsa, you find little Stillwater. Stillwater looks a lot like Huntington. It’s closer to other airports, and traffic basically disappeared when the pandemic hit. Loads plunged to a mere 18 percent in May. Like Huntington — the home of Marshal University — Stillwater is a college town with Oklahoma State dominating the landscape. Now those kids, if they ever come back, will have to just drive into town from elsewhere.
Roswell, NM – 177 miles from Lubbock
Roswell seemed like an odd choice. It is really far from just about anything else; El Paso and Albuquerque are both about 200 miles away. Other than aliens, there’s not much else around. Oh, except there are also a ton of shiny American airplanes parked there for good. But now, American has reversed course. In a statement from the airline, I was told this:
In the absence of an extension of the Payroll Support Program, we’re faced with a host of difficult decisions to right-size our airline, given the significant and sustained drop in demand we’ve seen during the COVID-19 pandemic. While our flights to Roswell are among those that have suffered when it comes to demand and profitability, we have been in touch with local officials and will defer our decision to suspend service to the market as those conversations are ongoing.
Like I said, this was a bit of a head scratcher. I’m just going to assume that the good people of Roswell threatened to blow up all the MAX aircraft American has parked there if service wasn’t restored.
Sioux City, IA – 89 miles to Sioux Falls, 96 to Omaha
I think we can all agree that this one just SUX. No really. American also had to file a notice to end service here, just as in Joplin. But this chart looks worse.
Nearby Sioux Falls, it should be noted, is getting a nonstop to Charlotte. So this is presumably a way to help shore up service there. while getting out of what looks to be a really ugly market for the airline.
Springfield, IL – 72 miles to Bloomington and Peoria
Central Illinois is one of those strange areas where, like the Florida panhandle, there are far too many airports in a small geography. You have Peoria, Bloomington/Normal, Champaign, etc. Springfield is the state capital, but United is the only one that flies to Chicago. American flies down to Dallas/Fort Worth, and it’s just not pulling its weight. Springfield saw load factors plunge 57.6 points in May vs January. That’s the third worst performance at American, tied with Honolulu which had a full quarantine in place that actively prevented visitors. Springfield had no such excuse.
Dubuque, IA – 81 miles from Cedar Rapids and Moline
Dubuque actually looks a lot like Springfield on the surface. It had similar SLA unit revenue numbers in both Q4 of last year and Q1 of this year. But there is a difference. Load factors held up better, only dropping 37.6 points. The bad news is that fares held up worse. SLA yield was 1.3 points better in Q4 but that gap had narrowed to 0.8 in Q1. I imagine things eroded from there. In the end, it gets to the same place and both cities are gone.
Lake Charles, LA – 66 miles from Beaumont
Poor Lake Charles. It gets walloped by Hurricane Laura and it’s losing American service. This one was somewhat confusing before the hurricane did its damage. Its SLA unit revenue is actually good, and it remained respectable into Q1. Load factors only dropped 18 points between January and May, settling at 54.6 percent. How does something like Lake Charles get killed when a similarly-sized place like, say, Central Wisconsin sticks around with consistently worse performance across the board? Both are close to other airports, so that’s not the reason. Maybe this is a political thing. I’m not entirely sure.
Overall, these are markets that look to be in rougher shape than most. That still doesn’t explain how some made the cut while others didn’t, but maybe it will become more clear when there’s more recent data available.
