The fight between American and the TWU/IAM unions representing mechanics and fleet service workers continues to get worse. Though American was able to obtain a temporary restraining order (TRO) preventing the mechanics from slowing down the airline, statistics show that things have not improved. American went back to court last week to force the unions’ hands, and the judge quickly agreed. Anytime you go back to court, the relationship frays further, so this certainly won’t help resolve the long-term issues. But whether this actually solves the problem or not in the short-term remains to be seen. So far, there’s not enough data to say if things are improving.

I first wrote about the escalating fight back in June when the airline’s operation was suffering more than it normally does. No, American’s poor operation isn’t entirely due to labor action, but the judge in the case agreed that it was clear that a slowdown was making things worse. A TRO was issued to prevent the unions from coordinating any sort of job action as part of their negotiations.
Back to Court, Armed With Numbers
What has happened since then? Nothing good. American laid out its case suggesting that the unions were not obeying the TRO as follows.
- Last year, mechanics completed approximately 78 percent of their nightly tasks. Before the court issued the TRO, productivity had dropped to 70.6 percent. Since the TRO, productivity has dropped even further to 70.0 percent.
- There is a list of open MELs (minimum equipment list items). These are minor things that aren’t working on an airplane, but since they’re not critical, airplanes can still fly passengers, at least for a time. The list of open items averaged 435 during the last two summers. Between May 20 and June 13 of this year, that sat higher at 526. As of July 8, it had spiked to 561. Even more damning is that the number of write-ups has remained steady, so work just isn’t being completed.
- Each morning at 7am, there is a measure of the number of aircraft out of service for maintenance issues (AOS). The average for the past two summers was 42, and American has set a target of 35. Before the TRO, this had spiked to 54.5. As of July 9, it had barely declined to 53.8 on average.
Those overall numbers are telling, but I’ve always struggled with proving a coordinated job action in a case like this. After all, it’s like when a witness says something in court and the other side objects. The judge may say it’s stricken from the record and the jury should disregard… but the jury definitely heard it and can’t really forget. In this case, it seems pretty clear that if there was any coordination before, it wouldn’t be hard for individual union members to carry on the work without actually following orders to actually do it. If I’m a mechanic and the union tells me to stop, I know why the union is telling me that, and it might not change my behavior. How can a judge really rule that something is coordinated if there’s no evidence. Well, in this case, there is evidence.
The Philadelphia Overtime Ban
Looking at some of the specific examples given in the court filings, it seems hard to refute that this wasn’t coordinated as some level. I won’t bother you with all the examples, but let’s just look at one week in Philly when the mechanics suddenly decided that they would take no overtime.

There was one shift on July 3 where the number of qualified mechanics wasn’t available, so I left that out. And the second shift on the 2nd wasn’t shown in the data. But the point is pretty clear. Not one line mechanic accepted an overtime shift between July 1 and July 7 despite there being a large pool of mechanics available.
You might think “well, it was the 4th of July and people wanted off,” but history doesn’t show that to be the case, and it certainly wouldn’t matter earlier in the week. Overtime pays well, and it pays even better on a holiday. It would be one thing if the numbers were down, but this is a complete elimination of overtime being taken. If that’s not enough, a manager overhead people talking on June 29 saying that there was going to be an overtime ban that week. This one looks pretty clear.
The Judge Agrees With American
Short of actually resolving the contract, how can American actually get the operation back on track? Well, the airline says it didn’t ask the judge to find the unions in contempt, because it doesn’t want damages. It just wants this to stop. So it asked for — and was granted — a modification to the TRO to force the following.
- Senior union leaders have to conduct in-person meetings with overnight line mechanics and union officials to “communicate a sincere and emphatic respect for the requirements of the [TRO].” It also requires them to tell the mechanics to return to normal behavior or be subject to fines by the union.
- If those mechanics and union officials can’t attend the meetings, senior leadership of the union has to call those people to convey the message. They must also record a video with the message and post it to a dedicated webpage.
- The union has to post the newly-modified TRO to that same webpage as well as on dedicated bulletin boards.
- Each union member has to sign and date a form “stating that they have read and understand their obligation to comply… upon risk of being disciplined or fined….”
- The senior union leadership has to send a notice to all mechanics with specific verbiage as approved by the judge that very sternly tells everyone to get back to work as usual.
This has to be done by this week, so in theory, American’s numbers should get back closer to normal. They haven’t yet, however.
The Immediate Impact
Even though the order only came down Wednesday night, I still thought it would be worth looking at operational numbers to see if anything changed quickly. Thankfully, masFlight was able to pull the Thursday-Saturday data for me yesterday.
American’s Thursday completion factor was a dismal 91.51 percent on Thursday. By Saturday it had climbed to 97.57 percent. That’s still not good. Meanwhile, on-time performance climbed as well in the same period. Thursday had departures exactly on time (D0) of 48.4 percent and arrivals within 14 minutes of schedule (A14) of 67.31 percent. By Saturday, it was up to 61 percent and 79.84 percent respectively.
There’s a lot of noise here considering we had Hurricane Barry messing with things down in the Gulf as well as the usual thunderstorms around the empire. And really, a three-day sample is far too small to glean any trends. So for now, I can’t really say whether this is going in the right direction or not. I can say, however, that when a day where an airline cancels 2.5 percent of flights and delays one in five starts to look good, you know the airline is not doing well.
Hopefully the operation will return from being poor to just being mediocre. (That’s the best we can hope for these days, it seems.) The permanent solution, of course, is to agree to a contract. That, however, seems pretty far away. I continue to look into that issue, so you can expect future posts.