Oral argument — Restricted Category v. Sturgell

My name is Chris Kilgore, and I am pleased to represent the Restricted Category Aircraft Association. You may have noticed two or three weeks ago folks from the association got a lot of airtime on the news and other things in conjunction with the fires in Southern California. They were intimately involved in fighting those fires. I'll begin my comments by going to the predicate issue of the standard of review since we have exchanged some briefing at odds on that issue. The two conflicting or potential standards of review are, of course, the standard in the Federal Aviation Act of substantial evidence and the default standard under the APA of arbitrary and capricious. In preparing for you... Before you go too far on this, is there going to be any substantive difference between the two standards? No, Your Honor, there's not. And in preparing the arguments, the possible arguments for today, I realized we wind up at the same place. So what I intend to do is argue strictly from the APA arbitrary and capricious. That and, of course, the direct case law that fits this particular case. But, again, we wind up at the same place. The FAA proposes or proposes in its briefing a very broad and deferential review. However, the factors, the specific factors that the Court has determined are contained in this Court's opinion in Hawaii Helicopter Operators, which has been cited by both parties in the case. They list four specific factors. The first factor, which has to do that the agency has relied on factors which Congress had not intended, I don't think applies under the facts of this case. The second and third factors, though, very definitely apply. And that is that the agency's determination is arbitrary and capricious when the agency fails to consider an important aspect of the problem. Also the agency's determination is arbitrary and capricious when the agency has offered an explanation for its decision that runs counter to the evidence before the agency. And I think that applies in this case in spades. The fourth element may or may not apply, and it has to do when the agency's explanation is implausible. In any event, pursuant to Hawaii Helicopter Operators, the agency decision must be based on consideration of the relevant factors. The administration, the FAA, has cited Starr v. FAA in support of its arbitrary and capricious argument. There is in the Starr case, as they say the rest of the story, it's the paragraph after the one that's cited that's a very interesting point made by the court in that case. Under the arbitrary and capricious test, however, the administrator's discretion is still limited. He is bound by the straight framework or the statutory framework of the program administered by the agency. The analysis that you undertake is governed by a very recent case from this court. This is a 2007 case. I've provided the clerk with a form and a copy of the case to opposing counsel, and it's a safe air for everyone case which was amended in May of 2007. Our review of an administrative agency's decision begins and ends with a reasoning that the agency relied upon in making that decision. With that standard, let us proceed to the reasoning the agency relied upon. Is it your argument that it is per se either arbitrary and capricious or not supported by substantial evidence to rely on proprietary analysis? That seems to be a big piece of the pitch that you're making, in your briefing at least, is that the error here is in relying on the representations of a self-interested party. That is, if it can get its way, it will make more money, and therefore it can't be trusted. But why can't an agency on independent review find those studies plausible? They could, Your Honor, and it's all fact-driven on a case-by-case basis. What we have in this case, and the part that it's not reliable or it may not be reliable, may be unique to this case in that there are other factors that underlie this particular lifing analysis by Honeywell. We know other things, but that wouldn't necessarily be true in every case. What we end up with, whether you agree or disagree with the validity of the lifing analysis itself, is you have a theoretical lifing analysis on one hand. On the other hand, we have a mature product. The product itself has been in service for over 50 years. The components being used today, the D979 alloy components, have been in service since 1968, almost 40 years, and they've amassed tens of millions of hours of experience. So we have other evidence in this case, and we say it is wrong, and it is the agency's decision runs counter to the evidence before it when it relies on a theoretical computer analysis and disregards all of the other evidence. How do you know how long operators kept these parts in service through how many cycles? You don't really know that, right? At least I thought. I read the briefs that we wouldn't know and the agency wouldn't know for sure, like how often, like how many actual cycles of flight each aircraft was used on on which a part was replaced. We would know each operator is required to comply with the type certificate for that aircraft. This engine, this airworthiness directive, and the related service bulletin apply to a very specific number of aircraft. They couldn't use it more than what the type certificate said. They may not. But they could buy a new part before the useful life was used up, right? They certainly may, Your Honor. They certainly may, and they can continue to use the D979 components. They are still free under the AD to use either the old D979 or to go to the PM Astroloy. That's their choice. Economically, for a good many of the operators, the new life cycle count forces them into the PM Astroloy because it would just simply be uneconomical to go the other way. It's kind of crafted that way, I guess, is our position, but that's the economic reality. The bulk of these aircraft, probably some 90 percent that are affected by this, are subject to a type certificate even though they're surplus military aircraft. When they're sold as restricted category aircraft, a type certificate is created, and it will specify in there what the maintenance criteria are that that aircraft will be maintained to. Now, some of those merely say you will comply with the Army maintenance standards that were in effect. The Army doesn't count cycles. That leaves these folks not counting cycles. Those are, today, a minority of our operators. But some people are still doing it because they can. Some of the operators, either because of their type certificate or voluntarily, have chosen to comply with one of the intermediate service bulletins, and in particular the Allied Signal Service Bulletin in September of 1994, where cycle counts kind of came into the picture with the service bulletin came out in 1972. But that was fairly generous, same cycle count, same hours. It really didn't have much of an effect on maintaining or taking cycle counts into the longevity equation, if you will. But in 1994, through the Allied Service Bulletin, they drastically reduced the number of allowable cycle counts, and they created a framework within which those could be considered. The operators, for the most part, are operating either under that service bulletin or under the Army requirement. Why is it arbitrary if the FAA decides we're going to rely on what the manufacturer now says is their best estimate of the useful life or safe life of the component? We've tendered their theory to our chief scientists, and our chief scientists say it's technically sound. Without more, why can't the FAA rely on a manufacturer's model that has been affirmed by the agency's own scientists? The agency's own scientists have confirmed that the analysis is technically sound and that the computer program does whatever it does. But the agency is required, because of the expense involved and other issues, to look at everything it has that it can look at. And the fact of the matter is the analysis, first of all, and I think this is an important point, is the lifing analysis doesn't address an unsafe condition. It attempts to create one. FAR 39 outlines the criteria for an airworthiness directive. And the first criteria, and this is underlined multiple times in the implementing orders and regulations on how to apply FAR 39, is that you first have to have an unsafe condition. The lifing analysis doesn't look out into the field and say, hey, we have a problem out there. What's the proper number of cycles? The lifing analysis creates a new number of cycles, but there's no problem. There never was a problem. They've created a solution for a problem that didn't exist. And the FAA knows that. This lifing analysis cannot be looked at all by itself. You've got to look at the field experience. And the FAA likes to, or has said in a brief, well, what the RCAA wants is there has to be an accident. Well, first of all, when there's an accident, it doesn't necessarily follow. There will be an AD because there's a second prong under FAR 39. And that has to be that this condition will be likely to occur in other parts. But be that as it may, there is a very complex system in place. And it's field experience. It's known by the manufacturer, known by the FAA. These components were tested under the standard they were certified at the time. And unless and until someone finds they aren't safe under that standard, they continue to run under that standard. And they're doing just fine. There is a system under FAR 21 where the manufacturer is required to report to the FAA the service experience they're having with these parts. That's how the FAA monitors these things. That's how they determine whether or not this is something that we need to look at. There's also a service difficulty report system governed by the FARs, the SDRs, where the users report issues. Collectively, these and other things amass some data from which the FAA can make the determination, well, we've got an issue here. That didn't happen. Everybody was sitting there and everything was fine. There were no failures. There were no unusual returns. No reports from the manufacturer. No complaints from the field. And out of left field comes Honeywell's lifing analysis. So why was this lifing analysis initiated? In March of 2000 and July of 2000, Honeywell made two reports. The relevant portions are on pages 338 and 322 of the record. They state collectively in these two things, one, there are no failures. This is analysis only. We've just decided to do analysis. Under the title, reasons for life assessment, why did we do this analysis? In other words, two reasons. One is UNIJESU, which stands for, it's an acronym for Universal Jet Air Start Unit. It's a Navy project that was begun and tested between 1993 and 1995. It's an air start unit that sits on the ground or on the carrier deck used to start aircraft without APUs. It uses parts from the T-53 engine. It is not a T-53 engine. It is a very different environment. It's a very severe environment. And the T-53 components wouldn't hold up. PM Asteroid was invented in the 90s to solve that problem. They still didn't get the contract. The second reason was the Army. The Army said, hey, we need to look at this cycle issue. The Army still refused at the end of the day to adopt the cycles. And in the briefing we make reference in October of 2000, the Army issued what they call the safety of flight message. That's the same as their AD system. And it's encapsulated in a table on page 296 of the record. And it lists what the Army came up with after getting all this information from Honeywell. And it most closely resembles, if you apply normal cycle count to it, it resembles the 1994 service bulletin. It's not even remotely close. Is it any part of your argument that the lifing model of Honeywell's was insufficiently disclosed in the record? In other words, I understand your argument that this model didn't, that the FAA didn't consider the field experience and didn't respond to a problem in the field. But do you have any problem with the fact that the proprietary model isn't put into the record? In the sense that we can't cross-examine it. We saw one little peak at the three-second issue, which is worth. Right. And that was all wrong. Even the FAA had problems with it. We have, in the underlying, we have various messages and communications. The aircraft certification, the ACO office in L.A. expressed concern about this whole thing. They went back to Honeywell and said, you guys need to re-look at this. You have in your record an e-mail from the DER at Honeywell. Karen Vail said, no, we're not going to do it. We have in the lifing, or the white paper from the FAA, the FAA's got all kinds of concerns. Your Honor, I see my time is up. Thank you, Counsel. Thank you, Your Honor. Ms. Lehman. May it please the Court, I'm Vicki Lehman. I am representing Robert Sturgill, who is currently the acting administrator of the FAA. The FAA's decision to issue this airworthiness directive was reasonable in light of the results of our exhaustive consideration of the scientific evidence, all the field experience, and the public comment. RCA may disagree with us, but that does not render our decision to issue the safety airworthiness directive arbitrary or capricious. Counsel, the issue that is of concern to me from your side of the equation is one that Judge Canby alluded to in his question of opposing counsel. It seems to me that there are at least a couple of ways in which it makes the absence of things in the record makes it difficult for us to review to determine whether, in fact, the decision is reasonable or is arbitrary on the other side of it. One thing that's missing is the data from which the conclusions were made concerning the manufacturer's testing. And the second is the cost of compliance, which you provided some information to us, but it was not in the underlying record for public comment. At least I think that's true. Is the record that was developed sufficient for us to review the rationality, rational connection between the facts and the rule that you adopted? Yes, your Honor, I believe it is. And I'd like to start with the cost of compliance first. It was actually an oversight that the information that I provided to the Court pertaining to the cost analysis was not entered into the record. But more importantly, in the comments, the commenters had an opportunity to explain what the costs of the parts were. And they were. There were numerous comments in which they explained what they thought the parts cost and how it would impact upon them. These were extremely general. But we didn't mislead them into thinking that perhaps this would be an inexpensive AD. The original cost that we cited in our NPRM was even higher than the cost that we eventually came up with. And the people were saying, the commenters were saying, well, it's even beyond that cost. So it's even beyond the original cost. So if I understand that piece of it, you're saying that the mistake turns out not to make any difference because you already listed a cost that was even higher than what you ended up with? That's right. Okay. And as for the data regarding the scientific evidence, you know, the FAA has a balancing act. We are absolutely dependent upon the manufacturers to provide us with information that may be proprietary. All of the lifing analyses that the various manufacturers use to develop their directives are proprietary. But we must have that information in order to enforce the safety regulations, to make sure that operators are removing these parts before the parts break apart. So how do we conduct review in a situation in which the record is opaque, in which we don't know? You say sort of trust us, we're from the government, and we've analyzed this, and we want everybody to be safe. But how do we know whether, for example, there's been another mistake? You said that one thing was an oversight. Suppose a decimal point gets put in the wrong place. How would we even know that? I think the way you would know that would be to review the reports of the chief scientists, which are in the record. In these reports, they explain how they very thoroughly came to their conclusions. We had many FAA engineers working on this case. The chief scientists were called upon to review those issues that were still remaining. So whereas we don't have information in the system's report, so the chief scientists report on all of the scientific issues that they reviewed, we do have information in those reports on the most critical issues that they reviewed. One of them was the question of the acceleration rate, which Mr. Kilgore was just talking about. But that is just totally a red herring. It just doesn't matter. I mean, this is an example of the thoroughness of our evaluation because our chief scientists were really concerned. I'm sorry. Some of our scientists were concerned that perhaps the three-second acceleration rate was too fast. Okay, fine. Let's find out. Let's find out how this engine actually works. So our chief scientists said, Honeywell, rerun those tests. Run them with much, much slower acceleration rates. Honeywell did do that. This allusion to an e-mail comment by Karen Vail from Honeywell about how we're not going to resolve any more questions, that was way before the chief scientists got involved. And once the chief scientists were involved, Honeywell provided all the answers, all the additional information, all the additional testing that the FAA requested. And we requested a lot. But, of course, there's no way we can examine the model and what went into it, what it's internal. I'm very sorry, Your Honor. I didn't hear you. There's no way we can examine the model that was used for lifing because it's proprietary. And we don't know what went into it, and we don't know how it works internally. And if we're supposed to decide whether the FAA's decision was arbitrary or not, I guess I'm repeating Judge Graber, how do we do that? You say review the chief scientist's report, but there's no way of knowing whether his report reflects accurately what the testing showed. Well, Your Honor, I would think that if that was really a concern of Mr. Kilgore and our CIA, they could have asked, they could have moved for us to supply the entire record to the court under seal, but they never did. And it's not that these records don't exist. All the records exist, and they've all been examined extensively by the chief scientists, as well as by other agency engineers. It's not that these things don't exist. They really are real. Another thing to ---- Ms. Lehman, I have a question for you. Okay. Does the record tell us how often in the aviation industry a lifing analysis of a part that's subject to stress is done by a scientific modeling or computer, some kind of calculation model method as opposed to eyeballing experience? No, the record does not reflect that. And that was something that I asked the scientists while I was working on this, is how often does that happen? If you look at, for example, the 1986 AD or the 1987 AD, which is in the record. I think it's attached to my brief. The 1987 AD is an example of an AD that lowered a life limit of the T53 rotating components, and that was based exclusively on scientific analysis and testing. It is my understanding that most AD life limit cases ---- I'm sorry. Most ADs pertaining to life limits are based solely upon testing and analysis. And that makes sense in light of FAA guidance. If you look at our advisory circulars pertaining to life limits, and those are also in the record, they say that these life limits should be computed primarily based upon scientific analysis and testing. And I'd like to correct something that my colleague said about how this is based on some sort of theoretical computer model. I'm not disputing that there are some calculations that are done in computer analysis, but at the same time, these findings were then verified by tests of actual engines and tests of actual materials, so the D979 components themselves. I think that what's also very significant here is that when the chief scientists reviewed all of the data, and one of the things that they found was a very simple explanation for why it is that the limits today are so much lower than the limits were in 1986 and then republished in 1994. And that is because Lycoming had not used, when they did their calculations, Lycoming was limited by what the science of the day was. And the science of the day did not say you have to factor in the effects of thermal transients. Now, that's a complicated concept, and the only thing you need to know about thermal transients is that thermal transients will deteriorate these rotating components. So here's Lycoming. They haven't factored that in. Honeywell comes in after they bought the turbine division, and they're relifing all of the parts. They're going through all of the various engines that they've now inherited, and they want to make sure that everything is being done properly. And that's when they realize that the thermal transients have not been calculated in. So that explains why the life limits are so much lower. I'd like to turn to field experience because RCAA says all the time that we have ignored field experience. Well, we really haven't ignored field experience. What we tried to do was we tried to understand the field experience. And I think, if anything, what has happened here is that we just take different views on what that field experience shows. First of all, let me talk about the military because they have the largest bulk of the largest They've been operating the largest number of T53-powered helicopters. And here we just don't have the reports. We do know anecdotally, or at least we know, that the military was very concerned about the life limits of these rotors. And as the fleet aged, they were experiencing problems. But we do not have the reports from the military. So we can actually point to them and say, look, this is how many hours this rotor was operated, or this is how many cycles this rotor was operated. And as a result, the part fell. The part broke up and the engine fell out of the sky. We can't draw conclusions from data that we don't have. The chiefs then looked at the next largest group, which, granted, is a fairly small group, but they looked at the next largest group, which would be the restricted category operators. And they sought to understand why the restricted category operators weren't, as they claimed, experiencing problems if they weren't replacing those parts until they reached the higher life limits in the 1994 service boats. And to understand this, what the chief scientists did was they went out and they interviewed mechanics and others with experience with the T53 engine and with restricted category operations. And what they found was that the restricted category operators were basically over-counting cycles using methods not published in the service boats. It's important when you think of cycles, it's not just how many cycles, but how did you compute the cycles, and you've got to use limits that correspond with cycle-counting methods, and that wasn't being done. The restricted category operators had unknowingly been replacing the parts prematurely, even before the parts reached the lower life limits in the 2001 service boats. RCIA has argued in their reply brief, gee, that their members have been operating just fine and just leave us alone. Well, it just doesn't work that way. It's just simply fortuitous that the restricted category operators have been replacing the components at safe intervals due to over-counting. Had they been using the 1994 limits as well as 94 counting methods, there might have been problems. And thank goodness there haven't been. As a result, it was reasonable and responsible for the FAA to conclude that it's time to require operators to implement these lower, scientifically determined life limits that were published in the 2001 service boats and to use the corresponding cycle-counting methods to prevent engine failures. And that's where the unsafe condition is. An unsafe condition does not have to be that we are experiencing terrible problems in the field. It's the FAA's job to be proactive, to see where problems are going to come, to predict where they're going to come, and to do something about it. And in this case, based on the very compelling scientific evidence, it was absolutely incumbent upon us to take the step. And that's what we did. We issued this A.D. and we urged the Court to find that we set that our decision-making process satisfied the arbitrary and capricious test, and we urge you to affirm the issuance of the A.D. and to deny the petition for review. Thank you very much. Thank you, counsel. Mr. Kilgore, you have used your time. It will give you a couple minutes for rebuttal. Thank you, Your Honor. The FAA is probably, in the greater scheme of things, a good thing. But proactivity has its limits. They do have guidelines. They have standards. And that's what the standard review is all about. They can't disregard some evidence and rely on a single thing. There's a concept here I think that's important, and that's the difference between safer and unsafe. Once the aircraft is certified, it lives under its original type certificate and the engine, unless a condition develops that indicates that it's unsafe. There are aircraft that were designed and built in the 40s that are still flying. Would they be safer if they were built under 2001 standards? Of course they would be. Anything would be safer under the newer standards. But are they unsafe? No, they're not. They're still flying. They're still certified. There is still this predicate requirement that they find an unsafe condition. This they did not do. Even within the existing FARs, we have standards of safety. It's that balancing act that Ms. Lehman referred to. We have FAR 23 aircraft, which are the general aviation aircraft, and aircraft with a certain seating capacity. Then we have larger aircraft, the jets and the ones with larger seating capacity used as air carriers under FAR 25. Totally different standards. Much higher standard. One example, we only have one snapshot, one little tiny piece of the pie in that lifing analysis, the three-second thing. They went back. They addressed the issue. They looked at helicopters. They didn't even determine it could be done in three seconds. I personally don't think it can. I couldn't do it. Our members can't do it. And they had Honeywell rerun the report, and it did change the result. Did they change the service bulletin? No. Did they change the AD? No. Well, they said it didn't change the result, that it was minuscule, the difference. It was small. But they didn't do anything about it. It only did was prove it wasn't right. And you know these effects are cumulative. And we only know one. What if we had two others? It could make all the difference in the world. My time is up again, and I thank you. Thank you, counsel. We appreciate the arguments of both parties. They were very helpful.