# Oral argument — Mid Continent Steel & Wire, Inc. v. United States > Briefs, arguments, decisions, and more. URL: https://www.frixlaw.com/law-library/documents/argument%3A101460 ## Record - **Collection:** Oral argument - **Document type:** Oral argument - **Published:** December 1, 2025 ## Text 24-1556, and I think as the order last week indicated, we are interested in the question of remand. Good morning, Your Honor. It's been ten years. We've had seven judicial decisions. My friends here have been on the case the same ten years, and, Your Honor, Judge Taranto, this is the third time we've appeared before you. Our case here is not like Stupp and Martin. We've never challenged the principle of Cohen's D. It's not an issue. In fact, we've said, we've affirmatively said, we agree, based on prior CAFC decisions and our own analysis, talking to our economists, that Cohen's D, that this type of differential pricing test, it works. It may not be perfect, but it's reasonable, and it works. However- It works if the conditions defined by Cohen's D are applicable, which, in my experience, typically in these cases, they are. We're talking conditions. There are three conditions that Stupp and Martin talked about. Normal redistributing, sufficient size, roughly equivalent variance. We don't know if those exist in our case. What we do know is that our case is focused on a simple average. It's focused on the denominator. It's focused on the yardstick, where the Commerce Department and Department of Justice have said the yardstick should be a simple average. We know that that doesn't work. And as an alternative, we've provided this court with a choice. We originally said weighted average, and I believe in Mid-Continent V, the court said, well, we can have a single standard deviation for the denominator. And that is the test. That's the critical factor in our case, based on our facts. Right. But we're now at a place, and I guess I just want to bring you back to this question of remand. So there's never been a dispute that the statute has this term, significant differences, and that the agency unavoidably has to figure out how to implement that. The agency has now said, having read our opinions, we no longer think that the whole Cohen's D methodology for the first step is the proper one. We want to do something now that's different. Very much in the ordinary course, we would say, okay, go back and do what you're now going to do. You will have an opportunity, if it's unfavorable, to say that is wrong as well. Why is that not the right thing to do here? We have ten, again, I know the concurring opinion, I think in the Stuck in Marmon case, talked about the history of the case. We have a ten year history of this case. In this ten year history, we've talked about a choice. Marmon and Stuck just said, throw out Cohen's D. We're saying, wait a second, we have a binary choice. You could use your simple average, you could use your denominator, your yardstick, or you could use our yardstick. We believe our yardstick is reasonable. Reasonable based on the factors in our case, on our record. That's a reasonable way of dealing with a very difficult problem. If we choose to remand, what instructions are you saying we should give on remand? I think on remand, if it was up to me, the instructions on remand would be to use a single standard deviation. Instructions to do what was reasonable based on the facts of our case. Also, instructions would be to do something, you may not call it Cohen's D, but to do something that makes sense under the statute to find if there's a significant difference between the test group and the comparison group. I know I'm not supposed to get there, but I could just see it. What Converse did wasn't even close. When you say did, do you mean what they're now proposing or what they have done in the past? No. In the past, when they used Cohen's D, I believe Cohen's D was a good faith attempt. You just don't like the simple averaging. Well, in Cohen's D, I just don't like the simple averaging. Right. I think Cohen's D works for us in our case, based on our facts, based on our record. It wasn't a per se attack. It was based on the record. Cohen's D or something, whatever you call it. And I think one of the problems was that Converse kept calling it Cohen's D. And I believe the court said, call it something else, but don't just marry it to Cohen's D all the time. And I think the test really makes sense. It's a reasonable, good test for 10 years in our case. And the simple average, and I know when I read the opinions, and I'm not a statistician, in the opinions, the Marmon and the Stubb opinions, you were talking about, well, you know what? If you use a single-stare deviation, maybe the normally distributed problem goes away, or maybe the sufficient-sized problem goes away. And I think it does, because in those opinions, you always focused on the fact that if you had a different yardstick, the test works. And that's the choice we were giving the court. And now, what Converse has done, and again, I know that it has to wind its way up, they've come up with something that I believe isn't close to being reasonable. They've taken, saying if there's a 2% difference in prices... Right, but that's not really an argument that we're positioned at this point to entertain. That's my problem, Your Honor. No, I mean, it seems to me you have to be making an argument that there's something about the history of this particular proceeding, and the way in which the issue has been framed until, you know... That's absolutely the point. Two mornings ago. That's absolutely... And the government is somehow stuck with that framing here, and that's... I mean, I think I understand that's the point. I'm not sure how you get there, though. Well, the problem is... Frustration isn't enough. I'm frustrated. I don't get there by being frustrated. I don't get there by being, you know, what's happened to me after 10 years. I get there by trying to tell the court, one, that you have discretion. When you have a binary choice, you could say, hey, you know, I had a choice for 10 years, and this is what was litigated for 10 years, and I could say, in the context of this case, I'm going to pick one of two. Again, Stump and Marmon didn't give you a choice. You know, we gave you a choice. We said this over this, and when you use a simple standard deviation, or if you used a weighted average, it's reasonable, and the other problems with Cohen's d, based on our facts, go away. And I think we've done that with the examples we've given you. We've done that with, you know, really precise analysis from the court. I mean, we've had two opinions from this court. I'm not a statistician, but I learned more about statistics in the last 10 years, you know, than I ever knew what I would. And it's because of this, and it made sense. If we remand, are you contending we need to give them specific options? Or what are you contending? I would say, based on the history and facts of this case, on remand, you have to use a simple standard deviation. And absolute instructions based on the facts. And I think the court has a discretion to do it, and I think the court has done it. We've had cases, I think I cited one case, where, you know, there was a binary choice, and instead of remanding and letting Commerce have another shot at it, the court said, hey, do it. We had a case many years ago where the court said, hey, this case has been going on forever. We're directing Commerce to do it this way. Again, because of the discretion the court has, and the unique circumstances of the history of this case, and you're doing something that works. I'll reserve my time. Okay, reserve your time. We'll hear from the other sides. Good morning, your honors. May it please the court. I thought when we argued this case, that I'd be arguing to sustain Commerce's remand, and unfortunately that is not what we're requesting that the court do today. We're requesting that the court remand this case in the same manner that the court remanded Marmon and Stuck. For Commerce to employ a differential pricing analysis that does not include the Cohen-Stee test. And the reason why we're requesting the remand is because we can't defend this case based on Marmon and Stuck. In Marmon and Stuck, the court was concerned with three assumptions. And those three assumptions, based on my understanding of those cases, related to the mean. And the mean, as the court put it, was the predicate for everything else that followed. So it's the predicate for the standard deviation. Can I ask this? Yes, sir. If we vacate the CIT's decision, remand with instructions to return it to Commerce for further proceedings, will it be open to MidContinent to argue that the new approach, the 2% difference in weighted average net prices, is comparatively more flawed than... That's not the right way to put it. That there are problems with that and that, among other things, an agency has to give a sound explanation for changing what it was doing or simply in choosing among two options. And MidContinent will say one option is the weighted average denominator for the mean comparison and your new option is this 2% weighted average difference between the test and comparison groups. And here's why it is actually unreasonable for Commerce to now choose this new approach. That's open to MidContinent to argue, right? That is correct. Whatever determination Commerce makes on remand, whatever the final determination is, they get to challenge that. They'll receive a draft of the remand redetermination and make their arguments and their case briefs before the agency and then, if Commerce continues to follow whatever is stated in the draft in their final determination, MidContinent will have the opportunity to raise those issues before the Court of International Trade and then, if necessary, before this Court. But we're ahead of ourselves on this one because we're not there yet. Commerce has not applied any new test to MidContinent. What is before the Court here is, initially, it was a simple versus weighted average, right? That was initially the single issue before the Court. However, that simple issue that related to the denominator and the Cohen-Stee coefficient is undermined when the two groups that the data was derived from did not meet the assumptions that were discussed in Marmon and Stump. I respectfully disagree with counsel that these... Right. Just to be clear. So, I mean, I understand that Commerce's position is now, at least based on the case law, it can't do what it did in this case up till now. But there's an alternative, which is to make a very small change in what it did, which is to use a weighted average denominator. That's obviously the alternative MidContinent wants. On remand, Commerce presumably will say, we're going to do in this case this 2% difference that we've recently articulated and MidContinent will be able to say, we think that that is more unreasonable. It is unreasonable to choose that in comparison to the version of, let's just call it Cohen-Stee, just as label, that MidContinent favors. And they'll get to make an argument about that under normal APA standards about making reasonable methodological choices. Is that right? That is correct. And Commerce's determination within the context of the differential pricing analysis will be evaluated based upon its reasonableness. I can't speak to what MidContinent will argue before the agency. I cannot speak to what choices the agency will make. What I can speak to is that this case should be remanded to the CIT with instructions to remand to Commerce to employ a test that is different from Cohen-Stee just as the court did in Marmon and Stott. What do you think are the limits, if any, on our authority in reviewing agency action to simply dictate an outcome rather than saying that the outcome reached by the agency was in some respect flawed? Thank you, Your Honor. This case is different than CS Wind and Olympic adhesives that were cited in MidContinent's response to our motion. This case, the court has determined numerous times it's based upon reasonableness. Was the agency's determination reasonable? And the answer to that question is either yes or no. Now, if this were based upon substantial record evidence, then maybe the court would be able to do as the court did in Olympic adhesives and CS Wind and tell Commerce what to do because the record is clear. Here... So CS Wind was a case, I think, and on this particular point, was a pure substantial evidence question about how to do the weight balancing of these towers on ships. And I think we concluded the evidence supported only one result and there was no other dispute. The other case, I thought, ultimately was about what was required under the regulation. So it wasn't really substantial evidence but also wasn't reasonable methodology. It was more driven by the statute of regulation or both. Thank you, Your Honor. And this is in this category of reasonable methodology. Correct. Thank you, Your Honor. It's been a long day. For all of these reasons, Your Honor, we respectfully request that the court remand this to the Court of International Trade with instructions to remand to Commerce to employ a test that does not include Cohen State. And as a final statement, because this is actually my last argument before the court, I would like very much to say it's been an honor to practice before this Court for the last 31 1⁄2 years. Congratulations. Thank you. May it please the Court. Good morning, Your Honors. Adam Gordon on behalf of Midcontinent Steel & Wire. Just an initial clerical correction, if I could. There was some colloquy about what Midcontinent might do on remand. I'm sorry, I meant PT. I just wanted for the record. I think that was focused on what the Taiwan respondents might do. I always get confused because the appellant doesn't get top billing here. No. Just to confirm and be clear, Midcontinent supports and has consented to the government's motion for remand. We believe it is the appropriate step for the court to take in this case. I would offer just a couple of quick points. Picking up on the question of or the suggestion that there should be a directed remand in this case, the court has not done so in its prior decisions so far. I think it's important to harken back to the language that in part was used in the last decision. I'm trying to recall the exact words, but the court sent the case back with instructions to the CIT to remand it to the Department of Commerce to either provide a reasoned explanation as to why the methodology it was using departed from the established Cohen's D approach. So are you arguing for or against a directed remand here? I know you are supporting a remand. Against, Your Honor. The key is then, it said, either provide a reasoned explanation or use a different method such as the weighted average or the standard deviation of the entire population. It wasn't giving those, as Mr. Marshak was claiming, binary options. It provided latitude for the Department to develop another, a different approach. And in fact, in light of this Court's precedential decision in Marmon, Commerce did seek public comment and has adopted a new methodology. That's something that distinguishes this from a case that is purely based on the record evidence. We're talking about a commerce methodology that's applicable across all cases. And that's where the Department has adopted a new methodology. Mr. Marshak comments on the length of this litigation and how long this case has been going on. It has been going on for over a decade. But that's not a function of the Commerce Department's actions or Mid-Continent's choice to defend the anti-dumping duty order that was in place. That's a function of judicial process and the administrative process, how long those things take. It's not uncommon to have cases go on for a period of time, longer than many would like, including us. But that is not the basis to take the extraordinary step of issuing a directed remand. I would also comment that in essence what the Taiwan respondents are asking the court to do is to issue a directed remand that ignores a presidential decision of this court in Marmon. As we read Marmon, it invalidated the use of Cohen's D writ large, not simply Cohen's D in that case or in a different case. And that's reflected in the agency's actions subsequent to recognize this court's decision to seek comment and to adopt a new methodology which we also support. And I would also comment on Mr. Marshak's attempts to raise comments concerning the new methodology that the Department of Commerce has adopted. Obviously, that is not ripe for consideration or addressing by this court. I would note that in the underlying remand in Marmon, the remand determination is due back at the Court of Appeals. I think it's in the middle of this month. And then in Stuck, the remand determination is due back at the CIT at the end of January of next year. So that is still very much an active matter in the trial court. What, if any, time limits are there on the CIT acting on the assumption we vacate the CIT, not the CIT, the Commerce acting on the assumption we vacate the CIT decision and order the CIT to remand it to Commerce. Is there a clock that Commerce is bound by? The CIT typically remands with instructions to return a remand determination within a set period of time. Roughly? I think it's typically 60 days. I've seen 30 days, but that's a bit unusual. And 60 to maybe 90 days might be more common. I wouldn't want to be held to that. And I also think it's important to bear in mind that the agency is currently, quite literally, unbearing itself from the consequences of the recent furlough. They're dealing with over 6,000 documents they have to process still and sort out in the ongoing cases as well as ongoing litigation. That's not an excuse. That's simply a practical observation of the agency's resources and strengths plus a lot of people they've lost. With that, are there any questions? Let me also say, if I may take an aside, I would like to thank Ms. Cote for her collegiality and professionalism over my entire career. We have crossed swords on a number of occasions. It's always been a pleasure to litigate both opposite her and with her, and I'm going to miss her dearly. Thank you. Anything further? I'll say something nice also. I'm not going to say it's fun, but we've remained friendly over time. I just want to make a couple of points from the cases. First, on the Stub case, the concurrence talks about and I'll just call from the concurrence, were we riding on a blank slate? I might well be persuaded by Commerce that its use of the Cohen's deed test is one step in a three step differential pricing analysis. It's reasonable. And then, the judge says, but of course, we are not addressing the question of first impression quite the contrary. The history of the case really should matter. Second, in Marmon, and I'll quote from the Marmon opinion, Marmon says, because there is no dispute that Marmon's data does not satisfy these assumptions. We don't know if our data satisfies these assumptions. There are three assumptions. We never addressed it. We never looked at it. We have one critical factor which we thought takes care of the reasonableness of the Cohen's deed test. And then, if you look at the first Stub opinion, when the court was focusing, it says, and I'll quote from the court, applying Cohen's deed to that hypothetical data seems problematic. As the variance with each test group approaches zero, the denominator in the Cohen's deed equation is greatly reduced. And then the court says, although the problem in that situation, the function of Commerce's use of the simple average pooled standard deviation, our concern is also related to the number of observations being compared and the distribution of these observations. So you're taking one critical factor that the court focused on, and we're saying, correct that factor, and you have something that's reasonable, that works. Nothing is going to be perfect, but we believe this is close to really reasonable, and it's been looked at for 10 years, and you have the discretion, and you should exercise your discretion rather than we have to go back to zero. Again, I'm talking about the future, which I know you're not allowed to consider, but I don't think we're ever going to get anything as reasonable as Cohen's deed with a single standard deviation. Thank you. Thank you. Thanks to all parties, and the case is submitted. --- Source: Frix Law Library, https://www.frixlaw.com/law-library/documents/argument%3A101460. Public record. Not legal advice.