Oral argument — Albemarle Corporation v. United States

We have five cases on the calendar this morning. A case from the Court of International Trade, a patent case from the District Court, a patent case from the PTO, a Veterans' Appeal and a Government Employee Appeal, the latter being submitted on the briefs and not being argued. The first case is Albremarle v. Shanze et al. v. United States, Calgon, 2015-12-88-89-90. We're going to hear from Ms. Kramer for 10 minutes and then Ms. Burke for a total of 10, Mr. Liberto a total of 5, and Mr. Menegas for 5. And Ms. Kramer will have her five minutes of rebuttal time. So let us begin. I think the clock should be set at 10 minutes for Ms. Kramer. Good morning. May it please the Court. In its third review of the Anti-Dumping Order on Activated Carbon, Commerce assigned our client, Gua Sui, a fully cooperative respondent, an outdated rate of 44 cents per kilogram, despite the fact that he had calculated a zero and de minimis margin for the two mandatory respondents. This 44-cent rate was not based on any data from the current review period. The government says it lacks the resources to make individual rate calculations. What's involved in that? The same surrogate values are being used for everyone. Are we just talking about collecting information about the U.S. sales price? Is that what's missing? An administrative review proceeding is an involved proceeding in commerce-expended cases. Try to answer my question. Yes. So what is involved is obtaining information from the respondent. Are we talking about just U.S. price information since the same surrogate values are being used for all of the respondents? For Hua Hui, the answer is yes, U.S. price information, because Hua Hui gave its information to Jacobi, and Jacobi then submitted that information to obtain its zero margin. Yet, Commerce disregarded the fact that Hua Hui had provided information through Jacobi in this proceeding and assigned Hua Hui the outdated rate of 44 cents. So the U.S. sales price information for Hua Hui was already in the record submitted for Jacobi? No, Your Honor. It was the other side of the dumping equation. The surrogate values? The surrogate values. So what the government would have to do to calculate individual rate for Hua Hui is get U.S. sales information? That's correct. That's correct. But here, it did not have such information. So Commerce reached back to outdated data from the prior year's review, and this 44 cent rate was not consistent with the statute, and nor was it supported by substantial evidence. A complete lack of evidence, there was no evidence on pricing data in the review, cannot meet any standard of substantial evidence or reasonableness to support that 44 cent rate. Commerce's rate also was not consistent with the statute because the agency did not use one of the statutorily sanctioned methods of calculating the anti-dumping rate, either the expected method or some other reasonable method. You don't attribute much significance, I take it, to the fact that the statute refers to investigations and not administrative reviews? Well, no party has seriously contested Commerce's practice to use that part of the statute, the part of the statute referring to investigations. Well, but the question really would be how binding is the statutory language? Commerce has said on several occasions, I gather, that in these kinds of administrative review situations, it will not do what the statute describes as the 735C5B procedure of looking to the mandatory respondents when they have a zero, or de minimis, but instead will use another process such as the process used here. They say that's one reason for that is because the statute isn't directed to administer reviews. And what's your response to that? Well, they say that they take guidance from that. Right, but they don't regard themselves as bound by that because the statute does not specifically refer to administer reviews. Even if they're not bound by it, though, the statute says any other reasonable method, and you get to the same place, really, if you do a Chevron 2 analysis. If the statute hasn't spoken directly to how the calculations should be made, then Commerce is given discretion to the extent that it's acting reasonably. Well, then that gets us to the question of whether this is reasonable. And what's wrong with Judge Stanciu's conclusion in a word that Commerce was faced with the question of whether to rely on outdated information, but information that was specific to Huawei, versus current information, but information that was not specific to Huawei? And they chose, given those two choices, neither of which was ideal, they chose the specificity over contemporaneity. What's wrong with that? Well, the 44 cents was not specific to this review period. Conditions change in each review period. No, that's true, and Judge Stanciu acknowledged that. So what he said was, I acknowledge that we don't have current information with respect to Huawei, but what we do have is Huawei's immediate preceding administrative review number, which was specifically calculated for Huawei, unlike Charisman and Shanxi, right? Well, the 44 cents, again, was not specific to this review period, but Commerce has to look at the entire record, and it had several pieces of information on the record. But it could also ask for additional information about the U.S. sales prices we were talking about earlier, right? I mean, that certainly was open to them under what the Corporate National Trade asked them to do. Yes, Your Honor, it was open to them. And, in fact, Huawei suggested that in the case brief stage at the administrative proceeding, asked Commerce perhaps collect quantity and value information, and backing up, Huawei actually asked to be a voluntary respondent in this proceeding. So it asked to have its own information used to calculate a margin, but due to limited resources, Commerce said it couldn't do that. Well, how difficult would it be for Commerce to make such a calculation? Well, Commerce would say difficult, because it says that it only had the resources to collect information from two respondents. But, as this Court has said in BESPAC and, I believe, Cheng Jiu-Jin, concerns for administrative resources can't trump accuracy. And what we had on this record that showed that that 44-cent rate was not accurate as to Huawei are several things. One is the – But answer my question. How difficult would it be for Commerce to make an individual calculation for Huawei, since the surrogate values it already has, and it's just talking about U.S. sales price? Well, it could have made it simple by using some reasonable proxy for U.S. price. One of those ways would have been quantity and value information that it, in fact, obtains in other cases when it's trying to select respondents. It didn't do so here, but it's routine for Commerce to collect that kind of information. As far as the information that shows that that 44 cents wasn't reasonable – Huawei did not apply to be a voluntary respondent in this case, right? Huawei did request. They requested voluntary? They requested, but then that didn't go any further once Commerce selected two mandatory respondents. Oh, I thought it was one of the others. Okay. So what shows that that 44-cent wasn't reasonable is, first, the trend in the margins. In the investigation, CCT and Jacobi were also mandatory respondents back then, and they had margins in the range of 60 to 70 percent. Then the margins went down in the second review. Then Commerce changed its methodology and went to a per-unit basis. And then between the second and the third review, margins also changed. Now CCT and Jacobi received zero margins. The surrogate values bolstered this story. I'm sorry. The financial ratios bolstered this story. So the financial ratios, which are a key component of the normal value part of the equation, went down by about 50 percent between the second and third review. So we know that all other things being equal, Huawei's margin would not have been the same as it was in the second review, because that key component of the dumping margin, the financial ratios, went down. So it's unreasonable for Commerce to pull forward that outdated 44-cent rate in view of the fact that we know that one part of the calculation would have gone down. Commerce says that it has no data on the margin to determine the separate rate company's pricing behavior, but nor is there any information showing that that 44-cent rate is consistent with Huawei's pricing behavior during the period of review. If there were such information, then it would be okay to use the 44-cent rate? Well, that would be speculation, Your Honor, because that information was not submitted on the record, and we know that the financial ratios went down between the second and third review. So we know that the 44-cent rate, all other things being equal, would not be the same. Another way we know that that 44-cent rate would not be the same is the surrogate values that were determined anew in this third review. We cite this in our brief, but if you look at JA101185 and JA200120, you see the surrogate value chart that Commerce relies on to determine the normal value. And that surrogate value chart shows that those rates are from the period of review, not from the last year's review when that 44-cent rate was calculated, but from this year's review. The statute says that Commerce shall conduct administrative reviews every 12 months, and that indicates that Congress's intent was there to be new analysis each review period. Commerce can spend considerable resources calculating those margins, as it did for CCT and Jacobi, and it essentially ignored those margins by pulling forward that 44-cent rate. I see that my time is almost up, but one important factor that I did briefly mention before is that Wahwee's information is part of Jacobi's zero rate. Wahwee is a separate rate company who is a supplier to Jacobi, and so part of the reason that Jacobi received that zero rate was because of Wahwee's own information. Thank you, Ms. Kramer. We'll give you five minutes to rebuttal later. Thank you. Ms. Burke, we're going to take eight minutes. May it please the Court, the trial court's original determination should be reversed, and I'll start with that part of our argument and then I'll obviously respond to the appellant. The only question for the Court is whether Commerce- How difficult would it be to calculate a separate rate, which is what they originally asked for? Well, as the appellant correctly stated, Commerce has part of the information. You have the surrogate value information. We have the factors of production information, but only for the portion of product that Wahwee supplies to Jacobi, not all of Wahwee's products. So we have part of that side of the calculation. We have none of the other side of the calculation. Which is the U.S. price. Right, we don't have any- So why, what's so difficult? If you don't want to use the de minimis rates, what's so difficult about getting the U.S. price information, which they are willing to supply to you and calculate a separate rate for? Well, as the appellant just said, it's actually quite an involved process. I don't recall that they said it was an involved process. They said it was a complicated calculations that Commerce undergoes. Commerce looks at information about customers, quantity, value, affiliations. It does all kinds of very complicated calculations. Sometimes it does verifications. It's a huge undertaking, and it is exactly- Huge undertaking? Yes, and it's- What is there in the record that shows it's a huge undertaking? Well, in the memo where Commerce determines that it can only- It only has the resources to take two mandatory respondents, and it does not have any resources to take on voluntary respondents. What I'm asking is what shows that it's a huge undertaking? I know they say we don't have the resources, but there's no explanation of that. Well, there's an- What Commerce says is that it doesn't have the resources to examine anymore, because examining more would be too difficult. I don't- I mean, it's- So, but I don't understand how it can be that Commerce can say, because we don't have the resources, we're not going to come up with an accurate rate. That doesn't seem to be very reasonable. Well, that's actually what the statute contemplates. You have a statute that contemplates that you are going to have no information about the separate rate companies. That's why the separate rate companies get, in a normal situation, an average of what the mandatories get. Yeah, and we've had prior cases where, in AFA situations, Commerce, in accordance with the statute, has used the prior rate. And we've said that that's reasonable under those circumstances, because it's a fair inference that the non-cooperating respondent didn't have a better rate or it would have supplied the information. Here, it seems to me that the basis for that kind of presumption doesn't exist, because this is not an AFA situation, and they've offered to give you the information. It's true that it's not an AFA situation, and obviously in the AFA world, you have in the statute that Commerce can look to other segments of the proceeding. But here, you arguably have something that gives Commerce even wider discretion. You have language that says- Wider discretion than in an AFA case? You have language that says Commerce can use any other reasonable method when you don't- when you're in a world where you only have either AFA rates or de minimis margins or zero margins. Ms. Brent, tell us about your cross-appeal. You want to get the rate up for the other suppliers? Right, so we think that the trial court erred in a couple of different ways. First of all, the trial court made factual findings that exceeded the court's standard of review. The court stated that the de minimis margins must be considered more representative of the industry-wide pricing behavior during the POR. Not only is that a factual finding that the trial court does not have the authority to make, it is inconsistent with this court's recent decision in NANYA. NANYA was very clear that the only thing that Commerce has to do is follow the statute. Notions of commercial reality do not fit and do not apply when you have a statute that gives Commerce such wide discretion. So NANYA overruled all the earlier cases that said there has to be some commercial reality? No, NANYA was very careful to explain that when this court has found that Commerce's decision is inconsistent with commercial reality, it's because the court either found that Commerce had made an error, a calculation error, or that it simply just wasn't consistent. Isn't the rationale underlying the whole scheme of having mandatory respondents and then separate rates, which are calculated not on the basis of an investigation of each of the separate rate respondents, but based on the mandatory respondents, isn't the rationale that the mandatory respondents are assumed to be representative of everybody else? No, not really. That's what Amanda says. Everybody agreed to that. Presumably Commerce was one of the everybody there. Isn't that the rationale? Otherwise, isn't the system irrational? The statute gives Commerce two choices. It can either use the volume methodology that it used here, or it can do a sampling methodology. Commerce doesn't use the sampling methodology that frequently. In 2013, it issued a notice for comment suggesting that it was going to start using the sampling methodology more, and it was going to refine that methodology because this volume methodology allows the smaller respondents, who know that they're never going to be investigated or reviewed, to basically do whatever they want. But you have a statutory structure that allows Commerce to set dumping margins for the separate folks at a rate which does not represent investigations of them. You have the choice. I assume the rationale for that is that those dumping rates are assumed to be fair for those separate respondents because they are assumed to be dumping at about the same rate as the mandatory respondents, correct? Isn't that the basis for saying you can hit these separate respondents with the same average dumping margins? I think it's partly correct. I think what it is is a crude way to allow Commerce to perform its job within its resource constraints while capturing as much of the industry as it can possibly capture given its resource constraints. I don't think we would concede, and I know that the Court of International Trade said that we conceded that it was representative in Amanda Foods. It's actually not quite true that we did, but leaving that aside, I don't think it's technically representative. Representative is a technical term. A sample is representative. Let's take the non-technical sense of representative. Is it a rough but roughly fair way? Yes. Okay. And then if that's true, then what's wrong with saying where you have comparison between the current attributed rate, which would be zero because of the two mandatory respondents are zero, you have that to look at versus a previous attributed rate, which was not Huawei's rate. I mean, not the Cherishman and Zhang Xi's rate at $0.28. Why isn't it better to look at the current attributed rate rather than the past attributed rate? Why isn't that more reasonable? Well, because you have a statute that assumes that you actually shouldn't do that. And then you have this court's case law that even further constrains commerce. But what in the statute says you shouldn't do that when you look at 735C5B, which says you can do that. And, in fact, the SAA says that's the expected method. It seems to me it's just the opposite. I'm a little bit over my time. I assume I can answer the question. So you have the statute that the first part of the statute assumes that you're not going to look at the de minimis and the AFA margins and the zero margins. 735C5A, which doesn't apply here. Right. The second part of the statute says when you're in this awkward situation, you can do anything else that you want, including averaging them. So you have the first part of the statute that assumes that these, I'm going to call them outliers, shouldn't be factored in. Then you have the second part of the statute that suggests that if you're in this situation, you might need to let them in. You might decide that it's reasonable to let them in, but you don't have to let them in. But the SAA seems to say that the expected method is to average the de minimis and zeros and that that will not be used but only if commerce can conclude that it would not be reasonably reflective of potential dumping margins and wouldn't be feasible. I don't see why it's not those two criteria are met in this case. So commerce went through that part of the SAA and determined that it wasn't reflective of, and I think potential is the key word, it wasn't reflective of potential dumping margins. There's the assumption. In the IND memo, which devotes a total of 20 lines to this whole issue, they went through that? To the extent that it quotes the SAA. But the exercise here is to figure out what the SAA assumes we don't know what these companies' dumping margins are, and that's why it says it should be reasonably reflective of potential dumping margins. Why is this reasonably representative? What is the basis for believing that these earlier margins are accurate with respect to these parties that weren't individually investigated? Well, that's where we get to this court's recent precedent. In BESPAC, the court said that whatever you, well, the first thing the court said was in theory. No, but forget about the theory. What is a factual matter suggests that these earlier margins are more accurate than averaging the de minimis margin? Because they bear, and I'm using language from BESPAC, because they bear some relationship to the actual dumping margins of the separate rate respondents. But not for the period in question. But it doesn't need to be for the period in question. What's the point of having administrative reviews if you're not trying to calculate an accurate margin for the period in question? We're not trying to calculate an accurate margin for the separate rate respondents. We're trying to give them, we're trying to figure out. That's a pretty damning concession, isn't it? I don't think it is. I think the whole point of the statute is that the separate rate respondents are not going to get their own margin. There's no way that they're ever going to get a margin that actually reflects their dumping, because we are never going to have that information. But you're not going to use the representative rates that you calculated for the individually investigated parties. And you're not going to calculate an accurate rate. You're just going to use something from the past that has no necessary relationship to the current period? Commerce does not think that it has no relationship. What's the basis for believing that it does have a relationship? Because BESPAC instructed Commerce to figure out a rate that was going to bear some relationship to the separate rate respondents' actual dumping margin. When we don't have any information, which we don't have, Commerce thought this was the next best option. And that's its stated policy. But you didn't have any information with respect to Cherishment and Chancy in the prior AR. So how is that information any more reliable than the information that's available to you with respect to the present AR? Well, at least that information was based upon dumping margins. Other than zero. Right. Of other people. I guess what I'm trying to say is that for that period of review, we were in Section A of the statute. Meaning Commerce was able to calculate the separate rate using calculated dumping margins. And so if you want to think about the separate rate or the volume methodology as being representative, which we don't really like that word. But if you want to think of it as a crude way to get to something fair, in the prior review, you had enough information to get to something fair. Whereas in this review, we don't. Ms. Burke, we'll give you two minutes for rebuttal. And we'll now hear from Mr. Luberda for the domestic industry. And you've got three minutes now. Thank you, Your Honor. Just to continue with the conversation you were having with the government, Your Honors. The rate that was chosen from the prior review, the immediately prior review, was reflective of the potential for margins. Again, nobody knows what the actual margins were going to be. But why is it more accurate than the current rate, which the mandatory respondents have? Since neither was calculated for, at least setting aside Huawei, neither was calculated for Cherishment or Shaanxi. It doesn't have to be more accurate. There's no way to know. How can it be even as accurate, since it's not contemporaneous, as the current mandatory respondents' dumping margins is? Well, it has to be a reasonable, they have to choose a reasonable method, which has to be based on some measure of substantial evidence. So there's no substantial evidence that supports a zero. But isn't what's going on here is that Congress does not like de minimis rates and that anything is better than a de minimis rate? Isn't that what's involved here? The statute disfavors both AFA rates and de minimis rates. Why is that, by the way? What is the policy underlying 735C5A of saying we throw out de minimis rates? I wasn't able to find anything that suggests why that is something that makes sense. The reason behind it is based on, so AFA, I think everybody acknowledges, is an exceptional circumstance. AFA, I understand. But I don't understand the focus on de minimis and zero as being somehow unreliable. It's not that it's unreliable. Why do we throw it out in the averaging? There may be cases in which you wouldn't throw it out. I understand. But I'm looking for you as an expert in this field to tell me why it is that the statute, 735C5A, says throw it out. Because it's presumed not to have been earned by the company that's applying for it. Suppose the company decides, you know, we got hit by this investigation with a dumping rate, an anti-dumping order, and we are going to comply. We're going to not dump anymore. And so they go to zero. Haven't they earned a zero rate in that situation? If they've actually gone to zero, and you can show they've gone to zero. But Congress has concluded that they went to zero when they give them a zero rate, right? Congress has concluded that the examined respondents got a zero. Yes, that's what I'm talking about. I'm looking at why it is that those examined respondents having a zero rate isn't an important data point. What's the policy behind saying it isn't? Well, in an original investigation, it would be an important data point, which is why I think the statute says that you can average them in an investigation where that's the only information you have. Because you have no other prior information you can rely on. And in a review, the agency has prior experience, and in this case had two previous reviews of less than fair value investigation, in which there was no evidence that any other respondent previously had ever earned a zero. But look, the world has changed between the two reviews. With respect to the individually investigated respondents, they go from having a substantial dumping margin to having de minimis. Doesn't that suggest that the same thing would be true of other importers as well? Not necessarily, because we don't know how those other importers price. But it certainly doesn't suggest that the other importers are continuing to dump, does it? It doesn't suggest that they're not, and if you look at the history of this case. In other words, you can't tell one way or the other? I think you can make a reasonable assumption. Your Honor, if you look at the previous review periods, where Jacoby's margin went down from 18.1%. But what I'm asking is why would you think that the previous review period margin would be accurate with respect to the non-investigated parties, when it's shown to be inaccurate with respect to the investigated parties? If there was some homogenous nature by how these numbers moved. I mean, Congress made a determination that if you calculate an actual margin for somebody, there are no outlier margins, we're going to go ahead and use those. But we would assume, or Commerce has the ability to go back and look at behavior of the respondents. So when the margin goes down for Jacoby in the second review, but it actually goes up for Hua Hui, to assume that because Jacoby went to zero, you get this outlier number. I mean, that wasn't my choice, it was Congress's choice. But because Jacoby goes from 18.19% down to 11 cents, Hua Hui actually went up to 44. So to make the assumption that these folks who've never demonstrated low or de minimis or zero rates are entitled to that, Commerce was entitled to make, under the statute Congress gave them, the ability to choose between the less contemporaneous, more specific, versus the more contemporaneous, less specific. You asked, did you not, that Commerce collect data for the individual respondents, right? We did, Your Honor. We asked that they do that because our choice was the trial court had left us with the option that based on the decision of trial court, there could be nothing other than use of the zeros unless other information was collected. But the statute was written specifically not to require Commerce to collect data. But our option for our side was, well, if we're going to get stuck with a zero or go get data and maybe get something else, we ask for more data, but we don't think that's what the statute requires. Mr. Lombardo, we'll give you two minutes of rebuttal, and let's hear from Mr. Menegas. Thank you, Your Honor. May it please the Court, just for sake of clarity, I represent Shaanxi DMD, and I'm making a consolidated argument for the various separate companies, including Beijing Pacific and CherishMet. Is CherishMet and Beijing Pacific, are they the same company? They've been collapsed, I believe, generally. Okay, so we're only talking about a total of three separate respondents in this case, right? We're talking about two groups, Shaanxi DMD on the one hand. Shaanxi DMD and CherishMet, right? Right. Okay. I could maybe pick up on Judge Bryce's question from earlier. Citing to page five of CherishMet's brief, they did also request voluntary respondent status. CherishMet did. CherishMet. Shaanxi DMD did not. And at the top paragraph in that, I want to quote what the government said in its respondent selection memo, quote, the department recommends not calculating individual dumping margins for any non-selected companies that may place full responses on the record. So the position of the government is we will shun anything you attempt to put on the record. And it does take a lot of effort and time to gather all your costs and sales information. And so in this case, apparently, Huawei and CherishMet selected not to submit that information that the government clearly said they were not interested in getting. But, you know, this court in Amanda Foods, the trial court in Amanda Foods and this court in BestPak said, look, you can always go back and get more information, but you can't make a decision based on this record based on no information. And so we just think, you know, they were given multiple opportunities. And Shaanxi DMD's position in its brief was that commerce was not compelled to do any particular thing on remand. The judge left it wide open to even open the record up in any way they saw fit. And they chose not to do it, and they delivered de minimis margins. So we hope the court could simply affirm on that basis. But we also, you know, I'm very glad to hear what I've heard from the bench today because it's covered a lot of the ground that I hope to cover. But what you've heard from the government, their position is the statute gives us two choices. Pick a representative sample or a non-representative sample. And I agree with Judge Bryson's suggestion that that's rather silly and totally unreasonable. It could never survive a Chevron 2 examination. Can you help me with the question I asked your opposing counsel? What's the policy underlying this hostility, if I can use that word, to zero margins? I just found that... As we typically represent respondents, we don't appreciate the policy. You may not like it, but can you tell me what it is? It does appear that there's a slant in the statute, but there is a savings clause, as we've all discussed here, that when all the margins are zero and de minimis, then the department has the opportunity to average those. And the SAA indicates, you know, unless they wouldn't be reasonable. So we think what the trial court merely did is go back and explain why the contemporaneous record is not reasonable. When you've set up the statutory scheme, you decided to select two respondents. You decided to shun all other information in the case. That leads to the only logical conclusion that they deemed that that selection of the two respondents was going to be representative of the parties they had no interest or ability to review in detail. And so they just don't like the result. I wonder if the court's aware that in an investigation, de minimis is 2.0 margin points, right? In a review, it goes way down to 0.5. So we have a very high standard to even get to de minimis, and the two largest exporters got to de minimis, and now they don't like it. But so if even one of them had a 0.6, they would not dispute that it was representative of Chauncey DMD and Cherishman's rate. But the minute it drops to 0.5 or 0.4, it's totally unrepresentative. And that happened, I guess, in this case with the preliminary investigation. It was 5 cents, and then it went down to zero. If it had stayed at 5 cents, then everybody would be at 5 cents, right? We would have certainly been happy, and probably the petitioners would have appealed. But so it's a very problematic situation we're in. And, you know, the government mentioned this sampling FR notice for comment to do more sampling, and they've done almost no sampling since the 2013 notice. And the criteria in there says, well, we have to first pick three respondents in the criteria. So guess what? Now they're not picking three respondents ever because they don't want to trigger their sampling methodology because they don't think they have the resources to review three or more respondents. So we're stuck with this two largest exporter formula in almost all of our cases. And we'd sure love to get some guidance from this court that says you can't say that's totally unrepresentative just because it dips to 0.5. And so, I mean, I don't know if the court has any questions for me, but I just think the statutory scheme contemplates you might not have information about the separate rate companies and that you're going to capture either a statistically valid sample or the largest group of sales you can gather, which is usually the two largest formula, and it usually captures 60% of exports. They usually try to meet that barrier. I can't say on this record whether they hit the 60%, but that used to be their policy, that we'll take the two largest or up to 60%. And over the years they've just winnowed it down to two in almost every case citing this lack of resources. They can't now blame us and apply an adverse inference against us for their lack of resources. This court and BESPAC said that's not okay. Thank you, Mr. Menendez. Ms. Kramer, you have your time, five minutes. Thank you. We have to remember that the overall purpose of anti-dumping laws is accuracy. This court has said that all the way back to Roan and perhaps before. So it is the case that the separate rate must be accurate, just as it must be accurate for the mandatory respondents. The expected method is the one method called out in the statement of administrative action that considered the authoritative interpretation of terms in the statute. And as the lower court, the trade court has said, it's unreasonable to take that one expected method and deem it the one method not to use. They said that in the Amanda Foods case. I think the logic there is noteworthy. Mr. Menendez touched on what happened with, and Judge Bryson, what happened with CCT. And that shows the illogical position of the government that, or really maybe of the statute, that when you have a five cent rate, that's okay. But if that five cent rate, if that rate goes below de minimis, then for our client, its margin increased eightfold. That can't be a reasonable outcome in view of the other objective criteria we have, which are the decline in anti-dumping margins since the order all the way down to the third review. And the decline in financial ratio is an objective factor. It's on the record. And also the surrogate values that are all new in this review period. Do you happen to know what the conversion factor is between cents per kilogram versus percentage? It's a good question. We've got apples and oranges floating around in this case. I was doing some back of the envelope to try to be able to give the court something that's concrete on this. Just an approximation, ballpark figure. If you import a ton of an R client. You don't have to work me through the math. Just give me a ballpark figure at the bottom of the envelope. Yeah, it's based on proprietary information, so I can't give it. But this is what I can say. Based on a ton, our client is paying $440 for each ton that it imports. So that's significant if they're importing. But that doesn't tell me what the percentage is. It doesn't tell me whether that's 1% or 80%. Yeah, I apologize that that's on the proprietary record, and I can't divulge it. But Judge Bryson, you were asking before about specificity. Commerce was favoring specificity. Well, it seems like to us that that's a results-oriented outcome in view of what happened in the Home Meridian case, where a respondent came forward and said, use our specific pre-POR data rather than contemporaneous data that was a basket category. And Commerce urged and was ultimately upheld here that Commerce could elevate contemporaneity over specificity. Here they're asking for the opposite, and it seems like that's really in order to shut the book on the current review period and to instead look back to what happened in a prior review period just to give our client a margin. Another factor I wanted to bring up is that it's not even really correct to call the prior review period temporarily proximate to this review period when you think about the fact that the first month of a prior review period is separated by 24 months from the end of the next review period. So certainly it's reasonable to conclude that prices change, that surrogate values change, as we have already indicated. And for that reason, because there's no evidence on the record indicating that that 44-cent rate has any connection to the current period review, we ask that you overturn the trade court. Is it your understanding, based on Commerce's policy as articulated in the frozen shrimp and frozen fish fillets cases and so forth, and in this case, is it your understanding that if this case went on and had five more administrative reviews and Jacoby and CCT were the mandatory respondents in each of those, and in each of those, zero was the margin, that the margin for Huawei would remain at 44 cents? Even if Huawei was not investigated? Commerce said in the issues and decision memo that it looks back to the prior year's review in this kind of circumstance. So if we carry that forward, yes, it would seem that Commerce would do that each time. But it also says that it makes that decision on case-specific bases. So if we take that at face value, they are supposed to examine the record each time. But here they didn't examine the record, for the reasons I indicated regarding what we know about the objective factors that changed since the prior review. So your suggestion is that perhaps after an additional year went by, they would say, well, now, contemporaneity is not so close. We're going back two years, in effect, even though the 44 was assessed for the immediate preceding year. That is what they've been doing in these cases, is just ignoring contemporaneity and looking back to where we find the margin. Are there any cases, to your knowledge, in which they have double jumped, in effect, and had gone back to a year such as, in this case, the prior administrative review, used that for the next administrative review, and then justified the very next administrative review as having a particular margin, based on the fact that the attributed review for the previous year, which came from the year before that, was positive? Do you know of any such cases? Yeah, I can't name such a case, but I do believe there are any, but I don't want to misspeak, so I won't comment. Thank you. Thank you, Ms. Kramer. We'll hear further rebuttal on the cross-appeal from Ms. Burke, two minutes, and then from Mr. Liberta. Thank you. Just picking up on where you left off, in the brake rotors determination, Commerce had a history, Commerce was looking at a dumping history for the order over the course of eight years, and I'm not totally sure this is accurate, but I think it was almost a full eight years of no dumping. And in that situation, Commerce looked back and said, okay, it's not reasonable not to average the de minimis margins. Now, I don't know where the tipping point is, but there's obviously a tipping point between a three-year history, where it looks like there's been dumping, and an eight-year history where there hasn't. Is it accurate to say that if Commerce had determined a five-cent margin here, that it would have used that for the non-investigated respondents? Yes. If that's true, that that's not de minimis. That was the preliminary investigation number that they came up with for maybe Shaanxi, I think. Yeah, if that's not de minimis, then that's correct, yes. No, CCT, I'm sorry. I'd like to just say a couple words about the expected method. I don't know the answer to the court's question about what the rationale is, about why the statute is written the way it is, why all we have in terms of legislative history is the Statement of Administrative Action. I will say, though, that although we use the statute and the SAA by analogy for these administrative reviews, it is important to remember that it was written for investigations, and it technically applies for investigations. So when you're in a situation where you're trying to figure out whether there's been dumping, maybe it makes sense as a mathematical proposition to throw out the high and the low when you're at those initial stages. And when you get to the SAA, you also have to remember that it's quite possible that Congress and the administration were assuming that there would not just be de minimis and zero margins. There would also be AFA margins. Because if you don't have AFA margins and you just have de minimis and zero margins, you don't have an order. There is no anti-dumping order if everything's been zero. Well, if that's the result of the investigation, but if that comes up in an administrative review, you certainly can have those numbers come up. That's right. And the SAA applies to reviews as well as investigations, I believe, does it not? No. I mean, the SAA is the – this part of the SAA is the explanation for the part of the statute that just applies to investigations. So why in an investigation, when the investigative respondents all have zero margins, is at the end of the investigation, but in an administrative review where the same thing happens, you don't use the de minimis margins? So in an investigation, if there are rates calculated and there are no dumping margins calculated, then there is no basis upon which to issue an order. I'm not sure I understand the question. Well, I'm just saying that in an initial investigation, if the investigative respondents end up with de minimis rates, you don't try to impose – There is no order. But in an administrative review, if the investigative respondents have de minimis rates, you think, well, that's not going to lead us to assume that the other parties aren't dumping. Well, I mean, in an administrative review, even if there are a series of zero years with zero rates, Commerce still has an obligation every five years to do a sunset review and determine whether there should be an order. So it sort of accomplishes the same thing. Thank you, Ms. Burke. We'll hear two minutes, if we need to, from Mr. Liberton. Thank you. And looking at the statute quickly, the statute taken as a whole is disfavoring using AFA and using the de minimis. And just picking up on Judge Dyke's last question, there is sort of a shift of burden. In an investigation, when representing petitioners, I have to prove dumping. Our petition has to get shown that we've made all the allegations necessary to show dumping and injury. The burden is on the government and the domestic industry to demonstrate that dumping has occurred. In a review, there's something of a shift. Now the respondent data is being looked at to determine, having already been found to be dumping, whether that dumping has gone away. But Commerce isn't giving the non-individually investigated parties the opportunity to show that their margins have changed. And Congress said they didn't have to give it to everybody to do it, recognizing that it would be an impossible burden. I mean, I have cases where there are hundreds of respondents. There aren't hundreds in this case. There potentially could be dozens. But everybody who's here today but one of them has had a chance to demonstrate that they weren't dumping by looking at their actual margins, and not one of them has in the past. So Commerce, recognizing that Congress gave them a choice. But what I mean, none of the non-investigated respondents was given the opportunity to submit data. Not in this review, but in previous reviews. It's not like it's completely unfair that they never get chosen. It doesn't seem to make sense to say that the failure to consider their data in this review is reasonable because they didn't submit information in some previous review. The statute is set out in such a fashion that they don't have to consider all of it, recognizing that it would be impossible for Commerce to consider everybody's data. The outcome, based on what the lower court did here, would be that in every single case, you would always have to use the current data no matter what. But the statute specifically says Commerce can use any unreasonable data. No, I don't think that's the point. I think the point is that if you're going to use the prior data, you have to have some reason to believe that that is representative of what the current data is. And on this record, Commerce hasn't been able to point to any reason to believe that the margins for the non-investigated respondents would be accurate for the current period. There is also no evidence, and the court below said there is absolutely no evidence, to tie the pricing of these non-investigated respondents to the pricing of the respondents who were looked at. Would you agree that the reason, the justification for using the mandatory respondents' numbers with respect to the separate respondents is that there is an assumption subject to rebuttal, but an assumption that those rates are representative? Yes, except in the case where you have AFA. So you disagree with the government's position, which is that representativeness has nothing to do with it? No, I don't disagree. It's not that representativeness has nothing to do with it. They are not specifically representative. But they are assumed to be good prospects. There is no information that you have on this record except in the case where there is a zero de minimis or AFA. Thank you, counsel. We'll take the case under advisement. Thank you.