# Oral argument — Apex Frozen Foods Pvt. Ltd. v. United States > Briefs, arguments, decisions, and more. URL: https://www.frixlaw.com/law-library/documents/argument%3A28803 ## Record - **Collection:** Oral argument - **Document type:** Oral argument - **Published:** March 8, 2017 ## Text As far as the zeroing issue, and tell us what the differences are between these two administrative reviews. Yes, Your Honor. Again, for the record, Robert LaFranchi on behalf of plaintiffs, this is my rebuttal? Yes. Okay. So, yeah, I'd like to make a couple points in rebuttal. And I think, to Your Honor's point about the statute, and I would like to mention one of the most interesting things mentioned was a shortcut. I agree that what they're doing is a shortcut. And it's not following what the statute says they should do. It's simply a shortcut. If you look at the statute, the administering authority must explain why such differences, and that's differences from the pattern of prices in the U.S., right, cannot be accounted for between the two, between the normal methodology and the alternative methodology. Well, taken into account. Let's stick with the actual language of the statute. It's taken into account. What does it mean, taken into account? For what purpose? Well, and that's a good point. But the question is, yes. I mean, what's the answer to my good point? Well, our answer is based on what is the purpose of this, which Congress has said is to unmask dumping, what Congress has indicated in the SAA, which is to alleviate concealed dumping, that which is hidden. And so the question is, can it account for that? And in the examples I gave you, sometimes it can and sometimes it can't. And that's whether it's hidden. So when you break out the average price to the individual transactions, you can reveal dumping that's hidden. In that situation, the A to A cannot account for that because it can't see it. But in other situations, it can account for it as between averages because there's no masking. There just might be no dumping. And so by just comparing— How do you know that? You can only know by breaking all the transactions out, right, in every averaging group, right? But what Commerce is doing is they are simply taking a shortcut and assuming that no matter where the sale is, whether it's within an average or whether it's in different averaging groups, any negative dumping margin is masking a positive dumping. And what we're saying is that's not necessarily true. You have to actually look at the evidence. You have to look at the record. You have to look at the patterns. And you can't just automatically provide those offsets on one side, which then eliminates all of your positive dumping. And I take their point that that's the way they normally do the calculation. But that doesn't mean that they need to do that calculation the same way when they're doing the comparison because it's a different question. It becomes self-fulfilling. And I realize Aztec has listed some cases where the two comparisons do not yield a meaningful difference. While it looks like a lot of cases on one piece of paper, compared to all the cases that are conducted, it's a minuscule portion. And yes, there could be some cases where it's not going to make a difference. If every sale is dumped, which doesn't happen very often, it's not going to make a difference. But I don't have statistics to back it up. Most of the time, it's going to make a huge difference because you're just automatically taking a shortcut and assuming that any time there is a non-dumped sale, which generates a negative dumping margin, you're just assuming that that is masking dumping in another sale. And we're saying that's just not necessarily true. It depends on the pattern of prices. And it depends on whether you're talking about masking within an average or outside of the average. And so just the last point to rebut is the US government makes a point. And Commerce has said this in their decision memos, that they equate offsets to masking. And they use the terms implicit masking and explicit masking. What does that mean? That tracks the example I gave you. Implicit masking is within the average. And that is when you break the transaction out. You can reveal individual prices. So like the 10 and the 14. You then unhide the dumping from the average. We agree that that implicit masking is masking. That's what, if you look at Union Steel, that's what Union Steel says. Mass dumping within an average. They have never talked, Union Steel has never talked, nor has the Federal Circuit talked about the other type of masking, which is the explicit masking, which is across the average in groups. And Commerce has taken that position. So if nothing else, I just want to clarify. I realize maybe I should have gotten this 30 minutes ago. Averaging groups is where in the statute? Averaging groups is in a different part of the statute. It's in the regulation as well. Meaning what? Meaning you, the shortcut version is you calculate average prices by a specific product, a control number. So you're only going to compare prices within a specific product that's identical to another product. So if you had a piece of steel that was eight feet long and one that was two feet long, you wouldn't compare them. You'd do them separately. So that's the averaging group. And then you might have all these different products compared in one case. So the averaging group tends to be within a specific product. So then you would have only specific products compared to each other. But that's the point, though, is that the implicit masking is within an averaging group. When you're taking two prices of identical products and averaging them together. And then you split them out, you might actually reveal the dumping. The explicit masking is going across averaging groups. And the groups here are different products. Different products. In this case, they're all shrimp, but you might have some with a head on, with a head off. But that's actually an interesting point. What they're doing, what the Commerce Department is doing, is assuming that negative dumping on the other product across the averaging groups is masking dumping on a different product. And we're just saying, that's not necessarily true. And even their term, explicit masking, if you think about the difference between implicit and explicit, that gets to the whole point. The whole point here is masked, concealed, hidden dumping. Even the term they use, explicit, means it's not hidden. The A to A method shows that there is negative dumping. Commerce has chosen to abandon zeroing under the A to A method. So that's a policy choice. So back to the statute. We would argue that A to A accounts for the difference. It accounts for it. Other than within an averaging group, we concede A to A cannot account for it there. It's hidden. You can't see it. It's 12 to 12. Maybe I didn't notice it. Do you make an argument in your brief about the difference between averaging on the same product and averaging between different products? I don't think we make that specific argument. I'm explaining to you here. We make the argument that legally it's distortive and it's measuring zeroing. So I'm just explaining now how it works. So my closing point on that is the notion of concealed and hidden. And the point we made in the brief is that they are just automatically assuming that every time there is negative dumping, which is no dumping, wherever it is that it is masking dumping somewhere else. And we're just saying that's not true. So this court should decide, should there be a difference between explicit and implicit? Is that really what the statute was intended for? We say no. We say if you look at the whole concept of masking. In fact, when Congress enacted this provision in 1994. Why is it that the 14 sale in your hypothetical isn't masking? Well, we would say as a pure matter. You just said that it wasn't always. They assume that in your hypothetical where you had the 12 at home, 14 at home. Why isn't the 14 masking? Well, we would say it's not masking because it's from a different averaging group. It has nothing to do with that other group. That true masking. But it's an event. It's a data point. It is a data point. It's a transactional data point. It is, and it's obvious. It's not concealed. And it's obviously masking. Well, I disagree with that. Masking up the lower. It's offsetting. There's an offset. What's the difference between offsetting and masking? Well, an offset is a legally sanctioned allowable negative dumping margin. It's got nothing to do with masking. And the point I was making originally was that. I was just calling you out because maybe I'm just too dumb to deal with this. But it just seemed to me like the 14 data point that you identified is masking the 10 in every case. Well, it's offsetting. I guess it's a matter of semantics. Where do you want to use masking? Is it just a terminology issue that's different between us? No. We would say it doesn't matter what word you use. You want to use mask. You want to use offset. The question becomes, what are we trying to get at here? And so masking. But if your pattern so-called comes back to the statute, and the pattern shows that the 10, 12, or the 10, 14, then why isn't it fair to assume that the 14 is masking? Because that, in that situation. That's a front stage. Yeah. Because in that situation. Or suspicious. It raises a suspicion. Well, there could be a suspicion if they find enough pattern differences. But I think our argument, Your Honor, is that the whole point of the targeted dumping provision was to get to the difference in A to A and A to T because of averaging within an averaging group. And what's most telling about that is at the time that Congress enacted this provision, zeroing was used on both sides. A to A used zeroing. A to T used zeroing. So really, the only difference, that's what we're advocating as a test. The only difference between the two was the averaging within that averaging group when you split out the 12 to the 10 and the 14. If you look at the SAA, that's what it talks about. Concealed by averages. That's what Union Steel says. Concealed by averages. Is that the difference between the 7th and 8th administrative review and the change procedures? Let's focus on what's different in this appeal. No. That would be the same basic issue. And so we did make the same basic zeroing issue in the 8th administrative review. But we made other arguments as well. And one is, it's related to this, but it's slightly different. So in the 8th administrative review, we made the same zeroing issue about the disparate use of zeroing. We've covered that. We made another argument separately, which is that also during the threshold meaningful difference test, not only was Commerce using zeroing in an unequal manner to distort the results, they were also running their test on all sales. What I mean by all sales is those that were found to be targeted and those that were found to not be targeted. And so we said that that was not only unlawful, it's distortive. When you say those that they found to be targeted, what does it mean to be targeted? Well, in the 8th administrative review, Commerce adopted its new differential pricing test. Different methodology. Different methodology. I'll call it the Cohen test. So specifically, a targeted sale, or they refer to it as differentially priced sale, is one that passes the Cohen's d test. Now, we're not challenging Cohen's d. It's horribly complicated, but basically it's used to measure a spread in prices using pooled standard deviations. And so what they do is they test all U.S. sales. And what kind of sale passes the Cohen test? It has to be... So it has to deserve the label targeted. It has to pass what's called the large coefficient in Cohen's d, which just means there's a really big difference in the standard deviation of the prices, and they do it separately by customer, region, and time period. So essentially... It's all still simply with regard to the activity in the United States. That's correct. No relevance at all to the home market. That's correct. It's just looking at U.S. prices and interest. So they run the test three different times against every sale. I've used an example of if you put a bunch of quarters on the table, every sale has a chance to pass this test three times, either by customer, by region, or by time period. And once it passes, they take it off the table and they put it in a box. So they collect all the passing sales. So some pass and some don't pass. Even when they do it three times, some may not pass. And that, again, is looking in the U.S. only amongst... And the whole purpose of this is to ascertain whether there's some meaningful difference in prices? Yes. Or timing or location? Well, to look for significant differences in prices, which then signals that there may be targeting going on, which would allow you to mask a higher-priced sale and a lower-priced sale. But the interesting part about this is... When you run the test on the individual transaction, you determine that it's not significantly different. It then passes the test and it's no longer considered. Actually, it fails the test. But yes, it goes into the pool of non-targeted sales. So ultimately, they'll create two pools of sales. But they do the test within the same averaging groups that I was talking about before. So they'll do it by a control number, by a unique product. And I think that's significant because what they're looking at in their pattern analysis, they are comparing for high and low prices within the averaging group. And they're deciding for the averaging group as a whole, right? Okay, there's a pattern or they're not. So they compare. I see my time is up. Okay, let's hear from the other side. And we'll save some of your time. So does that end the argument on the second case? Well, we didn't really get to it. But I was going to ask Mr. Perlin to elaborate on the second case. We'll see where that takes us. I think the way this was set up, Your Honor, is now we're in the second case. So he's had his opening argument in the second case. Come to the microphone so it's recorded. He wasn't up 15 minutes. Well, let's see what happens. Before I start, Your Honor, I think he was up for his 12th. My recollection is Mr. LaFranchi saved three minutes for rebuttal time in the second case. So he's been up for his 12 minutes in the second case. Now we're going to do our… That's all right. We still need to know your view of the difference between the 7th and 8th Administrative Review. Sure. Well, there's a straightforward difference between the 7th and 8th Administrative Reviews, which is that Commerce moved from its former test, which was called the Nails Test, to what is referred to as the Differential Pricing Test. One important thing that was not clear in Appellant's comments is that the Differential Pricing Test involves whole groups of sales. So that colloquy at the end where he was talking about if an individual sale doesn't pass a bucket, that's not quite how it works. Commerce compares whole groups of sales to a particular purchaser, region, or time period to the rest of the sales. And if that group of sales exhibits a significant price difference, then that group passes the Cohen's d-test. And I think that's important to note in the context of these two cases being before the court together, because the differences in the test is what results, for example, in a much higher percentage of sales that pass Differential Pricing versus the Nails Test. And the Nails Test is a bunch of very stringent criteria, so uniformly you get much lower numbers of percentages of sales that pass. It might look strange in the context of this case where you have much higher percentages that pass. So you're saying there's no zeroing in the calculation for the group? There is no difference in the Meaningful Difference Test. The Nails Test and the Differential Pricing Test are the first stage of the calculation. The answer is there is zeroing. There is zeroing. The Meaningful Difference Test works the same way in the Seventh Review as it does in the Eighth Review, where Commerce applies the A-to-A methodology without zeroing and the A-to-T methodology with zeroing to determine whether there's a meaningful difference. But I've never understood. The numbers are there. They have to be there in order to be zeroed. Why don't you just push a button on the computer and say, or I treat these as a group with the actual numbers, not the zeroed numbers? Well, in some sense, that's what Commerce does in the sense that with the Differential Pricing Test, it runs the analysis automatically. I think to get at the core of Your Honor's question, throughout these proceedings, Apex's position has been that it's somehow unfair that we end up with a rate with the A-to-T methodology and not a rate with the A-to-A methodology. No, it is unfair. You get a dumping margin. You say, never mind. We'll treat it as zero. Well, the reason that it's fair, Your Honor, is because this isn't just math. These are all sales that were actually dumped. The reason that you get a rate with the A-to-T methodology that you don't get with the A-to-A methodology is because there are a whole bunch of sales that were sold here at less than fair value, that were dumped, but that get masked because of the offsetting function of the A-to-A methodology. You end up with- How can you justify that other than as to make the calculation feasible? Well, it's justified for a couple reasons. First, because Commerce only switches to looking at the A-to-T methodology as a potential comparison if it reaches that precondition of there being a pattern of significant price differences under 1677 F1D1B. If you have a situation where Commerce says, hey, wait a minute. We're looking at this and we see a pattern of price differences to a particular purchaser, region, or time period that shows that there's potentially a mass dumping situation going on, then it's appropriate for us to look at the A-to-T methodology, which the statute authorizes us to look at in those circumstances. And all the A-to-T- Can I ask, is Commerce- I don't know that it did this in this case, but in some other case, I was trying to, I guess, in my own mind, construct a simple-minded market economic story in response to Judge Newman's question. If there's a- if prices that Apex- if Apex is selling in California at 14 and it is selling in the rest of the country at 27, that's probably telling you something and the rest of the country. And therefore, probably telling you something about domestic sellers, the beneficiaries of this whole regime, so that in some actually meaningful antitrust type sense, this would be a separate market segment. And therefore, when you find a pattern, there's reason to say, if you can otherwise show dumping, which now introduces the home market price, that you can compensate for that because the group of harmed sellers, domestic sellers, is probably different if there's a sufficient difference disparity within the U.S. by either by customer or by time period or by region of the prices. Is that- That's largely what's going on here. Yeah, but did Commerce ever, or Congress in coming up with this regime or somebody say, here's why there's actually a connection between intra-U.S. price disparities and what we are going to do about targeting, about dumping the unit that we're going to use in deciding whether there's dumping nationwide or sub-area? Sure. The statute does not set up a regime and that's something that this court specifically addressed in its JVF RAC decision, which concerned the nails test targeted dumping or mass dumping methodology. Commerce doesn't look at the reasons behind a pattern of price differences. It's kind of a bramble bush. Instead, Commerce looks to see whether there is such a pattern and then whether the dump sales that are causing that disparity in price differences are getting masked by other non-dump sales. And that's the whole point. In the absence of a pattern, masking is just perfectly fine under this statute. Well, I wouldn't quite go that far, Your Honor. That's why you don't do zeroing in A to A unless there's this pattern. It certainly doesn't provide a remedy. There could well be underpriced sales. By underpriced, I mean, of course, positive. It's sort of nice to get the mirror out of this picture. Right, I know, I know. Underpriced sales. And there's no remedy because elsewhere in the country you're overpricing. But where there's a pattern of intra-U.S. price disparity, the statute says that's one we can single out and actually get at the sub-market, to use an antitrust term, where there is dumping. Sure, and the A-to-T method is a statutorily prescribed methodology for doing that. There's one other very important statutory argument for me to articulate here. We've mentioned this a couple times in our brief and the court picked up on it in this case, in the POR8 case, which is that it's been widely recognized that zeroing is an inextricable part of the A-to-T methodology. But it's beyond that. If you removed zeroing from the A-to-T methodology, it really becomes the A-to-A methodology. We've said it slightly differently, which is that the results would be the same. But as I've thought about it in preparing for the argument here, what really happens is if you take zeroing out of the A-to-T methodology, it becomes the A-to-T methodology. And I'll explain that in a little more detail in a moment. The important statutory point is that the statute provides for both of these remedies. And so it's a well-known canon of statutory construction that Congress isn't providing for multiple remedies if they're all really the same thing. And that's been one of the rationales, or maybe not a rationale, but a way to explain why the A-to-T methodology still has to remain separate from the A-to-A methodology. Now, to get a little bit in the weeds on that, the reason that A-to-T without zeroing becomes A-to-A, I can demonstrate with numbers. You know, if, let's say, the U.S. price and the normal value is all 10 and you're comparing averages to averages, 10, 10, you have zero dumping margin, right? If you're using the A-to-T methodology, if you have one at 2 and one at 18, with zeroing, you find the dumping on the 2 and you zero out the sale that's masking it on the 18. If you start offsetting, as you do in the A-to-A methodology, with the A-to-T methodology as well, you're really going back to... It's turning the A-to-T methodology into an averaging methodology because you're going to have an 8 and a minus 8 and you're going to end up with 10 again. So it's another way of explaining why zeroing is an inherent part of the A-to-T methodology. If the A-to-T methodology didn't have zeroing and Congress has not articulated any intention to stop zeroing when it uses A-to-T, in fact, it's repeatedly said in the cases on zeroing that this is why we always use zeroing with A-to-T to date. If I wanted to get myself a little better educated and know how Congress dealt with these situations before NAILS... NAILS, I think, was 2006, something like that? Because we've gone from NAILS to Cohen in a one-year time period. Is there someplace I could go look to learn about how these cases were treated before NAILS? I believe there's... I don't have the site. It's not something we've cited in these proceedings, but there's a Court of International Trade case called Mid-Continent NAIL, which upheld the NAILS test, which goes through the whole history of Congress's targeted dumping. The POR7 decision in Judge Goldberg's case also provides a reasonable amount of background about Congress's at least starting with NAILS... Yeah, but the question was really about pre-NAILS. There's a puzzle here about what you did under A-to-A before you stopped doing zero. Right. I mean, there was not a lot of targeted dumping allegations pre-NAILS test, so it was a rarely used phenomenon, and in fact, you know, this goes to some of the limiting rule issues that we haven't even discussed today and probably don't need to, but, you know, Congress explained when it withdrew that rule, which only applied to investigations, that it hadn't really had a lot of experience with targeted dumping at the time. It promulgated that rule. What's the status of the regulatory, the second attempt at a regulatory withdrawal of the limited rule? That happened in 2014. I'm not aware of any... It's completed? I believe so. I'm not aware of... You know, I see cases. I'm not aware of any litigation that has said that that was wrong or there's something... I don't even know if it's been challenged. I tend to doubt that. But if it would affect at least the 8th Administrative Review, I suppose it's relevant. Would you comment briefly, again, on the difference between the 7th and 8th? Yes. Since this is a separate appeal. Yes. There are both differences... Several differences. I'll try not to draw it out too far. Different tests, the NAILS test versus the differential pricing test. Very different records. In this case... On the specific issues presented to us, my understanding is we do not have an issue about the difference between the NAILS test and the Cohen D test for determining if the pattern existed. Do we? No. I don't think we do. No. There's no challenge in the ascertainment of a pattern in either one of the two cases. That's right. That's right. That's right. Their legal arguments are slightly different in that... The argument that they were held... Judge Goldberg said they had waived because they hadn't sufficiently made it. That's the applying to all. That's a factual difference between the 7th Review and the 8th Review? I understand that, but then they repackaged that argument in their petition for rehearing before Judge Goldberg in sort of legal question. He rejected that again, saying you can't dress it up a different way. They've made that same argument in the second case. Yes. And that's where there's an important factual difference because the facts are against them in the second case. Because in the second... Their argument that it was waived... I understand that, but I mean their argument that it was wrong to apply to all the goods they're making in this case, that's two different tests, one under Cohen, one under NAILS. Does it make... Is that legal argument that they're making different in this case than it was in the earlier case? It's slightly different. First of all, they didn't... In the first case, they said you can't apply the A-to-T method to all sales at what they call the remedy stage. Once you decide that there's a pattern and the A-to-A method can't account for it, they're saying at the remedy stage, you shouldn't be able to apply to all sales. And there are a number of good arguments why that was wrong. Judge Goldberg explained why it wasn't consistent with the relevant part of the statute. Commerce explained why you're not getting... In the NAILS test context, you're not getting the whole pattern because the NAILS test is only looking at the lower sales and it's only looking at individual sales as opposed to whole groups. And so in order to get up... In the Seventh Review NAILS test context, in order to get the whole pattern, commerce at that time applied A-to-T uniformly. And there are a whole set of arguments in that case. In this case, and this is at page 51 of their brief, they're conceding that commerce can apply to all sales at what they call the remedy phase, but they're saying you can't look at all sales during the meaningful difference test. And that's a bad argument. What they're trying to do is kind of manipulate the data to make it look the way they want it to look. What do I mean by that? They're saying, well, if you just look at the dump sales and don't zero, then we find similar amounts of dumping with the A-to-A and A-to-T. But that just proves our point, right? Because what they're doing is saying, hey, there's a whole bunch of dumping going on. In fact, there was, as we know in our brief, seven-figure amounts of dumping going on when you do it that way. And second, that's not the way the A-to-A method ever works. The whole point is that the A-to-A has this offsetting function, which in the context of when you have this pattern, offsetting is masking. It's taking sales with positive margins and wiping out the dumping margins that result from sales that are actually dumped. And so that's their argument here. It's wrong because you're obviously, if you're saying, well, is there a difference between the way we apply these two methodologies, you want to look at the way they actually would be applied, not saying, well, if we take the pinky of this thing and the pinky of this thing, then there's not much of a difference between them. I mean, that just proves the point. Can you clarify this? And I think there was some material in your briefs about this and I'm just not remembering. So once you've identified the customer, the time period, or the region that, under the pattern provision, and I don't remember, in the first case, it was a time period. This is also the same time period. It was a time period and a customer in the first case. And in the second case? In the second case, because they're just pulling groups together. Anyway, so once you've identified the special group, when did commerce in either of these two proceedings do averaging within that group or did it zero, that is non-zero averaging, just take the entire group of these sales and say, let's take the average of the US prices in those, compare it to the normal price and figure out if there's dumping, or did it say, well, this is a group, 80% of the sales there are seriously underpriced, 20% not. Let's zero out the 20% and then do that. Yeah. So unfortunately, the answer to your question is slightly complicated and I want to make sure I give it to you so I appreciate the leeway to do so. I think there are two different ways. There's during the actual differential pricing test, which deals with groups. So commerce is saying, hey, if this group passed the differential pricing test, then all of the sales in that group are going to be considered part of a pattern of significant price differences. Right. My question was after you've identified the group. After that. So California, bad pattern. So the T of the A to T is averaged to individual transactions on the export price side. So the whole part of the point of the A to T is that they're not averaging. They're looking at an average of the normal values and comparing that to the individual transactions in the United States. And that's part of what helps reveal what's called, what commerce has termed the implicit masking and explicit masking. When they say implicit masking, they're talking about masking that occurs because if you're using A to A, you're doing averaging. So you're already smoothing out some of the differences. There might be dump sales that are averaged with non-dump sales. Then there's the explicit masking, which is in part what zeroing takes care of, where you have, you know, the number on the U.S. side is coming out dumped, but it's getting masked by non-dump sales in other comparisons. So at that final remedy stage, the whole point of the T and A to T is that they're looking at individual transactions. They ultimately aggregate the results. That may be the one final difference between this case and the previous cases. There's this so-called double zeroing argument at the tail end of Apex's breach. So I think I'm, I hope this is, I've got this right. If California were identified as the bad pattern area, you would not actually zero out any of the overpriced sales in California. You said you were doing individual transactions. So every single sale in California, you would compare to the, I don't know, the $12 price in the home market, and presumably, you'd find a lot of California prices, U.S. California prices, under $12, but there'd be a few above $12, and they would all count in the aggregation. Well, no, Your Honor. With average transaction, you're looking at individual transactions. So if there were a bunch of sales in California and a bunch of sales in the rest of the United States. Forget about the rest of the United States. You want to know what happens to the 17 sales in California of which 13 were underpriced and 4 were overpriced. That's right. You understand what I mean? I do. I do understand, Your Honor. Compared to the home country. There would be zeroing there in the sense that you would... So it's not actually transaction specific. Well, it is, and maybe I just need to figure out how to best explain it to Your Honor. If you had 10 sales in California and 8 of them were, and let's say the normal value is 10. If 8 of the 10 sales were at $5 and 2 of the sales were at $15, Commerce would identify the dumping and the average transaction method would compare the 10 to the 8 sales that would have 5 and each of those would be dumped by $5. It would compare the 10 to the 2 sales that were at 15 and find that there was no dumping. That's what zeroing is. You're essentially finding that there's no dumping on those 2 sales. And so rather than those sales partially masking the dumping on the first 8 sales, it would be set at zero and Commerce would identify that there were 8 sales that were masked by... There would be a dumping margin of like $5 on those sales because you're comparing to individual transactions. If you're using an average-to-average methodology, then you would average all those sales, come up with the average number and compare that to the normal value of 10. I hope I answered your honest question. Okay. Thank you. We'll hear from Ms. Rolig to explain it further. Good morning. In the late arrival, may it please the court, I'm Whitney Rolig on behalf of the Ad Hoc Shrimp Trade Action Committee. And I just wanted to highlight one additional nuance that appears in the 8th review as opposed to the arguments raised in the 7th review, and that is the exhaustion argument that we raised in our brief where Apex is trying to belatedly claim that it has a Chevron 1 argument with respect to the differential pricing analysis. That is incorrect. They are trying to avoid the question in all of this of whether commerce acted reasonably by trying to say that the statute was unambiguous. They failed to articulate that theory below to the agency and commerce should have had the opportunity to refute that reasoning at the outset. This is particularly important because Apex's entire Chevron 1 argument relies on extra statutory terms such as targeted dumping, targeted sales, as opposed to the statutory term of significant price differences. We just want to make sure that the court rejects that attempt outright and goes directly to the question of whether commerce interpreted the statute reasonably. And again, this goes to the idea of whether commerce can determine whether A to A accounted for the mass dumping, the significant price differences with respect to all sales versus just with respect to the targeted sales, as Apex has put it. So we want to make sure the court stays focused on that. If there are no other questions, that's all I wanted to raise. Well, so tell me how this would affect the remedy. Assuming that after all these complicated AA and AP transactions, there's found overall that there is a margin of dumping. Then the remedy is assessed against all imports? Exactly, exactly. So again, what commerce is measuring is how if when they apply A to A to all the sales, as they do when they measure dumping overall, how does that compare to when they apply A to T? If there is a meaningful difference, as there was in this case, because the margin crossed from zero to an actual measurable margin, that is considered a meaningful difference. So then they do apply A to T. If Apex had its way, commerce would only compare A to A and A to T on the sales that are identified as having significant price differences. So it's a much smaller subset of sales. And then Apex makes other adjustments to that comparison. For example, having zeroing or not zeroing in both sides, which as has been explained extensively today just doesn't make sense as a logical matter. So again, commerce acted reasonably in evaluating the significant price differences and how A to A accounts for those in the context of all sales because that's how it's actually going to measure dumping overall. Apex has tried to say the statute unambiguously prohibits that, and that's simply not the case. That's an untimely argument and it just does not square with the language of the statute. Thank you. Thank you. Okay. Mr. LaFranca, three minutes of rebuttal. Yes, Your Honor. May it please the Court. A couple points in rebuttal. One point I'd like to clarify is the counsel for the U.S. government insinuated that A to T and A to A would become the same depending on what you do with zeroing. We're here to say, again, that is not true. If you look at their brief, what they say is they would be the same with offsets if you did offsets on both sides. But if you do what we propose, which is zeroing on both sides, they would not be the same because that is our point, that that is the mass dumping that's revealed. Now, I'll try to spare you another example other than to just say that that is not true, that they would not always be the same. If you zero on both sides, they would not be, number one, and number two, if you do what we propose, which is to limit the test to the targeted sales, the results would not be the same. Help me again. How do you zero in the A to A? Under A to A, you would zero... In transactions, you know, you've got four different prices, 10, 11, 12. How do you zero? You zero, again, not within the averaging group. Zeroing with respect to the home market price, is that what you're looking at? No. You do the comparison with the home market price and you take the difference, right? And so our point is that you cannot zero under A to A within the averaging group because it's hidden. Where you can zero is across the averaging groups because that, if you have a negative margin in a different averaging group, you would just set that to zero. That's how you would do it. Across the averaging groups, not within an averaging group, which is actually our point. How many averaging groups were there in each of these two cases? Well, there's a lot of control numbers, so I don't know exactly, but there were a lot. I mean, in some cases, there could be thousands. Is that mapped anywhere in the brief so I could look at it? You mean that specific point? I think it would be... Yeah, because what you're saying, now I understand how you're going to... You're not averaging inside of A to A. You're saying, well, you've got an A to A analysis on this group, this group, and this group, and this group. And you're saying that depending on the results, if some produced a positive, some a negative, you want to throw out the bad ones. That's right. I haven't seen the math anywhere in this case to how that works. And you're making that argument in both cases. Same argument in both cases. It's essentially the same argument, yes. But to answer your question, the data is on the record in the confidential appendices in terms of the detailed calculation or worksheets. It's all buried in the computer. So it is... It's a complicated subject. There's complicated calculations. I'm trying to just give you simple examples, but in some cases, there could be thousands of averaging groups. And so the point we're making, if nothing else, is that what is masking? That is the limitation of A to A. It's within that averaging group. And you can't actually zero in there. But you can zero across the averaging groups because that's what Garmer's did for I don't know how many years before they abandoned zeroing. 20 years. So the one point I want to make is the results would not always be the same as was articulated. I wanted to pick up on another point that your honor made, which was that sometimes masking is fine. Yes, sometimes masking is fine. Wherever A is the accepted rule, masking is going to happen. Masking is going to happen within that average. That's correct. Now, across the averaging groups, the point we're trying to make is not all masking is bad. Not all masking is prohibited under the statutory approach. Only that tied to the significant price differences. And the one broad point we're trying to make is that Garmer's isn't doing that analysis. It's taking shortcuts. It's making assumptions. It's simply using a zeroing methodology to amplify the differences between these two margins and thereby assuming that every non-dump sale, every negative dump sale, is masking impermissibly a dump sale. And that's not always true. They need to look at it. And what we're saying is it's not only the zeroing. They're also combining that distortion by running the calculation on all sales, including those that are targeted and those that are not targeted. So at the end of the day, the comment was made it's about dumping. Actually, this case is about non-dumping. Can I just double-check something? Do you happen to know, is this so-called limited rule, limiting rule, whatever? Limiting rule. Limiting rule. Has that been finally repealed? It was finally repealed. Okay. After this case, our view is the effective date of that repeal was after this case. Okay. And had that rule, that specific rule, ever been applied in a review context before? It had not. Okay. So your argument is that because at the time relevant here, it was on the books and by its terms applicable to investigations, it should be applied, in fact, for the first time to a review. Yes, that's correct. Or at least explain why it's not being applied now. Because it was a bad rule. I mean, for the same reasons that they repealed it. We would argue it was a good rule. Right. But it is at least a reason, as with precedent, to say we'll leave it alone for where it is, but we actually think it's not that sad, so we're not going to extend it. It's a reason, but they never gave the reason. Granted, they thought the rule was repealed, but they had acted illegally in repealing it. And be interested to hear what their reason would be, why they would treat a review differently now than an investigation when they're now treated the same. So, I mean, to that point, we think there's complete unfairness in that, right? I mean, Commerce had promulgated, they had interpreted the statute at one point to say, okay, we should only apply this remedy. Now I'm talking about the remedy phase. Should only apply the remedy to the targeted sales. Then they withdrew the rule. And for a small time period, including both of, at least, no, no, the first case here, the AR-7 case, they then started applying the remedy to all sales. Now, when they switch to the differential pricing analysis, they've sort of thrown out that across-the-board approach. And now, if you're within a certain pattern, they will only apply the sales to those targeted sales. So they have a hybrid approach. But the point is, they've only used that for a limited time. Okay, okay, to move on? All right, I think that we need to sign it to ourselves. Thank you all. It's a helpful argument. --- Source: Frix Law Library, https://www.frixlaw.com/law-library/documents/argument%3A28803. Public record. Not legal advice.