Oral argument — KY v. EPA

Case 23-3216, Kentucky v. EPA et al. and Case 23-3225, Kentucky Energy and Environment Cabinet v. EPA et al. 20 minutes to be shared by petitioners, 15 minutes for respondents, 5 minutes for amici. Mr. Bentley, you may proceed for the petitioner. May it please the court, distinguished counsel and our attendants here today. My name is Jared Bentley, and I rise today on behalf of the Kentucky Energy and Environment Cabinet and this challenge to EPA and their administrator Michael Reagan's arbitrary and unlawful disapproval of our state implementation plan. Is there any difference in the positions of the two of you, or are you dividing your argument in some particular way? We are dividing it, Your Honor. I understand the time, but substantively, what are we going to hear from one another? Frankly, it's awkward to have the same people dividing stuff because sometimes we want to cross between issues. What's the division here? Mr. Cooner is with our Attorney General's Office. He is representing the Commonwealth at large. As the representative of the Energy and Environment Cabinet, we have the statutory authority to promulgate and submit state implementations. What's the reason for a substantive division between your arguments here? The arguments are very similar on the challenges that we've brought. Go ahead. It's really disfavored to do this. Go ahead. Take your best shot. Thank you, Your Honor. We would like to request four minutes for rebuttal from Mr. Cooner. All right. Whoever wants it can have it. Go ahead. Your Honor, when an agency initiates rulemaking, it is fundamentally unfair for that agency to keep moving the finish line, and that's exactly what EPA has done here with Kentucky State Plan. EPA provided modeling, they provided guidance, and they provided express directions to the Cabinet to submit an approvable state plan with an approvable good neighbor provision analysis as well. Can I ask, before we get into the merits, can we talk a little bit about the venue provision? Sure. So we have case law suggesting we should defer to a motions panel, but it's not clear to me why that is the case. So we say it's quite conclusory. Yes. It's an unpublished precedent. So presumably the next case that comes up could just say that's wrong. So precedent wouldn't be a basis to defer. I suppose it might be law of the case, but you guys didn't argue law of the case principles. You just cited this case law, and the case law itself doesn't necessarily argue law of the case. So it's not clear to me why we should defer to the earlier panel. I believe that when, and I'm recapping the briefs that the parties have submitted, and I believe that Mr. Coon did a fantastic job when he actually went through and broke down the different areas where a motions panel would defer to the merits panel with instructions to rule on that motion with the merits. There have been other instances, of course, where they can dismiss that with or without prejudice to the merits panel. And I think that what we've cited here are both of the rules for rehearing or rehearing en banc as it pertains to what was in the merit brief provided by EPA. And the case law essentially says that unless there has been some manifest injustice, that the merits panel would not overturn an earlier motions panel decision. Where do we say that? Get the case citation for you, Your Honor. I mean, you know, motions panel, typically, I mean, for a significant matter like this, we're frequently doing this in like a 96-hour period, trying to get an order out. And it would seem very strange to me if after a motions panel, typically, you know, time pressure situation is going to render, you know, some opinion or decision. Then the next stop after that is potentially the U.S. Supreme Court, that a merits panel with the time that we have and the ability to have argument would just stand down. And if the motions panel kind of in, you know, in their quick decision get it wrong, then the Supreme Court has to fix it, right? Why would that make sense? Thank you for the question, Your Honor. I want to unpack that a little bit and think about how that would work. Mr. Kuhn can talk about it too. I understand. And I would point to the court the Wallace v. FedEx decision, which also cites to the two Sixth Circuit Rules 35 and 40, and also Glymphoff v. Rickard and Bowles v. Russell, which was affirmed by the Supreme Court in 2007. And that is looking to later panels cannot simply choose to disregard motion panel decisions, and if a litigant wishes to challenge a motions panel decision on a dispositive motion, the proper course of action is to request a panel rehearing or rehearing on file. So those cases were the ones that we were alluding to there on the venue arguments that were made by EPA and the Mayor of Virginia. Looking back at what was provided to the Cabinet, and I think the crux of the situation looking at the Wisconsin decision, the gravamen of the claim, one being the timely action by EPA in reviewing the state plan that's been submitted, and the second is EPA's reliance on modeling that was updated after the plan was submitted. And the legitimacy of a SIP disapproval lies with both of those factors. What's wrong with their harmless error argument? Your SIP submission was quite conclusory, to be honest. You conceded that you had one even over the one parts per billion in Maryland, and then you're just like, yeah, but we're making other changes. We're not going to engage in any technical analysis as to why those other changes would actually bring that number down, but just kind of we think it might. I understand your question. Well, that strikes me as, I mean, if I'm the US EPA, I'm like, well, maybe you should do some analysis about whether those other changes are going to affect this Maryland, I don't know if it was a maintenance or an unattainment, but whatever, it was a maintenance. Okay, that's right, that's right. Whether it would affect this Maryland maintenance receptor. Well, on that point, your honor, the Maryland monitor was the only monitor that Kentucky impacted above the one part per billion threshold, and EPA's specific comment to our draft SIP was, use our modeling, it's a more straightforward approach, and rely on the one part per billion standard and analyze the Edgewood monitor, the Maryland maintenance monitor, that you impact above one part per billion. We did so. We went through, and you can see in our SIP submission, there's a couple of pages where we actually go through new measures that were taken by the state of Maryland to reduce their emissions. We also pointed to comments from Midwest Ozone Group and Alpine Geophysics to say, the reductions that Maryland plans on taking by the time the 2023 attainment deadline hits, that receptor will be an attainment, and we rightly predicted so in our SIP submission. In what you just said, I heard you say that EPA made some kind of response where they told you to use the one part per billion. Did I hear that correct, and what document are you referring to? Thank you, your honor. I am referring to our preliminary SIP comments, our draft SIP comments from EPA. So when we promulgate our state plan, we send a draft of that state plan to EPA for comments, and one of their comments back to us was a more straightforward approach, and be to rely on our modeling. Do you know the date of that document or the JA reference? I do. I'll give you just one second. I might be able to come back to that one, your honor. I know that I've got it. But what I'm focusing on is the way you state it is they told you go ahead and submit it using the one part per billion, and that seemed to be a pretty strong statement, so I wanted to see what the basis for it was. I understand, and it is? If not, one of you can submit it. The SIP comments at page three, and it is in our brief. What? On our SIP comments at page three. Of your, what's the? 126 in the joint appendix. 126 of the joint appendix. All right. We'll check it out. All right. Thank you. All right, we'll hear from Mr. Kuhn. Thank you. Thank you, sir. Thank you. May it please the court, Matthew Kuhn for the Commonwealth of Kentucky. Judge Boggs, if I can start with your question, and then I'll discuss venue and then harmless error, as Judge Murphy has asked about. Judge Boggs, it is joint appendix 126 is the language that we're talking about. They told us at joint appendix 101 that the two contribution thresholds were generally comparable. That's in the March 2018 memorandum. We then submitted a pre-draft of our SIP to the agency, and at joint appendix. At different places, each of you say something about comparability, and how .7 can be comparable to 1.0 when it's nearly 50% larger, I find very odd. So you might, I mean, you've got an order, so go ahead with it, but keep that in mind and try to give me an answer at some point. I guess what I would say is the agency with the technical expertise in this, those are their words, that they're generally comparable. And what they did is they looked at the monitoring sites and the linkages and the amount of out-of-state upwind pollution that's captured, and that they found them generally comparable that the 1% of the NACs only captured 7% more than did the one part per billion. And if you actually look at the new data. That's more on the national total of all the sites. That's correct. On any given site, I mean, that's what we're fighting about. Here we've got a large difference where they're dinging you for things that are exactly in that area between .7 and 1.0, correct? So, no, I don't think that's the justification that they gave in their final approval. Their final approval said that we did not provide an adequate technical justification. And our technical justification was what they told us to rely on. I'll point out that their 2016 data that they imposed on us, the difference between the two standards actually shrunk to 5%, so it became even more generally comparable. But to give you the language, Judge Boggs, that we relied on, the comments from the EPA said encourage Kentucky to adopt what it called the, quote, straightforward approach of relying entirely upon EPA's projected design values and contribution data and apply the one part per billion screening threshold. That's the 126 JA? That is JA 126, correct. To shift the venue if I can, Judge Murphy, about your question, I agree it's a non-published order. The best published case out there about an unpublished motion panel ruling is the Wallace v. FedEx case. That is a published case. What was the rationale? I just don't understand why. So Wallace didn't really give a rationale for why we should defer. I think the way to think about it is Wallace said that a dispositive motion, your only recourse is to go to the rehearing procedure. So to think about this, if we had lost the motion to transfer venue, our only recourse would have been to go through the rehearing route. We won it, and so they're getting two bites of the apple that we would not have gotten. If we had lost the motion, we wouldn't have gotten to brief the merits. We'd just been to the D.C. Circuit. And so I think that's the rationale of it when it is, in fact, a dispositive motion. EPA pointed out in Docket 19 that their awareness of the rehearing route, they sought rehearing in the Fourth Circuit when they lost the venue issue there. So this I don't think is anything new to them. But even if the court's not, we've argued both, even if the court is not constrained by what the motion panel did, it is the considered views of both your colleagues. Judge Kaplitz, this was not a 96-hour issue. This is a 22-page ruling that was written over a dissent. We waited, you know, I guess we got an administrative stay in June, and it came in July. So we do have a reasoned decision. But can we turn to the merits of venue? I'm sorry, the merits of venue. I thought it was somewhat weak based on a determination of nationwide scope or effect. Why wouldn't you say under ordinary meaning the EPA has adopted the determination to use this updated data and they've applied that to all these different states? Or the one part per billion. So both of those. That's what the EPA has relied on, both of those. Why isn't that like an air standard to determine that now they're going to apply it to everything in front of them? I would urge the court to look at Judge Neubauer's opinion from the Fourth Circuit. He addressed this very directly. Yeah, I read it. So if you look at page 9356 of the Federal Register, that's the Kentucky-specific part of the final rule. That is the determination of whether Kentucky is linked above the 1% of the max. Well, in the jargon of 7606, it's really the action for which you petition for relief, right? Correct. So the final action. This provision uses determination separately from the word action, and so it would seem pretty clearly that determination in the phrase we're talking about is an intermediate decision on the part of the agency, and why couldn't that be something like we're going to use 1% not 1 part per billion going forward, at least with every SIF that's before us now? So I don't think that's what they did, though. They analyzed our specific arguments for a 1 part per billion, a 1% of the max, and they denied it based on our specific arguments. Well, I agree with you there was an action. Let's think of this in terms of IRAC, okay? So, you know, like the 1% standard is the rule. It seems like that is the rule that they're applying, and the A of the analysis as to each SIF is the application of that rule to the particular facts of the state. We would say that's the determination. I mean, if we go back to what I said earlier, determination is clearly something intermediate. It's not the action. I mean, 7607 uses the word action very clearly. That is the thing as to which you're petitioning for relief. The action is based on a determination, right? The determination is something intermediate. It's not the bottom line. Your SIF is inadequate. It is an intermediate thing. It doesn't say it's standard. If it said based upon a national standard, I think that would be right. A determination is some sort of reasoning, and so we understand determination to be taking reasoning and applying it to Kentucky. Determination is a conclusion of some sort, okay? I mean, I don't think this should be terribly controversial. It's a conclusion of some sort, and they're saying that their conclusion that they're going to use the 1% standard, all right, they're saying that is a determination on their part. Linguistically, that doesn't seem like a crazy thing, and so that's their determination. We're going to use that. That's what you're unhappy about, and they're saying that has nationwide effect because they're applying it to all the SIFs in front of them. So there are two parts to that. Is it a determination? I mean, I think you're going to have a tough argument, with me at least, to say that's not a determination. Why don't we talk about, or I'll ask you, why isn't that of nationwide effect? So one part per 1%, all right, and he's focused, I mean, he's talking about the data and the shifting data, but, you know, why aren't those determinations of nationwide effect when they are, I mean, conspicuously using the same analytical framework in every single one of these? So nationwide is obviously 50 states. We've got 21 state SIFs here. So we don't even have half the country. So even, I think, if we're agreeing with your description of determination, I don't think we have nationwide. I think this is 21 states. It's regional. The second point, though. Scope, any different from effect? I think here they're probably the same, as I would think about it. And, Judge Kessler. Scope means that they're going to apply it nationwide if people happen to come up. I mean, there may be places with no pollution, so it's not relevant. The other pushback, too, I have, that I have this idea of nationwide, is I think the court, if you're reading determination that broadly, you're going to allow the exception to swallow the general rule. What Niemeyer, Judge? That's what he said. And so the EPA has to apply consistent standards every time, right? The EPA requires that or this court is going to say that's arbitrary and capricious. And so I do think you have to look at are you going to allow it to swallow up. One thing that I think would fit regarding a SIP that is of nationwide scope or effect, and it's before the language that we're talking about here was amended, but it's a case of this court that I urge you to look at, is Dayton Power. That's 520F2D706. And what happened there was they amended every state's SIP in the exact same way. And when you do that, that is obviously of nationwide scope or effect. If you look at footnote. Every single state. Correct. And I think footnote 24 of the Texas 2016 decision from the 5th Circuit talks about what footnote 24, I don't have the case site for you, but Texas 2016. 2016 Texas. The Texas one. Not major. As opposed to the 2018 or 2020 one. But I would urge the court to look at that. Just a quick word, if I can, on harmless error for Judge Murphy. Before we jump off, so you disagree with Southern Illinois Power, Judge Sykes' opinion there? No, I don't. How is that different? That was an attainment determination. And yes, it only applied to particular areas, but every state has to consider an attainment determination in crafting its SIP so it was nationally applicable. I view that just like ATK from the 10th Circuit, where they talked about how these determinations were of nationwide scope or effect. I think a SIP call can be a nationwide scope or effect. We're just talking about a SIP, which then Judge Kavanaugh called the prototypical state action. I just think that we're in the 6th Circuit for that. So I wanted to clarify something you said just as a factual matter. So the denial only applies to what, 21 states, 22, somewhere around there? It's 21, yes. But wouldn't the rule apply to everybody? So the rule might not be applicable to everybody. If we think it is a rule, you have to use 1%, and we're going to use the 2016 modeling. It applies to everybody. It's just the states that were, I assume, approved, that's why they're not here, because everybody had to file a new SIP with the new national air quality standards. So I suspect that EPA would say, yeah, we applied the 1% to everybody. It's just that it's only relevant to those that we denied. But that doesn't mean that it's not a rule of nationwide scope. I think the premise, though, of your question runs into swallowing the scope of the general rule. They're going to apply a nationwide rule every time. And it doesn't say application of a national standard. It says determination, and the determination was taking that standard and saying that for Kentucky-specific SIP, it didn't meet it. And so that's how we think the based on determination nationwide scope effect would apply here. Just harmless error quickly. I do think the harmless error rule is very, very narrow when we're talking about an agency. CALCUT is the U.S. Supreme Court decision that establishes that. I think a couple of points here. I think this is somewhat similar to the determination in CALCUT. This is an iterative four-step process. We think the two through lines of their analysis are the data set and the contribution threshold. If you pull all of that out, we think everything falls apart. We also think that if you look at their criticisms of our step three analysis, which is what you were asking about, Judge Murphy, that you called quite conclusory, their criticisms are bound up with the step one and two, the data set, and the contribution threshold. So if you look, they criticized, for example, that we identified some unit retirements that would decrease emissions in Kentucky. Their response to that was not to say, well, even assuming the 2011 data applies, or even assuming, they actually used the 2016 data to then criticize and say, well, we accounted for these retirements or shutdowns there. So I just think that everything is bound up together. If you look at the page I cited you to in the Federal Register, 9356, which is the final rule as applied to Kentucky, they group all of their criticisms together and they say it's denied for all of these reasons. They don't say, you know, as an independent basis, even assuming you adopt the 2011 data, even assuming you adopt the alternative threshold, we deny. They group it all together and under the sort of the narrow scope of harmless error review. Judge Fox has a question. If I may, on the merits, at page 18 of their brief, they have a table one that shows the different modelings and the different receptors. Do you have any quarrel as a factual matter with that table? The quarrel that I have with it, I guess, would be both factual and legal. I'm asking you factually, in other words. Are the numbers there anything that you disagree with in terms of that's the numbers that they were using and that's the numbers that the receptors show? I don't have a quarrel with the numbers. Full stop. That's my question. Okay. From that, I understand the argument between what I call .7 and 1.0. That makes sense to me. If you lose on that point, if they are entitled to enforce the .7, then doesn't the table, especially the two Connecticut ones at the top, show that you lose no matter what data set you use? So, no. I think that when we drafted our SIP, as you can see from that table, we focused on different linkages. And so our analysis, as EPA has directed us, was context-specific to our linkages. And so if the court says it was not harmless for them, or that it was within the agency's discretion to impose the new data set, then we have new monitors that we haven't addressed yet. I guess I'm arguing a little bit with the premise of your question. Well, I'm hoping you wouldn't. I'd hope you'd answer it, which is, as I say, I'm not saying I'm going to go that way or my colleagues are going to go that way, but if they can use the .7, then isn't it true that no matter what set of modeling and data set you use, you're over .7 for those Connecticut monitors? So I would agree with that, but I think that the alternative would be, in light of the cooperative federalism framework to allow Kentucky to address those linkages, in the first instance we had not gotten to address the linkages and explain how they would apply. By that you mean that these other arguments about closing down plants and other contributions you didn't think you needed to address the Connecticut ones? So we did address part of the Connecticut in our SIP, to be frank, but again, as you can see, the numbers have changed slightly. At least so far as I'm concerned, they've beaten it to death. All right, we'd better let the agency have some time here. Thank you, Your Honors. Jeffrey Hammons for the United States. This case is about whether EPA reasonably disapproved Kentucky's submission under the Good Neighbor provision of the Clean Air Act. As Your Honors have already noted, that even if putting aside Kentucky's challenges to the screening threshold and the use of updated modeling, EPA still had a reasonable basis to disapprove Kentucky's submission. Even under that 2011-based modeling and under a one part per billion screening threshold, they were still linked to downwind air quality problems in Hartford County, Maryland, and as Judge Murphy pointed out, their submission and its analysis to that receptor was entirely conclusive. But with respect to Hartford, why didn't you model it under your 2016 modeling? I mean, again, I look at that Table 18, and I'm all ready to see, you know, okay, what happened to Hartford under the new modeling since you're relying on it, and you don't have any answer. I mean, Hartford could have been shut down. Hartford County still exists, but what happened there was the update of the new modeling. So when EPA periodically updates its modeling, it does two things. It does more than two things, but the main things here is they update the emission inventory, so where all the power plants are across the country, all the other ozone precursor emissions. They update where they are. Are they closed? Are they planned to close? How much are they emitting? And then it also updates weather patterns. And so what happened when EPA updated the modeling is that Hartford County was modeled to no longer be a nonattainment or a maintenance. Doesn't that suggest that there may have been a vague suggestion without technical analysis, but your own modeling suggests it was accurate? If you rely on the new modeling, which Kentucky says the agency cannot, but if you rely on the new modeling, it does show that Hartford County, Maryland, was predicted to be in neither nonattainment nor maintenance. That seems particularly weird then for you to say, oh, we're going to ding you for something that's no longer bad. Well, in EPA's proposal, they provided multiple reasons why under Kentucky's own submission. So it's important to note, EPA did not simply say, well, you had this old modeling, you submitted this submission, you had barely any technical analysis, and you based it on legal conclusions that would conflict with the Clean Air Act. Putting all that aside, you also are now linked under this new modeling, and because you didn't analyze your linkages to that new modeling, we're disapproving it. That's not what the agency did. The agency took every substantive position that Kentucky made in its submission and analyzed it on its own merit. It still analyzed the submission itself both under the 2011-based modeling, which is Kentucky's preferred modeling, and under the agency's. And the agency noted that you're still above 1% at these Connecticut receptors over all iterations of modeling, and so we consider you linked because we are using a 1% threshold. They say that JA-126, which I have not read, says use the 1.0. Is that accurate? What's going to happen when I go read JA-126? What's going to happen is what Kentucky's pre-draft submission did was they actually had two types of modeling, one by a third-party contractor, Alpine, and then EPA's 2011-based modeling. And their original draft submission sort of merged the two together, where they were using one set of modeling to figure out contributions, another set of modeling to figure out what areas are in attainment or non-attainment. And EPA's comment was suggesting that you shouldn't do that. You can't really take two different types of models and try to merge them together. By their standard, though, I mean, isn't it just egregious that EPA apparently specifically tells them use one part per billion and cite our memo? And they do those things. And then EPA says, no, now it's 1% and you're out of compliance. Not to mention blowing the deadline by a year and a half. I mean, how is that not capricious? To change the guidance in the way that the agencies seem to have changed it, unless I'm misunderstanding the facts here. So the agency's guidance, the August 2018 guidance, which we refer to as a threshold memo in our briefing, it caution states that these are non-binding guidance with recommendations. EPA did a nationwide analysis of all receptors under its 2011 modeling and found on average there is a similarity between using 1% and one part per billion. But it still caution states that following this guidance does not mean automatic approval if you're below an alternative threshold. This is the document where you say in appropriate circumstances you could use the other. Is there any guidance as to what appropriate circumstances means? No. The guidance assumed that the states would do a receptor-specific analysis of whether it is appropriate in the context of that state. The guidance also advised the EPA regions as well as the states in formulating their plans that any analysis of alternative thresholds need to take place within the facts and circumstances of the submission. Because some receptors, there might not be much of a difference between 1% and one part per billion. For other receptors, there might be a large difference between the amount of upwind contributions that are screened out if you use a higher threshold. EPA's position is simply that Kentucky itself just cited to this guidance, which does not mean it's automatic approval. Yes, from the EPA. They cited the guidance. Is the EPA at all contrite about telling them apparently use one part per billion and then saying here in this final action as to 21 states, we're using 1%, period? I understand the concern. I completely understand the concerns. It's a tough issue. I'm not trying to give you a hard time. From a rule of law standpoint, it gives one pause. You're right. Here's our position. It's that footnote two of that threshold memo links to past actions under the Good Neighbor Plan for the 1997 ozone standard and the 2008 ozone standard, noting that historically the agency has always used a 1% threshold. But we're now considering potential alternatives, and we want you states to consider it too and in your submissions explain why you think it might make sense in the context of your submission. After EPA reviewed all the submissions and issued its proposed decisions on those, it determined that there might be good reason in the good neighbor context maybe to not use a higher threshold. And there were two key reasons. One, consistency with past actions. So the agency noted that it would be a little anomalous here if the 1997 standard and the 2008 standard used 1%, but then the agency strengthened the ozone standard in 2015, but then allowed a less stringent threshold. So what you end up having is those older ozone standards are actually more stringent from a good neighbor perspective than the 2015 standard. It's consistent with one thing, but quite inconsistent with what they specifically told the states. Yeah, and what they... Because of this very set. And I can quote directly from the document. But I understand your point. You can go ahead. And so the other reason was there are equitable concerns with having alternative thresholds. One, it means as a meaty for downwind states, we'll talk later. It could mean that some upwind states using a higher threshold are able to screen themselves out so they don't have to analyze any emission reduction opportunities, whereas other upwind states do. So then the downwind states are faced with some additional pollution from some states but not others purely because of a different threshold. So there's equitable concerns there. And even among upwind states, there's also equitable concerns. If some states are saying, well, we can screen ourselves out by using a higher threshold and then other states like the state of New York, who is a meaty in this case, uses a 1% threshold. And so there was... I've got an argument for your not having been able to give that guidance, quote, guidance at all. That's where I'm... I mean, I understand a little bit about pollution rules and environmental rules. And it seems to me that there is such a difference from my math between the two standards that I never saw. I'm not saying which one is right, but I find it hard for you to say that it's comparable for any given state. That is, I read that national chart as on a national basis aggregate, it may be sort of comparable. But for any individual state, it's a huge difference, isn't it? There can be large differences depending on what specific receptors. And so it's important to note that what we are... what's being challenged here is EPA's final disapproval. And so while the guidance had potential alternative thresholds, it assumed states would do some analysis, receptor-specific analysis. Kentucky, for their part, did not. And EPA has explained in its final action why there might be good reason to stick with 1%. Okay, let me just go back to when we started with this question about JA-126. And I said, if I read it, what am I going to find? They're saying that it flatly says, you know, next time you submit this, use 1.0. Is that what I'm going to find? So what it exactly says is, this is in the context of them suggesting that you do not mix and match different modeling. It says, quote, an alternative, more straightforward approach would be to rely entirely upon EPA's projected design values and contribution data and apply the one part per billion screening threshold. And then it cites the memo. Okay, and so did they do that? They did. They did rely on EPA's modeling and they did use the one part per billion threshold. And then for the reasons explained in EPA's proposal, they were still linked above one part per billion to Hartford County and that the lack of technical justification for whether the, you know, power plant closures they point to, they don't explain things like, well, how much is it going to reduce your emissions? How much do those power plant closures comprise all your statewide emissions? When a power plant closes, the electricity has to come from somewhere. So does that mean existing sources are going to increase their emissions or a new plant is going to be built? All of those are information that the agency would need before it can approve Kentucky's conclusion that these power plant closures are going to be enough. None of that information and analysis was in their submission. So that's why we argue that even under their own modeling and their preferred threshold, that they still sent a submission to the agency that they just could not approve. If I could turn to Vinnie, is there anything wrong with my reading? So I'm talking about the exception for local and regional, which is based on a nationwide determination, a nationwide scope for effect, based on a determination of nationwide scope for effect. The Supreme Court has precedent that has interpreted the phrase based on and other contexts to adopt, at a minimum, but-for-cause principles, you know, like common law torts. So in this analogous action would be the action here is the denial. So you'd ask, the denial is based on a determination. If without that determination, it would have been approved. Is there anything wrong with thinking that that but-for-causation concept should apply here? No. And here, as your Honor has already pointed out, there are a few determinations that the agency made. But if that's true, all your harmless error arguments are tied to incredibly fact-specific items about Kentucky's SIP. And so I struggle with seeing whether the two, what you call generalized arguments, both the use of the 2016 data and the use of the 1%, how are those but-for causes of the denial, if you're getting up now saying the denial would have happened in any event? So let me count the ways, I guess, is the one thing. There's a lot of reasons why EPA could not have approved Kentucky's submission. One of them was that their submission had no technical analysis on these closures they claimed would reduce their emissions. So how is that possibly a determination of a nationwide scope? That would be a factual determination. But there were also other determinations. EPA determined that for all states subject to disapproval, they would use a 1%. Okay, but if you denied it on that fact-bound reason, the broader reason is not a but-for cause, because your point is that it would have been denied anyway. So, I mean, if we dissect the final action, and I guess it would first require, EPA put forward a bunch of independent basis for its disapproval. Some of those might go away. For instance, EPA also said that it was using 1% threshold. That might go away, but then the backup would be a one part per billion, and EPA said, okay, in that instance, you were still linked, and so you're still connected. And so the venue question is sort of separate from the merits. If we're discussing the merits, then I assume that means venue is here, because otherwise the merits would be heard in the D.C. Circuit. And so I understand. I don't see that. I mean, the venue provision references the merits of the action. It's based on a determination of nationwide scope or fact, and so we have to look at the merits for that purpose. And if this determination is basically gratuitous for purposes of rejecting the SIP, which would seem to be what your harmless error argument is telling us, then it's not based on it. So the determinations were not gratuitous. So at least under the framework, EPA's four-step framework, you have step one, which is identifying the areas that are in either nonattainment or borderline, and those would be maintenance areas. So that is a determination you must make that then informs step two, which is, okay, what states contributing to those areas have a de minimis amount, and then that's the screening threshold. So then you make a determination of what's the appropriate screening threshold, one percent, one part per billion. But once you've made that determination, then you look at step three. Okay, if you are above de minimis, you then have to show, okay, what are cost-effective emission reduction opportunities. States don't have to do that. They could do other things like, okay, we're going to do these other things that are going to reduce our emissions. They have to prove it, but then you make a determination, okay, was the step three analysis adequate to show that they are going to reduce their emissions or otherwise reduce their impact to those downwind areas. So all of those independent determinations inform the ultimate action, and so it's difficult to separate out and say only that step four determination. Step four. So you interpret determination to mean any analysis that leads to the conclusion. Black's Law Dictionary defines it as the decision of a court or administrative agency. It implies an ending or finality of a controversy or suit. So I don't know what the – I mean, determination is an opaque word, and we have to fit it in the context of this very long statutory scheme. And at least in your rule, you interpret determination to be essentially everything, including, for example, the use of the four-step process. But if the mere use of the four-step process itself is a determination of nationwide or scope or effect because you've always used it and you use it everywhere, doesn't that essentially suggest that every SIF approval or denial will be of nationwide scope or effect based on it? So no offense to Black's Law Dictionary, but I would say that that definition makes action and determination coterminous, and obviously they have to have different meanings. Not necessarily. It could be – there could have been a determination that was made independent of the local action, but then it's incorporated into the local action as the reason for the denial. An independent determination, whether a regulation – so say the first step is you issue a regulation. The state decides not to challenge it at that point, but then the state tries to get around the venue provision and then challenges it in the context of the denial of its SIF. You could argue that that determination, the original one, was final at that time. They should have gone to the D.C. Circuit. But instead, they tried to get around that by suing in the local circuit because of its application to the SIF denial. That would be one example. It seems to me that under your view, don't you essentially read out the local circuit scope? Because I think every action the agency takes, unless it's just acting with caprice or arbitrary caprices every time, just like doing whatever it wants every time, is going to be based on some type of national standards. And once you say that those national standards are the basis for the determination being of nationwide scope or effect, I just don't see what's left local. So I think that is not our position of what a determination is. A determination is a decision made in the context of an action. Here, the context was reviewing all good neighbor SIFs. Why isn't that the context? Why shouldn't we look to the Clean Air Act itself to suggest what the context is? The context suggests it should be a state-by-state basis. I don't understand why you can take what should be a state-specific action, the approval or denial of a SIF, package them all together, and then say it's a nationwide action. So the reason the disapproval applies to 21 states is because all of them were being disapproved, and there were common reasons for disapproval. All states that were disapproved, for example, all found they weren't linked at step one because they all used alternative contribution thresholds. That is one of the issues that we're now arguing about now, as well as the Fifth Circuit and the Eleventh Circuit and the Fourth Circuit and the Eighth Circuit. The same issue is raised there, the same thing with the use of updated modeling. The same issue is raised there, and it's our position that if there's one determination made in an action that is being challenged by every petitioner across the country, that's a quintessential determination that has a nationwide basis. I fully understand that argument, but then there's the rejoinder which Judge Niemeyer makes in the Fourth Circuit case, and it's basically what Judge Murphy just said, which is that the EPA can make some determination of federal law. It could be the most simple, uncontroversial determination that 7607 says thus and so, therefore, ultimately, we reject your SIF, or some other provision. The reasoning you just offered as to, say, the 1% standard would apply equally, it seems, to some pedestrian recitation of federal law, which is then part of a syllogism after which a SIF is denied. And so it kind of goes back to Judge Murphy's point that that would mean that there is no jurisdiction over SIF denials anymore in the circuits, and it just always goes to the D.C. Circuit. So our position would be road application of statutory or regulatory law, that is not what we mean by a determination nationwide. A good analogy would be like a SIF call. So a SIF call is when an agency finds that a state implementation plan is deficient in some way. Texas 2011 is a case that we cite that involved a determination was made that state implementation plans should have control or should have greenhouse gas as a pollutant under their prevention of significant deterioration programs. But there were only 10 states that didn't. And so when EPA issued that SIF call, it only applied to 10 states. So I would respectfully disagree with my colleagues that the action itself needs to apply to 50 states, whereas a determination nationwide scope is only applicable where relevant. I'm not tracking why you think the 11-state Texas thing is a determination nationwide scope in effect. Well, in that case they found it was actually nationally applicable even though it applied to 11 states. But it draws a good analogy because if you're going to look at determinations instead of actions, if they are applicable nationwide where relevant, that is an example. How is it nationwide but where relevant means it could be like three states, right? I mean, here we have 21. It could be two. It could be theoretically one. Why is that where relevant? Why doesn't that defeat the nationwide character when it's not nationwide, it's a lesser number of states, unlike a national air quality standard, which by its terms, that's like a rule, and that's federal law as opposed to determinations that are by nature state-specific. Yeah, anyway. Well, yeah, a national air quality standard is defined by the statute as nationally applicable, so in that instance you would never get to the determination nationwide scope in effect. But I think one thing that's important for the updated modeling, that was a determination that EPA applied not just to disapprovals but also approvals. For example, under the old modeling, it was going to propose to disapprove Wyoming's state implementation plan, but the updated modeling showed that they're no longer linked, so they withdrew that and then approved the plan. So the updated... I was going to ask you, suppose I interpret nationwide, not like some circuits have to like, you know, if it's close enough to all this nationwide, it is nationwide. If I interpret nationwide to mean across the stage, the rule that applies across the stage, is this... Would you still be here arguing that DC Circuit is the appropriate venue because either the 1% or the use of the updated modeling applied nationwide meaning it applied to every state? I can clarify. Yeah, I don't think I... Okay, so if I interpret nationwide to mean applicable across the country, is it your position... I think you disagree with that interpretation, but assume that that's what I think. Would you still be here arguing against venue? Do you think that that would doom your argument or would you still say that some of the determinations in the final rule had that type of nationwide scope? So I don't... I don't know how to divorce the applicability. You know, as your honors have noted, you know, think of that as a rule under IRAC. Well, whether it's applicable in that case will then depend on the facts and circumstances of that case. So, you know, the determination to use updated modeling because it represents state of the science and it's the best prediction, it provides the best predictions. Where it's relevant, it would apply to all states and that's what it did here. And I know where relevant is doing the heavy lifting, but that's because some states, they have... were not linked under any iteration modeling, whether used old or new. So like for in those states, it doesn't... It applied. It applied, it just doesn't... It just doesn't, yeah. I mean, there was nothing to be done because it applied. So that's the point I'm trying to make is a rule can apply nationwide, but if you fall within the rule, then fine, but it still doesn't mean it doesn't apply in this case. Yes, and I do think it's important to clarify when you say... Why do you think the use of data? So you use nationwide modeling data. To think of that as a determination struck me as strange. Well, in order to... The good neighbor provision is aimed at solving a collective contribution problem where you have a bunch of upwind states all contributing amounts of pollution to common downwind areas. For example, the New York City Metro non-attainment area has contributions from Kentucky, from West Virginia. From a lot of states. And so you need to start from a common baseline. And so you need nationwide modeling to figure out where are the areas facing air quality problems and what are the contributions. If we interpret the nationwide... The fact that you use nationwide modeling itself to be a determination of nationwide scope, again, I think we just obliterate the difference between the two provisions because will you ever not use nationwide modeling in the context of determining good neighbor obligations? I think that's an important caveat. In the context of good neighbor obligations, which is unique, most state implementation plans are just about the in-state effects of pollution within that state. The good neighbor provision is unique in that it is specifically about interstate pollution. I agree with that, but when you adopted the change in the air quality standard, didn't their state implementation plan include everything including good neighbor? Wasn't good neighbor a part of the state implementation plan? States revise their SIPs, their state implementation plans, for different programs at different times depending on when new standards come up. Kentucky has only ever submitted one good neighbor submission, which is the one it submitted in 2018 that led us here today. Was that only a good neighbor submission or was it the entire update including just for all the other... It did not include... It was not like a wholesale revision. Their SIP is huge. I just meant for the specific 2015. Oh, yes. It was limited to the 2015. But that would include not just the good neighbor provision, it would include their own in-state updates that are necessary for Kentucky to be in compliance, right? I'd say like 95% of the submission was good neighbor. I can't tell you even what the other 5% is. Maybe they can't, but one clarification before I leave, I know I'm way over time, is that it's not the use of nationwide modeling it's determination of which modeling is the best predictor for purposes of making decisions. So I would say that it's not our position that just anytime you use nationwide modeling, boom, you're in the DC circuit. It's if you're using it, but then there's different types of nationwide modeling and you have to make a decision and apply it to all states, which the agency did here, that is the determination. So it's not just the existence of the nationwide modeling. One hopeful final question. You're very helpful, Table 1, at page 18 of your brief. When I pressed you about Hartford and why I said you didn't model it and I thought your answer was the reason it says NA is that Hartford is no longer either out of compliance or even maintenance. Is that true for the other ones that are labeled as NA, Richmond and Bucks County and Milwaukee? Sorry, Your Honor, what page was it? Page 18 of your brief. It's your very helpful, Table 1, which gives the different receptors and the different modeling. Yes. Earlier in this enterprise, I pressed you as to why Hartford, which you're relying on, now is listed as NA and your answer, I think, was that, oh, it's no longer either a non-attainment or a maintenance area. Okay? And my question is simply the other ones, Richmond, Bucks, Milwaukee, that are also NA, is that the same answer? Yes. So where you see NA, it's... The answer is yes. That's fine. Thank you. I have one final question. You've been really, really helpful. All right, so back to venue just real quick. It's Part 2D3 of the final rule and this is regarding Step 2 of the four... I don't have a page number. Just Westlaw page number. Step 2 of the four-step framework, this is where the agency's talking about the 1% threshold and it says... It's the second full paragraph of this Part 3. It says, in this final action, the EPA relies in the first instance on the 1% threshold. And that's the only reference I can seem to find as to the potential scope, like temporally, I suppose, of this determination, if we want to call it that, to use 1%. And so if that's true and the agency has said in this action we're using 1%, then why wouldn't that be confining this determination to a scope that is not nationwide? Well, I think that's just because the action applies to 21 states. Right, okay. But the determination to what your screening threshold you use is dependent upon... I think Judge Ketlech's point is, did you use that screening threshold for the 29 other states when evaluating their plans? And is there any evidence that it was of nationwide scope? So even if it was easy to meet, where in the record do we find the notion that this was used generally across the country? So where in the record? I don't know, but I believe every state that was approved was under 1%. Yeah, so 1% was sort of why all those other states that were approved were screened out at the first step. So 1% was used, whether approved or disapproved. This action is collecting all the disapprovals. And so in the record, yes, we include all the approvals as well. And so if your honors would like, we can go pull... Sure, okay. We can submit a letter. That's fine. Okay, thank you very much, Mr. Hammonds. Very helpful answers. And we'll hear from Ms. Walthall. Oh, okay. This is a switch-up. That's fine. From the form, at least. May it please the court, Claiborne Walthall from the New York State Office of the Attorney General. Oh, I'm sorry. State and local amici. Yeah, no, that's all right. I was just looking. What's the other name? I thought it was the order in which you were sitting. So the good neighbor provision's core purpose is to protect downwind states, our people and our industries from pollution from upwind that we cannot control. This pollution has major impacts. Emissions from upwind states account for more than half the ozone at certain receptors in our jurisdictions. And significant amounts of that pollution could be reduced or eliminated through the application of controls. For example, states like Kentucky, including Kentucky, requiring their sources to operate pollution control equipment they've already installed or to make upgrades to modern equipment that many downwind state sources have used for years. If they don't do that, then downwind states bear the burden, which is they have to regulate our own industries, which are already highly regulated at even greater cost and difficulty. Or we risk not attaining the ozone standards and being subject to potential further reclassification and unattainment and more severe levels of unattainment, additional regulations of even more and potentially smaller sources. Just as a matter of first principles, don't you think... Are you defending the federal agencies kind of flip-flopping? I would think that any state, just as a matter of neutral principles, would think that's not a basis for cooperative federalism, to tell states one thing and then to do another. So it's hard for me to understand why New York's interest would be not aligned with the idea that perhaps the first standard was too lenient, but the way to remedy that is to give notice and tell them they have to do better. I believe you're speaking of the thresholds... And both, frankly, and the use of data that the states had no ability to incorporate into their plans because it postdated their submission. Well, I think with respect to the contribution threshold, which, again, here, we view as not determinative since Kentucky's own threshold and the modeling it preferred showed it was contributing significantly. But our point is that the framework for identifying and good neighbor contributions and the analysis that occurs once there's a linkage is not challenged here. Our real point is that the 1 PPP threshold does not make sense. And we have consistently... We've consistently commented against that. In Kentucky's submission itself, New York objected to the use of a 1 PPP threshold as opposed to the 1%. The use of that threshold would be a major inequitable shift in part for the reasons that my friend... But their argument is not on the merits. Their argument is, fine, but you changed course on us unfairly. And that seems... I mean, just as a matter of kind of a Rawlsian veil of ignorance, it seems like the states, all states, would have an interest in kind of a fair process. And that's what this case is about and not so much about the substantive standards. I think that there's too much made of the purported unfairness of the process because, really, the process is what EPA directed states to follow here, which was to examine the memo and apply the memo. And there was a citation to JA-126, but at JA-127, the following page, which is the comments from EPA on Kentucky's initial SIP, it says not only to examine the memo and look to it, but at 127, it said, and then apply it as part of the rationale, essentially for considering this one threshold. You said apply the memo, meaning the previous memo that has this ambiguous when appropriate and so on language. That's right. The threshold memo, the August memo. And its process, if at all, directs states to analyze their contributions and what the effect of applying that memo would be. Kentucky essentially just referred to it as a citation, but did not engage with any of the analysis that would be necessary. And the issue is that it would just allow states to choose their own. And you'd have a patchwork of upwind state standards. It would prejudice downwind states because it's essentially just screening out pollution that then downwind states have to deal with. The reason that's really... But having any of these standards is going to, quote, screen out some, correct? Yes. It's intended to screen out just a minute. And help me here because this is so technical. But when I read this in the context of Kentucky, the standard is 70 parts per billion. We're fighting about 1% of that or a little bit more than 1% of that. And we're also talking about, at least in most of these, your states are barely over the .70. So the sort of gestalt of Kentucky is going to put all this evil pollution on us seems really quite minor. Am I missing something? That is, when I looked at the nationwide one, there was something like, California had like 30 sometimes. I mean, are we... Is there that big a difference? Or is there something technical that means Kentucky is really harming you or you, New York, and others can really damage Kentucky in regulating them? Am I missing something general? I'd say two quick points. One is that the nature of ozone transport is many small contributions made up of multiple contributions from upwind to multiple downwind. It is complicated. The Supreme Court calls it a thorny causation problem EPA is fast with dealing with. Nice turn of phrase. And the issue is, if everyone says my contribution is too small and there's every incentive to do that, then all it does is push that increment of pollution onto downwind states. And downwind states are already controlling their emissions. That whole argument applies equally whether the standard is 0.5 or 0.7 or 1.0. Really, doesn't it? Yes, Your Honor, it does. In fact, they even talk about a standard of 2.0. I'm not saying any of those are right, but the rhetorical force of your argument is the same at every standard. Yes, but the issue here is that the proposal was to make it less demanding for upwind states while it was becoming more demanding for downwind states. And again, remembering that downwind states are controlling pollution. For example, for New York, $5,500 per ton of NOx. Connecticut, $13,000 per ton of NOx. Kentucky sources, in many cases, can reduce their NOx pollution for $1,600 a ton. That's a third of New York's cost. That's almost around a tenth of Connecticut's cost. And if that pollution increment goes away, it just shifts more onto downwind states, more and more. And so it becomes more difficult and more costly to reduce those same amounts of pollution and burden already highly burdened. That makes a great deal of sense, but I take it that the EPA standards and all of this has in fact nothing to do with that. That is, if Kentucky's out of compliance, it's out of compliance whether it costs them a zillion dollars per ton or whether it costs them $10 per ton. Well, you know... In other words, they don't give any Kentucky any grace if it was extremely expensive, but neither do they ding them if it's extremely cheap as long as they're within the standard. I think that's not quite correct, Your Honor, in the sense that there's always been an analysis under this four-step framework that the EPA has applied fairly consistently for over a decade where the step three of that analysis, once there's downwind and out of attainment, then you're linked, and then you analyze, which is what I think was demonstrated Kentucky really didn't do here, is that that does often take into account the cost of controls that are available, and so there is some equity across states, and that's part of the reason for linking in all of the states that are significantly contributing. All right. Thank you, Mr. Walcott. Appreciate your arguments. We'll hear from Mr. Kuhn. Thank you, Your Honors. I have three points, one on harmless error, one on the merits, and one on venue. Start with harmless error. Judge Murphy, your question, I think, is the right way to put it to think about harmless error. They are using Kentucky's analysis of the Hereford monitor that EPA no longer thinks we're being, that we're linked to, to then criticize our analysis. So we said in our analysis that we're not going to be linked to Hereford, that it's going to come into compliance, turns out that our analysis, even though you called it conclusion, we were right about that. So they can't use their data that says we're not linked anymore to criticize our now obsolete analysis. On the merits, I think it's clear that the way they've switched the data, switched the contribution threshold, it's made approving a SIP an impossible task for a state unless we're lucky enough, like Wyoming was, as was mentioned, that under their new data, we're not linked above the contribution threshold. There's been a lot of discussion about what they told us at JA-126. I'll point out, Kentucky was not alone on being told by EPA to rely on the March 2018 memorandum. After they submitted a pre-draft page 10 of our brief, we collected all of the advice that they gave to other states that's very similar to the advice that we got at JA-126. Let me take one more swing at venue and explain how I think all the various pieces of 7607b1 fit together. So we've got the general rule talks about state implementation plans in the singular, not the plural in the singular, approval, disapproval. And so we know that, generally speaking, the rule is that state implementation plans are locally or regionally applicable. And I think then Judge Kavanaugh said that in the American Robes. So we start with that presumption that we're locally or regionally applicable for SIPs. So then we get to the exception. Judge Murphy, I think you're right to focus on based on when we haven't talked about that language yet. You agree that it's at a minimum is but for cause? At a minimum, I think it can be even more substantial. It must be a meaningful or substantial basis. I think a good case for the court to look on that issue is Saudi Arabia v. Nelson. That's U.S. Supreme Court 507 U.S. 349 at Pennsite 356-58. The language there is based upon, but I think it's similar analysis to what we think would be. Does that undercut your argument though? If the conclusion is it wasn't a but for cause because it would have been denied for other reasons, doesn't that just kind of suggest that it was harmless? I don't think it would have been denied for other reasons. They never said that. The way that I think that a case like Saudi Arabia v. Nelson helps me is that merely saying a one part per billion threshold should apply and merely saying that the 2016 data should apply, that doesn't do anything after Kentucky. You have to then do analysis that's specific to Kentucky's receptors to whom we are linked. That's what I think the determination is. It's taking that and the substantial basis for the determination is applying that national standard. I do think, Judge Kethledge, that's the way we link the general rule and the exception. I think you asked a great question. Does a pedestrian recitation of federal law apply that would get you within the exception? The answer was no, but I didn't hear any clarification of how we draw that line. What application of federal law, which, by the way, applies nationwide. Which application is nationwide and which is not under EPA's paradigm? I think they told the court from the lectern today that every single SIP disapproval that touches on the Good Neighbor Plan is automatically going to D.C. Circuit. I think that would swallow up the general rule, which has always been the case, that SIP disapprovals are local or regionally applicable. That's no longer going to be the law because we have this robust exception. Thank you. Real quick, I understood from some of Mr. Hanlon's last comments that, okay, this rule aggregates the disapprovals and that there might have been another rule that aggregated approvals. Let's say it aggregated something like 29 and also applied a 1% standard. I don't know if those things are accurate, but if they were, what effect would that have upon the issue, whether it's nationwide scope or effect? Sure. I want to see what the submission says about the state's behavior. Sure. One thing that I will say... We add hypothetically if that were a scenario. I don't think you're going to get to that hypothetical because if you look at what happened to Tennessee, Tennessee was linked only to what's called a violating monitor reception or receptor. They tried to make that argument. As to us, you saw a little bit in the brief. They created a whole new type of violating monitor based on later than 2016 data that they applied to Tennessee and I believe it was Kansas as well. I don't think it's going to be nationwide because you're going to look at a state like Tennessee and see that there's a whole different type of monitor that they applied to disapprove them. Let me ask it as a genuine hypothetical because our decisions are just ad hoc. If we had something similar to the facts here but then we had an adjacent rule or action that approved the complement that equals 50 based upon application of the same standard and similar circumstances as here, sort of late in the game or whatever and they give you a 1%, something like that. Would that decision to use 1% be a determination or a nationwide scope or effect for purposes of that? I think that's closer. I would still rest on the idea that the determination is the application of the standard and that's the way that the two parts fit together. Just real quick. I'm curious. I meant to ask this of you, but does the case pending at the Supreme Court have any, I know it doesn't have anything to do with venue, but does it have any impact on the merits of this case? Will it have any impact how the court rules on, I guess it's Ohio versus EPA? I think the only place it came up in our brief is the EVA has asked for remand without vacater and there are two factors under this court's precedent. I do think that the fact that the Supreme Court could or could not vacate or stay the subsequent FIP that would be imposed on Kentucky, I do think that could be relevant to that issue of whether to remand with or without vacater. We think that cuts in our favor. But on the merits itself, the merits of the argument that Ohio made to the U.S. Supreme Court relate to the SIP disapprovals. The idea is that they didn't factor all of those SIP disapprovals into the FIP, but I think the merits here stand regardless of what the U.S. Supreme Court does. All right. Thank you, Mr. Coon, and thank everybody who argued today. I think that you really have helped us try to understand this complicated and important case. The case will be submitted. Clerk may adjourn the court.