Oral argument — State of Arizona ex rel. v. USEPA

I would like to begin by saying that to have some usefulness of these arguments with many parties and many issues, first of all, we have increased each argument to 30 minutes. And second of all, I thought it might be useful to tell you which issues we are particularly interested in. You are free to argue what you like, but we conferred a bit just before the argument and agreed as to which issues we would like to hear about. So in Arizona v. EPA, the issues that seem to us complicated enough that they are worth hearing an argument as opposed to relying on the briefs are the overnight cost issue, the unit aggregation issue as to visibility. And I hope I'm clear enough about what I'm saying. If you have a question of what I'm saying, let me know because I'm not sure I'm using the right technical words. The three-year averaging regarding visibility improvement and the complex of issues around the FIP authorization and the consent decree. And you see that one of the things we're not all that interested in hearing because it's a very complex issue. So I'm going to turn it over to you, Mr. Chief Justice, Glaser, for the State of Arizona. First of all, we appreciate the additional time for argument this morning. I notice that the clock hasn't started yet. Thank you. Good. I will be addressing, and we'd like to reserve out of the 30 minutes 3 minutes for rebuttal, if we could do that. Thank you. Could you speak a little louder? I'm having a little trouble. Certainly. I am going to address the issues around the EPA disapproval of the BART determination for the Coronado units and, in particular, the issues that the Court just informed us that it is most interested in addressing. I will be using the lion's share of the time. As the Court knows, the CHOA and APACHE issues have been held in abeyance, so we will not be addressing the issues around CHOA and APACHE. Mr. Fichthorn will be using the balance of the time to address the EPA imposition of the Federal implementation plan, the federal BART determinations for Coronado. One issue that I think on your list that we could pretty much eliminate right away is the issue, if I think that I understand it, the unit aggregation in terms of the visibility improvement analysis, if that's the issue, because that actually was not an issue on Coronado. EPA for APACHE and CHOA said that you have done the visibility improvement analysis wrong because you've done it unit by unit and you should have done it on a full facility basis. But for Coronado, that's not the case. We know Kagan. In Coronado, they actually did it that way. Fichthorn. They actually did it that way. Kagan. Okay. That's what I thought. Fichthorn. Right. Okay. Well, if that's clear enough, then we can skip that. Let me turn first then to the overnight cost issue that the Court inquires about. For Coronado, you know, the main issues in the case were degree of visibility improvement versus cost. The bang for the buck is something that you might call it. EPA had two criticisms on cost, specifically for Coronado. The first issue had to do with the fact that the Coronado unit cost determination was based on a quote received from the vendor. EPA said that that quote did not have sufficient line items in it in order to be evaluated. The second issue was the overnight cost issue. I take it from the Court's specification of issues that on cost you're more interested in the second issue, so let me address that one first. The issue on overnight costing has to do with EPA's regulations that go to the cost of compliance of BART factor. Under the statute, in making BART determinations, we are required to take into consideration the cost of compliance. Under EPA's BART guidelines, the BART guidelines provide that the States have flexibility in terms of how they determine cost of compliance. The BART guidelines say that States have the option of using the cost control manual, but they also have the option of turning to more site-specific information, and it specifically references vendor quotes. EPA takes the — Just ask a question. With regard to Coronado, I think, but I want to check this, there's actually no breakout of the costs, including these costs that are in dispute. Is that right? There are. The bid information that was received from the vendor specified a general cost category, like O&M, like capital cost. That became relevant because EPA said that we need to evaluate whether you're including costs that are excluded by the cost manual under the theory that the cost — Is it possible to do that with the data in the record at this point? It is. I mean, even if you — I mean, is it possible to tell what was included? In other words, whether these things that were not in the cost manual were included, can we tell that? Well, we would — let me just say two things. First of all, we don't think that those costs are excluded in the cost manual. I know, but what I'm asking you is can we even tell whether they were included or not excluded? Well, we can — yeah, we can tell you that the bid or the quote that was received from the vendor did include interest during construction. I mean, that's what the major issue in terms of the cost that EPA says that you excluded, that you included, but you should have excluded. And that goes to overnight costing. EPA's view is that the way you should — But what exactly does that mean? I'm having — I have a hard time understanding this. If you — if they go out and get a loan to build it, and the loan has a, you know, interest, and the interest is whatever it is, is that — is the notion is that it includes the interest as if it were paid with dollars as of the date of the lending of the contract, not the carrying costs during the period of the — of the construction. In other words, they don't take — they don't say you can't use it, you can't include interest, do they? No. They say that you cannot include financing costs that are incurred during construction. That's where the phrase overnight costing comes from. But what does that mean exactly? Frankly — If you — if you get — if you do get a loan and there is interest to — I mean, there's an aggregate amount of interest due. You're going to get, you know, $500,000 and you're going to have to pay $50,000 in interest over the entire period. Is that the $50,000 doesn't count or it just doesn't count? It is — it is in a lot of ways kind of incomprehensible, because as I was saying, the underlying assumption here is that you build the facility overnight. But even if you build it overnight and you borrow money, you still have to pay interest. Well, you pay interest, but you're not paying interest for the period of time. It could be 3 years, 4 years, 5 years if the facility is under construction. I take a million dollars out on day one. Well, I have to start paying interest on it. Even if I'm building my house, you know, and it takes me 5 years to build the house, I'm paying interest during those 5 years. And the reason I'm saying that is because EPA's position is we recognize that that is a real cost of compliance, but we have to exclude it. That somehow violates what's in the cost manual, which we don't think it's true. But even if it were true, then the cost manual must be wrong, because ultimately what triumphs — what trumps here is the statute. The statute says cost of compliance. But there — I mean, the biggest chunk of this cost is this AFUDC, whatever that's what that is. But what is that? That's not real money, as I understand. No, of course it's real money. It's attributed money. No, no, no. Of course it's real money. It is, as I said, if you borrow money on day one to build something, you pay interest from day one. EPA is saying that you will not get credit for that. You have to deduct that real — there's no issue here that that's not real money. I mean, EPA won't stand up here and tell you that's not actually real money. What we want you to do is assume that you built it overnight. I can understand why this would be difficult to understand, because it seems so far-fetched. But that's what overnight costing assumes. Assume that the facility is built overnight, and therefore you — But you're still paying interest even if it's built overnight. Well, all right. So I have another question. As I understand the EPA's explanation, it's essentially — ultimately what we're trying to do is a cost-benefit analysis or a cost-benefit analysis. And what matters in that is the ratio. And so in a way, if we back out a category of costs across the board, it doesn't matter because we — because it just makes it easier for us to look at cost ratios across different places or different projects. And what matters ultimately is what cost ratio we're going to accept. That's their argument. Is that right? That is their argument. They say we have to do this because it's consistent. What's wrong? Because that makes them consistently wrong rather than consistently right. I mean, how can you — It doesn't make them consistently anything. It just means that the — that the — because this is a fairly consistent — no, these are not site-specific costs. You say there are, but they're really not. If you look at the way they're used in one of the — as to one of the plants, it's just a flat percentage that they just, I don't know, 12 percent or something just say it. And it doesn't seem to come from anywhere. It's just an attributed amount. And so if you just take that out, then that just affects the ratio, and it affects what ratio you're going to accept, but it doesn't — it doesn't seem to matter otherwise. No, I'm not sure, Judge Berzon, what you mean by the ratio. The ratio that we're solving for here is cost of controls versus — Tons per dollars. No, we're actually solving for visibility improvement. We're not — we're not trying to reduce— But that's secondary. I mean, as I — it's not secondary. It's primary. But in the way that the calculations are done, it's a separate calculation. It doesn't have to do with this cost. Well, we have a statute that in very general broad terms says the States are required to consider five factors. One of them is cost of compliance. The other is — the second one is visibility improvement. Those are the two main factors. And so what you're solving for is real dollars that are spent for X amount of actual real-world visibility improvement. That's what we're trying to find out. And so for EPA to say we disapprove your cost analysis, your cost analysis violates the statute, which is really what they're saying here. Your cost analysis violates the statute because you included real-world costs seems to us to be the essence of something that's just nonsensical. And so that is really the sum and substance of that entire argument, whether EPA can say that because they want to have consistency, they can exclude costs that, Judge Berzon, as you said, actually apply to any utility unit. Any utility will experience these costs. So if we're interested in a consistent application, it would seem to make a lot more sense to say we should include these costs for everybody, because everybody incurs those costs. I would also say that it is very clear from the statute and also from the D.C. Circuit's corn grower decision that these BART determinations are supposed to be made on a unit-by-unit basis. Consistency is not actually something that you find in the statute. There could be different site-specific costs in different situations, depending on the nature of the plant, how much room they have to install control facilities. There might be higher costs at one plant versus another. And EPA can't try to force all of that into a single box in order to – because they want some consistency from BART determination to BART determination. But ultimately – I mean, this is a problem. I've had other cases with this. I mean, once you know the emissions and you know the cost, then you have to make some determination about how they relate to each other. And if you don't have any criteria or at least consistency across projects with regard to that ratio, then you're nowhere. You have simply no way to make a viable, non-arbitrary decision. And this has happened sometimes. Sometimes you say, oh, well, that's too much. Well, how do we know it's too much? Yeah. Well, you know, I would say, Judge Berzon, in this case, that you don't really confront that issue. And the reason why I say that is because, first of all, we do know that Arizona – whichever standard of review that you apply here, Arizona gets to make the first decision on weighing the BART factors and making a judgment as to whether the costs are worth the visibility improvement. EPA actually criticized Arizona for not adequately explaining its decision. And in that regard, if – I would just direct the Court to a portion of the joint appendix where Arizona set forth sort of its – sort of the penultimate weighing of the factors as to the Coronado unit. That's joint appendix page 710. And there's a table there. And in the table, Arizona sets forth four different control technologies. And for each control technology, Arizona sets forth what the cost of each would be and what the visibility improvement of each would be. The costs range from about a million dollars in control costs, actual control costs, up to about 17 million for the most stringent BART control measure. The visibility improvement numbers range from – I think it's 0.11 decibels, which is about a tenth of what is humanly perceptible, up to 0.34 decibels for a difference of 0.25 decibels. But all our – all you keep saying in your brief about that is it's a lot of money for an imperceptible improvement. But, of course, the response to that is, well, this one may be imperceptible, but the – but they add up. Yeah. Actually not. Actually not. Because these numbers are actually grave exaggerations of what the visibility improvement in the real world will actually be. We submitted to the Court recently – There are exaggerations on Arizona's part? No, no, no. Well, exaggerations is not quite the word I think that I'm looking for. They overstate by a great deal, yes, and on Arizona's part, not inadvertently but deliberately, according to EPA requirements, what the actual visibility improvement will be. And the reason why I say this is, as we pointed out to the Court in our recent Rule 28J letter, EPA took another action. It's called Phase 3 in this case. We're dealing with Phase 1 now. The next argument is Phase 2. There followed Phase 3, where EPA promulgated a second FIP. The first FIP here had controls for these big electric generating units. The Phase 3 FIP had additional controls that resulted from the disapprovals in Phase 2. EPA, in that Phase 3 FIP, compared the total visibility improvement you get from all the bar control measures in their two FIPs, including here for Coronado, with the State's plan. And the difference, what they call the FIP effect, all the measures, not just 0.08 to 0.16 deciviews. Those numbers are obviously much less than what I just talked about for Coronado. So how is that possible? I'm sorry. That was 0.08 to a 0.what? One-sixth. One-sixth. And that, I'm sorry, that is the rate, that's the marginal rate of improvement, or that's the absolute change in the, that's what the deciviews are now? That is the marginal rate of improvement, all Federal measures versus all State measures. So that's completely unresponsive to what I said before, though. But, no, what I'm trying to say is the reason for the difference is that in the, what EPA did in the Phase 3 FIP is they said, now that we know all the Federal measures, we need to figure out all the visibility improvement we're going to get. And we're going to model that improvement by looking at a background condition. We're going to model actual real-world background conditions. The BART determination modeling per EPA requirements is required to assume that the background is natural conditions. And so it overstates the visibility improvement because the actual background is not natural conditions. So when I say, look at this page on the joint appendix with the table, and I say, and as Judge Berzon, you said, well, you're just saying that, you know, it's just a little bit and lots of little bits add up. Well, actually, it's a really, really, really, really, really little bit. It is virtually nothing. Two sort of contextual questions. One is the, in Apache and Shola, reconsideration has been granted and we've stayed the proceedings. In Coronado, my understanding is that there was a motion, there has been a reconsideration petition, and the EPA has agreed to hear it. Is that right? We have counsel for Coronado here who will be speaking right after me, and I think he could probably address that with more detail than I will be able to. All right. You don't want to tell me. So related to that, maybe, or not related to that, is another procedural question. We gave you a list of issues, I guess. Suppose we thought that as to one of them, let's say we disagreed with you on the overnight cost, or let's say we disagreed with you on the other issue, haven't addressed the three-year averaging question, and we agreed with EPA. At that point, does EPA prevail because they had a basis for disapproving the SIP, even if everything they said wasn't right, or do they lose? What happens? Well, if we, there were, we have three disagreements with EPA on the bar determinations specifically for Coronado. Cost, visibility improvement, we've covered both of those, and the adequacy of the explanation, and we've covered that as well. Obviously, if you agree with EPA on all of those issues, then these specific disagreements on Coronado go away, but arguments three. That's not what I was asking. I'm sorry. I don't agree with them at all, but I agree with them on some. Oh, you don't agree with, you don't agree with, I'm sorry, Judge Berger. I agree with them on some, but not all. I think that at that point that you would have to look at the materiality of each one. Remember, you know, for instance, on adequacy of explanation. There's something odd about the whole issue about adequacy of explanation just in the sense that, you know, if this were court, if this were a court reviewing and said the explanation was inadequate, it would get sent back to the agency to provide the adequate explanation. But since this is EPA reviewing the alleged inadequacy of the explanation, what happens is EPA imposes a FIP. So our contention would be that it's a little odd for EPA to say what was missing here in your SIP was a paragraph, two paragraphs, explaining what we think is the obvious, and as a result of that, we now get to impose this $17 million a year in costs on you. That seems extreme, and so. But, for example, if we disagreed with you and agreed it was EPA and overnight cost, the rest of it wouldn't really matter, would it? No, I disagree with that, because I don't think on the total cost issues that, and I think, Judge Berzon, you know, you intuited this, that the amount of money at issue, you take the actual construction cost of the unit, including the financing costs on that unit, and you separate out just the piece that's financing during construction, I have a 20-year loan, let's say, and so for the first three years, you know, that financing cost, that's not really material. So we do have a punishment fitting the crime here. I would say that that is the case for all three of these particular disputes that we have with EPA. None of them are serious. Visibility improvement, their issue on visibility improvement on Coronado is that we took an average of the visibility improvement at nine Class I areas, rather than, and they said, well, including the five most impacted Class I areas, and they said, well, that could dilute the effect that is happening at the most impacted area. But, of course, we In fact, as for the most impacted area, which was the Grand Canyon, in the last year, the number was actually quite significantly above the cutoff. Well, the most impacted for Coronado was Petrified National Forest. Right. All right. But EPA said in its brief that the total dilution was only 0.05 decibels. In other words, the average of the nine Class I areas was 0.34 decibels, and at Petrified, it was 0.39. This is not material. Even if you go in their direction and you say, well, really, there should have been, you know, something else that you looked at here, it's not the criticism is not material. So we would suggest to you that, you know, if there's one of these that you disagree on, then it would be up to the Court to determine materiality. And I think, in fairness to my colleague, that I should sit down and let him talk, unless the Court has further questions for me in particular. Thank you. May it please the Court. Norman Fickthorn for Petitioner Salt River Project. I understand that the Court is particularly interested in the consent decree issue, so I do want to address that. I do want to preface my remarks briefly by noting that if the Court determines that EPA's decision to disapprove the SIP as to Coronado is — if it affirms that decision, there's no need to go to any issues having to do with the FIP itself, because if EPA had no authority to disapprove the SIP, under the Clean Air Act, it has no authority to promulgate a FIP. And I think all parties are in agreement on that issue. Okay. Could you just tell us what's happening with the reconsideration, or is there one? Yes, Your Honor. On April 9th — well, let me back up a second. The Salt River Project filed a petition for administrative reconsideration with EPA on February 4th, 2013. This was about two months after the final EPA rule was published in the Federal Register. EPA did not act on that petition for reconsideration in any manner until April 9th, 2013, at which point it submitted a letter to counsel for Salt River Project stating that it had decided to grant reconsideration on certain specific issues raised in the petition for reconsideration. Not all the issues, by any means, but certain specific issues. And by granting reconsideration, EPA meant, and it explained this, that it had decided that it would initiate a process by which it would solicit public comment on those particular issues raised in the petition for reconsideration. Again, that letter was dated April 9th, 2013. It is now one year and 11 months since that letter was received. EPA has taken no action, no public action to solicit public comment on any of the issues that it addressed in the petition for reconsideration. With respect to the consent decree issue, our position is that EPA arbitrarily refused — and again, this applies to the FIP for Coronado in particular — EPA arbitrarily refused to apply the provision of its own BART guidelines that states that emission limits in settlement agreements that are designed to resolve disputes under the Prevention of Significant Deterioration Program, which includes a BACT requirement for certain sources, BACT meaning best available control technology, that those settlement agreements generally should be deemed to meet BART requirements. This is relevant with respect to Coronado Unit 2, but also to Coronado as a whole. Coronado Unit 2, under the consent decree resolving PSD allegations that was entered in 2008, Coronado Unit 2 was required to install SCR to meet a 30-day rolling average emission limit of .080. It has actually done that now as of June 1, 2014. In addition, the consent decree set a plant-wide NOx emission limit for Coronado of 7,300 tons per year. SRP has made clear in the rulemaking as well as in the briefing that the effect of a plant-wide NOx emission limit of .020 pounds per million BTU, which is actually below the presumptive BART limit for the types of boilers at Coronado that is specified in EPA's own BART guidelines. There is a range, EPA says, of presumptive BART limits for Coronado. Kagan. To be honest, you're losing me. All right. I thought that the consent decree for the unit that was covered had the — what was the emission assumption there? In the consent decree itself? Yes. For the consent decree required .080. Right. So what's this other number you're talking about? Okay. The other number derives from an additional provision in the consent decree itself, and that provision covers the entire plant, both units, on a plant-wide basis. And that limit is a tonnage emission limit. It's not expressly articulated in terms of pounds per million BTU, but it is a 7,300 ton per year limit on NOx emissions from the plant as a whole, which SRP pointed out in the record and in the briefing, equates to a plant-wide emission limit of .020 — I'm sorry, .2 — I'm sorry, excuse me, .20 pounds per million BTU, which, as I said, is below the presumptive BART limits for Coronado. And therefore? And therefore, our position is that the consent decree under EPA's BART guidelines should have been accepted by EPA in the FIP as BART. Now, we actually take the position that it's more stringent than BART in our view, but at a minimum, EPA should have concluded that the requirements of the consent decree satisfy BART for BART. And therefore, our position is that the consent decree should have been required by EPA in the FIP. Where do you say there's a provision in the guidelines that says you're supposed to — that the consent decree is binding? Where is it? It's at Joint Appendix 4573, Your Honor. It's 70 Federal Register 39164. It's at 39164? Correct, Your Honor. And what is that? At that page — Yeah, but what is that? Is that the final rule or is it something else? This is — what I'm referring to here, Your Honor, is the BART guidelines themselves, which EPA says are binding on plants such as Coronado. Now, what we have actually starting at Joint Appendix 4572 at the very end of that page is a discussion of EPA's treatment of so-called MAC standards, which are standards to address hazardous air pollutants. And EPA goes through that discussion by saying if you have a MAC standard and it's in place at a facility, you can assume that unless there have been full technologies since that MAC standard was established, you can assume that that MAC standard will satisfy BART. Now, it might actually be more stringent than what you would determine BART to be if you didn't apply the MAC standard, but nothing more stringent than MAC would be required. Where is the discussion of consent decrees? Then on consent decrees, it's in the first full paragraph at JA 4573. There, EPA says — And that's at 70 — Federal Register 39164? Correct, Your Honor. First full paragraph on that page, we believe — That's not what you said at all. It says we believe that the same rationale also holds true for emission standards developed and for many NSRPSD determinations and NSRPSD settlement agreements. However, we do not believe that technology determinations from the 1970s and early 1980s should be subject to consent decrees. Well, Your Honor, we believe that this is properly read to say that there is a conclusion that the consent decree will apply if it resolves PSD issues, which — and there's no dispute that there is — It doesn't say that. Unless there's — unless there are advances in control technology, which no one argues in this case there have been. One or the other of you. Two minutes in rebuttal. Yes. Thank you, Your Honor. Good morning, Your Honor. Angeline Purdy for EPA and with me is Charlotte Withee of EPA's Office of Regional Counsel. I'd like to begin with the cost issue and just to clarify, and I believe the Court already indicated this, the — whether the analysis for Coronado complied with the cost control manual or not was not the basis for EPA's disapproval. EPA stated that it simply didn't have sufficient information. The cost analysis submitted did not contain sufficient detail. So we are happy to answer any questions the Court may have about EPA's approach to cost in general, but that was actually not a basis of EPA's disapproval of the SIP for Coronado. It was the failure to document the cost analysis. So what — coming to that conclusion, were you relying on Arizona's SIP report, a SIP statement, or were you also looking at whatever underlying information they got from the power plant? It was a combination of both. There was a cost analysis submitted by the power plant that the State then used as the basis for its SIP determination. And because there's been discussion of the SIP's — excuse me, of the State's finding, I think it's useful to look at what the State actually said that appears at JA-710. It said, after reviewing the BART analysis provided by the company and based on the information above, the State has determined that BART control at Coronado is a particular control technology, which was the least stringent available, by the way, on a — you know, that's it. That is the balancing. That is the analysis. That is all the information the State gave EPA about how it had reached its conclusion that this was BART. What the State is now saying about, oh, well, it was too much cost for too little visibility improvement, that doesn't appear. They just laid out the information on each factor and then said, well, after reviewing all of this and after reviewing what the company did, this is BART. So it's that explanation of how the State reached its conclusion that's missing. And that was — Can you tell me again where that is? I'm sorry, that's JA-710. It is also cited in the Federal Register at 77 Federal Register 42851. And that appears at JA-19. EPA specifically cites and quotes that statement by the State. That's what the State told EPA about why it was ultimately making the BART determination it did for Coronado. And EPA concluded that the cost information the State had looked at and in turn provided to EPA simply was not sufficient to allow a reasoned analysis of costs, but also that the State had failed to explain its basis for the BART determination. I would just like to make one quick point on the cost issue because I think that it sort of plays into the way that we look at what EPA did here. The guidelines that EPA applied, that it applies to cost determinations, that it also applied, for instance, with regard to the consent decree issue, those were promulgated at Congress-expressed direction. And so when we talk about statutory language, as counsel was pointing back to the language of the statute, that is language that Congress expressly told EPA to interpret through regulations and through guidelines. But this cost question is not exactly in the guidelines. It's in the cost manual, and it's sort of by implication. Is that right? The guidelines favor use of the cost control manual. They say that cost calculation should be safe. And the cost control manual does this. It doesn't exactly use the terminology, but it does it. Correct, yes. Yes, that's correct. It's from the guidelines to the cost control manual. But the guidelines, again, were promulgated at Congress direction and were specifically intended to aid States in making these determinations. But your determination was that you didn't have enough information even to see whether they satisfied the CCM, and so you were just going to promulgate your own FIP and deny the FIP. Is that correct? For a number of reasons. Because they had not supported the cost calculation, because they had not explained the basis for the ultimate BART determination, because they the visibility – because of the visibility issue that the Court has also indicated it's interested in, because they had understated the control effectiveness of a particular technology, all of these went into it. How difficult would it have been for Arizona to have supplied the figures that you wanted? Your Honor, I don't know offhand, but certainly these are – this is information that States are capable of providing to EPA in the context of a BART determination. Did EPA then attempt to do its own cost calculations? I mean, it just seems like somebody had to do the right cost calculations. If you didn't have the right information, somebody had to get it, and that was either Arizona or else EPA. But somebody needed to do this in order to figure out what the right answer was. EPA did do cost calculations in connection with its promulgation of the FIP. It looked at cost information that's been gathered for use in EPA modeling. It's EPA's integrated planning modeling, has a great – one of the modules of that has a great deal of information on the cost of this particular control technology. And so EPA used that information. It also used site-specific costs submitted by Coronado to the extent that those were consistent with the control cost manual. EPA used those specific figures. So where Coronado sent a figure that was not consistent with the control cost manual, EPA would not incorporate that, but it incorporated whatever site-specific costs it could in addition to the preexisting information in the agency's possession regarding the costs of the FIP. Do you want to see if you can – you're essentially saying this, how overnight cost you think doesn't matter. For the purpose of EPA's decision here, no, it does not for Coronado. So while I'm happy to answer – Are you going to tell us about it nonetheless? It's very mysterious. We're trying to do all sorts of accounting theories. We are certainly happy to answer any questions the Court may have, but I think for purposes of the Court's information and decision in this case, it really isn't pertinent because that was not a basis for the SIP disapproval on Coronado. It was not a basis for, I'm sorry? It was not a basis for the SIP disapproval on Coronado. Is that true? It just wasn't mentioned at all? Or was it an alternative basis? The fact that the State had not adequately documented its cost analysis, that was a basis for disapproval. The fact – one of the bases for disapproval. The fact that whether or not the overnight methodology had been used in that cost analysis, that wasn't a basis because EPA couldn't even tell. So the use or nonuse of the overnight methodology, the allowance on funds used during construction, those are simply nonissues with respect to the SIP disapproval on Coronado. But once you promulgated your FIP, again, my question was you had to do some kind of calculations. So you excluded these costs. Is that correct? We would have excluded them to the extent that they were provided by Coronado, and I confess I do not know, and I don't know if the record clearly indicates exactly what costs Coronado provided. But I do know that EPA used whatever site-specific information. It would have excluded them. It would have excluded them, yes. It would have excluded them. Yes, it would have excluded them. So their challenge on the FIP is that you've excluded costs that ought not to be excluded. That's my understanding, yes, although I don't think so. We're going to have to figure out what we think about overnight costing one way or the other. It may not be relevant to determining whether EPA was justified in turning down the FIP, but it would be relevant to determining whether we have to, whether the FIP is arbitrary and capricious. To the extent that Coronado submitted information that included that costing, which I admit I'm sorry, I simply don't know off the top of my head, that with regard to overnight costing, then, as I've already discussed, this has required, or excuse me, the guidelines that EPA promulgated at Congress direction recommend use of the control cost manual. The control cost manual. Okay. But what does it mean? I mean, we had this dialogue before. I don't really understand what it means. Does it mean, it doesn't mean, I take it, that you don't count interest. It actually does, Your Honor, in the context of this allowance on funds used during construction, which is something that, as I believe counsel indicated, any utility is likely to experience. This is a cost methodology that is typically used by utilities. By what? By utilities. To spread the cost of interest out over a number of years and work that into what they're able to recoup through utility rates. Okay. So this is something that's typically used by the utility industry. The guidelines apply to all industries. And in looking at BART determinations and in looking at what controls are cost effective, which is the metric that EPA uses to assess cost, cost of compliance, EPA strives for consistency, strives for consistency across industries. This is a method of accounting that is, as far as we know, particular to the utility industry. So in developing an analytical methodology that can be applied consistently and that EPA can look consistently across industries, EPA uses what the manual recommends, what is referred to as the overnight methodology. That's a colloquialism. That term does not appear in the manual itself. But what the manual sets out is the costs that are allowed to be included do not occlude that allowance on funds used during construction. And that does not mean, as counsel has suggested, that site-specific costs are never appropriate to consider. We talk about this in our brief. An example would be a facility that really, truly has unique considerations due to where it's located in the country, for instance. Our particular emissions control would require it to acquire a great deal of water, and it's in an arid region. Certainly, those sorts of truly unique facility-specific costs can be considered, and the manual contemplates that. The guidelines contemplate that. What they don't contemplate is including costs that are unique to or, excuse me, using a cost methodology that is unique to the utility industry, as far as we are concerned. But I'm really having trouble understanding why it's unique to the utility industry. If somebody wants to build a major control mechanism and they need to go out and borrow money, what's unique about that? Your Honor, I apologize. I do not have an answer for that. What I do know is that it comes up primarily with utilities. I believe that it has to do with the utility industry. Overnight, you still have to pay interest. And I did not understand that you were excluding that. I thought you were excluding some projection of that over the construction period, but you weren't saying if you have to borrow money to build this thing, the cost of borrowing, it doesn't matter. I mean, I guess, I mean, this may be just to demonstrate why we shouldn't get into in this case, because you can't tell. But it doesn't make a lot of sense to me to say that, I mean, for example, if this year it costs, you're going to have to pay 10 percent interest, and 5 years ago you only had to pay 2 percent interest, then it does seem relevant to how much it actually costs to build this thing or to buy this thing, to buy it, leaving out the building, the buying. Counsel, isn't the reason why FUDC is, we use the word unique, to the utility industry is because of the unusual relationship between utilities, which are privately owned, they have shares and so forth, you can, on the exchanges, but still the public, the rate payers, pay part of the project. So it's a very unique SIS setup, and I guess the question here is that that's not the setup we have here, I take it, that these power plants, or are these power plants like what we have here in California? Because in California, when Edison has its projects, AFUDC is talked about all the time. Is that a similar setup that we have with these power plants in Arizona, or is it the same? As far as I know, it is similar, but you're correct, Your Honor, as between power plants here and power plants in Arizona, but you're correct that it does have to do with the way that utilities function as opposed to other industries. And the guidelines apply to all industries, all facilities at which EPA is looking at BART determinations. AFUDC has to do with the way that costs are spread over years and ultimately work into the rates that rate payers pay, which is a utility-specific consideration. In the interest of ensuring that determinations are made consistently, reviewed consistently, the guidelines don't contemplate including that sort of industry-specific cost allocation as part of the cost of compliance. So, for example, if a large power plant was at a car factory, and this power plant at 750 megawatts, would, I take it, the analysis would be the same as far as EPA is concerned? Correct. And so that would be a situation where, because that's not a utility, there's no AFUDC issue, whereas in the utility industry here in California, I know California, I used to do energy law, so in California, then, they're specific to utilities because there's this relationship between the state, the utility company, and the taxpayers. Or not taxpayers, I'm sorry, the rate, the rate, the rate, people who pay for their every month, they have their energy turned on in their house, whereas you're saying BART, though, is, in a sense, a nationwide, industry-wide standard. It happens to apply to utilities here, but it doesn't always have to apply to utilities. Well, I want to be careful with that. BART is a source-specific determination. It is always going to be a source-specific emission limit. The analytical methodology used to get to BART, though, EPA strives to make that as consistent nationwide, as consistent across industries, as consistent across states, as it can be. And so in your hypothetical, where you have a power plant and a car factor, each of which is over 700, you know, the power plant's over 750 megawatts, both are subject to the guidelines. You, or excuse me, you want to have a consistent determination. I apologize, I misspoke. The guidelines would be mandatory as to the power plant over 750 megawatts. You want to have a consistent methodology of those two. You want to be looking at the same factors. You don't want to be building in interest costs that are unique to the utility part while you're looking at the associated car plant completely differently. And that's what the guidelines and ultimately the cost control manual are trying Sotomayor, we've now taken a lot of your time up about this. Do you want to talk about the consent decree authorization? The authorization question? Certainly. With regard to the consent decree, I think as the Court has already recognized, the guidelines state that consent decrees or PSD settlement agreements may be BART. They aren't necessarily. In this case, EPA took the consent decree into account. And the emissions limit ultimately imposed at Unit 2, which is the one actually subject to the consent decree, is consistent with the consent decree. So for Unit 2, EPA did not go in and adopt a lower emission limit. EPA did conclude, however, that first of all, it was possible to install additional controls on Unit 1. And second, that even at Unit 2, the emission control system that Coronado was installing in response to the consent decree, that there was technology that would enable that to be even more efficient than it actually was. And that could be installed consistent with the stage where Coronado is that wouldn't require them to go back and start over, for instance. Based on that, EPA adopted the plant-wide emissions limit. So it took the consent decree into account. It took the guidelines statement that it may be BART into account, but ultimately determined that in this particular case, no, the consent decree was not BART for the facility as a whole. And I'm happy to. Ginsburg. There's a very complicated argument in one of the briefs, opposing briefs, trying to demonstrate that this is all not doable because of the fact that one or other of the units may be shut down for some period of time, and therefore it would be impossible to obtain the plant-wide numbers. And I understand that you then attribute some emissions to the period that they're actually closed down. It was very, again, hard to follow, but it was at least facially plausible that there might be problems in some time periods. Just as a point of information, that is actually an issue on which EPA has granted reconsideration, although as indicated. I'm sorry, what? That is an issue on which EPA has granted reconsideration, although as indicated, we have not yet proposed action. My understanding is EPA anticipates doing that relatively soon, although I don't have a specific date for that. Well, I mean, we've gotten, you know, a little roast pulled out from us in this case. Should we be waiting for that? Your Honor, I would have to consult with my counsel on specifically what EPA's plans are, but at this point I would be hesitant to represent that action is anticipated in the near future. However, we will certainly, if we learn of anything that suggests that this issue might become moot or might be resolved through reconsideration, we will certainly promptly inform the Court. With regard to the plant-wide emission limit, that was adopted in an effort actually to give the source some flexibility in being able to, that was something that was actually requested in comments as an option to be able to average emissions across the plant. EPA adopted that as simply a plant-wide limit. So, again, though, that is an issue that EPA is going to be addressing on reconsideration. So I'm happy to discuss the details of that further if the Court has specific questions. But EPA's conclusion on the record was that, in fact, Coronado could meet this on a plant-wide basis. That, yes, some reductions from Unit 2 would be required beyond what is contemplated under the consent decree, but the consent decree sets a maximum. It doesn't set a floor. And that the technology was available that would enable Coronado to reduce those emissions further and meet the plant-wide limit when you look at both units together. And finally, before you sit down, what about my question about what if we agreed with you on some but not other things? I mean, I gather on the FIP that would probably invalidate the FIP, wouldn't it? But as to the disapproval of the SIP, it would run the other way. If the Court, take those two in stages. If the Court ultimately concluded that EPA's SIP disapproval was arbitrary and capricious, then, yes, there would be no basis to go on and do a FIP, because EPA. Sotomayor, but if you had a basis for disapproving it, but not another basis, you would still have a basis for disapproving it. Yes, you're right. And that's the other piece, is on what basis would the Court conclude that that was arbitrary and capricious? There were a number of bases here for EPA's SIP disapproval. And I don't think that knocking any one of them out would be fatal. So if the Court disagrees with us on one sub-issue or another. But what about on the FIP, then? I'm sorry, Your Honor. Perhaps I'm not following. All right. So that's the SIP. Okay. The FIP, the Federal plan, presumably would run the other way. I mean, if it disapproved part of it. I think if by what you mean run the other way is that the Court can uphold EPA's promulgation of the FIP, even if it doesn't necessarily accept every single basis for that promulgation, then, yes. I mean, ultimately the question is whether EPA's FIP limits are arbitrary and capricious or whether they're supported by the record. And if the Court finds that ultimately those were reasonable and supported by the record, then partial disagreement. But usually that's not the way you review a direct Federal agency action. If the Federal agency had, because of the Chenery issue. So if you had an erroneous reason for something, we don't then go make up our own reason. I'm sorry, Your Honor. Perhaps I'm not. If your basis for a conclusion was incorrect, in our view, or arbitrary and capricious, then we would remand. We wouldn't go looking for another reason that would be okay. No. No, of course not, Your Honor. And I apologize. I was not intending to suggest that the Court would invent its own reasons. Only that there is a great deal of information in the record supporting the FIP and a number of issues that have been raised. Well, you have two reasons. Multiple basis. Multiple basis, yes. And disagreement on one of those would not necessarily be fatal. That's ultimately the Court's assessment of whether the record still supports the FIP, even with a conclusion that one of the basis was arbitrary and capricious. But no, of course, we did not intend to suggest that the Court could, you know, create reasoning that it finds absent from EPA or establish its own basis. If the Court has no further questions, I think that I have addressed everything I intended to. So I had promised interveners that I would yield a bit of time to them, and I have time left, so. Thank you, and may it please the Court. My name is Michael Hyatt, and I represent the interveners. Arizona's BART determinations violated the Clean Air Act and failed to provide the Grand Canyon and 17 other national parks and wilderness areas with the protections they are due. The act required Arizona to select the best available pollution controls as BART for Coronado, which was one of the worst sources of visibility impairment in the nation, according to the National Park Service. Now, Arizona's plan didn't select the best available controls as BART or even the second best pollution controls. Instead, it selected the least effective pollution controls as BART, and it did so based on a flawed analysis and without providing a reasoned explanation of how it reached the anomalous conclusion that the worst pollution controls are the best available. There have been several questions about the cost issue so far, and there's a few quick points I'd like to make on that, is that EPA is correct. The reason they disapproved the cost analysis for Coronado was not because of a failure to comply with the cost methodology, but because the state's cost estimates were just so overly generalized and lacked such sufficient detail, there was no way that the state could have reasonably and rationally evaluated whether those costs were reasonable. And the Joint Appendix pages 19 and 20 explains that. Another issue is that EPA explained that the cost analysis for Coronado was such lacking in detail that EPA itself couldn't even begin to determine whether or not the state did or did not apply the cost methodology. And as that page, Joint Appendix – Did EPA have the figures in order to calculate it itself when it did it to FIP? EPA – I'd have to leave it to EPA to say exactly how they use the site-specific information in the FIP. They did have to conduct their own cost analysis to support the FIP. They got some site-specific cost figures from the utility in order to conduct that cost analysis. For the three coal plants as a whole, EPA used some of the site-specific figures, and they didn't use others because they were the type of cost disallowed by the cost manual. But I'm not sure on the record whether or not they used those disallowed figures for the Coronado determination in particular. But did the – was the – did the proposed rule point out this problem with the cost – the inadequacy of the cost figures for the SIP? Yes. And that's in the Joint Appendix at pages 19 and 20, Your Honor. So in other words, they were on notice and they didn't bring it forward at that point. So Arizona was on notice of the problem. Yeah. It was in the proposed rule. Another question here is about the whole sense of overnight costing and why it might be required or not required in different instances. Do you do a better job of explaining it? I don't know if I can or not, but I think just quickly one example might help. EPA has long used the cost manual and its overnight methodology to calculate cost-effectiveness figures for sources under the Clean Air Act. This is not a new application to the regional HAZE program, first of all. And EPA has insisted on a consistent and equitable cost methodology, which is the overnight methodology, in order to provide an apples-to-apples comparison. And I think the example is if you have two sources that are identical power plants, for example, that are both looking at the same pollution controls for a cost analysis, well, then those cost figures should be, if not the exact same, then very close to one another in order to provide consistent treatment. But if you start allowing sources to consider financing costs, then, say, poorly financed utilities. You're basically saying how you pay for this is your problem. Yeah. Poorly financed utilities might have a higher cost of borrowing, or some utilities might have certain revenue requirements or costs that don't apply to other types of sources. And if you allow those financing considerations to be. . . But those are still real-world costs. They are. Somebody's going to pay for this, and somebody's responsible for it. It's going to get recouped from the consumers at some point. So the fact that one might have a slightly different financial profile from another plant, one in New Mexico might have a slightly different profile from one in Arizona, doesn't mean that they're not real-world costs. It's still a cost. Is that not supposed to be accounted for here? You're right. It is a real-world cost. I don't think any of the parties here dispute that. If it's a real-world cost, even if there's a difference between, let's say, twin plants being constructed in two different states, surely there's a way of discounting and saying, well, you've accepted some additional risk, and that's going to cost you. We're only going to give you up to the amount that's done by your twin plant over here on the other side of the border. Well, EPA has explained that if you start allowing those considerations as costs, it results in inequitable and inconsistent determinations because sources that have a higher cost of financing might find that the exact same set of pollution control is not cost-effective. Well, the same source- Part of the problem that then you'd have to deal with the fact that some people, some of the plants aren't going to borrow money. They're going to actually have the money. So then they want credit for the, essentially, cost of not using the money for something else. So then you start having to elaborate. I mean, that's what I understood part of the problem was, is that you'd have a sort of non-real world equity attribution, i.e., what you could have done with the money if you hadn't done it for this and used it for this. Yeah, I think- That seemed to be part of their concern, is that you could look at financing in two different ways. I think that's correct, Your Honor. I think it just goes to show why EPA has said we just don't look at a financing cost. We look at the overnight costing methodology. That's how we get an equitable apples-to-apples comparison of what types of pollution controls are or are not cost effective. Now, petitioners have claimed in this case that EPA has usurped the state's authority to make BART determinations. But Congress gave EPA, and not the states, the final say on whether or not state BART determinations and state HAYS plans comply with the Clean Air Act. Section 110 of the Act gives EPA broad oversight authority to review state plans and to ensure they fully comply with the Act's requirements before EPA approves them and they become federally enforceable law. And there's nothing different or unique about the regional HAYS program in this regard, as both the Eighth Circuit and the Tenth Circuit courts have recognized in the North Dakota and the Oklahoma cases. EPA has the authority and the obligation to disapprove inadequate state plans. And Congress gave EPA this final say in order to make sure that recalcitrant states don't undermine the purposes of the Clean Air Act through their state plans. And that is exactly what would have occurred here had EPA not disapproved the state plan, where Arizona selected the worst pollution controls as the best available on a coal plant that is one of the very worst sources of visibility impairment in the nation, according to the National Park Service. For those reasons, EPA's actions were proper, and the petitions for review should be denied. Let's see. I have 50 seconds if you have any questions. Thank you. You're back. You have 44 seconds. Thank you very much. You have two minutes. I don't get to use his 50 seconds, too? Let me just quickly address cost issues and the adequacy of Arizona's explanation. It's news to us that overnight costing was not a reason to deny the Arizona bar determination. Let me refer the Court to 77 Federal Register at 42-841. This is where EPA talks 77 Federal Register 42-841. This is the proposed rule. This is where EPA talks about generically its reasons for disagreeing with Arizona on the cost analysis. It's generic. Yes, it is generic. And I will give you that when we get around to the specific Coronado cost discussion, they only refer to the adequacy of the information. Right. So that's fine. That's fine. Of course, when we get around to the FIP, when they impose the FIP, here comes these costs that they say that they would have excluded, and we thought that they But at that point, they have a lot more authority to do it their way, don't they? Well, it is a different standard. But the point I want to make about the adequacy of the cost information, if we just focus in on that, you have to understand here that the cost information came from the vendor that does all the pollution control work at the station. There seems to be some implication that maybe it was puffed, maybe it can't be adequately examined. But you also have to understand that by the time that information was provided to Arizona and EPA evaluated, work was well underway pursuant to the consent decree on an SCR for Unit 2. So what the State knows when the State makes its determination is that the vendor, which is going to be doing the work and has now begun to do the work on an SCR on Unit 2, is now telling us what the costs are. That seems to us to surpass the indicia of reliability and ought to be a reasonable basis for approving the plan. But how does a vendor know what the financing costs are? The vendors – well, this is what a vendor does. I mean, a vendor provides an overall bid. The vendor is providing the service to the utility. To sell them something. Yeah, that's correct. And they're going to borrow money and go borrow it. Yes, that's correct. One other point I'd just like to make. I don't understand it. No. The vendor's quote provides, as a part of capital and O&M, provides a total quote for what the – Because it's not a real number. It's an attributed number, right? Which would account for – which is whether they actually go out and borrow money or whether they use their own money. No. The vendor provides – Is the vendor financing this for them? The vendor is not financing this for them. A good explanation of AFUDC and why it is a real cost is supplied in the Arizona Corporation Commission amicus brief. I would refer you to that. These are real costs. This is not methodology. It is a methodology unique to the utility industry. But it is a methodology that is used to account for real costs during financing. Okay. Thank you very much. Thank you very much. Thank all of you for making a real effort to help us with a very complicated case. And the case of Arizona v. EPA is submitted. And we will now go on to Stage 2, Phoenix Cement v. EPA.