Oral argument — Pacific Gas and Electric Company v. FERC

Case number 23-1041 et al. Pacific Gas and Electric Company Petitioner v. Federal Energy Regulatory Commission. Ms. Goldenberg for the petitioner, Mr. Ettinger for the respondent, Mr. Bain for the intervener. Good morning, Ms. Goldenberg. Good morning, Your Honor. May it please the Court. The class-based approach that FERC adopted violates the plain text of 824KH2, which is incorporated into the tariff. That approach would vastly expand the grandfathering exception to the ban on mandatory retail wheeling, allowing, as the Commission has previously said, grandfathering covering every consumer in San Francisco based on some kind of unspecified resemblance to a consumer that San Francisco served in October 1992. There is no textual My opponents have not identified one, and that's because the statute refers to a discreet, particularized consumer provided service on one particular date. Before we get into the merits, can you help me understand what injury you currently, your clients currently, are suffering because of the tariff? Because the tariff is no longer in existence. I mean, it's expired, right? The particular tariff at issue here, yes, there's another tariff, successor tariff, that's come into play. And so the reason that this is very much a live dispute is that, as to that other tariff, there are still legal disputes outstanding that the parties are working through. And in particular, the parties stipulated in the FERC docket with respect to that other tariff in August 2022, they made a joint motion to stipulate that the commission granted saying that what was decided in this case would determine whether it was just and reasonable for PG&E to remove from that next generation tariff, the language that's at issue here about grandfathering referring to the eligibility criteria statute. So that's one of the reasons that this case is very much alive. And then I would just add that the court rejected a moot disargument in its last decision in this case in 2022. One of the things it said is that for prior tariff violations. So that would be another reason why this is very much a live dispute here, even though this particular tariff isn't currently in effect. What is the redress? I mean, if we were to vacate the orders that are before us today, how would that redress PG&E's injury? Well, the PG&E made a compliance filing before the commission that changed the way that it was treating various delivery points based on the commission's order in this case. And that compliance filing is something that the commission hasn't acted on yet. And so if this court were to rule for PG&E, that compliance filing would go away and PG&E could return to the way that it was treating delivery points before FERC's order in this case. And in addition, as I say, in the next level, the next generation tariff, the dispute would be different. And would PG&E be entitled to if this court were to agree with PG&E it wasn't required to provide service? PG&E hasn't asked for compensation, but again, a change in the way that PG&E was treating these delivery points has occurred as a result of the commission's order. And when this court rejected the mootness argument in the last appeal, one of the things it looked to was not only the effects with respect to this particular tariff, but also the fact that the next tariff is not yet fully in effect because there are legal disputes between the parties that haven't been resolved. And the same thing is true here. Those legal disputes remain. So in that sense, the next tariff isn't fully in place. It's not fully accepted by the commission because there are these legal disputes between the parties, one of which is the question of whether this particular language that's in dispute here is something that can or cannot be in the next tariff. So the... Sorry. Can we have jurisdiction simply because FERC indicates that it's waiting for our decision? Does that make something a live case or controversy? I don't think that alone would make something a live case or controversy, but that's not at all the only thing that's going on here. There'll be real effects in the world with respect to this tariff, which is still a tariff before the FERC fully accepts the next tariff, while there are still legal disputes ongoing with respect to the next tariff. And there would also be effects with respect to that next tariff because of the stipulation that I mentioned. And in addition to all of that, and this was something else that this court relied on in rejecting mootness in its last decision, FERC could order equitable relief of some kind for prior tariff violations. So there's an effect in the world for that reason as well. And San Francisco in its complaint here did ask for equitable monetary relief. So for all of those reasons, I think the same reasoning that the court gave when it rejected mootness last time around applies here as well. The situation with respect to the next tariff really hasn't changed significantly since this court's prior decision in 2022. So I'd like to start with the text if I could, and then I would like to spend a little bit of time talking about our arbitrariness and capriciousness argument, and then also spend a little bit of time talking about what's really at stake here in this case and what's not at stake in terms of what the effects would be in the world of this court's ruling. So with respect to the text, the word class and its equivalent doesn't appear anywhere in this provision. Let me interrupt you for a second. I'm assuming that everyone has triple checked this, but there is no definition of consumer in this text. Correct. What do you think consumer means? Rather than debate about what class means, tell me what you think consumer means. Consumer is the end user of the power. The ultimate consumer is the end user of the power. So not somebody who's going to transfer the power, sell the power. There's a person who's actually consuming or using the power, and I don't think there's any dispute in this case. There has never been any dispute between the parties. FERC has never said otherwise. That's everyone's understanding of what an ultimate consumer is, the consumer of the power. And you can see from the statutory provision that that's what Congress is getting at here because the provision requires for the grandfathering to apply, the ultimate consumer to have been provided power on a particular date, October 24, 1992, and to be the consumer of that power, to be the ultimate consumer of the power. So the statute has, it says, an ultimate consumer. So read literally, that would be one person. And so once we all agree that we're not talking about one person, we're talking about classes of consumers, whether it's the class that you propose or the class that the other side proposes. I think there's a big difference between saying that the term an ultimate consumer can apply to many different ultimate consumers and saying that we're talking about a class or category that's defined in some way by characteristics, which is what the agency has done here. But the definition is also a class of consumers, the ones who were being served at that time. Not as classes being used by the agency here. I think this is a really important clarification. Class is sometimes used to mean a defined group of people who experience something. If you were provided power on October 24, 1992, if you were doing a class action, you would use class that way. What the agency has said here is something extremely different, which is that class is a category, a broad category of consumers defined by some characteristic into which people may or may not fall. But the characteristic is not the statutory. I understand what the dispute is about that, but your textual argument, it seems to me, is that you must win because the statute supports your view. But once we get away, once we acknowledge and understand that the statute has to be talking about some class, whether it's the class they say or the class you say, your textual argument is greatly weakened. I don't think so, Your Honor. What we're saying is that an ultimate consumer is singular, that Anne, the Supreme Court has told us this, refers to an unspecified particular or discrete countable thing. And yes, when you look at the Dictionary Act, you can see this as well. When a statute says singular noun, it can refer to a lot of different people or things. And you can also see this from the Supreme Court's decision in Mi Chavez, which we've cited in our brief, where it was a notice to appear. And what the court said is that is a singular thing. A notice to appear can go out to many different people. So if we think the statute just means ultimate consumers, because we're talking not about one person, how does that affect your argument? So I think, again, our view is that an ultimate consumer who meets the statutory criteria, who was provided power on October 24th, 1992. But my question is, if we read the statute as an ultimate consumer, it's just a way of saying ultimate consumers, which I think is a very logical way of looking at this. How does that affect your argument? I don't think it affects it very much in the sense that it's still very different than the class-based approach, because if you're talking about multiple ultimate consumers, you can look to see, was this ultimate consumer in front of me, somebody who was provided power by San Francisco on October 24th, 1992, and was this ultimate consumer, someone who was provided power on October 24th, 1992? And everyone, you can think of that as a class if you like, but everyone has to qualify. I just don't see why we necessarily should. I just don't see what supports your argument. Well, again, I think the way that the statute is written, it's very clear that it says that the consumer, the ultimate consumer and ultimate consumer, as to whom the wheeling of power is going to go, right? The power is going to go to that ultimate consumer. As to that ultimate consumer, such ultimate consumer, there has to be a showing that such entity, which is here, San Francisco, was providing the service on that date. So you have to map them up. In your view, there's that sort of hypothetical from Suffolk County, where, you know, two neighbors are receiving service, but on the day in question, the power was out for one of them. They would not qualify for grandfathering, in your view. I think that particular hypothetical, there would be a question of what it means to provide service, actually. That's not an issue in this case. I don't think that the statute would be too brief to argue that. You could argue, I suppose. And you would say that that person could get service as a matter of being an ultimate consumer, even if they weren't receiving service that day. I think if that person was a customer of San Francisco, and they were supposed to be being provided service, and it's just that a tree came down on the line and interfered with the flow of electrons that day, you could construct an argument, and I'm not endorsing it or not. You're saying yes, they would be entitled to power that day. I think there would be an argument that they could, but that is so far away from what the agency has done here in saying that you can have these broad, amorphous categories of consumers, and that you have to look at the people who are being provided service in 1992, and they agree, you have to look at the specific people being provided service in 1992, and try to construct from that the characteristics that would define this broad, amorphous class. But you think that there needs to be an analysis of every person who is being served in 1992, and then analysis of every person who's being served now, and a matching to see if they match, and then those people be provided service. You think that's what's required here? I think that you need to ascertain whether the ultimate consumer in question was provided service, and that is not hard to do. In the record, in this case, the list of... Is that a yes to my question? I'm asking you how would this work. Is that what they're supposed to do? Yes, but in the record, in this case, there is a list of customers from San Francisco from October 1992. It's not particularly long. And I just want to understand, what are you relying on to argue that that's what was intended or what Congress wanted to have happen here? Well, we're relying on a number of things, a number of aspects of the text that I've talked about already. We're relying on the fact that this is a grandfathering provision that is supposed to agency did. We're relying on the underlying... I'm interested in your winnowing argument, too. Why do you think this is supposed to winnow? Because we have this Raker Act that says that San Francisco is supposed to compete with PG&E, and it's supposed to provide electricity cheaply to San Francisco residents. So the Raker Act is absolutely irrelevant here. It does not say that San Francisco is supposed to compete with PG&E. It gives San Francisco rights of way to PG&E. But it doesn't say anything about wheeling of power. It doesn't contradict Section 824. No, but it doesn't suspect your winnowing argument. It seems to be to contradict your winnowing argument. I don't think so, Your Honor. That's a separate statute that seems to contemplate that San Francisco would have the ability to serve consumers. And it seems that the endgame of what you're arguing is that San Francisco would have to stop doing that at some point. Not at all. PG&E would take over their customers. Not at all. That is not at all what's at stake in this case. And I'd really like to explain this, because I think it's critically important. Without this broad-based, class-based approach to grandfathering, San Francisco will have ample opportunity to serve consumers in lots of different ways. It will continue to generate power in the Hetch Hetchy Valley, hydroelectric power, just as it does now. And it will get that power to San Francisco consumers. And there are several different ways that that can happen, all of which would mean that the power would flow essentially as it does now. So it's not like new facilities or anything that would necessarily need to be constructed, certainly not by PG&E. So one of them is the Community Choice Aggregation Program under California law, which gets power from municipalities to consumers using the utilities facilities. So the power flows as it does now. San Francisco would generate the power, and the customer would pay San Francisco for the generation and pay PG&E for the services it's providing and getting that power to the consumer. It's used all over California by many, many municipalities, and San Francisco uses it as well. It is intended to get power generated or purchased by municipalities to consumers. Another thing that San Francisco could do would be to put in intervening facilities between PG&E's distribution facilities and the end user. I think San Francisco has argued in the past that that would somehow require duplication of PG&E systems. It would not. It doesn't require them to duplicate every PG&E has. All it requires them to do is put some equipment in between PG&E's distribution system and the ultimate consumer, like a protective device or a disconnect switch. And again, many municipalities have done this. There are four cases about it, and it is not this massive, difficult undertaking. Another thing is that PG&E has made a concession in this case previously, and it's talked about a little bit in this court's prior decision, that grandfathering under certain statute. And that's PG&E trying to make sure that this is all working well. Why is it that PG&E doesn't want to wield this power? Well, what San Francisco is trying to do here is create an unfair playing field. It wants to expand the grandfathering exception so they can cover every customer in San Francisco, and then so that San Francisco can use the wholesale distribution tariff to get the power that's PG&E's own facilities to get the power to the consumers. And then San Francisco engages in rate arbitrage because it pays the wholesale rate to PG&E, and then it sets the retail rate for customers slightly under PG&E's retail rate. And the difference between the wholesale rate and the retail rate is free money to San Francisco. So the effect, the combination of those two things, the vast expansion of grandfathering under a class-based approach and this rate arbitrage that San Francisco can engage in, means that every single customer in San Francisco could become a San Francisco customer, even though it is PG&E's own facilities that are being used. That's synthetic competition. It's not real competition. It is not San Francisco making things more efficient, more cost-effective, building anything, adding anything, transforming anything. And that is what PG&E feels is a very unfair playing field that would be created by this decision. And that, I think, is the practical effect of the decision. But I don't want to get too far away from the text. I'm interested in understanding how you think our previous decision in city and county of San Francisco constrains our decision here. Because, of course, the previous panel didn't say what the statute means, and it asked FERC to do that statutory analysis. But there is, in that opinion, language about the importance of fairness and competition in this context. And, you know, the panel admonished FERC to think about its Suffolk County precedent. And, arguably, FERC kind of picked up on both the holding and maybe some of the dicta of that panel opinion in issuing the order that it did. I think it's right that the panel certainly asked FERC to take a look at the Suffolk County orders and take a look at the statute in the first instance, because the agency had not done that at that point. So that is certainly true. That's what the panel instructed. But I think that's as far as the panel's instruction went. There is that discussion at the end about fairness. And I think I've just explained why we feel like it is actually quite anti-competitive, what has happened here, and that the Raker Act doesn't really speak to this and doesn't suggest that there should be some kind of mandatory wheeling of power in any way, shape, or form. And I think there's also a lot of language in this court's prior decision that we think shows that when the panel, which was not interpreting the statute, and I understand that last time, but when the panel just looked at the face of the statute and looked at its plain text, the panel was reading it the way that we read it. And in fact, that's actually doing some work in the decision. So the place that I would point to most notably is page 664 of the 2022 opinion, where the court said any end-use customers San Francisco did not serve in 1992 could not qualify. And the work that that's doing was the court was using that to reject the argument that San Francisco did not serve in 1992. And I identified in our brief a couple of other instances of language like that, that I think just helps emphasize that when you just read the statute, it's plain language, shows that there's nothing about a class in there. There's nothing about a category type approach in there. It's just not discussed at all. And in addition... But they've moved across town. Are they grandfathered in their year interpretation of the statute? So that's something that the agency did not address in the orders here, and I think it's not really presented here yet. I think that would be an issue on remands for the agency. We're trying to come up with a construction and one that is sensible and reasonable. Understood, Your Honor. I just wanted to flag that that's not something that the agency has decided, and they didn't decide it because they thought the class-based approach precluded any kind of location limitation. But our position about that, I'm happy to tell you, is first of all, you have to look at the point of delivery language in the tariff in conjunction with the tariff's reference to the statute here. So the tariff talks about points of delivery qualifying for grandfathering under the statutory criteria, and a point of delivery is a place. It's a physical location in the world. And so we think that when you look at tariff language and you look at it in conjunction with the statutory language, that yes, there's the identity limitation, right? It has to be the same ultimate consumer that was getting the power in October 1992 that this case is about. We do also think that there is a locational limitation, so it has to be at the same place, the same point of delivery. But again, I think that's something that the party would... But your point is that the locational limitation comes from the tariff, not from the statute. Well, I think you can find it in ultimate consumer as well, actually. So it's not necessarily just from the tariff, but I think the tariff... How do you find it from ultimate consumer? Well, the statute talks about somebody consuming power and someone who consumes power is consuming it at a particular location where a meter is running and recording what's happening. And the commission acknowledged, as is at J56, that the utility industry commonly links customers to their delivery points. And then I think if you look at the other exception in the statute, not the grandfather exception, but the intervening facilities exception, which talks about San Francisco putting in the disconnect switch and the protective device in between the ultimate consumer and PG&E facilities, that that is plainly location-specific, actually. We're talking about technology, the physical things in the world that you can't just pick up and move around. So all of that, I would argue, suggests that ultimate consumer has this location aspect, but I do think that the tariff language puts the point on it, that it's points of delivery that have to qualify. But again, for purposes of this case, we're not talking about that. We're talking about the identity issues. And one thing that hasn't come up yet... Well, and so can a successor to the consumer as of October 1992 be an ultimate consumer? So somebody who moves into a place where somebody who was being provided service in October 1992 used to live? Is that kind of... A successor as far as, I guess, in the contract sense. The successor or a sign, someone who takes over the contract, so to speak. No, I don't think so. I think it has to be the same person or entity that was consuming the power in October 1992. But I think these questions about what exactly is an ultimate consumer and how do you decide exactly what an ultimate consumer is, there are some questions along these lines that are going to be questions for the agency sort of around the edges. But the bottom line sort of administrability question that we were talking about earlier, I think, is pretty straightforward because there is this list of customers. The list itself is not in the record, but it is discussed in the record at JA5... I'm sorry, it's not in the appendix. It is in the record, but it's discussed in the appendix at JA519. And so that's going to go a very long way to telling you who the ultimate consumer is, and who the ultimate consumer is. And so, in terms of the ultimate consumer, on October 24th, 1992, I also think that the other side's administrability problems are far greater. They too will have to identify ultimate consumers to some degree to figure out whether they slot into these classes that the agency has said should happen. But beyond that, trying to figure out what the class is is basically an impossible task. And that's because the word class does not appear in the statute. Congress has provided no guidance for how broadly or narrowly a class should be defined. And there's just no way to tell. And to sort of amplify on an example in our brief, if I may, on this point, imagine that the only consumer that San Francisco was serving on October 24th, 1992, was a seafood restaurant in San Francisco. And now you have to construct the class of people and entities that are like that seafood restaurant from that single consumer. But why is it that you think that they want to expand? It seems to me that they just want the status quo. They've been serving these consumers for all this time, for decades, and they just want to continue doing so the same way they were before. Why do you think this is an expansion? So, if we're looking all the way back to October 24th, 1992, the group of consumers they were serving was much, much, much smaller. And so, I don't think it's true that it's just sort of freezing in time as of the date that the statute specifies. I'm sorry, I didn't quite catch that. It seems they want to maintain the status quo of the classes that they've been serving. Right. And I don't see, and I'd like you to address this, because we didn't get to it before, why you think this is supposed to winnow? Because that doesn't seem to be supported by anything that I've seen. Right. So, let me just start with the expansion. If you have a class that's something like every entity that has operated in coordination with San Francisco has said that the class definitions the agency chose would describe, then they can keep adding and adding and adding and adding people and saying, well, now this person is operating in coordination with us, this person is. And so, the group of consumers that's being served by them. There's some connection to the municipal load or whatever. It's connected to San Francisco. Well, not necessarily, actually. I think it's actually quite unclear. So, what the categories in this 1987 agreement, which is obviously entered into before the statute was enacted. And one of the classes is this municipal class, which they have defined very, very broadly. But there's also something called non-firm Hetch Hetchy resources, which has categories in it like airport and river bank. And so, it's not just that it's purely this municipal category. And this is part of the problem with the class-based approach. It's really, really impossible to tell what are the classes, what are their boundaries, what are the categories, and how should they be defined. So, to get to your question about the winnowing, if I could, because I feel like that's really important here. A grandfathering provision by its nature, and we cited cases to this effect, it has a temporal limitation in it. And then people who are doing something or whatever the criteria in the statute is, after that date are included. And so, by its nature, the group of people who are grandfathered gets smaller over time. People die, entities close down. There's entropy in the world. So, your argument is that all grandfathering exceptions are intended to winnow. All of them are. I think that is the basic nature of a grandfathering provision. Could someone write a grandfathering provision in a different way, I suppose? That's the question, whether they did that here. And it seems that when they wrote this provision, they were trying to maintain the status quo. I don't see anything about winnowing. They're trying to just keep things the way they were. I think they were trying to keep things the way they were with respect to particular ultimate consumers who were being provided service in 1992. That's the question, isn't it? Well, I mean, I think we've discussed in other ways the reasons why we think that's correct. But with respect to your winnowing question, if you understand the statute that way, then yes, the people are going to die, they're going to move away, and the group's going to get smaller. And that's going to put the policy, the new policy that Congress enacted in this provision, more firmly into place. Is winnowing essential to your statutory argument? I mean, winnowing may occur, or it may not occur. But is the fact that winnowing may occur, or in your view, maybe should occur, I mean, is that necessary to rejecting a class-based interpretation? No, it's not necessary. I do think it is another reason to reject the class-based interpretation besides the plain text of the statute, which, again, does not say class or anything like class. And one thing that we haven't talked about yet that, to me, is very dispositive, actually, of the statutory issue, so I do want to make sure to discuss it, is the fact that there are other provisions, including another provision from the same 1992 Act that put this provision into place where Congress did talk about classes of ultimate consumers. So the main example we gave in our brief is Section 111E of the 1992 Act, which says ultimate consumer or group or class, sorry, ultimate consumers or groups or classes of such consumers. And I think that tells you two different things. One is that in 1992, when Congress enacted the provision that's at issue in this case, Congress knew how to talk about classes of ultimate consumers. It didn't do it in the statute. And the other thing I think it tells you is that ultimate consumer and class of ultimate consumer can't be the same thing because if they had been, then Congress wouldn't have said ultimate consumers or classes of ultimate consumers because that latter phrase would have been superfluous. So classes of ultimate consumers is something different. And the other thing that's really actually interesting about that other statutory provision is for that part of the statute, there is a definition of class, actually. It's in 16 U.S.C. 2602. Class is defined to mean groups of ultimate consumers with similar energy usage characteristics. Of course, that definition doesn't apply here because the word class is in our provision. It doesn't appear anywhere in our provision. And in addition, that's not the way that the agency has defined class here. The agency has defined class in this vague, amorphous way to say, well, we're going to look at the people who are provided power and we're going to pick out some characteristics having to do with them, and now we're going to somehow know what the class is. There is a massive level of generality problem with respect to that. There's no way to know how broadly or how narrowly classes should be defined. Can I ask you about that, though? What's been happening since 1992 until now when they decided to go with the tariff instead of whatever agreements? Like, how are you defining what would be covered up until this dispute arose? Well, so the class-based approach has never been in effect until the agency- No, I'm talking about what has practically been happening between San Francisco and PG&E because there has been wheeling. They weren't relying on the tariff, but how are they defining who could get wheeling? Right. So this dispute has been going on since 2013, I will say, the dispute in this case. I think I described before the accommodation that PG&E offered in this case, which was to extend the grandfathering exception in certain circumstances to mid-2015, way past 1992, the date in the statute, as an effort to resolve the dispute between the parties. And so the wheeling has happened under that concession for the vast bulk of the time that we've been talking about. I understand how you're defining the classes, who's getting the wheeling. Right. And so the people who are getting service in October 1992, San Francisco had disputed that. And now under this compliance filing, San Francisco has prevailed. But this has been a dispute between the parties back and forth for this entire period of time. So if you had a home in October of 1992 that was getting service and it was owned by, let's say, husband and wife, and they both passed on and left the home to their kids and their kids now occupy it today. Are the kids grandfathered or are they different ultimate consumers? So again, I think the agency can resolve some of these questions about exactly ultimate consumer supply, but I can give you my answer as I stand here today. And maybe the agency will think something different, which is, no, they're not ultimate consumers who were provided power by San Francisco on October 24th, 1992. They weren't provided power at that location. They weren't consuming the power. They may not have used it. They lived at the house in October of 1992. They've never left the house. Oh, I see, Your Honor. I'm sorry. I didn't understand the question. Right. So in that circumstance too, I think the agency will have to address this. But in our view, if somebody was in the residence, say, on October 24th, 1992 and was consuming the power, but their name just isn't on the bill, that could be as simple as having someone just check a box saying, I lived there then, I consumed the power, even though my name isn't on the bill, so I'm not on your list of customers. I'd also point out that with respect to San Francisco in particular, there were very, very vanishingly few, I think, residential customers that they were serving on October 24th, 1992. So this issue is not going to come up in this particular case in any meaningful way. But I point out again here that the administrability problems that the other side has, it's worse. Any administrability issues around the edges here that the agency can deal with, with respect to understanding ultimate consumer to mean a specific, particularized, ultimate consumer, and that's- Take your statutory argument to have a hypothetical. Is that correct? Yes, I think that's right because there are some ways that are not an issue with respect to whether the class-based approach is correct, but the agency could address, well, what is an ultimate consumer exactly in some of these circumstances? And I think that is the kind of job that's perfect for the agency, right? That's what agencies are supposed to do. But none of that suggests that you can take the statute that says, an ultimate consumer, such ultimate consumer, ultimate consumer provided with power on a particular date, which is a noble thing in the world, and transform that into a statute that means anybody who's kind of like that ultimate consumer, even if the person you're including in that class did not exist on October 24th, 1992, was not alive on that date, was not an entity that anyone had created on that date, was living in New York or London on that date. There's no way to read this provision to encompass those folks, and yet that's exactly what the class-based approach does. It just cannot be correct. And so what we're asking the court to do here is to vacate these orders, to reject the class-based approach as an impermissible construction of the statute, and to send it back to the agency for further proceedings, which may include some discussion along the lines of what we've been talking about here today. All right. Let's hear from the other side. Thank you. Thank you very much. Good morning, and may it please the court. I'm Scott Edgar for the Federal Energy Regulatory Commission. Thank you for hearing this case this morning. If I may jump in directly, Judge Pan, your questions about the naturally winnow over time, I believe that was in reply brief, page 12. I too was struck by that language, and I think this court will have a difficult time reconciling that statement in the reply brief with some of the legislative history in this case. And at footnote 124 of the rehearing order, the commission discussed the legislative history, including statements by Senator Wallop. And if you look at that page of the discussion of Senator Wallop in the congressional record, he specifically refers to the idea that the purpose of this legislation, it's the next to last sentence of this discussion on this section, and he says, by the same token, legitimate existing municipal wholesale sellers may apply and obtain wheeling orders that lowers the rates for their retail customers. That is something that there's evidence in this record that Congress wanted to preserve. If you look at page 12 of the reply brief, this naturally winnow over time, I think, is a... Let's deal with the text. Yes. Not what one senator said. The text says over and over again, directly to an ultimate consumer, and then even narrows it or specified a date. And it says providing service to such ultimate consumer on October 24th, 1992. So how on earth do we get from that providing service on a specific date in 1992 to class that includes entities that didn't even exist or that are successors to, or alike, or akin to, somehow, ultimate consumers that were getting service on that date? Thank you, Your Honor. I think the only... The way the commission resolved this, and this comes from paragraph 53 of the re-hearing order at JA-50, is to... The commission acknowledges the singular use of consumer in the section. However, the commission found that that cannot refer to individual. And I think there's support for that finding. And if we look at the date, what is provided to such ultimate consumer on a specific date? I'd assert, Your Honor, that the date doesn't preclude a collective use of the word consumer. And I think there's support for that. If we look at Black's Law Dictionary, 6th edition, which was the one in existence 1990, the latest version previous to the passage of the singular, the second and third both use a plural word to define consumer. What the commission was doing here was using consumer in a collective sense. And I don't... And to get to your specific question about the date, I don't think that precludes that understanding of the statute. And we can talk about some of the legislative history and what to the section is the idea that Congress was trying to prevent a sham wholesale transaction. And if we look at the statements of Senators Johnston and Wallop, the sham transaction comes up when it's a retail consumer that seeks to circumvent the ban on retail wheeling. Here is the very opposite of that. We have San Francisco, which has been a wholesale player in this market, going way back to the Raker Act of 1913. It seems like FERC's primary arguments are about some sort of purpose or policy or stepping back away from the text of the statute. So what is your best argument from the text of the statute that class of consumer, that an ultimate consumer can include a generalized class of consumers? I would refer back to what I said about consumer, that it can be used in a collective sense. And we find that from black slave issue. It can be, but was it in this statute? I think so. In this context and getting back to... But how do we know that from the text? There's nothing that precludes that. And I would assert that this isn't precisely answered in the text. And that's why... We're seeking some kind of deference. We are seeking deference and... What if deference vanishes in a few months? That is out of my control. And that's up to the court. Before we get to deference, don't we have to exhaust the tools of statutory construction? The words themselves are not clear. We still have to look at context and purpose and all the other tools of statutory construction. I would agree with that. I would agree with that. And I can make a case that the text doesn't answer the question, but there is support in the legislative history. And there is support in the notion of grandfathering that the commission discussed in these rehearing orders. So I do agree that you have to exhaust those tools. And when you exhaust them, then your claim is for deference. And if there's... I don't concede that after exhausting them that PG&E would win this interpretation. But if you do exhaust them and you find that there's a mixed bag, I think the current state of the law under Chevron is that the commission would get deference. But I'm willing to make that case that the plain text, as I said, does not preclude the interpretation that the commission gave. And I'll go back to what I was talking about, Lex. But that argument that it doesn't preclude is an argument that the text is ambiguous. It's not that it's the best reading of the text. You're saying it's a permissible reading. It is a permissible reading. And I think it's also the best reading that the commission gave after looking at... And we can look at previous orders. The Palm Springs order, for example, that discussed Senator Wallop's discussion of what really the commission should do here is to look at substance over form. And I think all these things together, the commission has given the best reading to the statute. The fact that elsewhere in the statute, it talked about ultimate consumers and classes of consumers. Right. I know that's not about this particular provision, but there is some indication. And then there's a specialized, as my colleague said, a definition of what class means in that context. Right. Gentile, you're referring to section, the EPAC 1992. That is a comprehensive bill that has, I believe, 30 titles to it. It affects virtually all aspects of the American energy economy. Section 111, that has the contrasting languages in title one, and that is a completely different context than section 722, which is the section that amends the Federal Power Act, something completely different. Getting back to 111, that's directing the Secretary of Energy to do something, i.e. write a report about integrated resource planning and how the states are affecting that. And I would assert that's a completely different context. And I would cite also this court's discussion in Clinchfield Cole at 895F2nd773 at page 779. And there, the court deals with this expressio unaeus, the expression of one implies the exclusion of the other, and suggests that that has a weak, that has a, well, it ignores other plausible explanations. Congress just might not have been thinking about this precise issue. It's not ironclad. It's just some evidence of context, maybe. And I would add also, it is evidence, but I think it's weak evidence here. And I think this commission, this court's precedent bears that out. And in particular, Clinchfield Cole also particularly calls out the Chevron context and that the statutory interpretation tool is particularly inapplicable in this context. Can you address, I'm still trying to understand, in a sense, why we're here. Why, from your end game, to winnow this down until they can take over all the end distributors? Or I just don't understand why this is such a, do you have a sense of what the... I believe you're right, Your Honor. And I take the words at page 12 of the reply brief at face value. I think that's exactly what... I'm sorry, what were those words? And actually winnow over time. That's exactly what PG&E would like to accomplish here. That would drive, would that drive San Francisco out of the market? It would, presumably PG&E would prefer to serve these customers through their retail tariff under the, which would have rates presumably set by the CPUC. Because San Francisco, in order to continue serving their customers, if they can't get willing, they would have to build distribution networks, which they're not going to do. That's right. My sense is that would be very expensive and unnecessary. And if I want to make one observation, that the question here isn't the idea that San Francisco is getting free service, it's just service under the tariff. There's a rate they're paying. If PG&E has an assertion that this rate is not recovering the cost for the service, that's a section 305 question that they can bring to the commission. And I don't know how the commission would address it. And we'll take up those issues. So this isn't a question of... Are any rate proceedings ongoing right now about this, sir? Yes, there are. This is a continually ongoing dispute. And I did want to refer to, this came up earlier, the commission did point out in the remand order, paragraph 31, note 66, J17, that this case only applies up until the filing of the second rate case afterwards. And that's a J17. So the power at issue that is being generated by San Francisco, is it being generated in California? Hetch Hetchy. This is Hetch Hetchy power. That's what was authorized by the Raker Act in 1913. And San Francisco not only operates that project, but has... The answer to the question is, yes, it is generated in California. Yes. So we have power that's generated in California. And then this is a dispute about the fact that, well, San Francisco has transmission lines. It doesn't have the distribution system to get it to each of these end users in San Francisco. That's right. And that's an interest state issue, right? That normally FERC wouldn't have any role, right? That is a... Well, jurisdiction hasn't been questioned here. I understand that, but shouldn't that be something that should kind of impact the way that we think about this statute in this grandfather? I mean, what does FERC even have to do with this under the way that kind of the cooperative federalism model of FERC was imagined? I mean, shouldn't that have some role in how we look at what this statute means? I'm not sure. This is a wholesale transaction that the commission has jurisdiction over. And I was trying to... On a case that discusses that. It's Detroit Edison. So even though it's all interest state because it's a wholesale transaction, wouldn't traditionally have jurisdiction. That's what you're saying. Yeah. It's a wholesale transaction for distribution service. And as this court explained, and at 334 F-38 and Detroit Edison, the commission has jurisdiction over this kind of transaction. Okay. So we shouldn't be concerned about that. But I guess what I'm still, I guess, having trouble with and would like to make sure that I give you ample opportunity to respond is how the text that says directly to such an ultimate consumer or to an ultimate consumer, and it says this in this particular subsection, I think it repeats that phrase three or four times, including in that last sentence. How do we get... How is the best natural reading of that text a class and not just a consumer who was provided energy on a particular day? I, again, I would go back to the statements about the definition of consumer and whether they can be in a collective sense. And I think, and in that sense, I would also invoke the statements of Senator Wallop and what the purpose of this provision is to allow a legitimate, not sham, wholesale purchaser of distribution to serve its retail customers. There's no question that San Francisco doesn't fall into that sort of category of, let's say, a high volume user that's upset with its electricity provider and wants to go out and to use the words of Senator Johnston, he talked about the XYZ procurement corporation. And so they create this procurement corporation that goes out and then creates a sham. That's the opposite of what San Francisco is. There's a long history of... Is there room in the text for a sort of intermediate position, which is the ultimate consumers were the actual, I guess, people and the points of delivery that were in effect on this particular date. And the consumer also being defined as if it's the police department, the police department can move. It's still the police department. I just think there is there some intermediate way that's more textually, I guess, supportable. That's not what PG&E is asking for and not what San Francisco is asking for. I'm not sure the way you've described that, that sounds like a class. If the police department is a class... But you can define police department as a consumer. Yes, I think so. And I do want to mention that every time in the brief, there's a reference to poorly defined class or something of this sort. This is their agreement. This is PG&E's agreement, the 1987 interconnection agreement. And all the commission has done was to preserve that relationship. So I would suggest that the agreement and their drafted agreement is sufficient to define. And so there's no question about it that that was the agreement that defined the relationship on October 24th, 1992. And again, they had a hand in drafting it. So it's not exactly sensible for them to say that it's a poorly defined class when it's based on their argument or their agreement. So if there is sort of an intermediate possibility that is more tied to the text, which would include consumers broadly defined and points of delivery that were being served at that time, if that were the best reading of the statute, are you saying that that is a class based approach? It sounds like a class based approach. And that's what the commission did. It said, OK, let's look at the date and the statute who was being served by the entity on that date. And there's a couple of ways to approach that. We can take the sort of particular individualized approach or we can take what the commission found was a class based approach. And the commission went through the steps of analyzing that and talked about how it would be unfair to the legislative history that is going back to this. Are you saying that regardless of what you call this, all they did was take the consumers that were being served at that time based on the 1987 contract? And that's what we're looking at here, whether whatever you call it. The commission was called on by the statute to look at ultimate consumer in October 24th, the commission believed that the best interpretation of that would be to look at the class that was served on that date. Defined by the 1987 contract. And in this case, absolutely. And in this case, that question is answered best by looking at the agreement PG&E signed and agreed to in 1987. Because that represents who was being served. Yes. I understand. Thank you. So, suppose we agree with you about the way that the statute should be interpreted. Your friend on the other side says that there's still in the terror that we are or that the commission was construing, there's still a locational restriction. Separate and apart. They argue that it's also in the statute, but let's suppose for the sake of this question, we disagree with that. What's your response to their argument that the tariff that we're construing has a locational restriction? I think, Judge Wilkins, that the commission resolved that in the very same way previous panel did. The commission took the position that this delivery point approach to the tariff that you're invoking here precludes the application of the class-based. And the previous panel had no problem with that. They said instead that all that language in the tariff means is that San Francisco will have to identify for each delivery point, yes, there is a reference to delivery point, but that doesn't preclude the ability of San Francisco to demonstrate that entitlement under the tariff by demonstrating that the customer was of the class served on October 24th, 1992. So, in that sense, there is no conflict there. It is possible to resolve that delivery point concept with the class-based approach. And I would just point out that my means it explodes. But again, I go back to the 87 interconnection agreement. And so, there were limits. It's not like the commission just said, you know, anyone and some free-wheeling sense of a class is very much tied to their agreement. So, it's not an unlimited class. So, in your view, unless PG&E has a reason to argument that the commission's construction of the 1987 agreement is arbitrary and capricious, PG&E loses. Am I oversimplifying things? I think that's right, Your Honor. I think the commission was approach here was very reasonable in defining the class. Once we get to that point and say, well, okay, what was the class on that date? Commission looked at the contractual agreement in place between the parties. And I think that's eminently reasonable approach to take. If I can just come back to this, and I'm sorry to belabor it. Are you saying, because this would address the textual argument, that the consumers that were served on the state in October 1992 are defined by the 1987 contracts? Here. In this case. And in the 1987 contract, they define the users by class? There's various definitions. There's municipal load. There's obligations to the districts. And there's a series of other things. But in a sense, that's class. In a sense? What do you mean by in a sense? Well, it is. There's our classes, I suppose. I'm not sure I follow that. I guess maybe that wouldn't solve like the larger statutory interpretation question necessarily. But it just seems to me that the text seems to talk about ultimate consumers on a particular date. And there's a debate as to whether that should be the actual consumers or the class of consumers served on that date. But in this particular case, it seems that the consumers served on that date, the specific ones, are defined by an agreement, which is the 1987 contract. And that 1987 contract defines them by class. I would say this, the statutory provision, the commission was a reasonable interpretation by determining that class was contemplated there, class of consumers. Then the next question comes under the tariff. Well, how do we implement that? What does that mean? Yeah, no, I understand that. I'm just saying this is a different route to analyze this case, which would leave open the question of whether in all cases classes can be determined. Because in this particular case, the actual consumers were defined by class, by the contract. I think that's a fair reading of the agreement, that they're in municipal load and the service to the district. I think it's fair to look at those by class, as a class. All right, thank you. Thank you. We'll hear from Councilor Brander-Wiener. May I please? My name is Jeff Bain, arguing on behalf of the City and County of San Francisco, intervener in support of FERC. I'd like to start with the text and the term ultimate consumer, but then circle back to some of your questions about the impact on the city and what's going on here. So I think there's some clarification I can provide on that. So I think a lot of the meaning of this term ultimate consumer, I take PG&E's position to be that the exclusive meaning, or at least the best meaning, is it has to be a specific, individually identified retail customer, and that it has to be at the same exact location that was served on this date. I think on its face and our friends at FERC explained, that's not the plain reading of the term ultimate consumer, but I think the other references to ultimate consumer in the statute confirm that. I think even if you look at section 111E from the Energy Policy Act, which we fully agree that's a totally irrelevant provision here, but even if you look at what that term's doing, it is not referring to specific individual retail customers. It's asking the Secretary of Energy to produce a study on how certain state policies affect the rates paid by ultimate consumers. And then it also recognizes there might be different effects on smaller, more narrow classes in that context, but ultimate consumers is a much broader term. And same as the saving clause at the end of subsection H that we're talking about here, where it says nothing under this subsection affects state law concerning transmission of energy directly to an ultimate consumer. The plain reading of that is not directly to Mr. Smith at 123 Main Street. It's to ultimate consumers distinguished from wholesale entities, but as a broad category of ultimate consumers. And the date, October 24, 1992, is doing work in this particular grandfathering provision because that's telling FERC, what type of consumer are you looking at for purpose of grandfathering? And that grandfathering provision functions to preserve or maintain legitimate wholesale wheeling arrangements as they existed on that date. And you might have different arrangements, and FERC applies this statute on a case-by-case basis. How do we understand the tariff language that talks about points of delivery, right, points of delivery for which the grandfathering criteria are met? I mean, are we here interpreting that full tariff term or just the meaning of the grandfathering term? I think the inquiry here is just on the meaning of the grandfathering term. I mean, the tariff, when we talk about point of delivery framework, that applies throughout the tariff. That simply means when someone's putting an application, they have to tell PG&E, this is the point where I'm requesting service. You have to meet various technical requirements, provide certain information for that point, and one of them is this eligibility requirement. But whatever eligibility requirement you apply at that particular point, that's independent from the fact that the tariff, right, has to identify points of delivery. I mean, you can imagine an eligibility requirement that says you're eligible only if it's an even-numbered street address. You know, putting aside whether that's just and reasonable at FERC, you could have whatever eligibility test. It just applies for whatever the particular point is in question. So, there's no inherent tension between the tariff's point of delivery framework and a class. And even PG&E's, what they call voluntary accommodation to provide non-tariff service to certainly these points, it seems, aren't tariff qualified. That's a class-based approach that it applies under the tariff with the same point of delivery framework. Can you explain why you think PG&E's not the best one? Can you say affirmatively why you think the class-based approach is the best one without any sort of Chevron deference? Yeah, absolutely. So, it comes back to that term ultimate consumer, which I think the various questions have pointed out. There's some question on how to define that. It requires FERC to draw a line. And going back to the function of statute to preserve wheeling that existed on that particular date, it asks FERC, who was the ultimate consumer? The statute was in part to limit, you know, mandatory wheeling. And then it provides an exception. So, those two things have to be understood together. It's not just the exception that we interpret. Right. Right. So, the question is, who was the ultimate consumer in 1992? And in the case of San Francisco purchasing for jurisdictional service from PG&E, it was under this 1987 agreement that didn't limit San Francisco's service to particular locations. It didn't limit San Francisco's service to specifically individually identified retail customers. That agreement allowed the city to serve its historical customer base, which is a small fraction of load in the city, and that didn't expand in this relevant period from 1992. So, you're saying that ultimate consumer means whatever the 1987 agreement meant? I think what I'm saying is you have to look at the arrangement in place at the time. So, you're going to have a different class in Suffolk County, where the county was serving all residential customers in a particular service territory. That's going to be much broader than here in San Francisco, where the city has and continues, it's about 15% of the load in the city, but served by the city instead of PG&E. So, what I refer to orders as saying, is you look at the actual customers that were served on that date in 1992 under that 1987 agreement, and as long as it's the same type of customer, there hasn't been a change in the order to classes, the same classes that FERC is relying on? So, the way I read FERC's order is FERC didn't address each individual point. So, FERC didn't rule that everything served under the 87 agreement up until 2015 is necessarily grandfathered. It asked PG&E in a compliance filing to go back, apply FERC's instructions, and classify those points. And that leaves open the possibility that PG&E could have said, there are some points, you know, these dozen points interconnected in the 2000s that are different than what was served on that date in 1992 under the 87 agreement. We think that's outside the class. Now, PG&E didn't do that in its compliance filing, and I think the reason was, is there's just been no dramatic expansion of the cities, the type of customers the cities have been serving. But I think the test FERC laid out in its decision is, look at the customers that were served on that date in 1992 under the 87 agreement, and as long as you stay within that class of customers, that type of customers, there's no change in the ultimate consumer, and grandfathering still applies under the statute. So, turning to the impacts, there are several points I'd like to hit. One is there was some mention of alternatives to how the city could fulfill its role under the Raker Act through community choice aggregation, which is a fundamentally different program that involves PG&E being the retail service provider. So, that's fundamentally different than San Francisco being the retail service provider. I think all that matters here is there's nothing in the record about this alternative of community choice aggregation. That wasn't an issue here. There's no record on why that's fundamentally different service, which it is, and I'm happy to get into. In any event, that's not a basis for how the statute should be interpreted. I think another thing that came up was, you know, what is the burden on the city? What would it involve if PG&E's view was to prevail? And so, again, PG&E's view, exact same individual retail customer at the exact same location. And for everything else, everything PG&E deemed non-tariff qualified, it's subject to these additional restrictions. Don't just apply under the tariff, but that PG&E's never sought to apply to its other wholesale customers, even those without intervening facilities. And once one of those restrictions is triggered, the city's forced into one of two options. It either turns that customer over to PG&E Retail Service, or it has to install facilities to meet the other option, ownership or control of intervening facilities. And what we're talking about in this case, these are all points that were actually served under the 1987 agreement, you know, when it transitioned to tariff service in 2015. They're already receiving power today. They're receiving it safely. There's absolutely no technical engineering reason for there to be facilities installed. But what that does do, it imposes costs on the city to install those facilities. And the cost of these unnecessary duplicative facilities raise exactly the kind of concerns this court previously identified with PG&E's anti-competitive administration of its tariff. And that's particularly a problem given the Raker Act, given Congress's intent not just to provide a source of clean power, but also to ensure competition in the power market in the city. So there isn't the equivalent of something like an open access tariff at San Francisco with access so that it would not have to... So outside of this wholesale distribution tariff, there's not another and a similar type tariff that might have different terms, conditions, rates, but that San Francisco could access to service these customers using PG&E's distribution. Right. So the options would be, this is the PG&E's wholesale open access tariff for access to its distribution system. PG&E's mentioned the CCA state law option. It's not for jurisdictional service. It involves PG&E being the retail provider. That's not a pass to the city being the retail provider. The other option is for the city to put in individual requests through section 210 to 11 of the Federal Power Act under the same statutory criteria, just on an individual basis. So instead of having this one case about the tariff, it would be individual point by point cases. And that's really, I think, why this statutory criteria is pulled into the tariff is because there's this option outside the tariff, but it would be through one-off individual applications, which obviously has efficiency, concerns on both sides, I think, on how that would be administered. But the tariff is the option for the city to be the retail service provider using PG&E's distribution facilities to reach its customers. And again, this grandfathering provision is only used for some of the city's points. The city does own, it has some transmission interconnections with PG&E that don't rely on this tariff at all. It has others that the city has intervening facilities at, often higher voltage ones that have more city facilities. And I'd also add, there hasn't been this expansion with the class of customers. From 1992 to 2015, the period we're relied on PG&E's distribution system at something like 2% annual growth. And that includes individual locations that grew, just general growth within the city, growth in electric usage. So there hasn't been some explosion of city service in that time. There's no reason to think when limited to that same class of customers, grandfathering would somehow expand. And PG&E hasn't pointed to any actual examples in the record where the city has somehow stretched the meaning of grandfathering, sought to use grandfathering to serve customers that are different than what it served in 1992. And there also hasn't been any expansion and use of this statutory grandfathering criteria since the commission first articulated its approach in the Suffolk County cases, which were in the early 2000s. So I think this concern about an expansion is contrary to the facts on the ground, and it just ignores Burke's limit that it has to be within the same type of customers. And in terms of the customers that are being served by San Francisco, can you explain that a bit? Because there was some talk about it being very unclear, like who we're talking about here. Right. So the city's description that we've used is it's city agencies, related entities, because they're long-time customers like the San Francisco Unified School District, technically a separate legal entity outside the city, tenants on city property, and some private entities providing services in coordination with the city. So it would be like a health clinic on city property. Now that doesn't mean the city can somehow use grandfathering to expand to any customer in the city. I think Burke clarified that in its rehearing order, where it talks about there was some reference in the complaint to classes of customers used for different California retail purposes, commercial buildings, industrial customers. And Burke clarified, we're not saying just because San Francisco served one commercial building in 1992, it can somehow rely on the grandfathering provision to serve all commercial buildings in the city under this clause. Because that's not within the same type of customers the city served under the 87 agreement on that date. That was limited to a much narrower, municipally related, some outside entities, but providing services in coordination with the city. And to the extent there were any disputes over particular points, whether they fit that, that would be an issue for PG&E's compliance filing, potentially a future case. But we actually haven't seen any disputes over the application of grandfathering to those kinds of entities. Thank you. Ms. Goldenberg, you were out of time, but we'll give you three minutes for rebuttal. Appreciate it, Your Honor. Just a couple of points. You can't provide service to an ultimate consumer on October 24th, 1992, if that ultimate consumer didn't exist on that date or wasn't in San Francisco on that date. But that is what the class-based approach does. It is contrary to the language of the statute. There's no broad category of consumers in the statute. It's an ultimate consumer. And so I think that's especially true because grandfathering provisions are strictly construed against the party trying to invoke them. That here is San Francisco, but it's true even without that strict construction principle. And I'd point the court in terms of grandfathering, this decision in the BP case that we cited in our brief where the court said, we have a grandfathering provision that has a temporal limitation. You can't say, well, somebody who didn't meet that temporal limitation maybe could have met that temporal limitation if they had been around at that point. And so we're going to put them in along with the other people who are grandfathered. That is what this court called imagining a world that never was. And that is not something that's permissible with a provision like this. And that's exactly what the class-based approach does. It imagines a world that never was. And that's several years before the statute. It's not intended to define classes under the statute. It's hard to know how to define classes under the statute. What do you think about my question about to sort of have a better textual link between the theory that San Francisco and FERC have provided that textually we're talking about the ultimate consumers on this date, on that date that consumers were defined by the 1987 agreement, and the 1987 agreement defines them by class? So I don't think that's correct, Your Honor. The 1987 agreement, again, not constructed for this purpose, has broad categories, and they're also sort of vague, of the kind of people that could be served, that could be served, could have been served on that date. Not that were provided service by San Francisco on that date, which is the language of the statute. And that is exactly the world that never was that the BP case is talking about. Because the 1987 agreement says river bank. There may not even have been a customer served by San Francisco in the river bank category. And that's especially true given that the 1987 agreement was amended several times over its life. It was constantly shifting. So I just think it's impossible to say that somehow this particular 1987 agreement solves the statutory problem in this case. There's going to be a lot of other cases where there is no 1987 agreement. This is going to apply in lots of other circumstances. And so I just think that what happens here under the categories in the 1987 agreement is the commission said this itself. This is a JA-123. It's from the commission's original order that this court vacated on other grounds in 2022. This will allow San Francisco to cover every single consumer in the city with grandfathering. The agency recognized that. And that's what this would do. The agency recognized that it followed the 1987 agreement that that's what it would allow. They recognized it with respect to the class-based approach in general. And again, the 1987 agreement categories are very broad. So individuals operating in coordination with the city is anything that San Francisco wants. And San Francisco has actually argued in the past that that means customers to whom they're providing power because those people are paying San Francisco and therefore operating in coordination with them. It is a- But you're saying theoretically they could do that. But what I heard your friend on the other side say is that that hasn't happened. And if that were to come up, it would be part of your account. And then there would be a decision on those. Well, again, though, I think the class-based approach is what is the problem here. The class-based approach does allow this to happen. And so we could make all the arguments we want afterward if this court approves the class-based approach. But the class-based approach would allow them to slot these people in to these categories in the future. And the fact that they haven't done it yet, I think, is not indicative. Everyone's waiting for this court to decide whether the class-based approach is correct or not. They have made very, very broad arguments. They're also trying to push PG&E out in San Francisco in various other ways that are not implicated in this case. So I don't think we can just- But PG&E has 85 percent of the electricity market in San Francisco, correct? It's approximately that, I think, yes. And what I'm hearing from the other side is that if you were to prevail in this case, all the options that you told us that were available would make PG&E the provider, or it would force San Francisco's cost to go up so much that they would have to give you their retail customers. Right. And I really disagree with that, and I'd love a chance to address it. So both with respect to the Community Choice Aggregation Program and the intervening facilities. So the Community Choice Aggregation Program, as I said, is used all over California by lots of municipalities. It makes the customer a San Francisco customer with respect to the power generation that San Francisco does in Hetch Hetchy Valley. Served by who? I'm sorry. And they'd be served by PG&E? Well, they would be served by PG&E in the same way that they're served by PG&E under grandfathering. PG&E's distribution facilities would be used to get the power to the customer, and PG&E would get paid for what it does. But San Francisco would also get paid for what it does. So they would be a customer of both, actually. And so that's not any kind of problematic thing. It is used by San Francisco now, and it's used all over the state. I'm just going to go back to why is PG&E fighting this so hard? Because what you told me before, I don't see how that harms PG&E. If they're making a profit on wholesale versus retail, whatever. Why are you fighting this? As I said, it is the rate arbitrage that is the problem. It's taking over all PG&E customers, essentially, because it's the combination of the broad grandfathering so that everybody's covered by the wholesale distribution tariff. And then it's interpreted to allow broad grandfathering, then you're okay. Which I think is exactly what the agency has done by pointing to these categories in the 1987 agreement and by the way it's described as a class-based approach in general. Yes. Do you believe that FERC misconstrued the 1987 agreement, and if so, how? I think FERC correctly referred to the categories that are listed in the 1987 agreement. But those categories, again, are not for this purpose. And there's no way to know whether they match up with what class means under the statute, because the agency has never explained how one would go about determining what a class is under the statute. And that's that level of generality problem we were talking about before. Not surprising, because the word class doesn't appear in the statute, and Congress hasn't given any guidance on it. So, and I would also point out that the agency really struggled here. Did PG&E agree to those classes in the 1987 agreement? Well, PG&E used those categories in 1987 to generally refer to customers. Do you agree to those classes? Well, again, I think it's not as simple as that, because it is a general reference. Those categories pop up in the agreement, but those people were not necessarily being provided power. I understand that, but those are categories that PG&E agreed to. But PG&E would never have agreed to them as classes for purposes of the grandfathering exception, because they're vague. They're amorphous. PG&E would have defined things differently. They would have put a lot more care into these categories. This was a simple agreement between PG&E and San Francisco, which, again, changed a lot over time, in fact. And so, I just think it's extremely unfair to say that this agreement that PG&E entered into before the statute even existed is somehow now going to bind it under this grandfathering exception when it was never intended for that purpose and wasn't written with that in mind in any way, shape, or form. And again, as I said before, the 1987 agreement won't exist in lots and lots and lots of other cases. Is that the reason that FERC gave in its order? I'm now not recollecting. I mean, I think it's the 1987 agreement was pointed to as a reasonable way for figuring out what these classes were. But I'm not sure that, did FERC rely on the 1987 agreement in terms of defining the meaning of ultimate consumer? Certainly not in terms of defining the meaning of ultimate consumer, no. FERC didn't say much at all about what ultimate consumer means. There's actually almost no statutory interpretation at all. They may have had a Chenery problem then with them relying on that here at oral argument. I completely agree there is a Chenery issue here, because I think the agency is not engaged in statutory interpretation in any meaningful sense here. They've just looked to the Suffolk County orders, which themselves engage in no statutory interpretation. The reference to the 1987 agreement came when the agency said, okay, there's a class-based approach. We've told you that now because that's what Suffolk County says, even though other FERC orders from that era say the opposite. But that's what Suffolk County says. Now let's talk about what the classes are. And that's when they looked to the 1987 agreement, although they struggled mightily. They looked to a lot of other things too. And only on exhibit to San Francisco's complaint, but maybe that's not quite right, even though those are categories that San Francisco came up with. So maybe we'll look at this 1987 agreement instead, and there was yet a third category. We're going to jettison that. They're struggling. They're struggling because there is nothing about classes here. Congress hasn't said anything about classes. Congress hasn't explained how classes should be defined. And that's the issue that they're having here is that they're making it up because it doesn't come from anywhere in the statute, and it's directly inconsistent with the plain language of the statute.