Oral argument — Citadel Securities LLC v. U.S. Securities and Exchange Commission

and we'll also hear from IEX. We'll hear first from Mr. Scalia. Good morning. May it please the Court. Eugene Scalia, representing petitioner. I'd like to reserve three minutes for rebuttal. The IEX Options Exchange is designed to operate in a manner that the securities laws intended to prohibit. Securities exchanges are meant to be neutral forms, level playing fields. Yet the exchange at issue here undeniably discriminates in favor of members of the exchange, these select group of market makers, and it purposely discriminates against all other market participants who are compelled to transact on this exchange even if they would prefer not to. In approving this exchange, the SEC relied repeatedly on a D.C. Circuit decision involving a different exchange that itself differed in fundamental respects. First, the D.C. Circuit found that the exchange there operated symmetrically as to both liquidity providers and liquidity takers. That is, it did not find the very discrimination that is the heart of Citadel's case here. On top of that, the D.C. Circuit in that case relied heavily on what it found to be substantial evidence that for whatever reason, IEX and SEC both refused to come forward with in this case. There are other differences between the cases as well in terms of the record itself, the alternatives identified by parties that the SEC unreasonably failed to consider, additional data that was requested by the parties, and other arguments that were unreasonably dismissed. For these reasons, what we ended up with was a rule that is unfairly discriminatory, that puts burdens on competition that are inappropriate and unnecessary, and it was adopted in a manner that was arbitrary and capricious. We ask that the rule be vacated prior to the exchange taking effect scheduled in August. Mr. Scalia, perhaps, I'm not sure if you can, but can you sort of walk us through how this actually works in terms of when does the match happen, when does the speed bump take effect, if you can? And then we're going to ask, obviously, the same side to see if they agree or disagree. Yes. There are essentially four things happening. It begins with a market maker posting a quote on the exchange. Market makers put a quote or price on the exchange. The first thing is when that price is the best displayed on the market, orders are mandatorily routed to it. Broker dealers must send to that posted price. That's step one. Step two is the pause, the delay, what the SEC itself calls getting stuck in the speed bump. What happens is the exchange, as a service to its sophisticated members, pauses that order, does not let it execute, and then the third thing happens. The exchange, again, for the benefit of its members, surveys the market to see whether it's moving in favor of its members or against its members. And that brings us to the fourth step. If the market is moving against the exchange's members, then the exchange cancels the quote that this all began with or it reprices it. Okay, so my question is with the repricing. Okay, if I ask I want to buy for $10 a share, and that's what I make the offer at, $10, and the market maker for IEX decides the market is moving, it's transitioning, and so now it's $10.99, and it gets repriced at $10.99, am I obligated at that moment to buy at $10.99 or can I go back and look for a different price? You're not obligated to buy at $10.99, Judge Lagoa, but you may have lost the ability to buy at $10. What would have happened in that scenario is the market maker posted at $10. Your broker dealer was compelled to send your order there saying, aha, that's the best. Your broker dealer may have known that's IEX. That could be a phantom quote. It could disappear, but you were sent there. If the market changes and the market maker is against the market maker and they cancel that, you can't get $10, even though it's possible that $10 would have been available somewhere else. But let me ask you a question about that because it seems to me that when this latency arbitrage is going on, there's no risk at all to the latency arbitrage. It's almost like they have a back to the future kind of book as to what happened in the past. They know what the price is going to be. They know the direction, and there's no risk to them. It's just money for the taking, and the result of that, it appears, has been a cost to the investors and a larger spread that the market makers go ahead and engage in in order to protect against that. And I guess it's estimated that there's about $5 billion in loss each year to investors as a result of this with no risk at all to the latency arbitrage. Is that a correct understanding, or am I missing something here? Your Honor, we wouldn't agree with that, but we also think that regardless, that doesn't justify the action taken by the SEC. Can you tell me how I'm wrong? Because markets are always moving. There's never risk-free trading. It's possible that the trader who has higher speed systems, which the record shows are not especially expensive to develop, it's possible that a trader with those systems is correctly identifying where the market is, but it's possible the market moves ahead again. Well, how much is lost in latency arbitrage to the latency arbitragers? I mean, I imagine it's not much, or they wouldn't keep doing it. Your Honor, I don't have a figure on that, but what's important is all of the data that the SEC was asked to develop and refused to, and this is a critical difference from the D.C. Circuit decision because Judge Rosenbaum, the parties in the rulemaking were saying, develop this evidence. What is the effect going to be on retail traders? How often will orders be canceled or repriced? This is the data. What is the impact here that the parties said the SEC ought to develop? And in the D.C. Circuit case – How would they develop that data? I mean, because they're on a 45-day clock, and you can maybe get extended another 45 days to a total of 90 days. I know that you had suggested the audits, I guess, but aren't there 400 billion records each day in the audits, and how would that be done in a timely way to meet that 45-day or even 90-day window? Well, first of all, Judge Rosenbaum, in the earlier proceeding, some of this data was provided by IEX. That's very important to understand, that the D.C. Circuit said there was substantial evidence, not only of latency arbitrage, but that this mechanism targeted latency arbitrage because in addition to talking about how frequently in the day it was triggered, IEX put forward evidence that 24% of the displayed liquidity was actually traded on during the period this mechanism was set. But they were running an exchange at that point, and this we're talking about they haven't yet started running the exchange. So they had the evidence, I guess, available in that case because they were already running the exchange. But here, there's not a record of their having run the exchange, so where would they get that evidence from? Your Honor, there's no evidence that anybody was unable to put forward this data. At no point did anybody say, oh, we don't have the data. We can't obtain the data. Were they required to review the CAT data? Our obligation is that they were required to have a good reason not to do so, Judge Marcus. I understand that, but were they required specifically to review the CAT data in assessing the viability of the IEX proposed option exchange? Judge Marcus, as much as I would like to go there, I won't go that far. All I will say is that when the rulemaking participants urged them to look at the CAT data, they at that point had an obligation to respond reasonably to commenters and to take reasonable steps, and we think it's a reasonable step to look at the data. What they couldn't do was give an utterly arbitrary reason. That's what they did. At page 164 of your report. Let me ask you the question this way, then. I understand your position to be they weren't obliged to necessarily review that data, but they sure should have had enough data or empirical evidence to support what they were doing, and they didn't. Well, and I would agree with that. Here's what's remarkable. Let me just ask you the follow-up question in that regard. They had comment letters on latency arbitrage and the problems surrounding it by my count from at least six people and companies that were involved in doing business. Matthew Boven Carrier, the director of all options, Gregory Bayard, James Cosentino, Steve Crutchfield, Benjamin Chifrin, Brian Donnelly, all involved in the market, some of whom claimed they were driven out of the market by virtue of if there was no speed bump here. Was that not a body of evidence, admittedly not empirical data, but was that not a body of evidence upon which the SEC could rely? Perhaps it couldn't just rely on that, but would that not be some data in support of the SEC's decision? They could rely on that, and I'm not here to quarrel that there's zero latency arbitrage, but back to what they refused to do. The only reason they didn't look at the CAD data, which would have told them how many retail investors were affected, Schwab, which is the largest broker dealer, it represents retail customers. It uses Citadel's facilities to execute trades. Schwab and others said, look at the impact on retail investors. The CAD data would tell you that. The only reason they gave for not looking at it was they said, well, it would be misleading, but the data that they said would be misleading, some of that is the same data that the SEC had relied upon in the earlier case. So you had the SEC saying in this case that it would be misleading to rely upon the data, and the D.C. Circuit had relied upon it. That is arbitrary. Isn't the equities market a little bit different, though, than the options market? I mean, it seems like there are a lot more trades that go on in the options market. The market makers are picking up a lot greater percentage of those trades. I mean, doesn't that have an effect on the analysis? That was argued both ways. The market makers said that because the rules proponents said that because we can't trade off exchange, we need special protection. Others said, well, precisely because there are no alternatives, this at least should be an optional mechanism. If this exchange is so good for investors, as the SEC claimed, because it would result in lower prices on the exchange, an alternative that was identified was make it voluntary. Let broker-dealers decide whether they trust IEX and use it. If it's a good deal, they will, but if it's a tilted playing field, as many believed, they won't use it. Call IEX's bluff on how great a product this was. That's another alternative that the SEC unreasonably refused to consider in finding that this was fair and also in finding that it was an appropriate burden on competition. Let me ask you a follow-up question. You're on our time now, not yours. Thank you, Your Honor. Isn't the problem of latency arbitrage a greater problem in the context of options trading and otherwise of the securities? Tell me about the difference between the two markets. Between an options market on the one hand and trading in stocks and bonds and so on the other. Your Honor, the principal difference that I think the rulemaking identified was that the options market, market makers are, they post there, they're obligated to post a certain amount of trades and also that there's not an off-market exchange that's available for equities. I think those are the principal differences. What none of that explains is why the commission wouldn't even look at the impact of its rule. If this rule is good, the CAT data would have shown it. If it showed the rule is bad, then we would have learned that. I've not seen an agency before say we have data, that data tells us what the impact of our rule will be, but we're not going to look at it because the data would mislead us. That refusal to even learn the impact of their action was arbitrary. Would it be fair to say that they looked at the data at least in the other case and relied in some measure on that? That would be half true, Your Honor. That's the problem. At pages 34 of the IEX brief, at 49 and 50 of the SEC brief, they rely on that part of the D.C. Circuit decision even though that part of the D.C. Circuit decision is relying on the data that here the SEC refused to look at. I guess what I'm asking is that would be one place to look. It may not be the only place. It might not be a complete examination, but you would agree that that would be one fair place to look, would it not? I would not agree with that. You think that it is of no relevance here? I think it's relevant in showing the kind of thing that they can and should look at, but because as we just discussed, these markets are different. I don't think it's... I understand that, but so is it your view that the data collected on the problem of latency arbitrage in that securities market, NASDAQ, is so profoundly different from what's happening here in the options market that it would not provide to the extent they looked at that data, that would not provide an additional foundation, maybe not a complete foundation, but a basis upon which to rule here? I was speaking to... I may not have put the question... I understand it, and Judge Marcus, if I could say, first of all, on the question of latency arbitrage, I have no problem with that. I was speaking to a different issue, which is the actual impact of this particular rule, because... You're talking about the ORP rather than the speed bump? Exactly. Well, I'm sorry, rather than the D limit, which was the early... Because in the D limit case... Rather than the crumbling, crumbling quote. I call it the mechanism, Your Honor, for simplicity, but in the D limit case, what was so important in the DC circuit was that, although this operated 0.007% of the trading day, 24% of the displayed liquidity was traded in that small period of time. It was putting those things together that clinched it for the DC circuit. And Judge Marcus, that's what never happened here. No, I guess, but let me just sharpen my question this way. Could the SEC rely on its prior findings of the latency arbitrage in the equities context here? Could they rely on that in some measure? I agree that they could have said, considering whether there's latency arbitrage in the market, we can also look at the record in the earlier case. Whether or not they did that, I actually don't know. And I think it's important to keep them to... Would you say the record does not reflect that they actually did look at that? Well, they... They say they did. Your Honor, the SEC has told this court that it declined to look at certain evidence. It had looked at before because in this proceeding, latency arbitrage was uncontroverted. That's just false. That's just not true. That's the only reason they gave. And again, it's false. What is your evidence that latency arbitrage doesn't happen in the options market? I mean, that just seems wild. Your Honor, again, our case rests not so much on the claim that there's zero latency arbitrage, but that the commission was obligated to determine the extent of it, and even more importantly, that it had to show what the impact of this particular rule would be. This rule discriminated. I'll be interested to hear what the government has to say about it. This rule discriminated in favor of one sophisticated set of market participants. That's unprecedented under the securities laws for an exchange, which is supposed to be a level playing field, to pick a winner and loser like that. And by the way, the SEC will tell you, oh, this was about stale quotes. Even that's not true. If there was a, quote, stale quote, but it favored the market maker, the trade goes through. It's only when the old quote is against the market maker that then this triggers. And so it was the failure to look at the extent of the latency arbitrage. Of course, I think they would concede that it discriminates in some sense against the market maker, but they would say the touchstone in the statute is whether it unfairly discriminates. And so you've got to establish not only they did, but they did it unfairly. Yes, that's right. And we would say that it was unfair in part because it lacked the evidentiary foundation in the very case that they rely upon, on the very pages of that case that they rely upon, and also because this record was different in terms of alternatives. Folks said if this IEX exchange is so good for retail investors, then make it voluntary. And the SEC never had a rational response as to why it would not do that. And then the last point that I want to make, because with all appreciation for the time you've extended, I understand that I should sit down shortly. The last point I want to make on that is that in all other circumstances, the SEC says you cannot back away from a quote. Once you've posted a quote and execution is sought, you've got to honor it. That is fundamental to the securities laws. And the SEC's only response to that, this is page A153, its only response is, oh, the market maker is not backing away. It's the exchange that does this. And our argument is that proves our point. But that's why I asked the question, which is how did the process work? The process, Judge Legault, is the exchange itself as a service to these members who pay to join, who bring trades that line the pockets of IEX. That exchange, which is supposed to be neutral, holds the orders at bay, checks out the market, and if the market looks good for its well-heeled members, bingo, we're going to trade. But if the market looks bad, then the exchange, it's making this assessment, then it cancels. That is a rigged game. It's not how exchanges should operate. All right. Thank you very much. Thank you. Mr. Scalia, you've reserved three minutes for rebuttal. We'll hear next from the SEC. May it please the Court. Emily Truparisi for the United States Securities and Exchange Commission. I have eight minutes and counsel for IEX interviewer has seven. The Commission's order in this case approving IEX's proposal to offer an options exchange that utilizes the tools it already uses on its equities platform as the product of reasoned decision-making was supported by substantial evidence and should be affirmed. Can I ask you a question? How exactly is it fair for the market makers on the IEX platform to be able to have that speed bump where they're able to assess whether or not they want to go through with a particular quote? Because if it's another exchange, the market maker on that exchange has to go through with a quote. A couple of responses, Your Honor. The market maker isn't deciding when it sees the order whether or not to go through. I think it might be helpful to take a step back as to your question. You can walk me through. Yeah, I'll walk you through how it works. So, again, a market maker opts into the ORP functionality before, and it's a risk mitigation tool because background about the options market, options are traded as opposed to equities, and this goes to some of the other questions, only on exchange as opposed to about 10,000 NMS equity stocks. There are over a million option series, and the options market is particularly dependent on market makers for liquidity. In the record, about 99% of open quotes and bids are provided by market makers, so it's a quote-driven market. And, again, options are a derivative reference asset. That means their price is pegged to the underlying stock price, and so when that price moves, the market maker needs to quickly update their quotes on all of these series that they're quoting. And so it's unlike someone trading in equities market. It may have to update one quote. They're updating hundreds, maybe thousands of series, and they have continuous quoting obligations on both sides. So the ORP is a risk mitigation tool that helps them update their stale quotes more quickly, and it's commensurate with other risk mitigation tools that lots of options markets use. I'm still not sure I understand how this works. Let me give you an example and walk me through it. War is raging in the Middle East. I decide I want to buy oil on the spot market. I want to buy $10,000 worth of WTI intermediate grade western Texas intermediate oil. It's now at $99 a barrel. I want an option that will extend until the end of June, and I'm betting that oil is going to go higher over the next three months and three days. And so I go to the market maker and I say, I want to buy $10,000 worth of WTI oil at $99 a barrel, and I want to be able to buy that at any point between now and June 30th. Walk me through it. What happens? So you could go to any exchange and they would have to give you the best bid or price. They'd have to meet that best price. And if you were routed to IEX, if you were a regular trader, you could interact with that quote. The only people who interact with the ORP are high-frequency traders because the ORP is only on for such a small period of time. Right. So presumably I would interact with a market maker. Right. I would say here's what I want to buy. Here's how much I'm willing to spend. The market maker creates the liquidity. How does that work? And what does the ORP do and the speed bump do to my offer to buy at $99 a barrel? So the market maker makes liquidity by quoting on both sides of the market. So it will quote offers to buy and bids to sell. And so your order would interact with one of those. And if you were a regular trader, it would interact as normal. Now, there are always reasons any time a trader – you may see a quote on a screen and by the time – but before it matches, this goes to your quote. Right, but isn't that one of the issues, which is that you have all these different market makers and these different exchanges. And my understanding, and you can correct me if I'm wrong, but my understanding is that the way the system works, Judge Marcus's bid that he wants to make is funneled to the best price or the person that's offering the best price. And in this case, you know, how I understand the allegations, IEX can do a very aggressive price because they have this speed bump so they can readjust. Their market makers can readjust as opposed to other exchanges. So Judge Marcus is going to get funneled to IEX. And if there's a transition, then that speed bump, that second or millisecond that happens, it could readjust what the initial offer was that the market maker was making. A couple of responses, Judge Lagoa. First of all, you're only routed to IEX if they're alone at the best price. So if they're not alone at the best price, if they're other best price, you would not necessarily be routed to IEX. And even then, you could go to another exchange if they met that price. But if you are routed to IEX, you're only going to be interacting with that repricing. As the commission found, .001% of the trading day that it's repricing. And the record evidence is replete that ordinary traders, long-term investors, pension plans, all the evidence in the record, they are not interacting with market makers' orders in that infinitesimally small amount of time when the repricing is occurring. So all of the regular investors are interacting with the ORP at that price. Explain to me, because maybe I misunderstood it, but I thought that I would not have the option that it would just funnel me to whatever the best price is. Are you suggesting that I have the ability to decide which exchange I want to go to? If there are multiple at the best price. You would only have to go to IEX if they were alone. Well, is there another exchange I would go to? If I want to buy $99 a barrel, wouldn't I go to IEX? Wouldn't that be the place where I go to purchase the option in oil? Again, only if it was at the best price. So just to follow up on Judge Legault's question. So I put in a bid, $99, termination date June 30th. It will go to a market maker. The market maker is going to look around and see the best price he can get me closest to $99. If I understand what's happening here with the speed bump and the ORP, that trade will not be executed instantaneously. There will be a delay. Do I have that right? So there is a delay of 350 microseconds. Whether it's a second, half a second, it's a detail. Maybe it's a moment, maybe it isn't. It's not going to be executed immediately. All orders going to IEX go through that 350 microsecond delay, similar to other geographic latency. Okay. Now what about what does the ORP do? Additionally to an infinitesimally small fraction of a second delay. What the ORP is doing, because remember options are reference prices, so it's looking to see has the underlying price of the stock moved in a way that the market maker now has to update all of its quotes on its options because they're pegged to that reference price. So it's looking for price dislocation. The prices that the market maker has already put out there, these are quotes that are already on the market that liquidity takers could come and interact with, are now stale. Let me ask the question slightly differently. Can they change my price? So the ORP functionality, they can either cancel or reprice within certain. That's what I want to ask about. Reprice is another way of saying I put in a bid at 99 bucks and they suddenly say, well, we're turning it into $104 a barrel. No, it reprices the market maker's quotes. My bid is there at 99. That doesn't change. Either they accept it or they don't. Right. It's either canceled because of the ORP functionality or it would go through. But this is the question I have and I asked Mr. Scalia, which is, so Judge Marcus makes the price for $99. The market maker, IEX, one of the market makers on the IEX exchange, it gets funneled to that market maker who initially had a $99 offer, but now it gets readjusted and repriced at $104. So Judge Marcus has lost the ability now to buy at $99, even though there was another exchange who potentially had $99. Well, everyone's updating their quotes. It's a race condition. But you are correct that if the price is updated. This is why the issue for me, the question is, is that it seems that the ORP or the speed bump, the ORP and the speed bump together really benefit the member of IEX and not the investor. So the commission found that the investor has now lost the ability to buy at $99 as opposed to now he has to think, well, do I want to buy at $104? So the commission made findings. The only investors who are hurt by the proposal are high-frequency traders because they're the only ones who can target that tiny bit. High-frequency latency arbitrage. Latency arbitrage, right, because they're the only ones who can trade in that small amount of time. So how does that somehow knock them out of the box? They're entitled to compete on the market too. Let me phrase my question this way. You've got to come forward with at least rely upon a substantial body of evidence to support the change. Plainly it discriminates. The only question is whether that discrimination is unfair. There's a problem with latency arbitrage. What's the evidence upon which the SEC based its determination that this was a reasonable way to go? In terms of cite me the evidence, chapter and verse that I could find in support of it. I cited to Mr. Scalia six statements from six folks involved in the marketplace, market makers, if you will. What other evidence was there beyond that to support the determination? There's the comment letters that we think are definitely substantial evidence, FA 68. There's the Volant letter saying that latency arbitrage caused them to exit the market. The commenters were almost universally. What is there beyond the comment letters? Then there is the record that was in the D-Limit case that showed latency arbitrage on the equities market. The commission definitely did rely on that. Then it said, and this is at A157, that in fact, because of the commission's expertise and understanding of market structure and the equities market and the options market, latency arbitrage is even more acute in the options space than in the equities space where it had all the evidence from D-Limit. Why is that so? It's because the options market is only on exchange, so there's no way to avoid the latency arbitrage. It's because of what I talked about earlier, because of the proliferation of series. You have 10,000 stocks versus a million series, so the market maker has many more quotes to update. So in that race condition, when the stock price has changed and the market maker needs to update a whole bunch of stale quotes before a latency arbitrage can pick them off, there's many, many more to update in the options space than in the equities space when you're quoting just the underlying one stock. Also, market makers have these continuous quoting obligations, so they have to quote on both sides of the market. The commission relied on all of this evidence at the A60 to say that that's why. Two things to me now. One, the comment letters. Two, the record in the equity markets issue. And I would say the commission's expertise of market structure as well. So was there any other data that you would have relied on? Well, then there's IEX. I'm not sure, just to make sure your question's not broader, there was certainly the data that IEX provided in terms of how infrequently the indicator would be on, and so that goes to your honors at one point had a broader question about why this is not unfair. And the commission critically reviewed IEX's data that they submitted about the infrequency of time that the ORC would be on. It was the same evidence that was at ND limit. It showed that the only people that can target these moments of price transition are latency arbitrage orders, and therefore it was not an unfair discrimination because it's commensurate with the risk that the market makers take on in this options context. It's also very of a piece with other risk controls that all the other options markets have. There's transaction-based controls, volume-based controls. Those are blunter instruments because they cancel entire classes of orders. Two questions, final questions from me. Why didn't you consider the CAT data? Yes, thank you for bringing that up. I wanted to address that. So on the CAT data – Wasn't that relevant data here that might have borne directly upon this issue? No, Your Honor, and the commission explained that. So two responses on CAT. First of all, it's raw trade and order data. So to get the information the commission needed, it would have had to essentially construct its own ORC, and that's just not something that the cases in the statute require. Remember, this is an SRO approval order. It's not the commission's own rule, and it's on the statutory clock that Judge Rosenbaum mentioned. But also and more importantly, the CAT data would not show you – it doesn't go to the relevant question, and the commission explained this, because the most relevant metric for figuring out, is this going to be narrowly targeted just to target latency arbitrage, is the time data. It's not how much of the trading day it's on, those percentage statistics that IEX gave the commission. It explained why that is what it called, quote, the most relevant and persuasive data. And the other data that Citadel points to, volume and fill rates, that doesn't tell you what you need to know, because it may be that there's lots of volume and quotes happening, but that means it's working. That means it is targeting latency arbitrage. So the commission explained both why the CAT data wasn't relevant, and also why it wasn't feasible to use it. I have a question just about sort of how the latency arbitrage works. So I asked Mr. Scalia about the risk to the latency arbitragers, and he explained that they have – maybe I'm misunderstanding, and he'll let me know if I am. I think he said they have risk – also they have just as much risk as everybody else. Is that your understanding of how latency arbitrage works? That is not my understanding. My understanding is their model is predicated on capitalizing on their speed advantage, and basically they capitalize on the stale quote and then turn around and sell it at the stale price. So does it eliminate risk? I don't think I would go – I'm not sure if it eliminates risk, but it certainly is not the same kind of risk as other markets. Does it substantially decrease risk? I would think so. Do you have any evidence other than your own thought on that? In terms of – I don't think the commission spoke to the risk to the latency arbitragers. It's more their effect on the market. No, I understand, but one of the things that the SEC is charged with doing is making it fair, right, making the markets fair. And so it seems to me that if there is a mechanism to eliminate risk, but you've said it doesn't, then that would be a relevant consideration. If there's a mechanism to substantially decrease the risk that's not available to all of the investors, that might also be relevant, and that's why I asked the question. Right, I think what I would say, Your Honor, is the commission is charged with looking at the market as a whole, and the commission had data saying that latency arbitrage is a legitimate disadvantage because it depresses liquidity and disincentivizes market makers, and there was complete evidence in the comment record. And again, it's not the commission regulating the entire market here. This is an exchange-proposed innovation on one options exchange out of many, and the commission's task was, was it consistent with the Exchange Act, and the commission had extensive findings supported by substantial evidence for why it was not unfair discrimination and not an undue burden on competition. All right. Thank you. Thank you, Your Honor. We'll go next from IEX. Good morning, and may it please the Court. Juan Shin for Investors Exchange. Now, Your Honor, I think it might be helpful to supplement the discussion earlier that you had with other counsel regarding kind of the mechanics of this and give a little bit more context because I think it's important to understand. So market makers are the participants who provide liquidity in the options market. So, in other words, they're the ones who make offers to sell or bids to buy, and then other participants, liquidity takers, come in and accept those bids or offers. Now, as has been noted, in the options space, there are many more listed securities than in the equity space. So for every one listed equity, let's take an apple, there may be dozens or even hundreds of different options that a market maker lists. Now, in the ordinary course, a market maker is entitled to update the prices that they're offering in connection with those options. They're not locked in once they initially post a quotation. There's no dispute that's permissible. And it's also true that some market makers expend incredible amounts of money to develop the fastest systems possible so that they can update those quotes incredibly quickly and therefore gain an advantage in updating those quotes. Can I ask you a question, though? When the market maker, let's say, sends out an offer to buy, and let's say it's 1,000 shares for, I don't know, $10,000, and I place an offer for those shares and you receive it, can you withdraw at that moment? I don't think the answer, I think the answer is no, but. No, once the market maker's quotation is matched with an incoming order, it's executed, and that's required under the rules, and that's what exchanges do. You can't back away, right? Correct. So you can't back away once there's a match. Correct. Okay, so you have to explain to me, then, how the match occurs with this ORP and the speed bump. I'd be happy to do so. I have a little diagram, but I won't. If I could just finish those, setting up a little bit more context, and again, I thank you for your patience. So it's not disputed that market makers post quotations and routinely as part of their market making activities update or sometimes cancel quotes. They're just doing that because they're scanning the market, they're seeing that the price that they posted is no longer relevant because they're seeing the market moving on that underlying security and, therefore, they need to update the options. So that kind of updating and canceling happens in the ordinary course all the time. And as I mentioned, some market makers, they invest tens of millions of dollars to develop very, very high speed, low latency systems so they can get that information more quickly and update their quotes more quickly. However, what we know and what's in the record is that some market makers are unable to invest those huge sums of money to update, to develop the highest speed systems to update as quickly as other market makers, like Citadel. Citadel is a market maker in this space and they have incredibly high speed systems. Now what IEX Options is proposing to do is to offer market makers the ability to update or cancel their quotes quickly, more quickly than they can using their own systems. They're providing a tool that market makers can use. And, again, all it is doing is proposing to help market makers update their quotes more efficiently. Now why do they need to update their quotes more efficiently? Because there's a presence of latency arbitrageurs in the market. And so what they do is they see quotes that are currently being offered by market makers and what the latency arbitrageurs can do is that they can detect movements in the price of the underlying security, which will flow through to the options. They can detect those movements incredibly quickly because they have those very low latency, high speed systems. Tell me why that's a bad thing. To the extent what we're looking to do here is promote competition, isn't it a good thing? Well, latency arbitrageurs is a bad thing, we would submit, Your Honor, because it depresses the provision of liquidity. So what happens, latency arbitrageurs come in incredibly quickly before market makers can update their prices. They come in and they pick off the quotes in those microsecond intervals, pick off the quotes, and then turn around and sell them, and they are able to extract for themselves the profit that otherwise the market makers would get. Presumably anyone else can do the same thing. Not anyone else. They have the skill and the money to invest. Respectfully, Your Honor, what the FCC has found is that an incredibly small number of these high-speed trading firms are capable of doing this. Ordinary investors are not capable of doing this. Even the risk here is you're going to drive out of the marketplace a certain quotient of people, right? You're going to drive some people out who can't compete. Is that effectively what you're saying? Not quite, Your Honor. What we're saying and what the FCC found is that in the options market, because market makers have this very unique role in the provision of liquidity, what the FCC found, based on the evidence, is that the existence of latency arbitrage, in other words, the risk that the market makers face of having their quotes picked off to their disadvantage, that causes market makers to offer worse prices. In other words, they offer worse prices because they need to build in the advantage on those particular trades. There's a bigger spread, right? Exactly, Your Honor. And so wider spreads, which are disadvantage ultimately, which are harmful ultimately for investors, they provide less liquidity. So not just worse prices, but sometimes they even pull back the amount of liquidity that they offer. Explain to me how liquidity would be affected. Say it would create less liquidity. Why? So because market makers ultimately, because when they offer more liquidity and they are on the hook for ultimately if an incoming order matches, they're on the hook to follow through with that transaction. And realizing that they're on the hook for those various orders. And again, remember, they're quoting hundreds or thousands of orders of quotations for each underlying. But I guess what I'm hearing is that you don't want market makers to lose. You always want them to win. You want them to always be the house. No, no, no, no, Your Honor. That's what it sounds like to me. No, Your Honor. What the SEC has identified is there's an imbalance currently in the system where latency arbitrageurs, they come in and pick off those sale quotes before they can be updated and they pick off that for themselves. And that ultimately, in those, in the first order, that harms the market makers. But what the SEC found is that ultimately harms liquidity because, as I was starting to say, market makers, as a result, they offer less liquidity. They offer liquidity at worse prices. Some of them are even driven out of the market. And so conversely, by combating the latency arbitrage and leveling the playing field, what happens is that it will encourage, and again, there's evidence of this in the record, it will encourage more market maker entrance. So in other words, it's not dominated purely by the high speed firms like Citadel, but more competition is engendered. But that's not necessarily true because the market maker, from what I understand, and the rule that we're talking about really is only IEX at this point, right? Because they're the only ones that have this ORP and the speed bump in question, the algorithm, right? So it has to be the members who are market makers in that exchange. Right. But there's evidence in the record where market makers have said the availability of this tool, the ORP on the IEX options exchange would either potentially cause the addition of entrance. So market makers to come in and post liquidity on IEX where they otherwise might not post liquidity at all, or they might be able to post improved pricing and more liquidity on IEX. What I'd like to go and talk about that I still don't understand exactly for you to explain to me about the match and where the speed bump happens and the ORP. So the order is placed, okay? So Judge Marcus has his order, he placed it. What happens now? And it goes to the IEX exchange to a market maker and IEX. What happens? Okay. So what the ORP does, and I want to stress this is very important, the ORP operates independently of that order coming in. So whether there's an order in the pipeline or not, what the ORP does is it scans the market. It scans publicly available data and other exchanges. It scans the pricing of underlying equities. And then it performs a calculation to detect what effect that will have on options prices. And if the market maker's quotes are dislocated from that predicted price change, and of course, if the market maker at the time of posting opted in to using the ORP, then at that time, the ORP will basically serve that function of changing the price of the equities and canceling it. And, again, that's all preselected at the time that the quote is posted. So the ORP operates independently of that quote. In other words, it's just helping the market maker do what it would do on its own anyway, which is scan the market and update its quotes. Now, what happens if a quote is coming in? Okay? So there may be a series of quotes coming in. And it goes through the speed bump, which is a 350-microsecond delay, what the DCs are called the 111th blink of an eye. So it comes in. And if the first order arrives before the ORP is able to do all that, it'll match and execute. The ORP is not a fail. It's not a guarantee of anything. Okay? But it makes it less likely that the incoming order will be able to pick off the fail quote. If, however, it gets through the speed bump, and it arrives at the matching engine of IDX, and the market maker's quote has already been changed or canceled so that there's no longer a match, then it won't match. And, therefore, it will not execute. But it will not to address the earlier concern. It doesn't change the incoming order's price. So if it doesn't match, okay, then you're not going to have a situation where how then if it – I don't understand then how is it repriced then? It's not the incoming order that's repriced. It's the market maker's offer that's repriced. So, in other words, the market maker may have said initially $100. However, in the overall market, what's happening is that the price is ticking up or down $1. And so – I guess my question is the way I understood it was that the way the system works or the engine works is that you're sort of funneled to the best price. And so let's assume that the market – the way I understood it is that the market maker, the IEX market maker, offered the best price. So my offer goes to that exchange. And then all of a sudden I get a repriced offer. So I'm now lost because it didn't go to someone else. Well, so I'd make a couple points on that here. So one is that losing out on a posted offer, on a posted quote, that is an ordinary phenomenon in the markets that happens all the time. So there are many reasons why an incoming order may ultimately not match with a posted quote. So, again, as I alluded to earlier, liquidity providers, market makers in the options space, they're permitted to change their quotes before they're matched all the time. They do that routinely. No one is saying that that's impermissible. And so it may well be the case in an ordinary functioning of the market that the market maker's quote has been updated or canceled on his own, again, without respect to the incoming order. And so when it arrives, the quote's no longer there. That happens all the time. Or is it possible that there's only a certain amount of liquidity at that price and there may be a lot more takers than the liquidity offered? That was exactly my next point, Your Honor. You anticipated my point. I alluded to earlier how there can be a series of many incoming orders coming in. And as I said, the first one, if it gets there before the ORP executes, then the incoming order will match with the sitting offer, with the posted quotation. However, further incoming orders behind it may be trying to get to the same offer. And if that's already all taken, they won't match either. So, again, the failure to match is an ordinary part of the markets. So, I guess maybe relatedly, can you explain to us what the effect of latency arbitrage is on retail investors when they're trying to take advantage of prices through the market makers? Yes. So, as I alluded to earlier, latency arbitrage is what they will come to. And the classic example of this is set forth in the D.C. Circuit's opinion. They give a very easy example. So, there's a posting at $10. And it ticks up to $10.01. And ordinarily, the liquidity provider, they would want to update their quote to the $10.01 to get that extra penny on the sale. Now, what the latency arbitrage does is, again, because it has these incredibly high-speed systems, it finds out about the $10.01 before anyone else, or before the ordinary liquidity providers. It comes in and buys at the $10.00 and then instantly turns around and sells at the $10.01. So, it gets that profit that the liquidity provider ordinarily would have gotten. Now, there's not that liquidity provided that would have otherwise been provided. Correct. And I also want to emphasize, however, Your Honor, that the SEC's order here was based not just on that kind of first-order harm, but the broader view of the market, which is that because of this incredible risk that market makers have in the option space, because they have such responsibility in first posting the liquidity and then updating it because they have obligations by the options exchanges, because of that unique risk that they face of having so many quotations picked off by latency arbitragers, that risk depresses the availability of options quotations overall. And that, too, harms other investors. And so, Judge Marcus, you had the question earlier about evidence of this decision. Yes, and I wanted to ask a follow-up. When I asked counsel for the SEC what it is the SEC relied upon, what substantial evidence? She referenced comments. We talked about that. Second, she referenced the record in the equity market in the earlier case. And third, she referenced IEX data that you had provided. What data did you provide to the SEC, and how does it bear on this? Yes, the data that IEX provided that was discussed earlier, that was specifically to address the question of whether the ORP is narrowly tailored. It wasn't about the question of whether latency arbitrage exists in the options market. What did you give them that bore on the question of narrow tailoring? We gave them an analysis of data that we conducted, in which we set the ORP to its most aggressive settings. So, in other words, where it would be activated at a lower threshold than it actually would in practice. But we wanted to be over-inclusive in terms of how often the ORP could possibly activate. And so we provided data over a number of trading days, and we analyzed two different periods of time during two separate months. One, I believe, was in February and the other in July. And we provided data to show over the course of a trading day, on average, what periods of time will the ORP actually be triggered by external changes in pricing data. And we analyzed that by different classes and series. A class is based on the underlying stock, and a series is all the different options contracts that are related to that underlying stock. And what we showed is that there is a little bit of variability in how often it would trigger in different classes and series. However, the overall average was that it would be just 0.001% of the trading day. A sliver, essentially infinitesimal slivers of the trading day. Even for the particular classes or series in which it was most frequently activated, it was just 0.2% of the trading day. And the reason that goes to the question of narrow tailoring, Your Honor, is that the SEC had evidence in the record which showed that ordinary investors, in other words, investors who do not have these latency arbitrage systems, ordinary investors are not capable of targeting those tiny slivers of time. Therefore, the ORP won't inadvertently block their trades. It's really only the latency arbitrage liquidity takers that are able to come in and detect those very tiny slices of time when the price is changing, that they're able to come in. And, again, going back to the example, they can come in at the 10, grab it before it's updated to the 10.01. Ordinary investors can't do that. And I just want to be clear, when I say ordinary investor, I'm not just referring to individual people trading individually. I'm also referring even institutional investors like T.R. Price and various state retirement systems. They put in letters as well. They said even institutional investors are not able to keep up with latency arbitragers. So it's really just a very small number of high-speed trading firms that are able to do that. And circling back to a point that I was making earlier, Your Honor, in terms of the broader harms to the market that the SEC discussed in approving this order, there was evidence not just by market makers who said this proposal would help us against latency arbitrage. Even institutional investors and pension funds and the like, they said that the IEX options proposal would help us against latency arbitrage as well. And it will help us and, therefore, by extension, ordinary investors on whose behalf we act in terms of the market. So the additional provision of liquidity that this is designed to encourage will actually have this broader effect on the market. It wasn't just market makers who submitted comments. It was other types of investors as well. All right. Thank you very much. Thank you. Mr. Scalia, you have three minutes of rebuttal time. We'll probably take you over. We're very interested. Thank you, Your Honor. We welcome that. Judge Lagoa, you asked the SEC why this was not unfair discrimination. And what you heard was that there were, quote, stale prices, that there was, quote, an update, that there was, quote, dislocation. But what the SEC continues to hide is the fact that they only care at IEX when the so-called stale price works against the market maker. If I'm a retail customer sending an order, I don't want to send it to an exchange that's biased for a sophisticated market maker and looks only at when the market is moving against the market maker and not against me. So this exchange discriminates first by favoring market makers over everybody else in terms of who can use it, second by working against latency providers and for that small group of latency takers. I'm sorry, liquidity providers in favor of a small group of liquidity providers, which is not what the D.C. Circuit found was happening before. And then when prices are changing, it's only helping one side. That is discriminatory. So same question for you as I had for Mr. Shin, and that is what do you think the effect of latency arbitration is on retail investors, institutional investors, things like that? First of all, we would know more if the SEC had been willing to look at the CAT data, which would have shown what the impact of this mechanism would be on retail investors. Do you have any evidence of any type that shows that there's a positive effect on those types of investors? I do in the sense, Your Honor, that the record undisputably shows that retail participation in options markets is increasing. Options markets have grown. If the options markets were bad for liquidity as the SEC maintains, then options markets would not have grown as they have. Well, do we know what percentage of those investors in the options market are not market makers like Citadel who have the availability of latency arbitrage? We know that there is an enormous amount of retail participation, and we also know this, that these retail investors, Judge Marcus, your order, will indisputably be affected. The SEC is behaving as if this operates just a little bit at a time. It triggers thousands of times a second, according to the SEC's estimates, against one option series hundreds of millions of times a day. And how does latency arbitrage affect the prices that retail investors and institutional investors have to pay? What the SEC maintains, Your Honor, as I think you know, is that it's caused increased spreads in the prices quoted. And what we say is if this is such a good deal for retail investors, then make it optional. This is a discriminatory, unprecedented, biased exchange. How could you make it optional, though? I mean, how could you make it optional and still have it fulfill its function? Well, Your Honor, and that's another way in which this is biased because the SEC keeps emphasizing, oh, market makers get to decide whether to opt into this feature. Well, what commenters said was let broker dealers and retail customers, others, decide whether to route to IEX. If the market makers get to opt in or out, why can't we opt in or out? And there was never a rational answer given to that. But as this mechanism triggers potentially thousands of times a second, it's true, Judge Marcus, that you may not, as a retail investor, be operating according to sort of the high-speed systems to discern prices as other traders might be doing. But what's undisputed is that if you send a retail order, and whether by good strategy or luck, you are about to get the best price on IEX, what's undisputed is this exchange, which you would have expected to be a neutral forum, has said, hold on, sit in the waiting room, we're going to look at the market and see if the market is working well for our market-making member. And so although you may not have been yourself using high-speed training, undeniably, your order is going to get caught there if it's there, as this triggers thousands and thousands of times a second and millions of times a day. With respect to the CAT data, which you've asked about a number of times, Your Honor, whether retail investors were going to be affected would have been disclosed in the CAT data. The CAT data is an enormous trove of data, but when you look at Appendix 164, I urge you to, the first column, it is an incomprehensible explanation from the SEC as why it refused to look at this data. And interestingly, the D.C. Circuit in the Tempanera case, which both parties rely upon, looked at a very similar question where one aspect of that rule was aimed at so-called professional traders who were accused of engaging in latency arbitrage. And Chief Judge Ginsburg said, yes, there may be some cost to latency arbitrage, but speed of execution is also a benefit. And he said, SEC, you have data. You can look at the data. You should look at the data. And that was a principal reason that the D.C. Circuit unanimously held that that rule that discriminated against professional traders because of supposed latency arbitrage was arbitrary and capricious. Mr. Scully, can I ask you a question? And I just want to make sure that I understand. If I'm an investor, institutional investor or retail investor, and I decide that I don't want to do IEX option market because I've done a couple there and the price changes on me, and I want to go to another exchange, do I have that ability? You do not. I do not. That's what I think. I do not. You're a broker-dealer, and your broker-dealer, who's obligated to be himself or herself a sophisticated market observer, may tell you IEX is a rigged game. I've sent so many orders there that didn't get filled because of this mechanism. It still gets routed there. Because it gets routed if it's the lowest price. If they have the lowest price, it gets routed there. So it's not like Amazon. I can't pick option A, B, or C. That's correct. And by the way, it was very telling as IEX counsel spoke. He spoke a couple of times about how this exchange is, quote, helping market makers, cancel quotes. That's exactly right. So can I ask you a question? Because we have to look a little bit at what happened in the equities market where IEX already has a similar type of program in place. Have there been any deleterious consequences of the type that you've spoken of today? Yes. What are those? Well, what Stiglitz commented, I think at approximately page 134, said that there's a high quote rate on IEX's equities exchange, but it's a low fill rate. And by the way, an exchange gets revenues even when quotes aren't filled. It gets the exchange member revenues, so the market maker have to pay fees to join. And then exchanges make a lot of their money from the data, the quotes that are provided. So a low fill rate, but a high quote rate. And, you know, as you heard from Mr. Shin, they are helping market makers even though the SEC claimed that the exchange was just operating independently. It wasn't facilitating backing away. You heard about how frequently this would operate. What was at issue was not what percent of the day was affected. It's what percent of orders were affected. If my 16-year-old daughter came to me and said, Daddy, my friends and I were forming a rock band, I would say, well, what's the demand in your time going to be? And suppose she said, nothing at all. It's four hours a week. That doesn't sound so bad. Lo and behold, they're practicing during algebra class. I wanted to know that, too. And so when the SEC tells you this is narrowly tailored, they told you the portion of the day affected. They didn't tell you the portion of the trades affected. That's what they purported to show at least. But aren't most of the trades latency arbitrage trades? I mean. I'm sorry. There's not evidence to that effect in the record, Judge Rosenbaum. What the evidence is, is this is a very robust, healthy market, which with a substantial amount of retail participation. Remember, they claim they're there to help retail investors. Is there evidence that it's not latency arbitrage? I mean, you could have provided that evidence. I think there's evidence that it's a robust market in which retail participation is substantial, and the SEC could have learned what the impact on retail investors was from that CAT data. If I could just respond briefly to the SEC's comment about risk controls. And, you know, the way the SEC talks about this case sometimes reminds me of the parable of the blind man and the elephant. We keep hearing about just one part, but not the totality. So it is true that trades get canceled in a variety of circumstances. But what's not true is that exchanges cancel trades at a benefit to market makers when the market is moving against them. That is a unique form of discrimination. Likewise, there can be delays in systems, but those aren't purpose-built delays in order to favor another person transacting on the exchange. And so with respect to the risk controls, yes, there are other risk controls. But to understand those, I would urge you to look at the clearance, the options clearing corporation rule that we cite at pages 22 and 38 of our brief. What that rule talks about is risk control measures to deal with erroneous inputs, erroneous trades. It's when people made mistakes. That's what's being addressed here. It's just the market is moving in one direction and not the other. And that's why it's being triggered. To wrap up, two brief points unless there are further questions. Again, you've not heard, I think, a clear answer why this discriminatory, experimental, unprecedented exchange could not at least have been made optional for broker-dealers as the mechanism was being made optional for market makers. And finally, just a point of clarification, we ask that this rule be vacated. What the IEX has indicated is that the rule, that the exchange will become operational in the third quarter of 2026. We ask that this court rule prior to that date. All right, thank you very much, counsel. We appreciate the well-prepared argument all the way around. Thank you, Your Honor. We'll be in recess.