Below is the transcript to Joe Saluzzi on High-Frequency Trading: The Equity Market Is Now Controlled By The Machines:
Chris Martenson: Hello, this is Chris Martenson of PeakProsperity.com, and today we’re going to be talking about a very important subject: how the stock market works. Specifically, about something called high frequency trading and its importance to you, its importance to the markets, and its importance to our future. We are very fortunate today to be talking with Joe Saluzzi, a partner, cofounder, and co-head of the equity trading firm Themis Trading, a leading independent agency brokerage firm that trades equities for institutional money managers and hedge funds. Prior to Themis, Joe headed the team responsible for equity sales and trading for major institutional accounts at Instinet Corporation for more than 9 years. He’s a frequent speaker on issues involving market access, algorithmic trading, other sell and buy side concerns. He’s provided expert commentary for media outlets such as 60 minutes, Bloomberg, CNBC, Wall St Journal and many others.
Joe, welcome, it’s a real pleasure to be able to talk to you today.
Joe Saluzzi: Thanks a lot, Chris. Thanks for having me.
Chris Martenson: Sure. So why don’t you tell people a little more about your background. I used the word “Instanet,” but people might not know what that means.
Joe Saluzzi: Sure, Instinet was one of the original ECNs, which back in the early 90s was slow going to get traction, but it became the original electronic trading forum for most institutional investors. I spent, as you said, about nine years managing their institutional desk as well as their broker dealer desk. After I left, my partner Sal Arnuk was also an Instinet veteran, and we decided to start our own agency shop. My job during the day is to trade for institutional clients to get best execution for them. For the first 3-4 years of our firm, everything was great and we didn’t notice too much of a difference in the markets, but we started noticing a lot of difference in the past few years, which kind of brought us into a new area which we’ve been developing and working on—market research. We’re delving into market structures and how that has changed particularly since 2007 when Reg NMS was fully implemented, and how now the name of the game isn’t necessarily fundamental analysis or technical analysis; it’s all about speed.
Chris Martenson: About speed. So, I really want to give you room to run on this story today. Because I think it’s a really important story that’s been given far too little attention. And it deserves maybe more time than the average 3 and a half minute TV spot can afford. So let’s talk about the stock market and how it functions, or maybe I should say dysfunctions. Arguably the stock market is the critical signaling mechanisms from which most people take their cues about how the economy is doing, how it’s likely to perform in the future, etc. So if the stock market is doing well, it’s going up, consumers are confident, more likely to spend—or so the story goes. The stock market is where people risk their savings, though, in the hopes of growing their wealth. Therefore it’s really vitally important that the market be perceived as free, fair, and equal, and confidence in the market itself is important to that.
So let’s go back to May 6, 2010. That’s a day that shook my confidence in the market to the core.
Joe Saluzzi: Well, it’s an interesting thing. If you read the report, the Joint CFTC panel, they’ll tell you one thing and they’ll kind of blame one particular firm. Most people took away from that report that it was one particular large institutional trader that sent in a large order that crashed the markets. So, we looked to them, and they said, no, that’s not exactly what happened here. Okay, what happened here—yes that order was in the market, and the order in question was a large 75,000 contract e-mini order which was set to trade for 9% of the volume. That is not that big of an order. Yes, it’s a big order, but the algo was set to be not more than 9%. It didn’t have a limit on or anything like that, but it was in the e-minis, which is the most liquid contract out there. Well, what happened that day wasn’t necessarily the order—that may have exacerbated a few things, but that day started breaking down earlier in the day when there was some news about the Greek riots and so on. There was lots of nervousness in the markets. Things started to get a little shaky. When the order did come in, it basically exposed the market and the major market structure flaws we have out there now in that there were really no liquidity providers, there were no specialists, there were not many humans out there to stop things when they went wrong. And things went horribly wrong. What happened was – one of the guys over at the CFTC refers to it as hot potato volume (his name is Andrei Kirilenko, he’s the main economist there). What happened is that the volume just started to feed off of itself. Normally, these “high frequency traders” will be on one side of a trade and look to flip it immediately just to scalp a penny or two. Well, they couldn’t get out fast enough. They were buying a little bit and then it just kind of cycled down. And before you knew it, all the bids were disappearing. The problem with the equity markets now is that it’s linked to so many other markets. All markets are linked together probably due to high frequency trading and going cross-asset and so on, but many of the ETFs – which are priced off of equities – started to fall quickly then disappear, and everybody just got the heck out of the way, bottom line. Even afterward some of the high frequency traders were quoted in the paper saying “Well, we just walked away. Why would we stand there and take a hit and get killed like that, then we wouldn’t even know if the trades were going to be broken afterward.” To be honest, that’s a valid concern. They have no obligation, no requirement to be there—why would they stand there? And that’s the problem with the market—no one is required to be anywhere. Yes, there are many guys out there who are getting huge benefits from the market structure, but they don’t have any requirements to stand there anymore. It’s a tough game—the spreads aren’t there for them. Why would you stand there for a penny spread? Back in the days when they were specialists they would stand there, but they had a bigger spread involved, and that’s why they would commit some capital—because they knew they had less risk. The game has changed now when you don’t even have a penny spread and no real liquidity provider who’s going to stand there? You get days like May 6th. The question I get asked a lot is “Can it happen again?” and that’s not really the question, the question is when will it happen again? It will definitely happen again.
Chris Martenson: So back on May 6th, we had that plunge of well over a thousand at one point, and then – really damaging to my psyche – was the way the trades were arbitrarily broken. I thought they were arbitrary, at least they were never explained to me in a way that I said “Oh, I get that.” They drew a line and said below that, forget about it, we’re just going to bust those, which to my knowledge, has not happened before, so that was kind of a unique and still poorly explained event. But so was the flash crash itself. As you said, I think the story that got circulated for my consumption through the WSJ et al was that there was that one trader in the Midwest who dumped a few e-minis out there. But really what you’re saying is that the problem out here is that the liquidity, such as it is, the trading volume (maybe I should use the words more precisely) is really dominated by these computer boxes that are trading.
Joe Saluzzi: There’s no doubt. I mean, the estimates from some independent research firms are as high as 70% – we’ve heard somewhere between 50% and 70% being automated trading. There are different strategies: there are all different types of these guys, so it’s sort of difficult, but if you add it all up, it’s dominated by computers now, and then the traditional institutional investor no longer has that price discovery power, where they were kind of just following along. Just to go back to your previous point of why did they break certain trades and why didn’t they – they kind of drew a line as you said at 60%, anything above or beyond 60%, which was an arbitrary number where basically a bunch of exchange executives got together that night and said “this is where we’re going to break them.” Nobody really questioned it. Heck, if I had gotten caught in one of those trades, I’d be pretty upset right now. If I’d had a stop loss order out there and then there was, say, 40% that got kicked out, and the next day it opens flat or something like that. Hey, you got taken advantage of. And they blame the stop loss market orders, saying “you shouldn’t have a stop loss market order.” Well, if you’re trading IBM or Microsoft or GE or Proctor and Gamble, who would have thought that the stop loss would drop it 14 points? So that’s not the problem – don’t blame the victim – which is what they were trying to do here – you have to blame the system. That is the problem. And to the other point, yes they’ve taken over the market – the equity market in the U.S. is now controlled by the machines, there’s no doubt about it. They’re controlling all the volume. If you look at Citibank, it’s a perfect example of a stock that’s basically controlled by rebate traders. They look to collect a small fee, the exchangers will pay them for posting liquidity, and they’ll look to sell the stocks for exactly the same price, and they’ll collect a fee on the other side. It’s ridiculous – it’s volume for the sake of volume. It’s not what the market is supposed to be about.
Chris Martenson: So, what we have is these boxes that are trading back and forth, so someone’s written a computer program, and this program is not looking at fundamentals per se, but really just market structure, like how many shares are short or long at this point in time, or what’s the bid/ask spread at this point, or are there orders coming in from institutions that I can front run a little bit by the shares a millisecond in front of them and then sell them two milliseconds later. That’s what we’re talking about here, right.
Joe Saluzzi: Right, and as I said, there’s different types. There’s guys who are traditional statistical arbitrage traders, who add a value to the system. You need statistical arbitrage traders, they smooth out prices. If prices were to get out of whack in one market vs. another, there’s a statistical arbitrage opportunity there. They’ll take advantage of that, they’ll make a small profit, but they’ll keep profits in line. Traditionally, that’s a good thing, but now you have pretty much arbitrage for the sake of arbitrage. Sometimes they’re charging one ETF vs. another or one exchange vs. another. It’s crazy, but the part of the trade that concerns us is more the predatory style which you just mentioned, where if there’s an institution out there and they’re buying stock and someone can detect that order. Again, traditionally trading is predatory – there’s a winner, there’s a loser. If you’re better than me, you’re going to do a better job than me. If your machine works better you’re going to take advantage of it. However, if you’re getting an advantage that not everyone has, then you get situations for instance like flash orders. Flash orders were a big controversy probably during May of ’09. They still exist in the market, and a lot of people are confused because sometimes they call high frequency trading “flash trading”. Well that’s not the case. A flash order is a very specific thing where an exchange will show an order to a subset of their users and give them basically a half a second look or peek at the order before they will send it out to the general lit exchanges or the places that are displayed. Well that doesn’t make any sense – you’re giving the subset an advantage, and that person doesn’t have to trade against it – they can just say “No, thanks; I don’t want to sell you that; instead I’ll buy the stock ahead of you.” And they can do that because it’s not their client, so technically it’s not a front running thing. There’s one exchange in particular out there now (which I don’t want to name right now), and they’re still doing it, and the other exchanges are up in arms that they’re losing business to them, saying “How can you still do this?” The FTC has had the proposal to ban flash orders for over a year now and they still haven’t done anything about it. That’s just one particular example of predatory trading that goes on. There are all sorts of other things they can do particularly in some of these dark pools hat a lot of people talk about and that’s a lot of where you’ll see these sub-penny prices happening, where an order can be sent to an internalized specialist who’s internalizing and crossing those facts. Well in a market that’s tight and working well, they love doing that because it’s free money for them, but when things get tough… And by the way, as I say, on May 6th, one of the biggest problems, which didn’t get any headlines, was that internalizers could no longer internalize the trades. They didn’t execute them, they sent them to the lit market, they sent them to the exchanges, and they said “you fill this order, I don’t want it.” Well, the exchanges weren’t prepared for it, there was no liquidity which would normally be out there, and that exacerbated the problem even more, so if you have an internalizer who’s making a profit off of these trades, yet doesn’t have a responsibility to trade when they come into them, they get in the free look window with no obligation, and that’s the bottom line there.
Chris Martenson: So, if I got that right – the public market has these high frequency traders and all these algorithms running, but there’s a private market running as well, the dark pools. It’s an off exchange trading, so there’s some volume happening off exchange. We’re not really seeing it, but when a crisis came like on May 6th, some of that private volume came squishing over into the public markets, if I can put it that way.
Joe Saluzzi: In 2008, the percent of trades that were printed to the TRF (the trade reporting facility) where not off-exchange stuff happens was 14% of the overall volume; in 2010 that number went up to 32%. Now, that tells you that there’s a huge increase in the dark pools, and we always like to make a distinction between different types of dark pools, and this is where the industry sort of shot themselves in the foot, they had these things labeled as dark pools, and of course that sounds very predatory, and most of them, there’s not necessarily any value, but there’s three of them that are good, they’re institutional. They’re these crossing networks, and I won’t give all the names now, but a good institutional crossing network with an average daily trade size of 35,000 shares or more, they’re serving a purpose to help institutions find each other to reduce friction in the markets so they can execute trades back and forth. That’s what a crossing network is. Well, they’ve kind of morphed into these dark pools, which are just internalization engines, who are just looking at basically profit off of retail flow and other things. They’re driving out a lot of the volume off the exchange, which basically then takes the price discovery away from the market, because the lit market can no longer feel the supply and demand. It’s hurting price discovery, and that causes a huge problem.
Chris Martenson: I see that problem, but that increase of 14% to 32% over two years, to me that makes sense if – let’s play devil’s advocate – I’m Calpers, and I know there’s high frequency skimming going on and I don’t want to be part of that, wouldn’t I want to participate in dark pool transactions?
Joe Saluzzi: The thing is when we go back to traditional crossing networks and they see that it’s liquid then ITT and Pipeline – those are institutional crossing networks – they haven’t seen an uptick in volume in those two years, so it’s not their volume going up, so it’s not that the Calpers of the world and the large institutions are finding more liquidity in those networks, the volume is going up from the dark pool internalization engines, which – some of them are fine, and some of them are more predatory. So you have to be very careful when you’re trading for an institution, and again, that’s what we do here, we trade for institutional clients, as to “where is my order being sent?” If it’s being sent to a dark pool, which I consider toxic, which houses a lot of predatory order flow, and the reason (and here’s another little secret that no one talks about), the reason these orders are being sent to these exchanges and dark pools is because of the pricing scheme that goes on amongst broker routers. So when you send out an order as an institution, it gets routed through a smart order router. Well there’s a pecking order. It will try to go to the cheapest version first. And most of the time those are dark pools that are free, so the broker has an incentive to go to these toxic pools and send the orders there, so that way his cost is lower, and he can say “OK, here’s my algorithm, I’ll only charge you half a penny per share,” because he knows his cost on the other side is zero, and even now in some cases now they get rebated to trade with some of these exchanges who are no doing this inverted pricing scheme, which is a whole other story. So the order has a disincentive – you have to be very very careful as an institution where you’re sending your order. Where is it going and how is it being routed? This is really deep into the microstructural world of markets, but you have to be careful.
Chris Martenson: Well that sounds – I hate to sound naive about this – but that sounds enormously complicated and opaque and an ever-shifting landscape. It sounds like the technology has perhaps run faster than our ability to understand the impacts, the risks, to manage it appropriately. I mean, I’m thinking about how the FTC having trouble managing obvious things. Things that seem completely obvious to me, like options backdating. It doesn’t take a genius to understand that, and still they struggled with it. So this sounds like an order of magnitude more complex to me.
Joe Saluzzi: It is. And like you said, it’s constantly changing. If they do catch on to something, they just kind of shift it around again. The other side, the high frequency trading world and the automated side, is enormous, and the money that’s being made there is also enormous, so they have huge resources at their disposal. They can kind of shift things around, if a strategy doesn’t work, but who in the end is always the loser? It’s the institutional and retail investor, because their orders, in ever so small amounts are getting a little bit of siphoning every time, and that money is being sent into the other side of the high frequency trading world. Now, it is a strategy, there’s nothing illegal about it. I always make this point very clear. High frequency trading is not illegal, certain strategies are not illegal. But some of them morph – it’s a very gray area. The FTC needs to understand that the biggest problem they have (the FCC) is they don’t have a consolidated audit trail. It took them five months to analyze fifteen minutes of trades. Right away, there’s a red flag. What’s going on here? Why did it take you guys so long? Well they had to piece together various exchange fees and various audit trails (there’s one called FOATs, which is the FINRA Order Audit Trail) and that will track NASDAQ securities. That’s a good thing. Well we don’t have one for the New York, and we don’t have one for the options. So they had to kind of put it all together like a big quilt, where they patched it together. They need to get – and there’s a proposal out – a consolidated audit trail at their fingertips to see what’s going on, who’s doing it, why this happened. It would be able to track a heck of a lot more, but it’s an expense. Things are billed. If you saw yesterday, there was an article about how the FTC is complaining that they’re underfunded – they don’t have the resources anymore. If you’re a bad guy, and you know the cops aren’t watching that bank, are you going to rob that bank? Well most people would say no, I’m not a bad guy, it’s not ethical, I’m not going to rob a bank. Well guess what, there are a few people in this world who will say “I’m going to go rob that bank until they’re watching.” That’s what we have going on right now.
Chris Martenson: Well, you know, this helps put some things in context for me. For listeners out there who maybe aren’t as deep into the market, I just want to put one point of clarification in which is that we’re talking about trading, which is different from investing. Investing, you’re exposed to market returns, and that’s all well and good. You buy GM a decade ago and hopefully it goes, and that’s your investing return. But trading is a zero sum game as I understand it, so you’re loss is my gain and vice versa. So what we’re talking about here is in the dark pools and HFT world, most of this is just about trading and securing trading gains. So help me understand two things: TradeBot reports 40 years of consistent returns trading, I believe it’s a win every single day. I see the big trading desks, the prop desks off the big banks, they’re turning in perfect trading quarters. How does that fit in to this thing you’re talking about?
Joe Saluzzi: Right, so most retail investors will say the exact same thing – “Why should I care, this sounds like minutiae.” Well, if you are invested let’s just say in a 401K, and your 401K happens to own some mutual funds or ETFs or is somehow invested in the stock market, well then your fund, if they’re not extremely careful – and the things we’re talking about when they’re trading is subject to leakage in their performance. So maybe instead of earning 12% that year, they only earned 11.2% because they lost a few basis points on leakage, and guess where that money went? That went to the high speed traders who were able to scalp a penny here or there or collect a rebate here or there, and it’s kind of like the movie Office Space where the penny scheme adds up, right? A penny here, a penny there, well before you know it, that’s some real money we’re talking about. So if you are invested in the market, yeah you are subject to lose a little here, because it is a zero sum game, Chris, you hit it right on the head – somebody wins and somebody loses. How can an industry be making reportedly billions and billions of dollars? It’s got to be coming from somewhere. It’s an almost an invisible profit machine, nobody really sees it, but there is leakage on the trades – there’s no doubt about it. That’s where you have to be careful.
Chris Martenson: That almost seems like a license to steal to me. I was a trader for a while, it was something I did from my home – just a smalltime retail guy – and I know that a really good year for me, I’d be hitting it like 60%/65%. That would be phenomenal. In fact, I don’t think I had a year with 65% in it. But you know, you’re shooting for over 50% somewhere. But these guys are turning in 100% win ratios. The statistics on that say they’re not trading, not like I was anyway – they’re doing something very different than I was doing. Anything where you have a guaranteed profit is not a trade – nothing’s at risk – it’s a completely different world than I’m familiar with as a small time trader.
Joe Saluzzi: Right, and the reason why they’re guaranteed (and I guess there’s nothing guaranteed), but they invest a lot of money in technology. It’s all about technology with these guys. It is an arms race when it comes to who has the best speed and they used to talk about milliseconds up to two years ago – now they talk about microseconds, which are a millionth of a second. That’s the kind of latency we’re looking at. Whoever has the fastest speed computer will beat the other guy. They’re always racing against each other to try to beat the other guy, to keep the institutional order. There’s always one loser, like I said before, and it’s usually the retailer. But if you’re out there trying to trade, you’re a day trader, you’re not going to beat the machines. You will always be too slow. If you’re an investor, that’s a different story. You don’t have necessarily the five minute timeframe or ten minute timeframe. You found your Apple or whatever stocks you like, you think it’s a good investment and you did your research, and you did your own homework, well then these things that we’re talking about probably won’t affect you too much. But if you’re a trader, as a lot of guys are out there, well you need to be very careful. I bet that win ratio is a heck of a lot lower now for most professional day traders because they can’t fight the machine – they’re not going to win.
Chris Martenson: Yeah, this tracks my experience perfectly. In 2003 – 04 things were fine, but something began shifting in about 2005. What I used to be doing didn’t work, and by 2007 I had to quit because what I was doing was failing. My experience is I would see these massive orders swamp the bid at critical junctures for example, or, for both individual stocks or the whole index, I would see levels leap over key resistant points meaning that a limit fill was no longer good enough – sometimes they would cruise right past your limit so you’d have to have these stopgap market orders, and those were horrible. So I just found myself getting my pocket picked, and it was a little bit here or there, and I understood that I was up against something I didn’t understand, I couldn’t compete with – it went off. So I had to quit.
Joe Saluzzi: We hear that a lot from professional day traders – that things have changed, no longer do the technical patterns – that have lasted for years and years and are written about all over – really work anymore. So what a lot of guys have done (and I give them a lot of credit) is they’ve adapted, and they know now that they have to live in a different world, where maybe a traditional breakout where you see a stock break down or break out, it doesn’t work anymore. So they kind of go alongside now, and what we’re in right now is a momentum-fueled market, and things are a little bit different. You do have to adapt your trading strategy, but you’re fighting a battle which is very difficult to win, and just to give an example: the type of traders that are out there now used to be dominated by institutional volume and the block trade was a big deal. When you sold a block, the stocks would kind of move accordingly. Well the blocks are minuscule now compared to the overall volume. Most of the volume is 100-share prints to an average trade size of 5200 shares. That’s because the type of trades have changed. 70% are hyperspeed trades. Trades that could care less about what a company does for a living or who the CEO is; they just care about making a penny or two and flipping it and getting flat by the day’s end because nobody holds positions, of course. Well, this is what you’re up against. So if you can adapt, good luck, but there are not many people out there who can adapt without that type of weaponry. You need the machines, okay? And it’s not cheap to get into. You need millions and millions of dollars to build your own system, and then you’re still fighting up against the big banks, the large brokerage firms, the large HFT firms – you’re not going to win. So, it’s a very difficult game as a trader, and I’m glad you made that distinction, because investing is still investing.
Chris Martenson: Right, but at the same time, what I saw after the flash crash, which by the way, broke my confidence for two reasons: one that it happened and wasn’t well explained, and two that it broke the trades. I guess now I have to add a third thing which is from my perspective, I haven’t seen any significant fundamental rule changes, systems implementations. Am I wrong there, has something fundamentally changed here?
Joe Saluzzi: You are right on the money. What we’ve gotten are a couple of band-aids that were placed on it because obviously there was outrage afterwards, and one of the band-aids was the circuit breaker. Okay, so they put in single stock circuit breakers where if a stock moves 10% or more within a 5 minute time period, they will halt that stock. Well, guess what? That’s not going to help very much. Because a flash crash could happen in five minutes, and the market will be down ten percent, and that is pretty much what we had on May 6th. That’s an extremely damaging event. Like you just said – when you lose that confidence in the market, it doesn’t come back, which is why the equity mutual outflows from May 6th until recently were out every week. Money was coming out of the market, because investors didn’t trust the equity market anymore. Now they’re starting to come back in of course, and it feels toppy to me, but now money is coming in. You kind of get scared there, you think “Well why are these people buying now?” but when you lose confidence in an equity market, it just doesn’t build overnight. So what the FTC has to come in and said is “We’ve got some major overhauls here, we’re going to be taking care of making this fair.” But here’s the problem, it’s about money – it’s all about money – and it’s all about who has the bigger lobby and who has the bigger funding lobby, and right now, running around Washington DC are huge HFT lobbies who are talking to all the politicians saying “nothing’s wrong here, guys.” They’re like the traffic cop at the accident saying “Keep moving, folks, there’s nothing to see here,” and pretty much winning the battle. We used to have one lone senator who would help and try to have a voice of reason – Senator Kaufman from Delaware, whose term expired and he didn’t run again. Now there’s nobody out there for investor protection. And one of the other biggest problems is that the exchangers, who normally, most investors look to the exchangers as protectors for investors, and their number one goal is to make sure everyone is treated fairly. Well guess what? Exchanges are all for-profit organizations now. Not non-profit. Some of them are public – two of them are public. The number one goal of a for-profit corporation is to make money – to make sure their shareholders are happy. And if that means cutting corners and cutting deals with guys who normally they wouldn’t have done – they’re going to do it. That’s why they still do flash orders. That’s why they have inverted rebate schedules. That’s why they’re doing all sorts of different things to make sure they course the most volume. There’s our problem – we point to it all the time. We say high frequency trading is just a strategy being used by certain people, and it’s allowed, but it’s being given to them by the regulators and the exchangers who either don’t know or are looking the other way, and there is no reform going on right now, and it’s scary. Like you said before – what’s to prevent another flash crash here? There isn’t anything out there. But if you hear the exchange heads, they will confidently get up on TV and say “there will not be another flash crash?” Oh yeah, can you guarantee that? I’ll take the other side of that bet.
Chris Martenson: Well, I’d like to talk about the risks for a moment, because one of the other things that sort of drove me out of trading – I decided my time was better spent elsewhere – were not only some of the technical signals – other things breaking down and I was seeing weird market behavior. But I used to count on having some sort of a hedging in the cross and anti-correlations between