WEBVTT - Interview With Keith Ross: Masters in Business (Audio) 0:00:03.240 --> 0:00:07.600 This is Masters in Business with Barry Ridholts on Bloomberg Radio. 0:00:08.320 --> 0:00:13.440 This week on the podcast, I have an absolutely fascinating conversation. 0:00:13.640 --> 0:00:17.880 My special guest is Keith Ross. He is the current 0:00:17.960 --> 0:00:21.400 CEO and chairman of p d Q Enterprises, which is 0:00:21.960 --> 0:00:26.880 a really interesting innovative automated trading system and dark pool 0:00:27.520 --> 0:00:30.880 that has a novel approach to getting X. I think 0:00:30.920 --> 0:00:34.280 it's novel. I don't know enough about dark pools because 0:00:34.520 --> 0:00:37.680 much of them are dark but his description of how 0:00:37.800 --> 0:00:42.159 they sort of put the auction process UH in an 0:00:42.159 --> 0:00:46.120 inverted form to obtain the best price and execution for 0:00:46.200 --> 0:00:50.239 their clients is quite fascinating. He has been in the 0:00:50.280 --> 0:00:53.280 front lines of high frequency trading for a long time. 0:00:53.600 --> 0:00:56.840 He used to be the CEO of get Go, sold 0:00:56.880 --> 0:00:59.200 for about a billion for to night Trading a couple 0:00:59.200 --> 0:01:02.640 of years ago, one of the biggest h f t 0:01:02.920 --> 0:01:07.800 s by volume. If you are at all a student 0:01:07.880 --> 0:01:13.000 of market structure, if you're interested in institutional execution, dark pools, 0:01:13.240 --> 0:01:16.360 h f t S algorithms, if you're a critic of 0:01:16.560 --> 0:01:19.200 h f t s as well, you will find this 0:01:19.280 --> 0:01:23.240 to be an informative and fascinating conversation. So rather than 0:01:23.280 --> 0:01:27.640 have me babil incessantly without any further ado. Here is 0:01:27.720 --> 0:01:35.000 my conversation with Keith Ross. This is Masters in Business 0:01:35.080 --> 0:01:39.400 with Barry Ridholts on Bloomberg Radio. My special guest this 0:01:39.440 --> 0:01:43.360 week is Keith Ross. Here's the chairman and CEO of 0:01:43.480 --> 0:01:48.920 pd Q Enterprises, a trading and finance firm. Previously he 0:01:49.040 --> 0:01:52.120 was CEO of get go Uh, one of the largest 0:01:52.200 --> 0:01:55.800 high frequency trading shops. Keith Ross, Welcome to Bloomberg. Thank 0:01:55.840 --> 0:01:57.920 you very much, Barry. Pleasure to be here. So so 0:01:58.000 --> 0:02:01.560 let's keep start out simple and will build in complexity 0:02:01.600 --> 0:02:06.320 as we go. First off, what is high frequency algorithmic 0:02:06.360 --> 0:02:10.160 trading and why should people care about it? Well, high 0:02:10.160 --> 0:02:15.080 frequency trading is just regular trading by computers instead of 0:02:15.160 --> 0:02:20.000 humans clicking a mouse. Computers handle everything from identifying the 0:02:20.000 --> 0:02:24.480 trade to executing it precisely, and it turns out I 0:02:24.520 --> 0:02:27.200 believe it's a it's a wonderful efficiency for our markets. 0:02:27.320 --> 0:02:32.040 Trading costs and commissions have come down dramatically over the years. Um. 0:02:32.080 --> 0:02:34.440 I know there's a lot of concern about the robots 0:02:34.480 --> 0:02:39.519 taking over the world, but um, the vast majority of 0:02:39.600 --> 0:02:43.760 high frequency traders are passive liquidity providers. So they are 0:02:43.800 --> 0:02:46.400 in the market virtually all the time, and we can 0:02:46.440 --> 0:02:49.000 talk about that in a little bit. But their goal 0:02:49.240 --> 0:02:52.240 is to capture spread the way market makers have done 0:02:52.320 --> 0:02:56.520 for centuries, and provide liquidity. So here's the here's the 0:02:56.560 --> 0:03:00.800 pushback on liquidity as always well, they like the banku 0:03:00.840 --> 0:03:03.080 who will lend you the umbrella when it's sonny out. 0:03:03.360 --> 0:03:06.320 They provide liquidly when it's not needed. But when things 0:03:06.360 --> 0:03:08.400 get very well, they just shut off the machines and 0:03:08.440 --> 0:03:11.440 step away from the market. So I want to respond 0:03:11.480 --> 0:03:14.800 to that and say it's actually a myth. Because a 0:03:14.919 --> 0:03:19.639 couple of folks were out loud about backing away from 0:03:19.639 --> 0:03:22.200 the market. The guys who knew what they were doing 0:03:22.240 --> 0:03:27.040 did not, And my understanding is they had very good days. 0:03:27.360 --> 0:03:29.320 So there's actually an incentive for them to stay in 0:03:29.360 --> 0:03:32.000 the market. With all that volatility and chaos, there's a 0:03:32.040 --> 0:03:36.640 lot of opportunity um and you know, those that are 0:03:36.680 --> 0:03:40.440 not as concerned about that would argue. And at other 0:03:40.480 --> 0:03:45.000 times in market chaos, market makers specialists didn't pick up 0:03:45.040 --> 0:03:48.160 the phone. That's the same problem. So you couldn't get 0:03:48.200 --> 0:03:51.440 to your broker. Maybe you wanted thought you wanted to 0:03:51.480 --> 0:03:53.880 do something, but you couldn't find anyone to take the 0:03:53.920 --> 0:03:57.840 other side. Again, I'll take the other side. Seven day 0:03:57.960 --> 0:04:02.240 was really a one off down. No one could get 0:04:02.280 --> 0:04:05.960 through to anybody. Let's talk about the flash crash, which 0:04:06.000 --> 0:04:09.320 is the modern version. So May ten, two thousand and ten, 0:04:09.400 --> 0:04:15.080 we saw a seven point practically instantaneous drop. What caused that? 0:04:15.240 --> 0:04:18.679 A lot of people insisted on blaming five frequency traders 0:04:18.680 --> 0:04:21.080 for that. Well, they're going to be the scapegoat because 0:04:21.240 --> 0:04:23.360 like the specialist in the old days, they're the ones 0:04:23.400 --> 0:04:27.919 that everyone believes are the operators in the market um 0:04:28.000 --> 0:04:31.000 and have that opportunity to either make the market or not. 0:04:31.880 --> 0:04:37.520 Both cases. The macro environment was very, very bearish, and 0:04:37.680 --> 0:04:41.880 so the question becomes, you know, what is the responsibility 0:04:41.920 --> 0:04:47.880 of the participants? Uh, they took time off and withdrew dramatically. 0:04:48.520 --> 0:04:51.840 The beautiful part about the flash crash to me is 0:04:51.880 --> 0:04:56.240 the following flash dash after a five second pause in 0:04:56.320 --> 0:05:01.680 the many futures, the whole market was able to recalibrate, reassess, 0:05:02.160 --> 0:05:04.200 and it was it was a hard event during the day, 0:05:04.240 --> 0:05:06.960 but it was over in an hour. It's funny because 0:05:07.000 --> 0:05:10.040 I was flying back from Dallas that day or somewhere 0:05:10.080 --> 0:05:14.200 in Texas, and I look at my screen when I landed, 0:05:14.240 --> 0:05:16.360 I'm like, that's got to be a mistake. And it 0:05:16.400 --> 0:05:18.839 was until I got in the cab and started speaking 0:05:18.839 --> 0:05:21.120 to my office that I found out what happened. We 0:05:21.120 --> 0:05:24.280 were down a thousand points. We closed down three, explained 0:05:24.320 --> 0:05:29.480 to me. So so that that's the big criticism UM 0:05:29.560 --> 0:05:31.920 in general. There are other criticism will get to later, 0:05:31.960 --> 0:05:33.839 but I want to set the table a little bit 0:05:34.360 --> 0:05:37.640 with what you currently do in some quotes of yours. 0:05:37.680 --> 0:05:40.000 Let me let me just start asking what is p 0:05:40.160 --> 0:05:43.840 d Q Enterprises and PDQ Trading actually do. So. P 0:05:43.960 --> 0:05:46.760 d Q Enterprises is the parent of p d Q 0:05:46.960 --> 0:05:51.840 A t S are broker dealer. We are a probably 0:05:51.839 --> 0:05:54.599 considered a dark pool, even though we believe we operate 0:05:54.720 --> 0:05:57.719 differently than all the others. We have a unique market structure. 0:05:58.560 --> 0:06:02.919 We can actually create liquidity on demand for orders, we 0:06:03.000 --> 0:06:08.400 can consolidate fragmented liquidity. And the beauty of our auction process, 0:06:08.440 --> 0:06:10.359 which is kind of the key aspect of why we 0:06:10.400 --> 0:06:14.640 are different, is that we can aggregate people with different 0:06:14.760 --> 0:06:19.880 time latency sensitivity. So whether the people who are either 0:06:20.040 --> 0:06:23.159 very high speed or high speed or normal speed correct 0:06:23.480 --> 0:06:28.039 all right, and how is PDQ different than your previous employer. 0:06:28.760 --> 0:06:31.240 Get Go, who is not too long ago sold to 0:06:31.920 --> 0:06:34.480 Night Trading, and at the time I think they still 0:06:34.520 --> 0:06:37.039 are one of the biggest h f T shops in 0:06:37.080 --> 0:06:39.400 the world. I believe they're one of the biggest traders. 0:06:39.400 --> 0:06:43.560 They did merge with night Um. I went from the 0:06:43.600 --> 0:06:46.560 liquidity providing side, which we were doing it get Go 0:06:47.279 --> 0:06:51.680 when I saw the PDQ opportunity. So the way PDQ works, 0:06:51.720 --> 0:06:54.479 an order comes to us and we pause it for 0:06:54.680 --> 0:06:58.680 up to twenty milliseconds. That sounds vaguely familiar, but in 0:06:58.760 --> 0:07:02.040 that time we do something very constructive for the order. 0:07:02.480 --> 0:07:05.680 What we do is we ping our liquidity providers with 0:07:05.800 --> 0:07:09.720 only the symbol and effectively bring back the question that 0:07:09.880 --> 0:07:12.160 used to be asked at the post of what is 0:07:12.200 --> 0:07:14.360 the market? So in other words, you don't say I'm 0:07:14.400 --> 0:07:17.360 buying a hundred thousands of selling a hundred thousand. You say, 0:07:17.400 --> 0:07:20.320 here's the symbol, what's your price, what's your best market? 0:07:20.800 --> 0:07:24.000 We have a dozen providers that are all responding. We 0:07:24.040 --> 0:07:29.160 can also include institutional passive liquidity providers, and we build 0:07:29.160 --> 0:07:32.280 a book and once we have the book, we take 0:07:32.320 --> 0:07:35.480 the instructions from our client to execute the order against 0:07:35.480 --> 0:07:38.880 the book. So we create competition for every order, and 0:07:39.600 --> 0:07:45.400 in this pause we're in position to take speed out 0:07:45.400 --> 0:07:48.320 of the equation. I'm Barry Ridults. You're listening to Masters 0:07:48.320 --> 0:07:51.600 in Business on Bloomberg Radio. My special guest today is 0:07:51.760 --> 0:07:56.600 Keith Ross. He is the CEO of PQ Enterprises, a 0:07:56.760 --> 0:08:00.800 rather interesting trading shop. Dark Pool goes by a number 0:08:00.800 --> 0:08:03.760 of different names. Keith, I want to start with a 0:08:03.880 --> 0:08:07.160 quote of yours and have you comment on it, and 0:08:07.240 --> 0:08:10.640 it goes something like this, there is nothing wrong with 0:08:10.720 --> 0:08:14.800 speed in any other industry. It's applauded. It's an efficiency 0:08:15.280 --> 0:08:20.040 discuss well. UM. One of my favorite examples is your telephone. 0:08:21.040 --> 0:08:23.840 Back in the day, we had telephone operators pulling cables 0:08:23.880 --> 0:08:27.320 and plugging them into the wall to connect you to 0:08:27.880 --> 0:08:31.400 your call. And today we have cell phones that are 0:08:31.440 --> 0:08:33.760 many computers and you can do everything on your phone. 0:08:34.360 --> 0:08:39.880 Incredible technology advancement proficiency. I haven't heard in the early 0:08:39.960 --> 0:08:43.959 days of of UM cell phones there were connection issues, 0:08:43.960 --> 0:08:46.280 but those have gotten so much better. You know, it's 0:08:46.320 --> 0:08:50.360 all about how many megabytes of upload and download and 0:08:50.400 --> 0:08:53.600 so forth. I believe the same things happened for trading, 0:08:54.280 --> 0:08:58.520 so spreads have gotten tighter, commissions have come down dramatically, 0:08:58.640 --> 0:09:02.960 the cost of execution and has never been lower. Um 0:09:03.000 --> 0:09:04.839 I think it was. Vanguard in two thousand and ten 0:09:04.960 --> 0:09:09.480 said our portfolios due to electronic trading have a one 0:09:09.559 --> 0:09:14.400 percent per annum increase in return. Given the fact that 0:09:14.400 --> 0:09:18.360 they're pension problems across our whole country. Anyone that's involved 0:09:18.400 --> 0:09:21.720 with the retirement programmer has pension money working in the 0:09:21.760 --> 0:09:26.640 marketplace is a beneficiary of a more efficient market. Jack Brennan, 0:09:26.679 --> 0:09:31.320 who's now chairman emeritus of Vanguard, when we spoke with him, 0:09:31.440 --> 0:09:35.439 he said something along those lines. But let me give 0:09:35.440 --> 0:09:38.559 you the pushback from the other side. Hey, h f 0:09:38.640 --> 0:09:41.120 T firms are taking out billions of dollars a year 0:09:41.160 --> 0:09:45.880 in profit. Trading is essentially a zero sum game. That 0:09:46.440 --> 0:09:49.320 profit has to be coming from somewhere, and it's coming 0:09:49.320 --> 0:09:54.600 from grandma's retirement account. Would actually vehemently disagree on almost all. 0:09:54.640 --> 0:09:58.319 I would hope you would disagree. Tell me why that 0:09:59.240 --> 0:10:03.600 description is an error, and I sort of have a 0:10:03.640 --> 0:10:07.360 foot in that camp that that makes some sense to me. Okay, 0:10:07.400 --> 0:10:11.520 first of all, let's start with the zero sum game problem. 0:10:11.559 --> 0:10:15.319 The capital markets are not zero sum the capital markets 0:10:15.320 --> 0:10:18.800 can create wealth and wealth can be destroyed in them. 0:10:18.840 --> 0:10:21.800 So I have a good friend, what's a thousand shares 0:10:21.800 --> 0:10:26.000 of apple for each of his kids? He sold it 0:10:26.160 --> 0:10:30.839 in two thousand and ten, he made times his money 0:10:31.440 --> 0:10:35.320 wealth creation. Anyone who's bought an apple product knows apple 0:10:35.360 --> 0:10:39.640 stock has gone to the moon. So that's not zero sum. 0:10:39.679 --> 0:10:42.320 But what happens if he goes to sell it and 0:10:42.400 --> 0:10:45.079 instead of selling that at a hundred one dollars and 0:10:45.160 --> 0:10:48.160 fifty cents, he only gets a hundred one dollars and 0:10:48.640 --> 0:10:52.199 five cents because someone jumped in the middle of his execution. 0:10:52.720 --> 0:10:57.960 So there could be that could be an issue. I 0:10:58.040 --> 0:11:00.560 don't think that it is, because in the old days 0:11:00.679 --> 0:11:03.600 it would have been a twenty five or fifty cent bid, 0:11:04.160 --> 0:11:07.439 a nickel much bigger. So you're talking commission, not spread. 0:11:07.559 --> 0:11:10.080 So in other words, the tradeoff is it used to 0:11:10.120 --> 0:11:14.600 be higher commissions and wider spreads, and now the spreads 0:11:14.600 --> 0:11:17.959 are tighter, the commissions are smaller, and if there's an 0:11:17.960 --> 0:11:22.440 execution charge that finds its way into this net net, 0:11:22.760 --> 0:11:26.160 it's still better for the UH traders that retail has 0:11:26.200 --> 0:11:29.640 never had it better. Bill McNabb, current chairman and CEO 0:11:29.840 --> 0:11:33.480 Vanguard and also a former guest on Mester's business. He 0:11:33.640 --> 0:11:39.360 said something very similar, and I was more vehemently questioning 0:11:39.600 --> 0:11:43.240 when I keep reading and hearing about market structure, and 0:11:43.240 --> 0:11:46.800 when I get a guy like McNabb saying, hey, you 0:11:46.840 --> 0:11:50.120 know this has made our trading costs lower, it forced 0:11:50.120 --> 0:11:54.120 me to at least rethink, well, now I have to 0:11:54.160 --> 0:11:56.440 take this with a grain of salt because he's got 0:11:56.520 --> 0:12:00.360 zero incentive to say anything. But what what is true 0:12:00.480 --> 0:12:03.480 true for Vanguard? So let's move to the other issue 0:12:03.600 --> 0:12:06.439 I always hear about, which is the front running. If 0:12:06.480 --> 0:12:10.360 somebody sniffs out what my order is and rather than 0:12:10.480 --> 0:12:12.960 executing that order, runs ahead of me, buys it and 0:12:12.960 --> 0:12:15.640 sells it back to me for whatever, and if we're 0:12:15.720 --> 0:12:20.480 petting higher cunitively, isn't that a huge drag on total 0:12:20.480 --> 0:12:24.440 returns for everybody? In the end, it's incredibly small relative 0:12:24.480 --> 0:12:27.560 to the markets to begin with. Second of all, the 0:12:27.679 --> 0:12:31.280 person doing that, the person the HFT that takes out 0:12:31.320 --> 0:12:34.280 the penny up offer, that's some price improvement for that guy, 0:12:34.760 --> 0:12:38.040 so he's a happy guy. The seller is paying a penny, 0:12:38.160 --> 0:12:40.760 is getting a penny more than they might have. I'm 0:12:40.800 --> 0:12:44.240 just saying that's certainly a possibility that's never talked about, 0:12:44.280 --> 0:12:49.520 that it's price improvement for a contra party. Certainly, the 0:12:49.600 --> 0:12:53.320 sniffing ability is only an educated guess, just the way 0:12:53.320 --> 0:12:55.840 people used to do in the market back in the day. 0:12:55.960 --> 0:12:59.720 So they can't see the order. That's a that's a myth. 0:13:00.760 --> 0:13:05.080 And the models understand supply and demand and they're trying 0:13:05.080 --> 0:13:07.960 to find that balance between supply and demand all the time. 0:13:08.000 --> 0:13:10.840 All of a sudden, if there's a big demand, the 0:13:10.880 --> 0:13:17.720 supply is gonna move their price higher. So, um, it's 0:13:17.840 --> 0:13:20.920 much ado about nothing in my opinion. And what I'll 0:13:20.920 --> 0:13:23.680 do is, I'll say this, most of the people that 0:13:23.760 --> 0:13:29.280 talk about that don't understand that liquidity provision is a service. Okay, 0:13:29.480 --> 0:13:34.160 So there to me, there's this mythical idea that when 0:13:34.160 --> 0:13:37.800 a trade takes place, the buyer and seller have this 0:13:38.280 --> 0:13:43.040 come bay, let's hold hands moment and exchange our whatever 0:13:43.240 --> 0:13:46.120 security is at whatever the price is, which is fine 0:13:46.160 --> 0:13:48.840 at the moment of the trade, but one tick later 0:13:49.240 --> 0:13:52.200 one of the parties has been disadvantaged. And when I 0:13:52.240 --> 0:13:55.360 was on the floor, used to trade with people and 0:13:55.440 --> 0:13:58.000 you'd confirm and you'd smile, and you'd say thank you, 0:13:58.280 --> 0:14:00.440 and then one second later you'd say, that's of a gun, 0:14:00.440 --> 0:14:03.240 he just picked me off, right, because the tick went 0:14:03.320 --> 0:14:05.880 agains well, But that's after the fact. You know, once 0:14:05.920 --> 0:14:08.720 the market moves either in your favor against you. Hey, 0:14:08.760 --> 0:14:11.000 that transaction is done, you can't blame the guy that 0:14:11.040 --> 0:14:13.959 you bought something you shouldn't or sold something too early. 0:14:14.080 --> 0:14:18.440 That it's it's you know, after the transaction is done. 0:14:18.960 --> 0:14:22.560 It's pre transaction or during the transaction that I think 0:14:22.600 --> 0:14:26.960 some people are are complaining about. Certainly UM. The counter 0:14:27.120 --> 0:14:32.440 argument to that was explained extremely well in Uh, a 0:14:32.520 --> 0:14:36.400 book by Covic. I believe Peter Kovic uh Flashboys Not 0:14:36.440 --> 0:14:40.960 so Fast. It was the the antidote to Michael Lewis. 0:14:41.160 --> 0:14:43.920 I'm Barry Ridhults. You're listening to Masters in Business on 0:14:43.960 --> 0:14:47.840 Bloomberg Radio. My special guest today is Keith Ross. He 0:14:48.160 --> 0:14:53.720 is the CEO of PDQ Enterprises, a trading dark pool 0:14:53.760 --> 0:14:58.400 as well as a broker dealer that Correct and Ats 0:14:58.640 --> 0:15:02.040 was formally CEO get Go, which is now part of 0:15:02.160 --> 0:15:05.280 Night Training and is one of the largest h f 0:15:05.360 --> 0:15:09.320 T s in the world. Let's talk about Michael Lewis's 0:15:09.360 --> 0:15:13.480 book Flashboys, which I read on vacation and found to 0:15:13.520 --> 0:15:15.840 be a couple of summers ago and found to be 0:15:16.320 --> 0:15:20.560 an entertaining romp that raised all sorts of questions about 0:15:20.720 --> 0:15:25.160 market structure. But I have heard that you didn't think 0:15:25.160 --> 0:15:28.840 it was especially accurate. Well, I actually wrote a review 0:15:29.000 --> 0:15:32.120 of the book the day after it came out, so 0:15:32.160 --> 0:15:35.080 I got it on mike Kindle and read it electronically overnight. 0:15:36.480 --> 0:15:40.400 Fascinating story. I'm a Michael Lewis fan. He writes great novels. 0:15:40.920 --> 0:15:44.680 But it's a novel. So the problem, the issue that 0:15:44.760 --> 0:15:47.960 I have is it's the story of Brad Katsiama, which 0:15:47.960 --> 0:15:51.520 is a fun story, terrific guy, really another guest on 0:15:51.560 --> 0:15:55.800 the show, absolutely and endlessly entertaining story is as told 0:15:55.840 --> 0:15:59.200 by Michael Lewis, and it helped me dramatically. There's a 0:15:59.240 --> 0:16:04.720 YouTube interview between Michael um glad Well. It's not Michael Gladwell. 0:16:05.880 --> 0:16:09.720 Malcolm Gladwell and Michael was a seventy minute love fest 0:16:09.800 --> 0:16:12.400 between the two of them. And what helped me the 0:16:12.440 --> 0:16:14.720 most was when Michael Lewis explained what I do is 0:16:14.760 --> 0:16:16.960 I find a person I'm interested in and I write 0:16:17.000 --> 0:16:19.600 their story, which is exactly what he did in the book. 0:16:20.360 --> 0:16:24.760 My issue is as a treatise on market structure, he 0:16:24.800 --> 0:16:27.720 never talked to any of the exchanges, any of the 0:16:27.760 --> 0:16:30.400 other a t s s, any of the high frequency traders, 0:16:30.840 --> 0:16:34.600 any of the electronic algorithm people that are in the business. 0:16:35.080 --> 0:16:40.240 So it's one view from one person, and I don't 0:16:40.280 --> 0:16:44.280 think it's a whole picture of the marketplace. So Lewis 0:16:44.320 --> 0:16:47.280 and Katsuyam were on sixty minutes, came out and said 0:16:47.560 --> 0:16:52.360 markets are rigged. Yes, the markets are rigged, and it 0:16:52.360 --> 0:16:55.800 it's a perfect headline. It's the perfect marketing way to 0:16:55.840 --> 0:16:58.520 get your book sold to hundreds of thousands of people. 0:16:59.320 --> 0:17:02.920 The issue is if you listen or read the book, 0:17:03.480 --> 0:17:07.000 what they really said was it's possible that the execution 0:17:07.040 --> 0:17:10.080 of an order in the marketplace is rigged. Later on, 0:17:10.080 --> 0:17:12.720 when Michael Lewis was asked you invest in the stock market, 0:17:13.000 --> 0:17:16.280 he says, oh, yeah, I've on dtfs for years and years. 0:17:16.320 --> 0:17:19.720 My family has always been an investor. So when the 0:17:19.760 --> 0:17:23.600 market is rigged, that headline screams it's a three card 0:17:23.640 --> 0:17:27.480 Monty game, there's no way you can win. Absolutely untrue. 0:17:28.080 --> 0:17:31.240 By your apple twenty years ago, you're a happy guy today. 0:17:31.400 --> 0:17:37.200 So that nuance is totally lost. In a sixty minutes piece, 0:17:37.880 --> 0:17:40.840 the world was in a tizzy. It appeared that the 0:17:41.200 --> 0:17:43.199 high frequency guys were making a lot of money. They 0:17:43.240 --> 0:17:48.200 had a two, but they're not making that much money anymore. 0:17:48.560 --> 0:17:51.840 Volumes have just completely dried up. And volumes have dried up. 0:17:52.200 --> 0:17:54.639 Let me share a quick story around efficiency that I 0:17:54.720 --> 0:17:56.920 like to tell. If you buy or sell a house, 0:17:57.240 --> 0:18:00.200 five percent commission. When you go to Starbucks and buy 0:18:00.240 --> 0:18:03.560 a latte with your credit card, you pay Master Carter 0:18:03.720 --> 0:18:07.560 visa three percent commission on a three dollar latte, So 0:18:07.600 --> 0:18:13.320 that's nine cents. If you trade three shares of Starbucks 0:18:14.040 --> 0:18:17.680 and you're an electronic market maker, they make about three 0:18:17.720 --> 0:18:21.280 cents per hundred, so they're gonna make the same ninth 0:18:21.280 --> 0:18:26.399 sense that you're gladly play Visa for your latte. And 0:18:26.480 --> 0:18:29.919 yet they're going to trade dollars worth of stock and 0:18:29.960 --> 0:18:33.640 take risk while they're doing the trade. Their margins are 0:18:33.760 --> 0:18:40.800 so infinitesimally small. There's no other capital transaction anywhere that 0:18:40.920 --> 0:18:45.400 comes close to that type of efficiency. So my colleague 0:18:45.440 --> 0:18:48.920 Josh Brown likes to say the people complain that high 0:18:48.960 --> 0:18:53.160 frequency trading uh and the markets are rigged. His responses 0:18:53.200 --> 0:18:55.760 the markets have always been rigged. There's always been a specialist. 0:18:55.760 --> 0:18:58.560 There's always been somebody taking the pound of flash. At 0:18:58.600 --> 0:19:01.879 least today we know who would is. Again, I'll go 0:19:01.920 --> 0:19:06.040 back to my comment earlier that liquidity provision is a service. 0:19:06.280 --> 0:19:08.160 If you want to be able to trade and transfer 0:19:08.280 --> 0:19:11.240 risk whenever you want, someone needs to stand there and 0:19:11.280 --> 0:19:14.760 take the other side. The way money has always been 0:19:14.760 --> 0:19:17.639 made in the markets is to capture spread between the 0:19:17.640 --> 0:19:21.800 bid and the ask. That's done very quickly now and 0:19:22.720 --> 0:19:27.240 very scientifically. So it's not that the game has changed, 0:19:27.280 --> 0:19:29.720 just the players have changed. So in the last thirty 0:19:29.760 --> 0:19:32.000 seconds or so that we have what do you think 0:19:32.200 --> 0:19:37.600 of I e X that's the shop that Bradcotts Yam 0:19:37.640 --> 0:19:39.840 opened up. What do you think of their proposal to 0:19:39.880 --> 0:19:44.760 actually becommon exchange as an A T S. I applaud 0:19:44.840 --> 0:19:49.240 them and appreciate their innovation as an exchange, their protected 0:19:49.320 --> 0:19:54.159 quote hidden behind the delay could be very problematic. And 0:19:54.240 --> 0:19:57.520 my concern is other exchanges imitate I e X if 0:19:57.520 --> 0:20:02.159 they start to get market share, and we would have 0:20:02.200 --> 0:20:05.200 the situation of quotes that are not valid because they're 0:20:05.200 --> 0:20:07.920 being canceled, but no one knows because the cancel is 0:20:07.960 --> 0:20:11.000 going through the speed bump. This is Masters in Business 0:20:11.119 --> 0:20:15.080 on Bloomberg Radio. I'm Barry Ridhalts. My special guest today 0:20:15.240 --> 0:20:19.480 is Keith Ross. He is the CEO of PDQ Enterprises 0:20:20.040 --> 0:20:23.680 on a T S dark Pool, And there are all 0:20:23.720 --> 0:20:29.080 sorts of other technical terms we can discuss. So earlier 0:20:29.080 --> 0:20:32.960 we were talking about how market structure has changed. One 0:20:33.000 --> 0:20:36.879 of the claims about issues with high frequency trading is 0:20:36.920 --> 0:20:42.159 that it has introduced new systemic risk into the markets. 0:20:42.200 --> 0:20:47.520 True or false, it's got to be true. Systemic risk 0:20:47.600 --> 0:20:52.120 is very hard to understand and know. UM. But there's 0:20:52.160 --> 0:20:55.479 always been systemic risk, So it's not as though it 0:20:55.480 --> 0:20:58.800 didn't exist before. Electronic trades just different today and it's 0:20:58.840 --> 0:21:03.560 just a different fashion. UM. I think the regulators actually 0:21:03.720 --> 0:21:07.159 actually after the flash crash with UH some of the 0:21:07.200 --> 0:21:10.320 things they put in place limit up, limit down, UM 0:21:10.359 --> 0:21:13.320 attempting to get the venues to synchronize with each other. 0:21:14.200 --> 0:21:17.440 UM are certainly going to go help us move forward 0:21:17.560 --> 0:21:25.840 to create confidence in the market structure. UM, but systemic risk. Interestingly, 0:21:25.880 --> 0:21:30.159 to me, during the crash, none of the listed and 0:21:30.280 --> 0:21:34.480 cleared futures options or stock exchanges had any issues. It 0:21:34.560 --> 0:21:36.880 was only the over the counter trades where there wasn't 0:21:36.920 --> 0:21:41.320 a central counterparty. The central counterparty is the to me, 0:21:41.440 --> 0:21:47.080 the lynchpin of creating eliminating systemic risk, and Dodd Frank 0:21:47.560 --> 0:21:50.920 all seven hundred whatever pages of it should be one page. 0:21:51.240 --> 0:21:54.320 Everyone has to centrally clear their trades, no matter what 0:21:54.400 --> 0:21:58.199 they are, through some exchange, through an exchange, or a 0:21:58.200 --> 0:22:01.360 third party, because then the third party manages the risk 0:22:01.920 --> 0:22:04.280 and and they know where all the risk is from 0:22:04.280 --> 0:22:10.000 all the participants. That's that's interesting. So let's talk a 0:22:10.040 --> 0:22:14.320 little bit about that structure in some um greater detail. 0:22:15.520 --> 0:22:18.439 At one point in our history, the exchanges, the New 0:22:18.480 --> 0:22:23.360 York Stock Exchanged, other exchanges were really almost public utilities. 0:22:23.359 --> 0:22:26.800 They were not for profits. We eventually run into trouble 0:22:26.880 --> 0:22:30.399 with compensation and other issues under Dick Rasso. There was 0:22:30.440 --> 0:22:34.479 that whole digression. What should the proper role of the 0:22:34.520 --> 0:22:40.200 stock exchange be in modern trading? Well, let's it should 0:22:40.200 --> 0:22:45.600 be risk transfer among consenting parties. And so besides the 0:22:45.680 --> 0:22:48.119 risk transfer process, which is the matching of the trade, 0:22:48.520 --> 0:22:50.840 you also want to know that it's going to clear properly. 0:22:51.240 --> 0:22:53.200 If you're a buyer, you're going to get your securities. 0:22:53.240 --> 0:22:55.400 If you're a seller, you're gonna get your money. So 0:22:55.480 --> 0:22:58.520 it's the first step in a chain of several events 0:22:59.080 --> 0:23:03.760 that worked lawlessly through the financial crisis. For the list 0:23:03.800 --> 0:23:06.960 that exchanges, the the issue was in the over the 0:23:07.040 --> 0:23:11.880 counter credit swaps and some of those issues where everyone 0:23:11.960 --> 0:23:16.760 became suspicious of their counterparties required more margin, and companies 0:23:16.800 --> 0:23:19.240 that had tens of billions of dollars two weeks later, 0:23:19.720 --> 0:23:22.840 we're needing to be rescued. So when I was coming up, 0:23:22.920 --> 0:23:26.480 we had the AMMAX, the Nasdack, the New York Stock exchanged. 0:23:26.520 --> 0:23:31.680 Today there's twelve exchanges, almost fifty dark pools, maybe even more. 0:23:31.720 --> 0:23:33.919 I have you lose track of forty is probably a 0:23:33.920 --> 0:23:39.560 better Okay, so forty plus dark pools. What about this fragmentation? 0:23:39.720 --> 0:23:44.000 What does this mean for getting orders executed? I've spoken 0:23:44.000 --> 0:23:46.800 to traders who have complained they go out to buy 0:23:46.920 --> 0:23:49.359 five or ten thousand shares. I don't need a hundred 0:23:49.400 --> 0:23:51.800 or five hundred, and they get forty seven here and 0:23:51.840 --> 0:23:54.760 a hundred twenty two there, and it takes a dozens 0:23:54.760 --> 0:23:56.840 of orders to get what used to be a simple 0:23:57.200 --> 0:24:01.639 transaction filled. What is this fragmentation and mean for for 0:24:02.000 --> 0:24:08.240 execution and structure? So too. Two parts to that answer. UM, 0:24:08.359 --> 0:24:11.399 One is it actually is redundancy if you look at 0:24:11.440 --> 0:24:14.119 it a different way. So a little while ago, this 0:24:14.280 --> 0:24:17.520 New York socket exchange was down for four hours. No 0:24:17.600 --> 0:24:20.720 one even new right, So that's actually a good thing. 0:24:20.920 --> 0:24:25.000 So you know, we get a start for that. Um 0:24:25.160 --> 0:24:29.440 the difficulty with the fragmentation. So the generic answer would 0:24:29.440 --> 0:24:32.919 be innovation. Innovation has come through a t S S 0:24:33.240 --> 0:24:37.280 has forced the stock exchanges to compete with each other again, 0:24:37.320 --> 0:24:40.080 has driven costs down, taking a lot of people out 0:24:40.119 --> 0:24:42.080 of the equation. There used to be many more bodies 0:24:42.119 --> 0:24:46.760 on the floor and on trading. Really it's it's going traumatically. 0:24:47.280 --> 0:24:48.840 It used to be like eight rooms. I think they're 0:24:48.880 --> 0:24:52.439 down to one room. It's a it's a that's right, 0:24:52.480 --> 0:24:56.639 it's a it's a television front. But here's how you 0:24:56.680 --> 0:25:02.120 solve that problem. If you have a structural pause, that 0:25:02.200 --> 0:25:05.480 can do good things for the order at PDQ, we 0:25:05.560 --> 0:25:09.720 can consolidate orders from other venues. So let's talk about that. 0:25:09.880 --> 0:25:13.600 So the complaint that some people have raised about I 0:25:14.080 --> 0:25:17.680 X CATSI Yamas Shop is that they introduced the speed bump. 0:25:17.760 --> 0:25:21.960 It's eight miles of coiled fiber. Optiction is right right, 0:25:22.000 --> 0:25:24.720 So that that introduces I'm doing the soft top of 0:25:24.720 --> 0:25:28.560