WEBVTT - How to Succeed at Multi-Strategy Hedge Funds 0:00:03.120 --> 0:00:16.120 Bloomberg Audio Studios, Podcasts, Radio News. 0:00:20.680 --> 0:00:24.000 Hello and welcome to another episode of the Odd Thoughts podcast. 0:00:24.079 --> 0:00:25.400 I'm Tracy Alloway. 0:00:25.200 --> 0:00:26.360 And I'm Joe Wisenthal. 0:00:26.800 --> 0:00:31.000 Joe, I know we did one episode on pod shops, Yeah, 0:00:31.040 --> 0:00:35.800 on multi strategy hedge funds, but it was primarily focused 0:00:35.880 --> 0:00:38.640 on their impact on the market, and I have to 0:00:38.680 --> 0:00:42.040 say I still came away from that conversation sort of 0:00:42.080 --> 0:00:45.920 wondering if I worked at a pod shop, what is 0:00:45.960 --> 0:00:48.240 it exactly that I would be doing all day? 0:00:49.200 --> 0:00:52.080 I would love to know the exact same thing. I mean, like, 0:00:52.120 --> 0:00:55.000 I guess I have this very vague sense of sort 0:00:55.040 --> 0:00:57.760 of they have a bunch of people all focused on 0:00:57.840 --> 0:01:00.600 their specific areas in the sort of ever out, and 0:01:00.640 --> 0:01:02.680 they net out a bunch of stuff and it's capital 0:01:02.720 --> 0:01:07.480 efficient and you know, it's like market neutral and theory 0:01:07.600 --> 0:01:11.080 and et cetera. But beyond that, like, I still don't 0:01:11.120 --> 0:01:13.160 really like understand. The only thing I know is like 0:01:13.240 --> 0:01:16.119 they've done really well and many people are launching more 0:01:16.120 --> 0:01:16.399 of them. 0:01:16.720 --> 0:01:18.959 Yes, they seem to be all the rage. They seem 0:01:19.000 --> 0:01:21.920 to be where everyone kind of wants to go in 0:01:22.040 --> 0:01:25.800 the quantitative finance space. At least everyone's sort of aiming 0:01:26.040 --> 0:01:30.679 for these big names, you know, places like Citadel Millennium. 0:01:30.760 --> 0:01:34.600 Maybe yeah, But my question is like, why was it 0:01:34.640 --> 0:01:37.840 just that they're minting money they're expected to continue minting 0:01:37.880 --> 0:01:40.319 money in the future, or is there something that's like 0:01:40.480 --> 0:01:45.560 fundamentally intriguing and attractive about working in that space that 0:01:45.720 --> 0:01:47.120 means lots of people want to get in. 0:01:47.440 --> 0:01:49.440 I mean, I think that could be two ways of 0:01:49.480 --> 0:01:51.720 saying the same thing. If they're minting money, then that 0:01:51.760 --> 0:01:55.520 probably is fundamentally attractive to people in that space. But 0:01:55.640 --> 0:01:58.440 I do think like backing up the questions, like what 0:01:58.480 --> 0:02:02.680 we know is that many and including apparently even like 0:02:02.800 --> 0:02:05.240 B tier C tier funds have done like very well. 0:02:05.560 --> 0:02:08.680 So I'm just like curious like how and why? And 0:02:08.720 --> 0:02:10.600 then yeah to the question of like what does it 0:02:10.680 --> 0:02:13.480 take to succeed in them or who is the type 0:02:13.600 --> 0:02:16.000 of person who can succeed in this environment? 0:02:16.080 --> 0:02:17.600 All right, well, I'm glad you put it that way, 0:02:17.639 --> 0:02:20.320 because today we're going to be speaking with someone who 0:02:20.360 --> 0:02:24.080 has done exactly that succeeded in this particular environment. We 0:02:24.160 --> 0:02:26.120 have the perfect guest. We're going to be speaking with 0:02:26.160 --> 0:02:32.000 Giuseppe Palielogo aka Gappy. He describes himself as a constant gardener, 0:02:32.160 --> 0:02:34.680 someone who's on gardening leave quite a lot. He is 0:02:34.720 --> 0:02:38.880 also the author of Advanced Portfolio Management, a quant's guide 0:02:38.960 --> 0:02:41.560 for fundamental investors, and I have to say it is 0:02:41.800 --> 0:02:45.600 one of the funniest books that I've read in quant finance. 0:02:45.639 --> 0:02:48.040 I can't say it's the funniest because I did read 0:02:48.200 --> 0:02:50.919 My Life as a Quant from Emmanuel Derman, but it's 0:02:50.960 --> 0:02:53.760 definitely up there. And Joe, I know you enjoyed it too. 0:02:54.160 --> 0:02:57.240 I did you know, I like skipped over all the numbers. 0:02:56.800 --> 0:02:59.200 And equations and you just looked at the jokes. 0:02:59.040 --> 0:03:02.400 And Greek letters. But it's very breezily written for what 0:03:02.520 --> 0:03:05.480 it is, and I did. Actually, I think maybe I 0:03:05.560 --> 0:03:08.400 learned a little bit even in my sort of basic 0:03:08.440 --> 0:03:12.119 reading of it. Extremely well written. I'm extremely excited about 0:03:12.160 --> 0:03:14.400 this conversation. You know, you mentioned that our guest is 0:03:14.440 --> 0:03:17.240 the king of gardening leave. If you look in his LinkedIn, 0:03:17.440 --> 0:03:20.519 it really is many different roles. Well. 0:03:20.600 --> 0:03:22.360 I also have to say he is the only person 0:03:22.440 --> 0:03:25.440 I know who has both an alpha and a beta 0:03:25.480 --> 0:03:27.040 tattoo on his shoulder. 0:03:27.200 --> 0:03:27.720 Oh wow. 0:03:28.240 --> 0:03:30.920 You know, some people do get the alpha symbol, but 0:03:31.080 --> 0:03:34.760 he has both. So you know, a well balanced portfolio 0:03:34.880 --> 0:03:38.520 of tattoos all around. Yeah, so gappy. Thank you so 0:03:38.600 --> 0:03:40.200 much for coming on all thoughts. 0:03:40.440 --> 0:03:41.680 Hi Tracy, Hi Joe. 0:03:42.240 --> 0:03:44.640 So maybe to begin with, I'm going to let you 0:03:45.120 --> 0:03:48.800 explain your previous job history because there is quite a lot. 0:03:49.000 --> 0:03:52.120 What is it that you've been doing in this industry. 0:03:52.280 --> 0:03:55.920 I'm not sure, I'm not sure, Okay, good question. Well, 0:03:55.920 --> 0:03:59.080 I got into this industry almost accidentally. I was for 0:03:59.360 --> 0:04:03.200 a few years a researcher in the math department at 0:04:03.240 --> 0:04:06.200 IBM Research, and then I got a little bit bored. 0:04:06.280 --> 0:04:09.080 So the only place that you can the only industry 0:04:09.080 --> 0:04:11.000 you can work in New York other than you know, 0:04:11.120 --> 0:04:14.480 IBM or tech, is finance. So I got into finance 0:04:14.480 --> 0:04:19.240 almost accidentally. And then again, there is no major plan 0:04:19.360 --> 0:04:23.000 to you know, to to my career choices. When I 0:04:23.080 --> 0:04:26.480 was getting bored for some reason, somebody called me and 0:04:26.800 --> 0:04:29.640 offered me a more interesting job. And so I have 0:04:29.720 --> 0:04:33.599 been working mostly on the so called buyside of the industry, 0:04:33.680 --> 0:04:36.760 so the part of the industry that invests, actively invests 0:04:36.800 --> 0:04:41.000 and takes risks. So I've worked for Citadel twice for 0:04:41.080 --> 0:04:43.440 a small edge fund as a portfolio manager, and then 0:04:43.600 --> 0:04:47.000 Millennium and Hudson River Trading, and I've kind of taken 0:04:47.080 --> 0:04:51.920 turns between doing quantitative research and risk management. So most 0:04:51.960 --> 0:04:54.440 recently I was at Hudson River Trading until the beginning 0:04:54.480 --> 0:04:55.120 of November. 0:04:55.760 --> 0:04:59.440 I think when people think about like multi strategy hedge 0:04:59.440 --> 0:05:03.280 fund or shop or whatever, maybe sort of Millennium is 0:05:03.320 --> 0:05:05.920 the first one that would come to mind for people. 0:05:06.279 --> 0:05:08.960 If someone asks you, how does Millennium make money? And 0:05:09.000 --> 0:05:10.480 they seem to have made a lot of money over 0:05:10.520 --> 0:05:11.880 the years, what's the answer? 0:05:12.440 --> 0:05:17.479 Okay, I hope without saying anything that is proprietary, but 0:05:17.600 --> 0:05:18.240 I think. 0:05:18.040 --> 0:05:20.480 That like the business model of Millennium, Yeah. 0:05:20.279 --> 0:05:24.880 I think that what Millennium has excelled at has been 0:05:25.080 --> 0:05:29.200 the ability to scale up, so to adapt its existing 0:05:29.240 --> 0:05:35.480 platform to accommodate new strategies and new portfolio managers and 0:05:35.520 --> 0:05:40.360 so sometimes actually in some of their marketing material they 0:05:40.400 --> 0:05:44.000 called it something like an investment operating system. So it's 0:05:44.040 --> 0:05:47.760 a system that is a firm that is willing to 0:05:47.800 --> 0:05:53.080 absorb some relatively new strategy and create an environment for 0:05:53.160 --> 0:05:56.600 that strategy to succeed. And so because of that, I 0:05:56.720 --> 0:05:59.279 think they might be having right now the highest number 0:05:59.279 --> 0:06:04.240 of individual pods maybe close to three hundred and hovering 0:06:04.279 --> 0:06:07.960 around sixty billion dollars of AUM of assets under management. 0:06:08.520 --> 0:06:12.080 But I would say, what is their superpower is really 0:06:12.120 --> 0:06:16.000 their ability to scale in number of pods. 0:06:16.320 --> 0:06:20.920 So you mentioned creating an environment for success there, what 0:06:20.960 --> 0:06:25.000 does that look like at an organization like that? What 0:06:25.080 --> 0:06:28.520 are the sort of like conduits that allow trades in 0:06:28.560 --> 0:06:31.120 that particular organization to be successful? 0:06:33.080 --> 0:06:35.880 So I would give a sort of an idiosyncratic maybe 0:06:35.920 --> 0:06:40.560 a story around, please the rationale for success of platforms. 0:06:41.040 --> 0:06:44.920 So I see platforms a little bit like managing an 0:06:45.040 --> 0:06:49.000 arbitrage or some kind of gap between the single platform, 0:06:49.080 --> 0:06:52.480 the single manager, or the small hedge funds and the 0:06:52.480 --> 0:06:54.919 fund of funds. So if you're a fund of funds, 0:06:55.080 --> 0:06:58.520 you do have the scale, but you do not have 0:06:58.720 --> 0:07:03.320 the ability to observe from a close distance the performance 0:07:03.480 --> 0:07:08.240 of your vehicles for investment. And let's say that they 0:07:08.240 --> 0:07:10.560 don't perform well, you have to wait a year in 0:07:10.640 --> 0:07:13.000 order to take your money back. In the case of 0:07:13.040 --> 0:07:17.200 a hedge fund platform, you could actually not only observe 0:07:17.680 --> 0:07:21.520 the performance of pms or volume managers their skill from 0:07:21.560 --> 0:07:23.680 a very close distance, but you can also help them 0:07:23.760 --> 0:07:27.720 perform better. So you can centralize some of the functions 0:07:27.720 --> 0:07:33.760 that make them better capital access, corporate access, risk management. 0:07:34.360 --> 0:07:36.840 If they perform well, to give them more capital. If 0:07:36.880 --> 0:07:39.760 they don't perform well, to take capital away from them 0:07:40.000 --> 0:07:43.680 or let them go. And at the same time you 0:07:43.760 --> 0:07:46.440 also solve two for two other problems. So one is 0:07:46.920 --> 0:07:51.160 there is a risk transfer happening because a platform almost 0:07:51.240 --> 0:07:54.520 by design otherwise is not really a platform, has a 0:07:55.200 --> 0:07:58.880 pass through fee structure that's fundamental for the existence of 0:07:58.880 --> 0:08:01.720 a platform that makes really a platform what it is 0:08:01.760 --> 0:08:04.680 instead of a just multi manager hedge fund like the show. 0:08:05.080 --> 0:08:08.560 So this means that a portfolio manager is not paid 0:08:09.120 --> 0:08:11.840 with the incentive fee that the hedge fund as a 0:08:11.840 --> 0:08:16.239 whole receives from the limited partners, but instead the portfolio 0:08:16.320 --> 0:08:19.000 managers are paid a percentage of their p and L. 0:08:19.560 --> 0:08:24.640 This payment is passed through directly to the limited partners 0:08:25.360 --> 0:08:29.760 to the investors, and this basically transfers the risk directly 0:08:30.440 --> 0:08:34.240 basically from the PM into the limited partner, And so 0:08:34.280 --> 0:08:37.160 this makes the system more robust in a sense, right, 0:08:37.720 --> 0:08:42.480 And combine this with the diversification across investment styles and 0:08:42.720 --> 0:08:45.200 the number of pms, and now you start having a 0:08:45.240 --> 0:08:48.640 mote around a platform that makes it successful. 0:09:04.640 --> 0:09:07.960 If a entity has three hundred pods and everyone's doing 0:09:07.960 --> 0:09:11.560 their own thing, et cetera, why doesn't the return just 0:09:11.679 --> 0:09:15.120 become the market return like it seems like because there's 0:09:15.120 --> 0:09:18.839 a right Like, one intuition could be that this model 0:09:18.840 --> 0:09:20.719 wouldn't scale. I mean, I know it does, but one 0:09:20.760 --> 0:09:23.600 intuition could be that this model wouldn't scale that the 0:09:23.640 --> 0:09:27.160 more you add, you overdiversify and then you just end 0:09:27.240 --> 0:09:29.280 up with like whatever, like you know, like buy the 0:09:29.360 --> 0:09:32.199 VTI ETF or something like that. Why doesn't it work 0:09:32.200 --> 0:09:32.480 out that. 0:09:32.679 --> 0:09:36.400 A simplest explanation for this is actually just to look 0:09:36.440 --> 0:09:40.280 at what a retail investor right would hold in their portfolio. 0:09:40.480 --> 0:09:43.559 So let's say that they are, you know, long Apple 0:09:43.679 --> 0:09:46.439 and IBM. Okay, they have a little bit of an 0:09:46.440 --> 0:09:50.320 imperfect version of the market, right, But what makes their 0:09:50.679 --> 0:09:55.080 skill is how different are the weights of their Apple 0:09:55.120 --> 0:09:58.480 and IBM holdings compared to the market. Okay, So you 0:09:58.520 --> 0:10:02.760 can decompose your performance in your personal account into the 0:10:02.840 --> 0:10:06.920 sum of let's say the market and your idiosyncratic bets 0:10:06.960 --> 0:10:11.319 into these stocks. Now, what the hedge funds do is 0:10:11.720 --> 0:10:14.600 they do the same, but they completely eliminate as much 0:10:14.600 --> 0:10:18.640 as they can their exposure or their investment in the market, 0:10:18.679 --> 0:10:22.640 So they run purely market neutral and factor neutral portfolios. 0:10:23.360 --> 0:10:27.680 So there is diversification, but these indiosyncratic bets don't get 0:10:27.679 --> 0:10:31.240 diversified away into a big market, but they actually become 0:10:31.760 --> 0:10:34.880 essentially a bunch of independent bets that by the law 0:10:34.880 --> 0:10:37.800 of large numbers, they tend to have better and better 0:10:37.880 --> 0:10:39.280 risk adjusted profiles. 0:10:39.640 --> 0:10:44.040 So I still see some platform heads describe like the 0:10:44.160 --> 0:10:48.080 overall tilt as market neutral. So what do they mean 0:10:48.160 --> 0:10:49.160 by that? Exactly? 0:10:49.240 --> 0:10:52.400 I mean they typically run a wide range of strategies, 0:10:52.440 --> 0:10:55.880 so let's focus because it's more relatable. Let's focus on 0:10:56.000 --> 0:10:59.800 discretion ory long short equities and systematic equities because everybody 0:10:59.800 --> 0:11:00.360 knows star. 0:11:00.679 --> 0:11:03.160 I've love that you think systematic equities is relatable. 0:11:03.440 --> 0:11:06.600 Yeah, yeah, I mean relatively to I don't know, treasury 0:11:06.640 --> 0:11:10.600 basis or sell involved. So they mean that typically they 0:11:10.600 --> 0:11:14.080 do have a so called factor model, and a factor 0:11:14.160 --> 0:11:16.920 model is a little bit like having a market model 0:11:16.960 --> 0:11:21.080 on steroids, So you have a market term, so you 0:11:21.120 --> 0:11:24.920 can see your portfolio as having exposure to the market, 0:11:24.960 --> 0:11:27.920 so behaving a little bit like a market. And then 0:11:28.400 --> 0:11:30.600 it's also behaving a little bit like a portfolio that 0:11:30.679 --> 0:11:34.520 has momentum okay, and then it also has maybe a 0:11:34.640 --> 0:11:38.360 tilt in terms of value. The platforms tend to run 0:11:38.600 --> 0:11:42.839 portfolios that have no market exposure whatsoever, and then they 0:11:42.880 --> 0:11:47.160 also tend to have controlled exposure in these more exotic factors. 0:11:47.280 --> 0:11:50.080 How do they know that? I mean, so there's someone 0:11:50.240 --> 0:11:53.800 up there at the center, there's all that three hundred pods. 0:11:53.840 --> 0:11:56.640 The data gets probably aggregated and sliced in various ways, 0:11:56.679 --> 0:11:59.640 but what is the job or how do they actually 0:12:00.040 --> 0:12:04.120 sure that on that their portfolio managers don't have that 0:12:04.320 --> 0:12:04.880 market beta. 0:12:05.040 --> 0:12:10.240 Yeah, they typically have at the very minimum. They will 0:12:10.280 --> 0:12:14.080 buy some commercial factor model, which is a model of 0:12:14.400 --> 0:12:18.719 the market, like of your investment universe, how the how 0:12:18.760 --> 0:12:22.040 a stock behaves. How can you decompose the performance of 0:12:22.080 --> 0:12:25.120 the stock in this various systematic or let's call them 0:12:25.160 --> 0:12:29.840 pervasive market wide factors and instead idiosyncratics. So you buy 0:12:29.880 --> 0:12:32.679 them off the shelves. I mean, they're really expensive and 0:12:32.920 --> 0:12:36.520 they do a job. And so once you've bought them, 0:12:36.800 --> 0:12:40.120 you create some kind of user friendly interface so that 0:12:40.200 --> 0:12:45.440 a portfolio manager can always see how the portfolio looks 0:12:45.520 --> 0:12:47.040 like at any point in time. It's a little bit 0:12:47.080 --> 0:12:49.800 like having an X ray of you know, your body 0:12:49.800 --> 0:12:52.080 in real time. You know, you can see, oh, well, 0:12:52.160 --> 0:12:54.760 my portfolio is is a little bit short, the market 0:12:54.840 --> 0:12:58.360 is a little bit long momentum, maybe there is some crowding, exposure, whatever, 0:12:58.960 --> 0:13:01.560 And so this is in the hands of the portfolio manager. 0:13:01.960 --> 0:13:04.800 And then there is another layer on top of that, 0:13:04.960 --> 0:13:09.439 which is very important risk management, which ensures that pms 0:13:09.480 --> 0:13:12.319 are behaving well, that they're not going out of scope. 0:13:13.080 --> 0:13:16.520 You know, they're not buying microstocks or you know, investing 0:13:16.520 --> 0:13:17.760 in crazy stuff just. 0:13:17.720 --> 0:13:19.479 Going or just going along in video. 0:13:19.559 --> 0:13:22.000 Or long in Vidia. Yeah, if their idea is going 0:13:22.080 --> 0:13:27.080 long in Vidia, probably that's not an ideal portfolio manager. Yeah. 0:13:27.440 --> 0:13:30.080 So the other thing I've been wondering is how much 0:13:30.360 --> 0:13:36.400 visibility are there between the different pods within one shop. Yeah, 0:13:36.440 --> 0:13:40.840 And I mean that like, I assume there's a centralized 0:13:41.040 --> 0:13:44.320 risk management system of some sort that is like netting 0:13:44.320 --> 0:13:47.880 out positions and trying to make use of capital most efficient, 0:13:47.920 --> 0:13:50.080 and that's where a lot of the edge comes from. 0:13:50.720 --> 0:13:55.120 But also, if you're just a trader pursuing your own strategy, 0:13:55.480 --> 0:13:58.480 do you know what the guy next to you is doing? 0:13:58.600 --> 0:14:00.800 Do you have that kind of vision ability or is 0:14:00.800 --> 0:14:04.679 the idea to keep everyone sort of intellectually separated so 0:14:04.720 --> 0:14:06.520 that they're not influenced by each other. 0:14:06.760 --> 0:14:10.400 Right, that's a good question. So there is no really 0:14:10.520 --> 0:14:14.480 black and white answer to this, because historically there was 0:14:14.520 --> 0:14:18.920 a time when platforms had more visibility and more collaboration 0:14:19.240 --> 0:14:21.920 among pods or at least pods in the same sector, 0:14:22.000 --> 0:14:25.080 for example. But I would say that the historical trend 0:14:25.560 --> 0:14:28.600 has been more and more to give them the tools 0:14:28.600 --> 0:14:32.880 to succeed, but not give them the ability to see 0:14:32.880 --> 0:14:36.160 into each other's portfolios for example. And the rationale for 0:14:36.200 --> 0:14:40.120 this is you probably prefer having independent bets to having 0:14:40.760 --> 0:14:44.000 maybe corredated bets that could be like maybe a little 0:14:44.040 --> 0:14:47.200 bit more informed. So that's the trade off. Let's if 0:14:47.200 --> 0:14:50.160 we talk, maybe we can come up with slightly better ideas. Sure, 0:14:50.200 --> 0:14:52.400 but yeah, I think that the trend is more and 0:14:52.440 --> 0:14:55.000 more towards you are not seeing what I am, what 0:14:55.040 --> 0:14:56.320 I'm having, what I'm holding. 0:14:57.120 --> 0:15:00.160 Talk to us more about the risk management component, and 0:15:00.200 --> 0:15:03.160 again I don't know very much. I understand that you know, 0:15:03.760 --> 0:15:06.360 stop losses are very tight and you don't get a 0:15:06.360 --> 0:15:08.440 long leash to lose money, and if you're not doing well, 0:15:08.480 --> 0:15:11.440 your capitals reduced. If you're doing well, I guess you 0:15:11.480 --> 0:15:13.680 get more, and if you do more, you get more, 0:15:13.760 --> 0:15:17.080 et cetera. But from how would you describe the sort 0:15:17.080 --> 0:15:20.800 of the essence of risk management at the hedge fund level. 0:15:21.080 --> 0:15:24.560 So there are maybe two or three core functions that 0:15:24.600 --> 0:15:27.480 can be described in a qualitative way, but you know, 0:15:27.640 --> 0:15:31.000 I think pretty comprehensively. And then there there is something 0:15:31.040 --> 0:15:33.920 that is a little bit more esoteric or like domain specific. 0:15:34.000 --> 0:15:36.920 So let's talk about the general principles. Okay, So you 0:15:37.000 --> 0:15:40.120 mentioned stop losses, so this is very important. You know, 0:15:40.200 --> 0:15:43.360 there are always stop losses, the ones that you know 0:15:43.520 --> 0:15:45.480 you have and the ones you don't know you have, 0:15:45.680 --> 0:15:49.240 but everybody has stop losses in life. Okay, So those 0:15:49.320 --> 0:15:52.240 are very important because you could imagine that a PM 0:15:52.360 --> 0:15:54.080 is a little bit like somebody who's holding a coll 0:15:54.120 --> 0:15:59.320 option and you you know, the PM who's losing money 0:15:59.360 --> 0:16:02.280 has kind of a incentive to go for broke maybe sometimes. 0:16:02.560 --> 0:16:06.760 But the stop laws is effectively at a sort of 0:16:06.800 --> 0:16:10.720 a primitive tail insurance TAE risk management tool on the 0:16:10.800 --> 0:16:13.240 left tail of a PM. So that's very important. The 0:16:13.240 --> 0:16:17.800 second principle is sort of self enforcing, is true diversification. 0:16:18.320 --> 0:16:20.120 So this is where you want to have some kind 0:16:20.120 --> 0:16:23.200 of risk model that tells you what are the hidden 0:16:23.280 --> 0:16:27.480 bets that kind of overlap and maybe compound at the 0:16:27.560 --> 0:16:30.400 aggregate level, so that if everybody takes a little bit 0:16:30.440 --> 0:16:33.240 of a factor exposure in the same direction and then 0:16:33.280 --> 0:16:36.600 you sum this across three hundred pms, it becomes a 0:16:36.600 --> 0:16:40.640 big factor exposure. So a risk management organization needs to 0:16:40.960 --> 0:16:44.280 get that right. The third thing is making sure that 0:16:44.480 --> 0:16:48.080 people stay in scope. Okay, so seems trivial, but actually 0:16:48.120 --> 0:16:52.440 that requires a lot of domain expertise. So understanding the trades, 0:16:52.960 --> 0:16:57.160 what can go wrong from an operational standpoint macrostructure standpoint. 0:16:56.840 --> 0:16:59.520 Is this factor drift risk as well? 0:16:59.600 --> 0:17:02.760 Or said that scope is more like factor drift or 0:17:02.800 --> 0:17:06.680 in general strategy drift, not only factor but whereas being 0:17:06.720 --> 0:17:10.480 in scope is more of a pure strategy drift or 0:17:10.600 --> 0:17:15.360 just taking risks that a portfolio manager would be possibly 0:17:15.400 --> 0:17:18.240 aware of, but that maybe the head of the hedge fund, 0:17:18.320 --> 0:17:20.280 because it's not an expert in that area, is not 0:17:20.359 --> 0:17:24.200 so aware of. So the risk manager has to know 0:17:25.320 --> 0:17:29.520 very well what's going on and an alert. Talk to 0:17:29.560 --> 0:17:32.760 the PM, talk to the business head and. 0:17:33.600 --> 0:17:38.160 Can you give us concrete examples from your experience of 0:17:38.240 --> 0:17:41.440 the kind of things that would set off alarm belts. 0:17:41.920 --> 0:17:44.560 So is there like, I guess you don't have to 0:17:44.600 --> 0:17:47.720 give us specific examples, but you know the kind of thing, 0:17:48.000 --> 0:17:51.320 the types of examples, Yeah, the types of examples that 0:17:51.400 --> 0:17:54.400 would catch your eye in a risk management position. 0:17:54.760 --> 0:17:57.360 So we covered a little bit the easy stuff, right, 0:17:57.400 --> 0:18:00.159 So the easy stuff is people taking too much risk. 0:18:00.240 --> 0:18:02.320 First of all, it simple, but you know, we think 0:18:02.320 --> 0:18:05.480 in terms of dollar volatility. Dollar volatility is a little 0:18:05.480 --> 0:18:08.399 bit like how much you can make or lose in 0:18:08.440 --> 0:18:09.040 one year for it. 0:18:09.119 --> 0:18:11.200 So like value at risk, those kind of kind of 0:18:11.240 --> 0:18:11.800 value risk. 0:18:11.880 --> 0:18:14.640 Yes, I mean most people think in terms of all 0:18:15.119 --> 0:18:18.720 value risk too. Okay, yeah, choose your risk metric you 0:18:18.760 --> 0:18:21.840 want to stay within that. Then factor exposures. Okay, that's 0:18:21.880 --> 0:18:24.800 also easy concentration. So if you take a mega bet 0:18:24.800 --> 0:18:28.920 in Nvidia, it has to surface. Okay. So these are 0:18:28.960 --> 0:18:31.919 relatively simple. There are things that are a little bit 0:18:31.960 --> 0:18:36.920 more complicated, like, for example, you take some true arbitrage 0:18:37.280 --> 0:18:40.960 positions where you think that something is running cheap versus 0:18:41.080 --> 0:18:44.919 rich in say bond versus futures, or you do some 0:18:45.000 --> 0:18:49.040 kind of funding arbitrage trade where different agents in the 0:18:49.200 --> 0:18:53.800 investing world have different funding rates for their assets, and 0:18:54.320 --> 0:18:57.160 those can break, like in a dislocation, that can break. 0:18:57.240 --> 0:18:59.640 And so the way that typically you manage these things 0:18:59.680 --> 0:19:01.520 a little it's a little bit like in merger arm. 0:19:01.600 --> 0:19:04.920 You give it a max size and you want to 0:19:05.000 --> 0:19:07.600 make sure that this is correct, that this size is correct, 0:19:07.640 --> 0:19:10.840 and it's monitored. So this is stuff that can go wrong. 0:19:11.800 --> 0:19:15.960 Two managers like, how much do they I mean, I'm 0:19:15.960 --> 0:19:17.720 sure there's sort of I don't know if it's accidental 0:19:17.760 --> 0:19:20.160 style drift or you know, drift is sort of a 0:19:20.200 --> 0:19:24.840 neutral term. How much does the risk manager have to 0:19:24.880 --> 0:19:27.840 watch out for I guess intentional drift or this is 0:19:27.880 --> 0:19:29.959 a working I know this is not quite my mandate, 0:19:30.000 --> 0:19:31.800 this is not quite what I was made to trade. 0:19:31.800 --> 0:19:33.840 But I could sort of justify it this way, or 0:19:34.119 --> 0:19:36.280 I just see all these lines up over here going up, 0:19:36.320 --> 0:19:39.000 I need to how much of a risk management concern 0:19:39.119 --> 0:19:39.280 is that? 0:19:39.359 --> 0:19:42.680 Okay, I think that in general the principle should be trust, 0:19:42.720 --> 0:19:47.040 but verify. I would say that the vast majority of 0:19:47.320 --> 0:19:51.760 portfolio managers are very responsible, and because they're in that role, 0:19:51.800 --> 0:19:55.480 they have been educated to control their risks, to understand 0:19:55.520 --> 0:19:59.760 them with occasional screw ups, and so that's why you 0:19:59.800 --> 0:20:01.320 need very fie got it? 0:20:02.000 --> 0:20:02.359 Okay. 0:20:02.680 --> 0:20:07.080 On the opposite side of screw ups, I'm curious how 0:20:07.560 --> 0:20:12.040 capital gets kind of doled out. And if I'm running 0:20:12.040 --> 0:20:18.040 a massively profitable, successful training strategy, do I automatically start 0:20:18.080 --> 0:20:22.359 giving start being given more money to you know, play 0:20:22.400 --> 0:20:25.560 around with or is there some amount of discipline here 0:20:25.760 --> 0:20:29.040 where you don't want people to be bumping up against 0:20:29.119 --> 0:20:33.080 you know, sizing positions or additional trading costs and things 0:20:33.119 --> 0:20:36.960 like that. Imagine I am the most popular trader, the 0:20:36.960 --> 0:20:42.320 most successful popular also popular, I'm both the most popular 0:20:42.440 --> 0:20:45.120 trader and most successful trader at Citadel. 0:20:45.320 --> 0:20:47.680 What is the process for traces getting more money to trade? 0:20:47.720 --> 0:20:51.040 How do I get more popular and successful? Probably not popular? 0:20:51.119 --> 0:20:54.399 Okay, assume that you're popular and successful, okay, So do 0:20:54.440 --> 0:20:57.080 you get more capital? You do get more capital up 0:20:57.119 --> 0:20:59.280 to a point. So there are a couple of factors. 0:20:59.320 --> 0:21:02.520 The first one is there is like a natural limit 0:21:02.840 --> 0:21:07.280 where somebody can be too successful. And without giving examples, 0:21:07.280 --> 0:21:10.919 but there are large funds whose daily p and l 0:21:11.240 --> 0:21:14.679 sometimes at points are driven is driven by a single 0:21:14.800 --> 0:21:18.679 strategy Okay, and maybe that's justified, right, But there is 0:21:18.680 --> 0:21:20.879 a point where there could be just too much because 0:21:20.920 --> 0:21:25.640 the concentration across strategies. Or think of pods as stocks, right, 0:21:25.680 --> 0:21:28.520 you don't want to have ninety percent of your savings 0:21:28.560 --> 0:21:31.160 in Nvidia, So okay, so that's number one. So there 0:21:31.240 --> 0:21:34.359 is some kind of basic heuristics. Then there is just 0:21:34.480 --> 0:21:38.680 a natural limit to growth for strategies, like there is 0:21:39.960 --> 0:21:44.440 a trade off because your market impact is very high 0:21:44.520 --> 0:21:47.200 and so, or there is just a hard size for 0:21:47.240 --> 0:21:50.560 your strategy, so you cannot scale high frequency, you cannot 0:21:50.560 --> 0:21:54.800 scale to infinity even index rebalancing. Or if you're a 0:21:54.840 --> 0:21:59.560 consumer PM, your costs increase faster than the size of 0:21:59.600 --> 0:22:03.720 your portfolio, so your P and L in the absence 0:22:03.760 --> 0:22:07.840 of costs goes more or less lilenar linearly, but your 0:22:07.880 --> 0:22:10.320 costs grow faster than linearly. So there is a point 0:22:10.359 --> 0:22:13.880 where you just don't want to grow all Right. 0:22:13.920 --> 0:22:16.919 On the flip side, Let's say Tracy comes in and 0:22:16.960 --> 0:22:21.760 she is a PM and she has her pod. How 0:22:22.119 --> 0:22:26.600 long is she likely to last and what would cause 0:22:26.640 --> 0:22:29.600 her what would be the threshold at which she gets fired. 0:22:31.800 --> 0:22:34.840 I don't have the statistics on the average tenure of 0:22:34.920 --> 0:22:39.080 a PM, Okay, if I had them, probably I shouldn't say. Well, 0:22:39.119 --> 0:22:42.400 and also depends a lot on the place. Okay, so 0:22:42.440 --> 0:22:47.880 how long I would say that it's like everything in life, right, 0:22:47.960 --> 0:22:50.840 So like ninety percent of everything is of poor quality, 0:22:50.880 --> 0:22:53.040 I'm sorry to say, but the same applies to pms. 0:22:53.160 --> 0:22:56.240 But this is another beautiful aspect of platforms, by the way. Okay, 0:22:56.280 --> 0:22:58.359 so let me take a quick ditchure about this again, 0:22:58.440 --> 0:23:02.480 because like a beautiful and under appreciated aspect of platforms 0:23:03.040 --> 0:23:08.560