WEBVTT - The Important Lesson a Quant Manager Learned in 2020 0:00:11.160 --> 0:00:15.160 Hello, and welcome to another episode of the Odd Lots Podcast. 0:00:15.200 --> 0:00:20.119 I'm Joe Wisenthal and I'm Tracy Hallaway. So, Tracy, it 0:00:20.160 --> 0:00:23.759 goes without saying, and we really don't need to uh 0:00:24.079 --> 0:00:27.720 be labor the point at all that from a market's perspective, 0:00:27.760 --> 0:00:32.080 the last year has been I guess unprecedented. Yeah, I 0:00:32.120 --> 0:00:34.040 think that's right when you say we don't need to 0:00:34.040 --> 0:00:37.080 belabor the point. Should I describe what we're talking about 0:00:37.159 --> 0:00:39.240 or should we just move on anyway? You know, it's 0:00:40.000 --> 0:00:44.360 it's just this crisis, this huge crash, this incredible rally 0:00:44.400 --> 0:00:47.320 that never seems to end in the middle of a 0:00:47.360 --> 0:00:50.640 pandemic and so forth. It's just like, there's no obvious, 0:00:51.000 --> 0:00:54.880 uh historical analogy to what we've just seen that I 0:00:54.880 --> 0:00:57.760 can think of. Well, I think what was really interesting 0:00:57.760 --> 0:01:00.600 about last year was that we basically saw an entire 0:01:00.800 --> 0:01:05.560 economic and to some extent, financial crisis and a recovery, 0:01:05.800 --> 0:01:10.960 so a sort of complete economic cycle all squeezed into 0:01:11.800 --> 0:01:14.880 less than a year. And if you think, I mean, 0:01:14.920 --> 0:01:17.400 the trajectory of everything was quite similar to the two 0:01:17.440 --> 0:01:20.480 thousand eight financial crisis. You have the sharp market sell 0:01:20.520 --> 0:01:23.399 off in March, and then you had the central banks 0:01:23.480 --> 0:01:25.880 come in and stabilize things, and then you had a 0:01:25.920 --> 0:01:30.080 slow recovery or a recovery in markets. But again, the 0:01:30.120 --> 0:01:32.720 big difference was that it all happened in the space 0:01:32.760 --> 0:01:37.600 of months versus years before. Yeah. No, I actually I 0:01:37.640 --> 0:01:40.840 think that's right. You could characterize two thousand and twenty 0:01:41.000 --> 0:01:46.000 as essentially a full cycle that really got compressed into 0:01:46.120 --> 0:01:49.920 the span of a few months early. Yeah, exactly. So anyway, 0:01:50.120 --> 0:01:52.720 it's still pretty unusual, even if even if you could 0:01:52.920 --> 0:01:55.560 sort of describe it using words that were familiar with 0:01:55.640 --> 0:01:58.040 it's pretty unusual. But what that means is, you know, 0:01:58.480 --> 0:02:02.200 for investors, obviously we have this incredible rally. Stocks are 0:02:02.200 --> 0:02:04.120 at all time high, but you know, I think a 0:02:04.120 --> 0:02:06.560 lot of people missed it. A lot of people throughout 0:02:06.600 --> 0:02:09.040 the summer saying, oh, we're going to get the second 0:02:09.200 --> 0:02:11.679 the second dip in the market that's coming in eight 0:02:11.720 --> 0:02:15.680 time soon. You talk to sort of wealthy investors and 0:02:15.919 --> 0:02:18.520 um sort of high net works, there was a lot 0:02:18.560 --> 0:02:22.040 of caution under exposure to the gains and so forth. 0:02:22.120 --> 0:02:25.160 So while the headlines were very impressive, I don't think 0:02:25.160 --> 0:02:27.760 there's any doubt that, you know, there probably aren't actually 0:02:27.800 --> 0:02:32.760 a lot of people that saw their portfolios really match 0:02:33.360 --> 0:02:36.000 or really you know, let's just say, I think there 0:02:36.000 --> 0:02:38.640 are probably a lot of people whose portfolios don't match 0:02:38.680 --> 0:02:42.280 the headline games. Yeah, I think that's right. And I 0:02:42.320 --> 0:02:46.000 think there's probably a sense in markets, or at least 0:02:46.680 --> 0:02:52.799 professional investors in markets, that what happened wasn't really what 0:02:52.880 --> 0:02:55.920 they expected or what they were necessarily prepared for. So 0:02:55.960 --> 0:03:00.560 we already mentioned that everything unfolded really really quickly. So 0:03:00.800 --> 0:03:03.839 even if you were well positioned going into March, you've 0:03:03.840 --> 0:03:06.520 taken a little bit of risk off the table, you 0:03:06.600 --> 0:03:10.040 might not have been expected expecting the markets to recover 0:03:10.200 --> 0:03:12.960 as quickly as they did. I think there's also the 0:03:13.000 --> 0:03:17.320 sense of dissatisfaction with what has rallied. I think a 0:03:17.320 --> 0:03:20.799 lot of people are not not upset, but I think 0:03:20.840 --> 0:03:24.240 there's a sense that the stupid stuff has gone up 0:03:24.880 --> 0:03:28.560 a lot more than some other things, which again, like 0:03:29.639 --> 0:03:34.640 that's the kind of thing that tends to upset professional investors. Right. 0:03:34.680 --> 0:03:36.840 In fact, that actually was one of the themes of 0:03:38.400 --> 0:03:40.640 which is that you know, like you know, the back 0:03:40.640 --> 0:03:43.040 of the spring, we got the we had the Virtual 0:03:43.080 --> 0:03:46.960 Buffet shareholder meeting. He was, like many people remarked that 0:03:47.040 --> 0:03:49.520 he was like very negative on the market in a 0:03:49.520 --> 0:03:52.040 way that he usually isn't. And then at the same 0:03:52.120 --> 0:03:54.880 time you had like the Dave Portnoy's of the world 0:03:54.960 --> 0:03:58.360 absolutely killing it in part by picking scrabble letters out 0:03:58.360 --> 0:04:00.840 of a scrabble tile is out of a box and 0:04:00.880 --> 0:04:04.000 picking stocks, and all these people are like, yeah, you know, 0:04:04.080 --> 0:04:06.520 I'm I'm not too busy these days. I'm gonna sign 0:04:06.560 --> 0:04:09.760 up our Robin Hood account making a fortune, you know, 0:04:09.920 --> 0:04:13.040 all kinds of confounding dynamics to make the market. You know, 0:04:13.080 --> 0:04:16.279 the market always is very efficient machine to driving people insane. 0:04:16.560 --> 0:04:19.599 But I think really took the cake at that. Yeah, 0:04:19.680 --> 0:04:22.280 I think that's exactly right. I've always said that. But 0:04:22.400 --> 0:04:26.400 the market is a great technology for humbling people, making 0:04:26.440 --> 0:04:30.560 a smart people feel h feel less smart. But anyway, yeah, 0:04:30.600 --> 0:04:33.560 you can be wrong every day. Yeah, I'm very excited 0:04:33.680 --> 0:04:36.400 about our guests. We're going to talk about some of 0:04:36.400 --> 0:04:40.680 the lessons learned from We're going to be speaking with 0:04:40.880 --> 0:04:44.160 Corey Hofstein. He has the c i O the asset 0:04:44.240 --> 0:04:49.039 management firm Newfound Research, really sort of active publisher of 0:04:49.080 --> 0:04:53.640 research publisher on Twitter talking about ideas in quantitative finance 0:04:53.720 --> 0:04:58.080 trend following various ideas such as that portfolio construction and 0:04:58.200 --> 0:05:01.479 a really interesting thread at the end of last year 0:05:02.000 --> 0:05:07.680 talking about essentially the costly lessons learned of what happens 0:05:07.720 --> 0:05:11.360 when you experience a market that is kind of basically 0:05:11.440 --> 0:05:14.240 unlike anything else before. So Corey, thank you very much 0:05:14.240 --> 0:05:17.560 for joining us, Thank you for having me here, really excited. 0:05:17.960 --> 0:05:22.400 What do you describe Newfound Research? What your general approaches, 0:05:22.440 --> 0:05:26.279 how you would describe it, and uh, yeah, like essentially 0:05:26.320 --> 0:05:29.960 the aims of your investment strategy when we say that 0:05:29.960 --> 0:05:33.279 Newfound Research is a quantitative asset management firm, So everything 0:05:33.320 --> 0:05:37.280 we do we try to focus on creating systematic investment strategies, 0:05:37.680 --> 0:05:42.200 but our predominant focus historically has been really on risk 0:05:42.320 --> 0:05:47.719 mitigation through diversification. But our view is that diversification is 0:05:47.720 --> 0:05:50.799 not just limited to what you invest in, but also 0:05:50.920 --> 0:05:54.640 how you make those investment decisions and when those decisions 0:05:54.640 --> 0:05:58.160 are made, and so we think better risk management is 0:05:58.200 --> 0:06:03.440 really possible through a greater breadth of diversification. Now, historically, 0:06:03.480 --> 0:06:05.479 for us, what that has meant is we've tried to 0:06:05.480 --> 0:06:09.839 play a more siloed role in investor portfolios by offering 0:06:10.600 --> 0:06:14.080 what we would consider to be resilient and robust. Trend 0:06:14.120 --> 0:06:19.479 following strategies predominantly focused on equity markets. So unlike most 0:06:19.600 --> 0:06:24.360 saying managed futures programs that will apply trend following strategies 0:06:24.400 --> 0:06:27.880 to commodities and currencies and rates and equities around the world, 0:06:28.240 --> 0:06:31.560 we really focused exclusively on the equity side of the 0:06:31.600 --> 0:06:35.599 equation because that's where we saw most investor risk in 0:06:35.640 --> 0:06:40.520 a traditional sixty portfolio. So, how did that work out 0:06:40.560 --> 0:06:43.799 for you in because you've been UM, You've been pretty 0:06:43.880 --> 0:06:49.320 vocal on Twitter about um poor performance last year? How 0:06:49.320 --> 0:06:53.960 exactly did that strategy fit into a year? And what 0:06:54.000 --> 0:06:57.040 went wrong? Well, So, so we won't we won't bury 0:06:57.040 --> 0:06:59.560 the lead here. It did not go well. But I 0:06:59.560 --> 0:07:02.400 think if we were to back up and say, how 0:07:02.400 --> 0:07:06.920 would we expect a trend following strategy to perform during 0:07:07.240 --> 0:07:11.640 one of the fastest market reversals in history, both from 0:07:12.280 --> 0:07:15.640 peak to trough and then trough to peak again, we 0:07:15.720 --> 0:07:19.120 generally wouldn't expect the trend following strategy to do very 0:07:19.200 --> 0:07:21.760 well in a in a fast reversal market. That's sort 0:07:21.760 --> 0:07:24.880 of the expectation going in. So when we walk away 0:07:24.920 --> 0:07:26.960 from the year and say, well, the trend following strategy 0:07:27.080 --> 0:07:32.360 didn't do well in I don't think, given how played out, 0:07:32.440 --> 0:07:35.840 that should be unexpected, but I do think it should 0:07:35.840 --> 0:07:38.400 cause us to pause and say, well, this type of 0:07:38.400 --> 0:07:43.320 strategy was really designed to help protect investors against meaningful drawdowns. 0:07:43.720 --> 0:07:46.760 We just saw a really meaningful drawdown and it didn't 0:07:46.800 --> 0:07:50.760 work particularly well. Are there lessons here that we need 0:07:50.800 --> 0:07:53.800 to think about? Has has market structure changed? Is there 0:07:53.840 --> 0:07:56.440 something we need to do differently in our portfolio? What 0:07:56.640 --> 0:07:59.800 is the next drawdown going to look like? I want 0:07:59.800 --> 0:08:03.640 to get into sort of like what you saw as 0:08:03.720 --> 0:08:07.400 the year progressed, but before we get to that specifically, 0:08:07.400 --> 0:08:09.280 and then of course the lessons from that, but before 0:08:09.280 --> 0:08:13.160 we get to that specifically. For listeners, how would you 0:08:13.240 --> 0:08:18.480 describe the theoretical basis for a trend following strategy, And 0:08:18.520 --> 0:08:22.160 whether it's in equities, whether it's in rates, currencies, commodities. 0:08:22.560 --> 0:08:29.000 Why is there theoretically premium to be harvested with strategies 0:08:29.040 --> 0:08:32.200 that are sort of built around following some measure of 0:08:32.200 --> 0:08:36.319 a trend? So I think there's really two primary arguments 0:08:36.360 --> 0:08:39.120 that are out there that could be supported. I think 0:08:39.440 --> 0:08:43.840 in many of the futures markets, there's an argument that 0:08:43.920 --> 0:08:49.200 trend followers are liquidity providers to hedgers, that ultimately hedgers 0:08:49.200 --> 0:08:52.240 are going to be trying to place trades that lock 0:08:52.320 --> 0:08:56.560 in their profit, and that trend followers are ultimately providing 0:08:56.600 --> 0:09:00.000 the liquidity to those players, and therefore, in providing that liquidity, 0:09:00.120 --> 0:09:02.760 they should earn a premium. So I think there's a 0:09:02.800 --> 0:09:07.000 reasonable argument there in the managed future space. I think 0:09:07.080 --> 0:09:10.560 the second argument that often comes around, and I certainly 0:09:10.600 --> 0:09:14.800 believe in this one, is that there are certain risk 0:09:14.880 --> 0:09:18.679 limits that firms tend to hit during periods of market stress. 0:09:19.320 --> 0:09:25.360 That market stress periods are just fundamentally different than normal 0:09:25.440 --> 0:09:29.320 market environments. And so what that means is that while 0:09:29.400 --> 0:09:32.320 markets may for the most part be a random walk, 0:09:33.120 --> 0:09:37.720 these non linear response functions that that firms have to make, 0:09:37.800 --> 0:09:40.440 that they have to cut risk, that they have to 0:09:41.160 --> 0:09:46.839 raise capital UH in certain market environments creates trends during 0:09:46.880 --> 0:09:51.040 periods of stress, and that they're very procyclical in nature. 0:09:51.120 --> 0:09:53.839 That as the market starts to sell off, firms are 0:09:53.880 --> 0:09:56.600 forced to de risk, which puts more downward pressure on 0:09:56.679 --> 0:10:00.560 markets and continues them to force the sell off UH. 0:10:00.559 --> 0:10:06.000 And so you get this sort of crisis trend type exposure. 0:10:06.480 --> 0:10:09.319 And that's ultimately what we're trying to exploit, particularly when 0:10:09.320 --> 0:10:12.720 it comes to trying to manage risk. So with that 0:10:12.800 --> 0:10:16.640 framework in mind, can you describe how you were positioned 0:10:17.040 --> 0:10:20.520 going into the big sell off in March and then 0:10:20.640 --> 0:10:24.440 how you started changing the portfolio. I'm assuming we started 0:10:24.520 --> 0:10:28.679 changing the portfolio or um the strategy, the systematic process 0:10:28.720 --> 0:10:34.559 that you were actually using as the month kind of unfolded. Absolutely, 0:10:34.640 --> 0:10:39.040 So as we went into March, what we really saw 0:10:39.600 --> 0:10:43.080 was that the majority of our trend signals remained positive 0:10:43.400 --> 0:10:46.240 late February. So what we're really looking at, just to 0:10:46.240 --> 0:10:48.520 be very clear for listeners, is we were looking at 0:10:48.559 --> 0:10:53.960 a cross section of different trend signals on US equity sectors. 0:10:54.120 --> 0:10:57.160 So that was really where the primary signals are coming 0:10:57.200 --> 0:11:00.480 from in terms of managing risk in our portfolio. And 0:11:00.520 --> 0:11:03.679 as we sort of went into late February, we started 0:11:03.679 --> 0:11:08.680 to see some sectors turn off, their trend signals turned negative. 0:11:08.960 --> 0:11:12.280 It really wasn't until say March thirteenth or fourteenth, that 0:11:12.360 --> 0:11:15.880 the majority of the trend signals turned off. And so 0:11:16.040 --> 0:11:19.760 for our portfolio, even if we implement those signals on 0:11:19.800 --> 0:11:22.800 a daily basis, ultimately, what it sort of looks like 0:11:22.960 --> 0:11:25.480 it was a dimmer switch that as the market started 0:11:25.520 --> 0:11:29.160 to roll over, we started decreasing our exposure to equities 0:11:29.520 --> 0:11:34.679 sort of ultimately minimize that exposure right around late March UM, 0:11:34.720 --> 0:11:36.839 so there was a buffer there. I should be very clear. 0:11:36.920 --> 0:11:39.920 I think our portfolio compared to the SMP five hundred, 0:11:40.160 --> 0:11:43.080 we had sort of reduced the draw down by about 0:11:43.080 --> 0:11:46.840 a thousand and fifteen hundred basis points, which was great 0:11:46.840 --> 0:11:49.720 for the time. The problem was, as the market rebounded 0:11:49.840 --> 0:11:52.960 very quickly, the trend types of trend following signals we 0:11:52.960 --> 0:11:55.760 were looking at which were a little bit slower in nature, 0:11:56.360 --> 0:12:00.520 looking for those sort of prolonged nine month type trends, Uh, 0:12:00.640 --> 0:12:03.400 we're slower to get back in. And so almost all 0:12:03.440 --> 0:12:06.160 of that draw down buffer that we had created was 0:12:06.320 --> 0:12:08.880 erased within a two or three week period, and the 0:12:08.880 --> 0:12:12.000 market kept ripping upwards, and by the time that the 0:12:12.040 --> 0:12:16.160 trends turned back positive in the portfolio repositioned back into equities, 0:12:16.200 --> 0:12:21.520 there was a meaningful delay in exposure. So do you 0:12:21.640 --> 0:12:24.840 think that you could have changed some of the parameters 0:12:24.880 --> 0:12:28.040 of the trends that you were following and maybe had 0:12:28.040 --> 0:12:32.880 a better performance, or was there something inherent in having 0:12:32.920 --> 0:12:36.920 this trend following strategy, the systematic process that made it 0:12:37.040 --> 0:12:42.520 difficult to be nimble in an exceptionally quick market sell 0:12:42.559 --> 0:12:44.880 off and then recovery. Well, I think that that's a 0:12:44.920 --> 0:12:47.360 really great question because there's sort of two sides of 0:12:47.360 --> 0:12:50.679 the answer. There's the what I'll say is the investment 0:12:51.160 --> 0:12:54.959 professional answer, which is with hindsights, certainly, and I think 0:12:55.000 --> 0:12:57.720 even in real time, we recognize that the nature of 0:12:57.760 --> 0:13:01.440 the draw down was extraordinary and very different than what 0:13:01.480 --> 0:13:04.400 we had seen in the past. Um, So that's gonna 0:13:04.400 --> 0:13:07.959 be my sort of portfolio manager answer, which is, yes, 0:13:08.120 --> 0:13:10.840 we were seeing things that made us think the way 0:13:10.880 --> 0:13:14.160 in which our trend following signals work this draw down 0:13:14.240 --> 0:13:17.520 may be very very different. If I sort of silo 0:13:17.640 --> 0:13:20.880 that and then say, well, what is my asset manager answer, 0:13:20.920 --> 0:13:23.160 which is I run an asset management firm, and I 0:13:23.240 --> 0:13:26.120 have clients who are allocating to me for a very 0:13:26.120 --> 0:13:30.400 specific style of exposure. They know that it's longer term 0:13:30.440 --> 0:13:34.480 trend following. I have a fund prospectus that says it's 0:13:34.480 --> 0:13:38.160 a certain type of trend following. Can I actually override 0:13:38.200 --> 0:13:41.240 that in an extreme situation? Well, yes, I do, as 0:13:41.280 --> 0:13:44.280 a PM have ultimate discretion, But there are a lot 0:13:44.320 --> 0:13:48.320 of hurdles that you need to overcome to really have 0:13:48.440 --> 0:13:51.480 the justification to do that, and you're gonna certainly have 0:13:51.600 --> 0:13:54.000 to answer to your clients if you're wrong, and so 0:13:54.080 --> 0:13:57.000 it becomes a difficult trade off to make. In many ways, 0:13:57.040 --> 0:14:00.120 you are prohibited from making large changes if you have 0:14:00.280 --> 0:14:03.720 painted yourself in a corner, both in the way you 0:14:03.840 --> 0:14:06.640 have legally described your mandate as well as the way 0:14:06.679 --> 0:14:10.760 you've communicated your mandate to your clients. This is this 0:14:10.800 --> 0:14:14.520 is super interesting because in theory, these signals that you've 0:14:14.520 --> 0:14:18.280 built your strategy off of it from years and years 0:14:18.360 --> 0:14:23.600 realms and realms of market data. To make the call that, say, 0:14:23.840 --> 0:14:26.520 you know, in a in a very fast moving market, 0:14:26.600 --> 0:14:29.040 to make the call to say, well, maybe these aren't 0:14:29.040 --> 0:14:31.960 working anymore is very bold either way, because, as you 0:14:32.040 --> 0:14:35.760 point out, you would look very bad if people invested 0:14:35.800 --> 0:14:38.200 their money with you because you had this sort of 0:14:38.240 --> 0:14:41.240 strict rule follow approach or the sort of general rule 0:14:41.280 --> 0:14:44.520 follow up approach and then you unplugged it and then 0:14:44.680 --> 0:14:48.320 they continue to work absolutely. And I think that's where 0:14:48.760 --> 0:14:53.560 there is this um potential problem between right the way 0:14:53.600 --> 0:14:56.280 you should manage a portfolio and isolation versus the way 0:14:56.320 --> 0:14:59.440 you're managing a portfolio when it's others, other people's money 0:14:59.560 --> 0:15:03.320 and you have potentially a legal obligation to them both 0:15:03.360 --> 0:15:05.560 in the way again, the way you describe your portfolio, 0:15:05.720 --> 0:15:08.280 or again if it's a it's a forty X mutual 0:15:08.280 --> 0:15:10.920 fund or something like that, your prospectus is going to 0:15:11.200 --> 0:15:15.880 be varying degrees of restrictions. And if you write an 0:15:15.960 --> 0:15:21.360 overly restrictive investment strategy summary, then it's really going to 0:15:21.400 --> 0:15:23.880 prohibit you from making changes even if you think you 0:15:23.880 --> 0:15:27.280 should make them. And what we ultimately found this year 0:15:27.440 --> 0:15:29.960 was when we did decide that there were changes we 0:15:29.960 --> 0:15:34.080 wanted to make, even enacting those changes and giving ourselves 0:15:34.200 --> 0:15:38.080 the flexibility took months and months and months of prospectus 0:15:38.160 --> 0:15:43.040 updates and operational updates to be able to implement those changes. So, 0:15:43.240 --> 0:15:46.280 you know, it's it's interesting you say that UM. You know, 0:15:47.040 --> 0:15:49.880 earlier you were talking we were talking about, okay, why 0:15:49.920 --> 0:15:52.960 does trend following work? And as you pointed out, during 0:15:53.120 --> 0:15:56.880 extreme periods, UM market has become less of a random 0:15:56.880 --> 0:16:00.080 walk and there are certain mandates that investors have of 0:16:00.760 --> 0:16:04.320 UM that you know, either catalyzed selling or for selling 0:16:04.440 --> 0:16:07.840 or prevent the sort of normal market reaction in what 0:16:08.040 --> 0:16:11.400 you just described about your own fund, just the mechanics 0:16:11.400 --> 0:16:14.880 of say, updating the perspective, the perspectives and the legal 0:16:14.920 --> 0:16:19.240 requirements around that. Are you a sort of microcosm of 0:16:19.960 --> 0:16:23.760 I guess why profits can be made at various times 0:16:23.840 --> 0:16:26.360 by different players in the market, because there's a range 0:16:26.360 --> 0:16:31.160 of players like yourself that in different times simply cannot 0:16:31.280 --> 0:16:34.760 you know, for various sort of structural regulatory things, client 0:16:34.840 --> 0:16:38.080 mandate things cannot change fast enough. I mean, is this 0:16:38.120 --> 0:16:41.080 sort of what we're talking about basically of why opportunities exist? 0:16:41.520 --> 0:16:43.680 I absolutely think so. I think when you look at 0:16:43.800 --> 0:16:46.840 different areas of the market, what you will find is 0:16:46.880 --> 0:16:51.840 that there are price and sensitive buyers and sellers, particularly 0:16:51.880 --> 0:16:56.160 around certain points. I would argue that our thesis played 0:16:56.160 --> 0:16:58.400 out exactly as we thought it would in March, in 0:16:58.440 --> 0:17:01.120 the sense that I leave that there was a true 0:17:01.360 --> 0:17:05.439 procyclical cascade of de risking that occurred across various market 0:17:05.440 --> 0:17:08.600 players that fed into each other in a very rapid 0:17:08.600 --> 0:17:11.960 and violent fashion. Our problem was that we were just 0:17:12.000 --> 0:17:15.800 expecting it to occur over a much slower time horizon 0:17:16.000 --> 0:17:18.400 than it did, and so the speed at which our 0:17:18.440 --> 0:17:22.080 trend models were going to adapt to the markets was 0:17:22.440 --> 0:17:25.200 more designed for a two thousand and two thousand two 0:17:25.320 --> 0:17:28.119 or two thousand eight type environment. To see the market 0:17:28.240 --> 0:17:31.960 unwind so quickly, and I will say our our trend models, 0:17:32.000 --> 0:17:34.920 which are dynamic in nature, did speed up very quickly. 0:17:35.400 --> 0:17:38.080 So I think we did a fairly reasonable job on 0:17:38.400 --> 0:17:42.520 helping mitigate the downside. But to then have central banks 0:17:42.560 --> 0:17:46.520 step in and unleash a playbook that took them two 0:17:46.640 --> 0:17:49.359 years to develop from two thousand seven to two thousand 0:17:49.480 --> 0:17:53.240 nine and unveil it in two weeks, well, you don't 0:17:53.280 --> 0:17:55.480 want to say this time is different, but certainly it 0:17:55.600 --> 0:17:58.040 is different. Right for them to be able to do 0:17:58.119 --> 0:18:00.760 that and have the confidence to do that and think 0:18:00.760 --> 0:18:05.120 about the policy implications from a narrative perspective and how 0:18:05.160 --> 0:18:08.720 that reinvigorates confidence within the markets, I think it really 0:18:08.800 --> 0:18:26.640 was different this time. I want to talk a little 0:18:26.680 --> 0:18:29.840 bit more about your market thesis, but before we do, 0:18:30.000 --> 0:18:32.840 can you I mean you have you have this overall 0:18:33.240 --> 0:18:36.440 view of the markets and this idea of a cascading 0:18:37.000 --> 0:18:42.240 effect depending on available liquidity and risk appetite, so you 0:18:42.280 --> 0:18:44.200 can see a bunch of players in the market kind 0:18:44.200 --> 0:18:47.680 of step away at precisely the moment you would argue 0:18:47.760 --> 0:18:50.840 that they probably need to be providing liquidity, and that 0:18:50.960 --> 0:18:54.320 kind of creates this bad feedback loop for the overall market. 0:18:55.280 --> 0:19:00.399 How did your models actually incorporate or where your models 0:19:00.520 --> 0:19:06.320 hindered by incorporating data from previous years? So the financial crisis, 0:19:06.359 --> 0:19:10.159 is that why they weren't able to adapt as fast? 0:19:10.280 --> 0:19:12.879 Or is that why you were sort of operating on 0:19:12.880 --> 0:19:15.439 a different time scale. I guess what I'm trying to 0:19:15.440 --> 0:19:19.440 get at is how much just having an investment strategy 0:19:19.480 --> 0:19:24.800 that's actually based on historical data, based on evidence, how 0:19:24.880 --> 0:19:29.040 much does that hinder you when it comes to big 0:19:29.080 --> 0:19:32.480 events that might be unexpected. I think there's really sort 0:19:32.520 --> 0:19:35.120 of two questions embedded in here. I think the first 0:19:35.160 --> 0:19:37.840 question is about the data itself, and I would argue 0:19:37.880 --> 0:19:40.120 for a type of strategy like ours, it's it's less 0:19:40.160 --> 0:19:43.280 about the data because the trend following strategy isn't necessarily 0:19:43.280 --> 0:19:46.359 going to incorporate information about two thousand eight. Where that 0:19:46.480 --> 0:19:50.920 information creeps in is in the design of the strategy. Right, So, 0:19:51.400 --> 0:19:54.960 as a PM, my personal memory of two thousand eight 0:19:55.200 --> 0:19:57.919 and environments like two thousand and other market environments that 0:19:57.960 --> 0:20:01.520 I've studied to say this is the sort of trend 0:20:01.600 --> 0:20:05.159 horizon that I think is going to be most efficient 0:20:05.680 --> 0:20:10.159 for trying to mitigate those types of draw downs without 0:20:10.480 --> 0:20:13.840 the foresight or expectation that the next draw down may 0:20:13.840 --> 0:20:17.200 be rapid, violent and only take a month or two 0:20:17.359 --> 0:20:20.159 to bounce back, right. And so I think it's not 0:20:20.200 --> 0:20:22.600 necessarily that it's in the data for a strategy like ours, 0:20:22.640 --> 0:20:25.639 but it is in the memory of the people who 0:20:25.680 --> 0:20:28.119 are designing the strategy. We can talk about quantitative and 0:20:28.160 --> 0:20:30.480 systematic all day long, but at the end of the day, 0:20:30.520 --> 0:20:33.400 there is a very human element that goes into designing 0:20:33.440 --> 0:20:37.159 these portfolios. Who are choosing what data is relevant and 0:20:37.960 --> 0:20:41.800 what sort of constraints or information they want to build 0:20:41.800 --> 0:20:45.560 into it. Philosophically, I mean this gets to one of 0:20:45.600 --> 0:20:49.080 the bigger questions is you know, what do you what 0:20:49.119 --> 0:20:51.639 do you do with the human element? Because you know, 0:20:51.680 --> 0:20:54.000 one of the most popular you know, it's a cliche, 0:20:54.080 --> 0:20:57.480 they always say, you know, your motions are a terrible guide. 0:20:58.000 --> 0:20:59.960 They tell you to do to buy at the time 0:21:00.080 --> 0:21:02.520 up to sell at the bottom. And one of the 0:21:02.560 --> 0:21:06.840 most popular strategies for a lot of people these days 0:21:07.000 --> 0:21:10.159 is a very simple algorithm when you have money by 0:21:10.200 --> 0:21:13.080 and never sell it. So basically this sort of this 0:21:13.280 --> 0:21:17.359 new dominant strategy is by every month by an ETF. 0:21:17.560 --> 0:21:20.320 Don't think about it, keep buying, keep buying, keep buying, 0:21:20.560 --> 0:21:24.000 keep buying. And so you know that's proven. That turned 0:21:24.040 --> 0:21:27.840 out to be a pretty good strategy in just absolutely 0:21:27.880 --> 0:21:31.200 whatever you do, don't sell. How do you think about, however, 0:21:31.359 --> 0:21:35.680 with more sophisticated strategies, the degree to which the human 0:21:35.760 --> 0:21:39.840 element either needs to be brought in or excised. Well, 0:21:40.040 --> 0:21:42.680 if you'll allow me a little bit of a philosophical tangent, 0:21:43.760 --> 0:21:47.080 i'll answer. You know, I've long thought about this difference 0:21:47.080 --> 0:21:49.359 between what's really the difference between a systematic manager and 0:21:49.359 --> 0:21:52.119 a discretionary manager, And I think if you sort of 0:21:52.160 --> 0:21:55.359 go towards the absurd and say, well, let's allow a 0:21:55.400 --> 0:21:59.560 systematic manager to follow around a discreptionary manager and just 0:21:59.680 --> 0:22:02.400 write down all the rules. Right, if you've ever looked 0:22:02.400 --> 0:22:05.560 at a discretionary manager's pitch book, there's always that upside 0:22:05.560 --> 0:22:08.120 down pyramid that talks about their universe and how they 0:22:08.160 --> 0:22:11.720 filter the securities they purchase, and so there is a process, 0:22:12.119 --> 0:22:15.040 and so that systematic manager could take that process and 0:22:15.080 --> 0:22:17.639 turn it into a set of rules and codify it 0:22:17.960 --> 0:22:21.119 and implement it, and then whenever those rules are broken, 0:22:21.160 --> 0:22:23.840 they could say to the discretionary manager, well, why do 0:22:23.920 --> 0:22:26.320 you break those rules? And Okay, there's a new rule, 0:22:26.320 --> 0:22:29.160 and they could just keep layering that in and ultimately, 0:22:29.160 --> 0:22:31.240 if you sort of take that to its absurd conclusion, 0:22:31.400 --> 0:22:36.240 what that really means is discretionary managers are systematic managers, 0:22:36.800 --> 0:22:40.119 with the big difference that their value adds should come 0:22:40.160 --> 0:22:43.560 in the idiosyncratic decisions that they make, that they are 0:22:43.760 --> 0:22:48.480 breaking the core rules for a truly unprecedented reason. And 0:22:48.680 --> 0:22:51.159 if you reverse that, I think you could also argue 0:22:51.160 --> 0:22:56.119 then that systematic managers are discretionary managers. But they're ultimately 0:22:56.160 --> 0:23:01.160 selling that optionality that they are foregoing ability to adapt 0:23:01.200 --> 0:23:05.639 their rules to an idiosyncratic environment. And the argument that 0:23:05.680 --> 0:23:07.520 they would make is the premium they're going to collect 0:23:07.520 --> 0:23:09.800 as they're avoiding the behavioral bias, and that there's an 0:23:09.880 --> 0:23:13.080 edge there. But the down side, the skew that they're 0:23:13.119 --> 0:23:15.760 they're really selling might be, well, this time is different 0:23:16.000 --> 0:23:18.679 and the rules are broken, and now they're foregoing that 0:23:18.760 --> 0:23:22.400 optionality to adapt. And so I think that's in my 0:23:22.480 --> 0:23:26.560 mind the ultimate distinction between them whether they're selling that 0:23:26.640 --> 0:23:30.240 optionality or retaining it. So I mentioned that I wanted 0:23:30.240 --> 0:23:34.120 to dig in a little bit to your markets thesis, 0:23:34.119 --> 0:23:38.800 your grand markets thesis before the big sell off in March. 0:23:38.880 --> 0:23:42.280 So you wrote this paper called Liquidity Cascades, and I 0:23:42.560 --> 0:23:45.119 read it today. It's really good, and it's sort of 0:23:45.200 --> 0:23:48.840 connects a lot of the different things that we've been 0:23:48.880 --> 0:23:52.280 thinking and discussing on all loots. But primarily it's about 0:23:52.280 --> 0:24:00.280 this self reflexive relationship between central banks, markets and risk 0:24:00.400 --> 0:24:05.600 taking behavior, and this idea that you can get feedback 0:24:05.640 --> 0:24:08.680 loops like deep in the bowels of the financial system, 0:24:08.760 --> 0:24:11.879 one of the obvious ones being um the treasury trains 0:24:11.960 --> 0:24:14.560 that blew up in March and then sort of fed 0:24:14.600 --> 0:24:20.000 into the equity market. Could you describe that pisis a 0:24:20.000 --> 0:24:25.000 little more or what inspired you to put that together 0:24:25.280 --> 0:24:30.280 in that way? Absolutely? So after March I took a 0:24:30.320 --> 0:24:33.960