WEBVTT - Maria Vassalou on the Small-Cap Effect 0:00:01.960 --> 0:00:06.040 This is Mesters in Business with very Rid Holds on 0:00:06.200 --> 0:00:11.560 Bloomberg Radio. This week on the podcast, I have an 0:00:11.640 --> 0:00:17.040 extra extra special guest. Maria Vassalu has a fascinating history 0:00:17.040 --> 0:00:21.080 and background London School of Economics to Columbia School of Business, 0:00:21.160 --> 0:00:24.120 where she actually was a professor for over a decade 0:00:24.600 --> 0:00:29.120 and started consulting to the hedge fund and financial services industry, 0:00:29.600 --> 0:00:34.559 and that led her to various jobs at Washerstein, Perella 0:00:34.880 --> 0:00:38.920 McKenzie's asset management group. She worked with George Soros, she 0:00:39.000 --> 0:00:42.920 worked with Steve Cohen at ZAC Capitol, and ultimately ends 0:00:43.040 --> 0:00:49.000 up joining Goldman Sachs Asset Management Group as cocio. A 0:00:49.040 --> 0:00:54.760 fascinating approach to macro, very quantitatively driven and very academic 0:00:54.840 --> 0:00:59.120 research oriented. She wants to know exactly when this, that 0:00:59.240 --> 0:01:01.600 and the other thing happens, What does it mean for 0:01:02.160 --> 0:01:05.280 this segment of the market. When do you own growth? 0:01:05.319 --> 0:01:10.280 When do you own equity? Why is certain anomalies persistent? 0:01:10.319 --> 0:01:14.040 And why do some seem to get arbitraged away fairly quickly. 0:01:14.440 --> 0:01:18.600 I found this to be an absolutely fascinating conversation and 0:01:18.760 --> 0:01:23.240 I think you will also with no further ado Goldman Sachs, 0:01:23.680 --> 0:01:28.120 Maria Vassalu tell us a little bit about the sort 0:01:28.160 --> 0:01:33.199 of work you did. How relevant was the academic research 0:01:33.560 --> 0:01:38.399 to what you're actually doing today. Well, actually, it sounds 0:01:38.520 --> 0:01:43.040 very unusual to go from academia to the industry, and 0:01:43.360 --> 0:01:48.000 usually it's not considered a very successful path. But in 0:01:48.040 --> 0:01:51.520 my case it was very helpful because I had the 0:01:51.560 --> 0:01:57.000 opportunity to spend over ten years doing intensive research in 0:01:57.120 --> 0:02:01.360 the intersection of macro and fine and as surprising and 0:02:01.560 --> 0:02:05.960 all these questions that I was trying to answer had 0:02:06.360 --> 0:02:12.560 direct applications to Hedgemund strategies and portfolio management, and so 0:02:12.760 --> 0:02:15.640 actually part of the reason I moved to the industry 0:02:15.840 --> 0:02:19.200 was because while I was doing this research and presenting 0:02:19.200 --> 0:02:23.400 it around and publishing it in academic journals, it was 0:02:23.440 --> 0:02:27.639 attracting attention from the industry, and I had the opportunity 0:02:27.720 --> 0:02:32.000 to be a retained consultant for Citadel, for Deutsche Acid 0:02:32.080 --> 0:02:36.480 Management and then eventually also for Saurusmund Management, and so 0:02:36.560 --> 0:02:40.480 along the way I was getting offers to join the industry, 0:02:40.520 --> 0:02:44.639 and finally I decided to join the SOURS. So it 0:02:44.680 --> 0:02:48.120 wasn't like a big Eureka moment. It just gradually became 0:02:48.160 --> 0:02:51.480 apparent that you were working in a space that was 0:02:51.680 --> 0:02:56.720 very valuable to people managing capital on a very let's 0:02:56.720 --> 0:02:59.760 call it aggressive basis, just a hey, we're looking for 0:03:00.240 --> 0:03:03.280 we're looking to outperform, and what Maria does could be 0:03:03.320 --> 0:03:06.240 really useful to us. That was certainly part of it. 0:03:06.320 --> 0:03:10.480 There was also an intellectual like curiosity aspect to it, 0:03:11.280 --> 0:03:15.440 because when I was doing that work, it was also 0:03:15.520 --> 0:03:21.600 the time where behavioral finance became more prevalent, if you like. 0:03:22.080 --> 0:03:27.400 And I was always on the camp of rational risk 0:03:27.480 --> 0:03:33.480 based explanations for various surprising phenomena. And my view was 0:03:33.560 --> 0:03:38.920 always if some if an anomaly persist and it doesn't 0:03:38.920 --> 0:03:44.000 go away, then maybe it's not an anomaly, maybe it's 0:03:44.160 --> 0:03:46.920 risk based. Then it's a risk factor that we haven't 0:03:46.960 --> 0:03:50.280 really accounted for. And so a lot of my research 0:03:50.560 --> 0:03:54.960 was related to trying to uncover what were the underlying 0:03:55.080 --> 0:03:58.680 risk factors. And the place where I was looking for 0:03:58.720 --> 0:04:01.840 those risk factors was in the real economy. So I 0:04:01.920 --> 0:04:08.240 was relating our surprises to GDP growth, to investment growth, 0:04:08.440 --> 0:04:17.280 to default rest factors like this, and so I was 0:04:17.320 --> 0:04:21.359 providing explanations for our surprising anomalists, such as the small 0:04:21.400 --> 0:04:25.239 cap effect or the value effect. Those were the first 0:04:25.240 --> 0:04:28.400 two that popped into my mind when you said, Hey, 0:04:28.480 --> 0:04:31.400 is this truly anomalous or is there a risk factor? 0:04:31.880 --> 0:04:34.880 Some people have said small caps tend to be more volatile, 0:04:34.920 --> 0:04:39.400 more risky. That's where the additional performance comes from. When 0:04:39.400 --> 0:04:41.600 we look at value, a lot of people say, well, 0:04:41.640 --> 0:04:44.760 they're widely disliked, that's why they're cheap. So there's a 0:04:44.839 --> 0:04:47.800 behavioral side. How do you crunch the numbers on that? 0:04:48.120 --> 0:04:51.600 And where do you come out on small cap and value? Yeah, 0:04:51.600 --> 0:04:54.400 it was actually very interesting because when I looked at 0:04:54.440 --> 0:04:59.880 the small caps, it's actually if you dissect the small caps, 0:05:00.160 --> 0:05:03.480 you see that the small cap effect always exists in 0:05:03.560 --> 0:05:06.760 the smallest of the small caps and it's related to 0:05:06.880 --> 0:05:10.520 default risk. Wait a second, so there's a small cap effect, 0:05:10.839 --> 0:05:14.600 and then within small caps there's a microcap effect and 0:05:14.680 --> 0:05:20.120 even smaller cap effect. Yes, and what happens is this 0:05:20.240 --> 0:05:24.480 small cap effect is related to the default probability. So 0:05:24.520 --> 0:05:28.600 I have a paper where I computed default probabilities based 0:05:28.640 --> 0:05:31.360 on Merton's model, and I did this for the whole 0:05:31.400 --> 0:05:34.840 Crows section of assets, and then I started them and 0:05:35.040 --> 0:05:39.320 created desiles and so on and tracked how the behavior 0:05:39.560 --> 0:05:43.320 is over time, and of course you see that depending 0:05:43.320 --> 0:05:46.920 on the part of the business cycle you're going through, 0:05:47.240 --> 0:05:51.640 the default probability varies over time, and it increases during 0:05:52.600 --> 0:05:56.159 downturns of the business cycle and so on. And when 0:05:56.400 --> 0:06:01.039 when that happens, then the small cap effect becomes much 0:06:01.080 --> 0:06:03.960 more prominent, And so you see it in the whole 0:06:04.000 --> 0:06:07.640 cross section of small caps. But when the default probabilities 0:06:07.680 --> 0:06:11.160 are lower and you look at the whole cross section 0:06:11.200 --> 0:06:15.760 of small caps, it's not so apparent. So people say 0:06:15.839 --> 0:06:18.719 that it goes away, but it doesn't really go away. 0:06:19.160 --> 0:06:22.240 It's just it's a matter of magnitude. Then where are 0:06:22.240 --> 0:06:25.599 you're looking for it? Well, that's really interesting. What about 0:06:25.640 --> 0:06:29.200 in the value space? Do you see the same issue 0:06:29.279 --> 0:06:32.960 of what used to be called Benjamin grab called stubs 0:06:33.120 --> 0:06:36.840 or cigar stubs. Is that the same default risk when 0:06:36.920 --> 0:06:40.320 stocks become very very cheap or is there something else 0:06:40.360 --> 0:06:43.719 at play there In the case of value versus growth, 0:06:45.160 --> 0:06:48.600 it was it's more related to the level of GDP 0:06:48.839 --> 0:06:54.560 growth and investment growth in different sectors of the economy, 0:06:54.839 --> 0:06:58.400 So it's not so much a default aspect, but it's 0:06:58.480 --> 0:07:05.720 more has to do with a variation of real GDP growth. 0:07:05.880 --> 0:07:09.600 So when GDP is growing rapidly, I would assume you 0:07:09.640 --> 0:07:12.920 would want growth stocks, and when things are going sideways, 0:07:12.960 --> 0:07:15.840 there's a greater margin of safety with value and that 0:07:16.640 --> 0:07:19.520 and that's why you So last year, for instance, when 0:07:19.800 --> 0:07:23.600 GDP growth started becoming a little bit more muted and 0:07:23.800 --> 0:07:28.480 expectations were for a lower GDP growth going forward, value 0:07:28.480 --> 0:07:32.400 stocks outperformed growth but by a huge margin, right, big 0:07:32.480 --> 0:07:36.200 big disparity. Yeah, So at that time I would go 0:07:36.320 --> 0:07:40.480 to conferences and publish papers and make those arguments, and 0:07:40.520 --> 0:07:43.360 then I had other colleagues that I would try to 0:07:43.600 --> 0:07:51.720 provide behavior explanations. And similarly with the momentum effect which 0:07:52.240 --> 0:07:57.000 I had related to corporate innovation as I was calling it, 0:07:57.040 --> 0:08:01.920 which was separate innovation, which was really a firm level 0:08:02.120 --> 0:08:08.720 total factor productivity. So how much innovation companies produce and 0:08:08.760 --> 0:08:12.960 how long they can remain leaders in that innovation to 0:08:13.120 --> 0:08:18.400 really maintain that momentum. So a company becomes very innovative, 0:08:18.440 --> 0:08:20.920 you get a little bit of a flywheel effect and 0:08:21.000 --> 0:08:25.200 that innovation DNA starts to spill over into everything they do. 0:08:25.320 --> 0:08:27.440 Is it just that simple? Right? And then but then 0:08:27.480 --> 0:08:30.120 it's it's a matter of being able to maintain this 0:08:31.280 --> 0:08:35.480 um and can companies maintain this indefinitely or is there 0:08:35.480 --> 0:08:39.920 a sly? Not? Usually not um, And so they go 0:08:40.040 --> 0:08:47.439 into cycles. And it also relates to when they are losers, um, 0:08:48.160 --> 0:08:52.440 you know, what's the probability of recovering And it really 0:08:52.480 --> 0:08:55.160 has to do with whether they have the ability to 0:08:55.200 --> 0:08:59.040 innovate and get out of that trap. So so you 0:08:59.080 --> 0:09:03.040 can see a very high correlation between losers and winners 0:09:03.400 --> 0:09:08.760 with respect to how they perform on that measure. But anyway, 0:09:08.880 --> 0:09:13.200 so we had all these ideas of about how all 0:09:13.240 --> 0:09:17.680 this different phenomena were formed and what was driving them, 0:09:17.800 --> 0:09:20.800 And of course my colleagues on the behavioral site had 0:09:20.880 --> 0:09:25.760 different ideas, and so we were always debating these topics 0:09:25.800 --> 0:09:30.320 at conferences and through publications, and at some time, at 0:09:30.320 --> 0:09:33.560 some point it became to me a little bit repetitive, 0:09:33.640 --> 0:09:39.040 and I felt like nobody could equivocally prove their point 0:09:39.320 --> 0:09:43.000 as to who is really right. And so at some 0:09:43.120 --> 0:09:46.400 point I thought, well, if I can go and manage 0:09:46.440 --> 0:09:51.440 money based on this risk based explanations and based on 0:09:51.480 --> 0:09:55.720 the way I understand how the world functions, how the 0:09:55.800 --> 0:10:00.880 market's functions, if that works, then that's one form of 0:10:01.040 --> 0:10:06.720 justification of what I'm doing really really intriguing. It's um. 0:10:07.040 --> 0:10:12.800 It sort of is like the John Sex poem about 0:10:12.800 --> 0:10:16.160 the blind men describing the elephant. They don't have to 0:10:16.200 --> 0:10:18.439 be one doesn't have to be right or wrong. They 0:10:18.480 --> 0:10:20.760 could both be right. You're just approaching it from a 0:10:20.800 --> 0:10:24.439 different angle. Is that fair? Or is clearly one is 0:10:24.520 --> 0:10:26.800 right and one is wrong? And that's that. I think 0:10:26.840 --> 0:10:30.840 it's much more nuanced, and as the time goes by, 0:10:30.920 --> 0:10:35.400 I think the two lines get blurred. Also because of technology, 0:10:35.480 --> 0:10:41.959 because of the increase presence of retail investors in the markets, 0:10:42.360 --> 0:10:47.280 the market microstructure has changed, and so UM, it's much 0:10:47.360 --> 0:10:53.599 more common now to see prolonged deviations from fundamentals in 0:10:53.760 --> 0:10:57.600 the in the market. And we've seen that recently as well. 0:10:58.040 --> 0:11:01.959 And so I wouldn't say that the one approach is 0:11:02.080 --> 0:11:04.319 right and the other one is wrong. But maybe it's 0:11:04.320 --> 0:11:07.959 a matter of timing. I think the risk based explanations 0:11:08.160 --> 0:11:14.680 need longer time to play out some of this behavioral driver, 0:11:14.960 --> 0:11:18.360 some more short term drivers. So you were consulting to 0:11:18.400 --> 0:11:22.240 the industry while you're an academia that had to make 0:11:22.320 --> 0:11:25.719 that transition. When you finally decided to jump in with 0:11:25.760 --> 0:11:29.440 both feet, I'm assuming you were prepared for what you 0:11:29.480 --> 0:11:33.400 were jumping into. It wasn't a big shock or am 0:11:33.400 --> 0:11:37.160 I wrong? Once you left the quiet confines of academia 0:11:37.760 --> 0:11:40.480 Wall Street is still a shock to the system. Well, 0:11:40.480 --> 0:11:44.800 it was certainly not exactly a shock, but I had 0:11:44.840 --> 0:11:48.480 to get adapted to it. But I am someone who 0:11:48.640 --> 0:11:52.360 is quite adaptable. I left my country. I lived in 0:11:52.400 --> 0:11:59.440 six different countries. I came to the US, and so 0:12:00.280 --> 0:12:03.960 you know, I'm used to changing environments and try to 0:12:04.000 --> 0:12:09.520 adapt to these new environments. Suddenly going to Sarus was 0:12:09.600 --> 0:12:14.440 a big eye opener. And also I was there during 0:12:14.480 --> 0:12:18.120 a very interesting time in the markets. What years were 0:12:18.120 --> 0:12:22.480 those I joined in the summer of two thousand and six? 0:12:22.800 --> 0:12:26.200 Were you there for the financial crisis? Pretty much? And 0:12:27.800 --> 0:12:32.480 I started, actually I developed my strategies and built the 0:12:32.880 --> 0:12:36.640 Quantitative Strategiest group from the summer of two thousand and 0:12:36.640 --> 0:12:42.920 six onwards. And I started running my strategies with money 0:12:43.400 --> 0:12:49.360 in March of oh seven, so soon before the quant meltdown, right, 0:12:49.960 --> 0:12:54.360 which was interesting, and so I suddenly I had a 0:12:54.360 --> 0:12:59.480 baptism by fire in the markets. But they was a 0:12:59.520 --> 0:13:03.880 great experience. We did very well during the quant meltdown, 0:13:04.440 --> 0:13:07.880 and it was also an opportunity to see up close 0:13:08.160 --> 0:13:11.760 what was happening behind the scenes in the markets, how 0:13:11.800 --> 0:13:17.120 the financial crisis was developing. And also it was very 0:13:17.120 --> 0:13:22.040 interesting because even though George Sarrows had retired from active 0:13:22.120 --> 0:13:25.920 in investing, when he saw what was happening in the markets, 0:13:26.000 --> 0:13:30.120 he came back and so I'm excited. Yeah, And so 0:13:30.200 --> 0:13:34.080 I had the opportunity to observe him up close, to 0:13:34.160 --> 0:13:37.640 listen to his views, to interact with him, and that 0:13:37.760 --> 0:13:41.240 was certainly a great experience I can imagine. So when 0:13:41.280 --> 0:13:46.080 you go through a substantial macro event, whether it was 0:13:46.120 --> 0:13:50.560 the quantz crash or the financial crisis or even the pandemic, 0:13:51.080 --> 0:13:54.160 does that send you back to your models to tweak them. 0:13:54.160 --> 0:14:00.360 Do those giant events affect how markets behave subsequently leads 0:14:00.360 --> 0:14:04.080 you to have to make some changes or hey, the 0:14:04.160 --> 0:14:05.920 model is going to do what the model does and 0:14:06.000 --> 0:14:09.560 it doesn't matter what happens out there. Well, que models 0:14:09.640 --> 0:14:14.680 always need to be evolved, so you kind of staled it. Yes, 0:14:14.800 --> 0:14:17.640 you can just build it and forget it, but it 0:14:17.720 --> 0:14:21.440 has to be done in a way that keeps up 0:14:21.440 --> 0:14:27.040 with the developments in the markets. So, for instance, when 0:14:27.920 --> 0:14:31.640 the British referendum happened, Well, we didn't have such an 0:14:31.720 --> 0:14:35.200 event before in the market, So that's not something where 0:14:35.320 --> 0:14:39.160 you want to make your model adapted to because we're 0:14:39.200 --> 0:14:41.880 not going to be having these events all the time. 0:14:42.560 --> 0:14:45.720 But that's an instance where you want to take your 0:14:45.760 --> 0:14:49.360 model and stress tested to see how it will behave 0:14:49.800 --> 0:14:54.040 depending on different scenarios that may transpire as a result 0:14:54.160 --> 0:14:58.480 of this event. So that's what we would do, and 0:14:58.520 --> 0:15:02.200 then we will ey'd whether to take down risk or 0:15:02.320 --> 0:15:05.760 leave the risk on and so on. If you have 0:15:05.920 --> 0:15:11.080 other phenomenon, like you know, changes and correlations between assets 0:15:11.200 --> 0:15:15.160 or changes in the level of volatilities, these are things 0:15:15.160 --> 0:15:19.800 that you want the model to adapt to going forward 0:15:20.440 --> 0:15:24.560 and incorporate this information into the model. So in that 0:15:24.640 --> 0:15:28.760 case you want to evolve it. Or there maybe factors 0:15:28.760 --> 0:15:33.760 that were not present before and you want to inform 0:15:33.800 --> 0:15:37.680 the model with it. For instance, how the monetary policy 0:15:38.360 --> 0:15:42.240 changes over time, the fact that we had QE for 0:15:42.320 --> 0:15:44.920 a long period of time. All these things are things 0:15:45.040 --> 0:15:48.600 you want to include in the model, but you have 0:15:48.720 --> 0:15:53.840 to be selective and really treat each case separative. So 0:15:53.880 --> 0:15:58.280 you're working with George Soros, known as a big macro trader, 0:15:58.400 --> 0:16:01.840 he makes big bets about these large events, you end 0:16:01.920 --> 0:16:05.760 up going to Steve Cohen and Zach Capitol. He's much 0:16:05.800 --> 0:16:10.200 more of a granular trader. He is not necessarily looking 0:16:10.240 --> 0:16:13.640 at the big events. He's looking at things on really 0:16:13.640 --> 0:16:16.120 where the rubber meets the road, so to speak. What 0:16:16.320 --> 0:16:19.520 was that transition like to go from a very top 0:16:19.560 --> 0:16:22.680 down approach to somebody who's you know, right there in 0:16:22.720 --> 0:16:26.120 the weeds with the rest of the trading desk. Yes, 0:16:27.040 --> 0:16:30.960 and now the great lesson. And I was still a 0:16:30.960 --> 0:16:34.880 global macro portfolio manager with my own silo at the 0:16:35.480 --> 0:16:39.880 SAC Capital. But as you said, at Sorrows it was 0:16:39.960 --> 0:16:46.840 all about big macro beds, and at the SAC Capitol 0:16:47.040 --> 0:16:51.040 it was all about risk management. So even though when 0:16:51.120 --> 0:16:54.800 I came from academia to Sorrows, I would look at 0:16:54.880 --> 0:16:58.000 how they were running the portfolios, and I was constantly 0:16:58.080 --> 0:17:00.600 scared because I felt they were taken in way too 0:17:00.680 --> 0:17:04.280 much risk compared to what I thought from an academic 0:17:04.359 --> 0:17:07.840 perspective they should be doing. Of course, I was novice 0:17:07.960 --> 0:17:12.720 at that time in the profession. Then I went to 0:17:13.480 --> 0:17:20.400 SAC and I realized that actually being careful with risk 0:17:20.440 --> 0:17:24.439 management is very much respected and even more than what 0:17:24.640 --> 0:17:29.840 I thought should have been happening at Sorrows, and so 0:17:30.359 --> 0:17:35.879 I spent the subsequent years trying to refine my models, 0:17:36.800 --> 0:17:43.480 make them much more smooth in terms of their return stream, 0:17:43.920 --> 0:17:48.760 focused much more on risk management downside the risk hedging, 0:17:49.359 --> 0:17:52.680 and I think the models became better as a result. 0:17:53.080 --> 0:17:55.280 So let's talk a little bit about how you ended 0:17:55.359 --> 0:17:58.399 up at Goldman. You were at Columbia School of Business 0:17:58.800 --> 0:18:02.760 where you were teaching, you were at Soros and zach Capital. 0:18:03.400 --> 0:18:07.720 What attracted you to government, Well, it's actually, um, the 0:18:07.800 --> 0:18:13.440 whole asset management business is changing. So we went from 0:18:13.440 --> 0:18:19.159 a period where hedge funds was really the hot area 0:18:19.280 --> 0:18:22.040 to be and of course there are all this big 0:18:22.080 --> 0:18:27.639 hedge funds that were developed over time. Um. But over time, 0:18:27.680 --> 0:18:31.199 as you know, there was this big shift or was 0:18:32.960 --> 0:18:37.520 passive investing, and so that was a big challenge for 0:18:37.520 --> 0:18:41.480 for hedge funds. At the same time, we had all 0:18:41.480 --> 0:18:46.359 this decrease in volatility and financial repression because of the 0:18:46.480 --> 0:18:51.240 QE and the extra liquidity that was in the markets 0:18:51.359 --> 0:18:58.040 that made them trading in in hedge funds much more difficult, 0:18:58.080 --> 0:19:02.560 if you like, in terms of provide superior returns. I'm 0:19:02.560 --> 0:19:06.000 glad you brought that up because if you look at 0:19:06.359 --> 0:19:10.400 hedge fund performance before the Financial crisis, there's a lot 0:19:10.400 --> 0:19:14.640 of alpha generators. The hedge fund industry generally is outperforming 0:19:14.680 --> 0:19:17.880 their benchmarks. I mean, not just the top decile. As 0:19:17.920 --> 0:19:20.360 a group, they seem to have done very well. And 0:19:20.400 --> 0:19:24.680 then post financial crisis it became very hard to generate 0:19:24.720 --> 0:19:28.480 alpha and there was a huge gap between the big 0:19:28.480 --> 0:19:32.920 winners and the losers. Are you attributing that to zero 0:19:32.960 --> 0:19:37.639 interest rate and quantitative easing or did things just change 0:19:37.640 --> 0:19:40.800 so much people didn't adapt quickly enough. Well, there were 0:19:40.800 --> 0:19:44.800 two things. I mean, my strategies were always in the 0:19:44.880 --> 0:19:48.919 space of relative value across SASA classes, so there there 0:19:49.040 --> 0:19:53.960 was all Yes, there was always some volatility to pick up, 0:19:54.400 --> 0:19:58.080 and so the strategies kept working. But by and large 0:19:58.119 --> 0:20:02.159 in the overall industry, if you look at long short equity, 0:20:02.680 --> 0:20:07.080 there was very little, you know, within us a class 0:20:07.160 --> 0:20:11.760 volatility to pick up. And also you have a period 0:20:11.960 --> 0:20:19.960 that because of this extremely liquidity and quantitative easing, equities 0:20:20.040 --> 0:20:23.920 were performing extremely well and so being passive and just 0:20:24.080 --> 0:20:29.199 holding the index you were doing great. So what was 0:20:29.240 --> 0:20:34.439 the point of getting into hedge funds having a zero 0:20:34.600 --> 0:20:40.399 beta exposure? Or going into other strategies. And so you 0:20:40.600 --> 0:20:44.520 saw that the hedge fund industry started changing over time. 0:20:44.880 --> 0:20:48.960 A lot of traditional macro funds actually started becoming more 0:20:49.119 --> 0:20:53.320 equity oriented funds or including a lot of equity exposure 0:20:54.080 --> 0:20:58.480 just to try to pick up beta in their strategies. 0:20:58.640 --> 0:21:03.200 And also there was an increased consolidation of the industry 0:21:03.359 --> 0:21:09.080 towards the bigger managers. But to me, at the same time, 0:21:09.359 --> 0:21:16.720 I was finding this uh, this uh concentration on passive 0:21:16.800 --> 0:21:21.960 investing also problematic because passive investing works when the markets 0:21:21.960 --> 0:21:25.760 are efficient, and the markets are efficient when there is 0:21:25.880 --> 0:21:29.919 enough trading happening for new information to be incorporated in 0:21:30.000 --> 0:21:34.320 the prices. If everybody is a passive investor, then you 0:21:34.400 --> 0:21:39.199 don't have this mechanism in place to incorporate information in 0:21:39.320 --> 0:21:45.280 prices right away to really benefit from them. So, so 0:21:45.320 --> 0:21:49.320 how much active management does there have to be for 0:21:49.480 --> 0:21:53.720 price discovery to really take place? And I've asked people 0:21:53.760 --> 0:21:56.919 like Andrew Lowe and MIT who said, you can have 0:21:57.040 --> 0:21:59.960 ninety percent passive, the remaining ten percent as we're all 0:22:00.160 --> 0:22:03.320 your price discovery will take place. Does that sound like 0:22:03.400 --> 0:22:06.040 it's a lot or do you agree with that perspective? 0:22:06.480 --> 0:22:12.399 That's Andrews answer. I think derives from the idea of 0:22:12.440 --> 0:22:16.560 the marginal investor, as we're saying academia. So all you 0:22:16.640 --> 0:22:20.000 need is the marginal investor is rational and always ready 0:22:20.000 --> 0:22:23.440 to take advantage of opportunities. Yes, but it's not very 0:22:23.440 --> 0:22:28.600 clear who the marginal investor is in practice. If they exist, 0:22:28.800 --> 0:22:32.879 then what I have noticed through the fifteen years that 0:22:33.000 --> 0:22:38.120 I've been managing my own strategies is that the markets 0:22:38.119 --> 0:22:42.720 have become a little bit less efficient over time, in 0:22:42.760 --> 0:22:46.399 the sense that they have become that you seem longer 0:22:46.440 --> 0:22:51.200 deviations from fundamentals. Eventually they do correct, but you seem 0:22:51.320 --> 0:22:56.080 longer deviations from fundamentals. Sometimes you see more intra day 0:22:56.240 --> 0:23:02.359 volatility in certain events, especially around announcements and so on. 0:23:03.080 --> 0:23:09.000 And so maybe this is attributable to an increased exposure 0:23:09.000 --> 0:23:14.480 to passive management. Maybe it's attributable to more noise traders 0:23:14.600 --> 0:23:17.959 what we use to call noise traders, which are effectively 0:23:18.040 --> 0:23:21.600 retail investors. Right, well, let me let's stay with this 0:23:21.680 --> 0:23:24.760 a second, because I'm intrigued by the concept of the 0:23:24.840 --> 0:23:29.960 market becoming less efficient. When I look at the sixties, 0:23:30.000 --> 0:23:33.440 the seventies, the eighties, and the nineties, it seems as 0:23:33.520 --> 0:23:37.800 if we've gotten more and more heavily focused on technology 0:23:37.880 --> 0:23:42.520 and program training and now algorithmic and high frequency training, 0:23:42.520 --> 0:23:46.000 and I would assume that that would make the market 0:23:46.119 --> 0:23:52.200 more efficient and harder to spot arbitrage opportunities and these 0:23:53.160 --> 0:24:00.280 various anomalies. You're suggesting passive is creating less efficiency. Does 0:24:00.280 --> 0:24:04.040 that mean there's more opportunity for active traders. I think 0:24:04.040 --> 0:24:07.479 there is more intradate trading now than it used to be. 0:24:07.520 --> 0:24:10.879 So you have the passive trade, the passive investors, and 0:24:10.920 --> 0:24:14.560 then you have a lot of intradate trading, and that's 0:24:14.600 --> 0:24:18.800 based on algos that are looking for short term trends 0:24:18.920 --> 0:24:23.840 to capitalize. Some of them are AI based, so they 0:24:23.880 --> 0:24:27.400 may be looking for particular words and then they will 0:24:27.520 --> 0:24:32.680 extrapolate from that. For instance, it was interesting to notice 0:24:32.840 --> 0:24:39.040 in the last FAT meeting, Jare Powell used the word 0:24:39.080 --> 0:24:44.640 disinflation a few times. And disinflation, yes, a deflation, just 0:24:44.680 --> 0:24:48.040 a slower rate of inflation. Yeah, So that means that 0:24:48.119 --> 0:24:52.639 the inflation is coming down and the markets will start 0:24:52.760 --> 0:24:56.160 rallying as soon as he would pronounce that, not because 0:24:56.240 --> 0:24:59.960 he was suggesting an inflation by and large is coming down, 0:25:00.119 --> 0:25:03.760 but he did say that in certain segments of the 0:25:03.840 --> 0:25:09.200 CPI we were observing disinflation, such as in the goods markets, 0:25:09.280 --> 0:25:13.640 and that could have been a case of you know, 0:25:15.160 --> 0:25:21.679 AI based the algorithms that were utilizing words to really 0:25:22.080 --> 0:25:26.120 take advantage of developments in the markets, and the following 0:25:26.200 --> 0:25:29.920 day the market will reverse the rally once people will 0:25:30.040 --> 0:25:34.800 digest what he actually said. So perhaps some of these 0:25:34.800 --> 0:25:39.200 algorithms are making markets less efficient then because they're keen 0:25:39.320 --> 0:25:41.879 on a word, but not necessarily the full meaning of 0:25:41.880 --> 0:25:46.679 the speech is that what we're thinking. They certainly create 0:25:46.800 --> 0:25:51.360 more intra day volatility. It's hard to say whether they 0:25:50.800 --> 0:25:54.840 maybe in some cases they make the more efficient, maybe 0:25:54.840 --> 0:25:58.159 in some cases less efficient, but I think what is 0:25:58.480 --> 0:26:02.359 likely the case is that they create more intra day volatility. 0:26:02.760 --> 0:26:05.880 So let's bring this back to how does this attract 0:26:05.960 --> 0:26:09.240 you to Goldman Sachs. You know, back in the eighties 0:26:09.240 --> 0:26:13.200 and nineties, it seemed like these young hotshots would start 0:26:13.200 --> 0:26:18.280 at Goldman, they put together a trading record, Goldman would 0:26:18.359 --> 0:26:22.200 basically seen them become their prime broker and send them 0:26:22.200 --> 0:26:25.080 out to be hedge funds. Now it almost sounds as 0:26:25.119 --> 0:26:27.760 if the opposite is happening. Hey had a big firm. 0:26:27.760 --> 0:26:31.399 With Goldman, we have so many different tools that you 0:26:31.440 --> 0:26:34.639 can use that you don't get at a small hedge fund. 0:26:35.040 --> 0:26:37.720 You're better off working at the big firm. Did that 0:26:37.800 --> 0:26:40.359 play into your thought process? Tell us a little bit 0:26:40.400 --> 0:26:43.679 about that. I think the future of the industry is 0:26:43.720 --> 0:26:48.160 really in the solutions space. Solutions space, Yes, that's really 0:26:48.240 --> 0:26:53.560 what institutional investors need and what let's define that a 0:26:53.560 --> 0:26:56.919 little bit. In other words, they're not just looking for alpha. 0:26:57.200 --> 0:26:59.640 We have a problem and we're looking for a solution 0:26:59.720 --> 0:27:03.840 to that issue. Well, yes, we're looking for particular solutions, 0:27:03.880 --> 0:27:09.320 whether that's a liability, whether it's a completion of existing portfolio, 0:27:09.520 --> 0:27:13.640 whether it's a particular return target they have, whether there 0:27:13.760 --> 0:27:18.760 is a particular liquidity profile that they need to achieve. 0:27:19.880 --> 0:27:24.760 There are all kinds of needs that institutional investors have 0:27:25.800 --> 0:27:30.040 that they cannot satisfy by just investing in the hedge 0:27:30.040 --> 0:27:34.960 fund industry because the assets they manage are many times 0:27:35.040 --> 0:27:38.760 larger than what the hedge fund industry can absorb. At 0:27:38.800 --> 0:27:43.359 the same time, just being passive is not really the 0:27:43.480 --> 0:27:47.119 way to go, and so what I think is happening 0:27:47.440 --> 0:27:50.800 is the two areas are merging somewhere in the middle 0:27:51.320 --> 0:27:57.480 where you're really what the demand is for creating holistic 0:27:57.600 --> 0:28:02.639 portfolios that incorporate as classes from the whole spectrum of 0:28:02.880 --> 0:28:07.240 assets out there, whether it's in public markets or private markets. 0:28:09.080 --> 0:28:16.120 Focus on portfolio construction with good the risk management framework 0:28:16.240 --> 0:28:21.560 and try to provide the right the profile of risk 0:28:21.600 --> 0:28:26.679 adjusted returns for the particular needs of the investor, incorporating 0:28:26.880 --> 0:28:32.399 alpha in there, but not just focusing on the alpha component. 0:28:32.560 --> 0:28:36.800 And I think this is interesting in many respects. You're 0:28:36.960 --> 0:28:41.760 really fulfilling and big need of this institutional investors. You're 0:28:41.760 --> 0:28:46.720 bringing together skills from the whole spectrum of the industry, 0:28:47.160 --> 0:28:53.680 and you get to create the bespoke customized solutions. So 0:28:54.760 --> 0:28:58.480 for someone like me who started my career in academia 0:28:58.600 --> 0:29:06.080 and spend my research years thinking about porfolic construction, asset 0:29:06.120 --> 0:29:11.440 the location macro as surprising, and then I went into 0:29:11.640 --> 0:29:15.120 the headge funding industry, this is an area that really 0:29:16.960 --> 0:29:20.600 straddles the whole spectrum of things that I have done, 0:29:20.800 --> 0:29:24.200 and I think it's really where the future is. So 0:29:24.480 --> 0:29:27.960