Today I am speaking with Russell Korgaonkar, CIO of Man AHL.
In his role, Russell oversees a large research organization and so we spend a large part of our conversation talking about research management. Russell provides his thoughts on topics such as determining which projects to take on, quantifying investments in technology, data, and people, how to avoid group think, and how to incentivize both researchers and reviewers. There is tremendous organizational alpha to be gleaned here.
In the back half of the conversation we discuss some of the research that Russell has published on dynamic risk controls. He explains how risk management signals are akin to alpha signals and how the practice of managing risk through 2020 differed from the theory of doing it.
We conclude with Russell’s opinion as the most important due diligence question he could ask, either of another manager or of his own researchers.
Please enjoy my conversation with Russell Korgaonkar.
Transcript
Corey Hoffstein 00:00
All right, Russell. Let’s do it in 321 Hello, and welcome, everyone. I’m Corey Hoffstein. And this is flirting with models, the podcast that pulls back the curtain to discover the human factor behind the quantitative strategy.
Narrator 00:20
Corey Hoffstein Is the co founder and chief investment officer of new found research due to industry regulations he will not discuss any of new found researches funds on this podcast all opinions expressed by podcast participants are solely their own opinion and do not reflect the opinion of new found research. This podcast is for informational purposes only and should not be relied upon as a basis for investment decisions. Clients of newfound research may maintain positions and securities discussed in this podcast for more information is it think newfound.com.
Corey Hoffstein 00:52
This season is sponsored by simplify ETFs simplify seeks to help you modernize your portfolio with its innovative set of options based strategies. Full disclosure prior to simplify sponsoring the season, we had incorporated some of simplifies ETFs into our ETF model mandates here at New Found. If you’re interested in reading a brief case study about why and how. Visit simplified.us/flirting with models and stick around after the episode for an ongoing conversation about markets and convexity with the convexity Maven himself simplifies own Harley Bassman. Today I am speaking with Russell core towncar CIO of man AHL. In his role, Russell oversees a large research organization. And so we spend a large part of our conversation talking about research management. Russell provides his thoughts on topics such as determining which projects to take on quantifying investments in technology data and people how to avoid groupthink, and how to incentivize both researchers and reviewers. There’s tremendous organizational alpha to be gleaned here. In the back half of the conversation, we discuss some of the research that Russell has published on dynamic risk controls. He explains how risk management signals are akin to Alpha signals, and how the practice of managing risk through 2020 differed from the theory of doing it. We conclude with Russell’s opinion as to the most important due diligence question he could ask either of another manager or have his own researchers. Please enjoy my conversation with Russell Korg on car. Russell, welcome to the show. Really excited to have you here. I have been reading all the work that’s been coming out a man group for seems like you guys sort of hit the accelerator on the publishing schedule. So I jumped at the opportunity to have you on and talk about a lot of the things you guys have been writing about. So welcome to the show.
Russell Korgaonkar 02:54
Thanks very much. I don’t know if we press the accelerator. We had rather less to do with us. But it’s one or the other. We’ve been publishing a few more pieces. Thank you for reaching out. Thank you for having me and for having read some of our work. It’s good to hear.
Corey Hoffstein 03:09
Well, my pleasure. And let’s dive in because we’re going to touch on a lot of that work and just the topic of research in general. But let’s rewind the clock a little bit. You actually really kicked off your career and began your career at man group just about 20 years ago, I think you might be coming up on your 20 year anniversary here shortly. And I want to start with the question around, in your opinion, what is really been the key evolutions in the quant landscape, and how has it affected the research process over the last 20 years.
Russell Korgaonkar 03:40
So I remember starting my career, it was about February 2001. And one of the key things was trying to think back to like what was work like at that time. And we were all wearing full on suits, first of all, jackets ties the whole lot. And we were doing research I spent all day in front of a computer in the office. And then you go back home. And nowadays of course, mostly like hoodies, and trainers. And most people don’t even bother going into the office anymore. A bit changed a lot. But the research processes obviously evolved. And I think one of the things is the bar has become higher and I think the bar is likely to get higher and continuously to get higher. I must admit when I started my career, I was a bit of a cynic. When it came to investment strategies or beating the market full stop. I had in my head that the market price encapsulated everything more or less. And we’d do some work and we’d come up with a strategy and it looked like it would win it would beat the market and I was thinking to myself like this is never gonna last. So I always had that cynicism in my head and I think I still do but the fact that I still have a career and I’m still here must mean something work. So these strategies, but in terms of change of approach, for sure, as I said that the bars got higher in efficiencies and markets get exploited. So when I started, I was working on a stat ARB system in equities. And it was compared to today very, very basic looks at reversion among pairs, neutralize the various factors, and it produces lovely returns, but you could just see it decaying through time. And in order to keep up the edge, you needed to invest more data for better execution, new signals. And that was obvious, then and it’s still the case. Now. And I think the net impact of all of that is, it really affects the way you think about research and the way you think about managing investment strategies, back then you could run a strategy, and it will still work for years, and you really need to do very much. These days, that shelf lives a bit shorter. And therefore, the value of the ongoing research process is more important. So I think that constant evolution, more data, more processing power, more signals, higher inefficiencies being exploited higher bar, all of that is pointing in one direction, which means you just need to work harder to stand still effectively,
Corey Hoffstein 06:16
sort of that red queen problem, I suppose run faster and faster just to stay in place. Right. I’m excited to dive into that research process with you. Before we do that, though, just as a little bit of table setting, man is a very large organization, I was hoping you could walk us through sort of the different strategy units that the firm is divided into so that myself and the listeners can get an understanding of maybe then how the research is processes broken down between them.
Russell Korgaonkar 06:44
So high level across man group, man group runs around 135 billion of total AUM, there are five investment units. Shell, which is my home, is the systematic macro unit. We’ve been around since 1987. Purely systematic, but we run long, only long, short, lots of different types of strategies. But in the Macrospace, mainly, then you’ve got new Merrick who are more focused on corporate assets. So equities and credit, not too dissimilar from HL, they started in 1989. So similar length of experience. And again, systematic, then you’ve got a discretionary division JLG, who have a number of PMS, they’re not constrained by a house view. But they run strategies across the space, we have a much smaller, newer real estate, private markets division, called GPM. And then the final of the content engines is frm, which is a effectively a fund of funds type division, that just one extra thing running across all of that is something called Man solutions. And we’ve found, particularly in recent years, the bigger institutions like to have access to various funds, and they like customization and that division works a lot with those types of clients.
Corey Hoffstein 08:07
So as the CIO of man HL you oversee was probably a massive research program. So I want to dive into that process, and how you sort of lay it out with all the researchers on your team. And I want to begin with thinking about setting the agenda, and get your thoughts on both how you think about setting the research agenda in the short term for your team, as well as sort of those longer term goals.
Russell Korgaonkar 08:35
I think you’ve divided it up. Well, actually, between long term and short term, I would distinguish between large scale projects, and general predictive research. And in the case of the latter, that, in my opinion, should be as freeform as possible to give as much freedom to researchers as you can. And there’s some accountability with that. And people know that that’s their job, they’re there to find new sources of alpha and to work on signals. In my opinion, the more freedom you can give in that setting, the better. The former large scale projects are completely different. So this is investing in capabilities. It requires planning, it requires technology, it requires commitment and investment from the firm. And our approach in that case, is generally to go kind of bottom up for ideas. So teams will brainstorm, think about what capabilities they could deliver, what new initiatives they might want to work on. And then we’ll percolate that up to team heads and we’ll do another brainstorming with team heads. And then try to put a value on these various projects. So the idea you get ideas from as many places as you can, and then try to put a number, like how much value can we put on that project?
Corey Hoffstein 09:52
So correct me if I’m wrong, but within a gel, there’s actually sort of four sub strategy units, I believe her core specialist equities and fast trading. And if that’s correct, I’ll have you explain those because you’ll do it infinitely better than me. But I do know there are at least a couple sub strategy units within man hl. And I was wondering if you could sort of expand on perhaps how the research process differs across those different strategy units.
Russell Korgaonkar 10:23
So the Strategy Unit is briefly core looks after all the various scalable futures and forwards strategies. So that includes things like trend, risk premia, more kind of esoteric macro strategies. But the key thing is, is all scalable, all apply to futures and forwards. That’s what we call core, the specialist is, for us more alternative, either asset classes or strategy types. So things like options trading falls in there, the machine learning efforts that we run different strategies on machine learning fall in that group. Things like credit trading interest rate swaps, more esoteric markets will fall in there as well. One commonality about specialist is, number one, you need specialist market knowledge in that area. Number two, you tend to need specialist technology and platform. So that kind of defines specialist, then you’ve got equities. And fast trading equities is probably the most self explanatory because it’s an asset class in its own right. And then fast trading for us is not super fast, high frequency, nanosecond stuff, it’s basically sub two weeks and holding period. And that means you’re looking for signals that may have a duration of hours and upwards. Fast trading runs a decent amount of capital, it’s not a huge amount per shell standards. But the idea, or the objective of that unit is not just to create alpha in its own right, but also to improve the way that we execute across the platform. So to improve the flow coming out from all of the other teams, we find when you’re running at scale, as HL clearly does, we run a large amount of capital, that execution bid becomes all the more important. So that’s the focus of fast from the perspective of research process across the teams, we try to make things as consistent as possible, one person comes up with a research idea, we always put two researchers to kind of peer review test the idea doesn’t matter where it comes from, what we’ll try to do as much as possible to get one person from in that unit, and another from a different unit just to give a sense of difference of opinion, different ideas, expertise, etc. And they work together, test the idea and make sure it’s sensible. And then it goes to an out of sample, which means it’s a dataset that they haven’t yet seen. And we have certain criteria that get agreed. And if it passes out of sample, it’s available to go into live trading. There’s basically a well trodden path for the research process. And it’s consistent whether it’s done in the fast team or equities team or Corp or wherever it is the same process,
Corey Hoffstein 13:17
as we’re talking about this idea of having to run faster just to stay in place. Seems like a lot of the new avenues of research are becoming more expensive, requires greater investment, either in time, data, technology, human resources. And it brings up a really difficult question that I think a lot of us have to answer in the research field, which is how do you determine when that investment is worth it when you don’t know what the results are going to be beforehand? Curious as to your thoughts on that.
Russell Korgaonkar 13:51
It’s a difficult one in research to a degree you accept r&d expense, and you accept that research is not always going to pay off. And that’s part of the game that we all know, not necessarily love. But we’ll know. I think the thing for me is, especially when you’re running large amounts of capital for your clients, you need to focus on projects that can have a payoff, you need them to have expected returns that are going to be multiples of the expected cost of running the project. And those expected returns have a decay rate and you need to kind of factor that in as well. It’s a difficult number to come up with and there’s never going to be a precise science. There’s a bit of guesswork that’s involved, but hopefully educated guesswork. And that is where experience of these things comes in handy and really getting a feel for what would the world look like with this project having been completed and what does it look like currently? And then can you scale it that for us is really important. Let’s say we have for example, an idea that’s focused on a really niche asset class or a niche market have very high turnover. And it’s not really scalable. It might be great. There might be a high conviction of alpha there, but it’s just doesn’t scale. We just need to work out alpha expectation relative to cost expectation, how does that work out. And then there are all the other questions, metrics that you might use. And one of those is what’s the chance of it working. And what’s the potential upside, but the chance of it working is, we will get views from various people, we need to be realistic about things. Often, it’s the case that, as you can imagine, those projects that require more resource more time, more commitment, more spend, they may go wrong, there’s lots, that doesn’t work. But if you do get there, the payoff is likely higher and likely more sustainable. Because of the competitive edge, we try to factor in all of those things, if I was to kind of look across the projects that we commit to, at any given time, we try to have a bit of a balance between those that are a little bit more blue sky. And those that we think look, we should be entering this space with fairly convinced as alpha there, we can see some of our peers doing it, there’s no reason why we don’t have that capability. Those are kind of more high conviction if you like, but just you got to commit to it and you got to do,
Corey Hoffstein 16:17
some of the costs associated with the research projects are always going to be up front. And then others are going to be ongoing, you might have these long term multi year goal projects that have ongoing costs. I’m curious as to your thoughts as to managing those ongoing costs and knowing when to potentially pull the plug, avoid that sunk cost fallacy and just stick with the project, because you’ve already put all the money into it. insights as to knowing when to say enough is enough, this probably isn’t going to work, you need
Russell Korgaonkar 16:49
that same mindset that you have with thinking about portfolio allocations. In theory, as long as you’re friction free, there should be no greater tendency to remove something from the portfolio as allocate to something new. In theory, you should be looking at these things fresh every time making a sensible decision every time. And honestly, that’s the case with predictors. And it’s the case with projects as well. Sometimes just because it’s there in the case of a predictor, or just because you started the project and spent two years on it, neither of those things is good enough to say, well, we’re going to spend another two years or we’re going to allocate to that predictor for another two years, that by itself is not good enough. Always you have to balance against the opportunity cost. In the case of projects, and the opportunity cost or the portfolio cost in the case of predictors and being relatively data driven people at HL, I think we’re okay with that, you know, in the instances where we’ve said, This projects, we’ve spent some time on it, we wanted to be here. And we’ve only got much less far than we’re intended to, this isn’t going in the right direction, it’s going to be another two years, even after two years, we’re not, we’re pretty good at saying look, let’s take stock, let’s look at the evidence and make a decision. And by and large people get on with it. Once you’ve kind of had a look at all the angles and you’ve made the arguments, people get on with it. And I suppose one of the nice things of working in an ongoing collaborative research facility or location or ensemble if you like, there’s always new ideas, there’s always new things to work on, you’ve never done. So stopping something means starting something else. And we’ve been okay at that. I can’t say there’s never been occasions where we should have stopped something. And we didn’t. But I can think of plenty of occasions where we should have. And we did. And as a result, we cracked on with something a bit more useful.
Corey Hoffstein 18:42
technology costs are one of those things that have come way down in the last 20 years, arguably one of the things that’s invited sort of the systematic arms race that makes this so much harder than it was, but a lot of other resources have become more expensive. There’s unique datasets that you need to get access to personnel costs are higher. And a lot of that, to me points to the fact that you need to find ways to leverage resources across different research teams and strategy units. And so I’d love to get your thoughts as to how you guys go about effectively making sure teams have access to all the resources they need.
Russell Korgaonkar 19:21
We’re in a fortunate position in man group that we’ve got a lot of teams and whether you’re looking at investment from a purely discretionary standpoint, or even a systematic standpoint, there’s lots of different angles to that. Everyone really is doing the same thing, which is you take in some data. And with that data, you make some investment decisions, and you send out some trades, and the bits in the middle can be fairly unique. I’m going to convert data to a signal and I’m going to put a bunch of signals together and create a trading strategy And I’m going to put risk management on top. And then I’m going to send some trades, those bits at the end the data coming in, and the trades going out. In theory, those bits are shared by whoever’s part of this ecosystem. And therefore those are examples of things that we will try to do well and do once. And data’s, I think, the best example, as you said, it’s becoming more expensive. You’ve got to question in many cases, the value of that in a world in which people are competing for the same piece of information, and that has limited value. But clearly, it’s the case that the more as a firm that we can coordinate that process, we can do it once. And then we can leverage that data set multiple times, the better value that we get collectively. So that kind of mindset, there are various things. Within the research process, you almost have to be individual about like you really cherish the ability of individuals to think for themselves have their own approach to be unique. And then there are parts of the process that you think like, we just need to get this done really well and systematically and consistently. And as much as we possibly can. We’ve been pushing in that direction. That’s not to say, we’ve found some Nirvana yet. You’re always seeking to improve these things. But I think we’re fortunate in the case that it man group, there’s a lot of potential for doing things collectively. And therefore there’s a lot of potential for teams to focus on the bit that they really add value and have the rest of this stuff just kind of done for them almost.
Corey Hoffstein 21:33
What areas of research do you wish you had invested in earlier? Or perhaps invested in more heavily? And sort of as a second follow up to that? How do you balance that idea of resource management with the regret minimization of not having invested early or not having invested more heavily?
Russell Korgaonkar 21:53
It’s funny how many projects that at the time of starting the project, there’s a deal of skepticism on what is this thing going to deliver, and even when it’s finished, there’s some skepticism about what value is going to add. And then, kind of years later, you think, thank God, we spent so much time and effort on that thing, because it’s really enabled us to do things better, not just in and of itself, but just in general. An example of that, for me is in execution. We spent a lot on a relative basis on having our own algos in the future space and the forward space. When you look at it at the time, and you think, well, maybe there’s a small edge there maybe and does it scale at all. And then as time goes on, I can remember at the time people going, Oh, you’re spending all this money on this one small part of the system? And why are we not doing more on signal research. But the signal research decays, the platform stuff, where you really invest in capabilities, that decay rate that has a decay rate as well, but it’s much, much slower. And therefore the kind of value of those projects can live on for years and even sometimes, decades. So that bit of resource management and regrets. I mean, the regret thing, I don’t think I personally feel that too badly. As long as I think we’ve taken all of the information that we had available at the time, and we made the right decision at the time and then you continue and it’s like investment decisions really as a regressing if you feel that you made the right decision with all the information at the time, like clearly, with hindsight, you could have done a better job. But just as with investments, that’s unrealistic in terms of areas, I think, now, had we been a bit quicker to a lot of our peers and even within my own group, our colleagues and numeric, have got into the credit trading space, I kind of single in corporate bonds and CDs a little bit quicker than we have. And I think that’s if I were to kind of pick an area that I think I wish I had done that two years ago. That would be one of them. Having said that, as I said, we tried to minimize living with regret as much as possible. One of the big
Corey Hoffstein 24:04
risks and research both within an organization I find across the industry at large is this idea of groupthink. Often we see certain trends become really popular. And you’ll see a rash of papers suddenly get published on the idea and it seems like everyone has a thought or opinion on it. And it might just become the next overcrowded trade. How do you avoid that groupthink in the research agenda that you guys set out to pursue?
Russell Korgaonkar 24:34
That is, without doubt, a challenge of our industry? I would imagine it’s a challenge of many research driven industries that many people are working on with similar datasets and they’re working on similar problems and I reached similar conclusions. Pharmaceutical is a classic example. But I strongly believe in my experience, in the value of trying to create diversity of thought, because there’s a real temptation, and a really easy path that people go down, that you end up hiring, putting your own skills to the fore and thinking that they’re the most important thing. And so for example, you quickly have these teams of four male, all mathematician, are from similar kind of educational backgrounds. And then you think, guys, come on, innovate, come up with ideas. And whatever spark there is, for idea generation, or difference of opinion, or really kind of challenging things is way lower in those teams that are less diverse across these different factors. And so you can think about, you can try to encourage that I will do, I’ll say to teams, look, if you have no gender diversification your team, that’s probably a problem. I mean, for me, personally, as a researcher, I don’t know your views and your experience. But for me, personally, the best teams I’ve worked in have been mixed from a gender perspective, they’ve also been mixed from a educational background perspective. So not just mathematicians, and not just economists, not just physicists, but a bit more breadth. And ideas can, frankly, come from anywhere. One of the things that I’ve really noticed in the last five or six years with a man group that HR has benefited from is the introduction of GLG. This is a discretionary pm based outfit, totally different culture, totally different approach joined with the quant division, different in every regard. And to begin with those two groups of people or investors, they just didn’t mix at all. And the JLG guys are based in one part of London, Mayfair and the NHL guys are in the City of London. And then at some point, the offices were brought together, and we all sat on the same floor. And we started inviting people that we got to know on that side to our research meetings, and vice versa. And all of a sudden, it was amazing how you could have a completely fresh insight on a signal or predictor or way their trading system worked by a person who lives and breathes work in those markets and estimate decisions about how to invest in those markets. So if you ask that person that kind of quantitative question, or how do you kind of think about testing a strategy or back testing or being rigorous, it’s clear that they don’t have that skill set or that training. But if you say, here’s a trading strategy, tell me what you think about it, tell me how you think it can be improved, what can go wrong, and they can be really, really insightful. So I think that it’s super important, I think you can actively encourage diversity of thoughts, by the way that you group people together the way that you select people to review, ideas, and research. And it’s definitely a good thing. There’s so much evidence from our perspective, but I’m sure just across industries, that diversity of thought is a good thing for innovation.
Corey Hoffstein 27:56
I suspect most of us in the research space are here, because we love the activity of research. But as someone who’s setting the agenda for a research organization, promoting the type of research you need to get done is really important to me, my brain immediately goes to how people are incentivized or how they’re compensated to get them to do the right type of research, you can try to promote blue sky research. But if you compensate them too much on outside ideas, they might keep swinging for the fences. And you might have no incremental advancements versus Do you compensate just for incremental advances, and no one’s really doing outside the box thinking? So I’d love for you to comment on how you think about compensation. incentivization how you create skin in the game, both for the researchers and peer review teams and anyone else involved in the process.
Russell Korgaonkar 28:48
It’s a big challenge. I think it’s super important to have a research team that they know what the endgame is, you can sometimes be so lost in the minutiae of what you’re doing, and that you lose sight of what the endgame is, actually, all of this, to be honest, is a work in progress for us. Even though we’ve been managing research teams for so long. It’s not as if there’s one answer to it, that necessarily works. All of this evolves through time. So even for us, and we’ve been a research team for over three decades, it’s still a work in progress. I think the most important thing, in all honesty, is that your people care in the first place, that they want to help their colleagues, they have a curiosity. They want to make trading strategies that are as profitable as possible. They just care from a pride perspective, that that works. And that actually is probably the most important thing first and foremost. Having said that, I’m also really keen that you do have an incentive structure that kind of emphasizes it and encourages you in the right direction. Because otherwise, as you said, you can get a particular group that are doing nothing that is going to really move the needle today, but it might enable other teams down the line to bend From the work that the first team did, and how do you share that out. So as much as possible, we’re really trying to be very data driven in evaluating research, evaluating how much value is added to our systems, not in back test, but in live trading. And then also, to try to get a sense of how good a job or the review is doing, as well as the researcher for any individual signal or project. And one thing we started to do is very early days, I don’t have enough data to give you a good answer. But I think it’s an interesting challenge, nonetheless, is to say to the researcher in the reviewers, what do you think the Sharpe ratio of this signal is going to be in live trading? And everyone can do that? And we do it for kind of 10s, hundreds of these signals, you start to build up a data set and say, Can we spot any patterns? Are there people who are particularly good at this, but the people who are particularly bad people are kind of over optimistic, etc, etc, and just start to build up some kind of stats on how good people are at doing that. That’s our kind of direction of travel. But to come back to the skin in the game bit. One thing that we’ve really encouraged and that people have responded to really positively is, we’ve recognized the value of the review bit, the Review bit for us, it’s not just a box ticking exercise where you go, Okay, I understand what you’re trying to do looks fine. No concerns here. Let me get back to my job. Thank you very much. It’s more How can you really stress test this idea is anything you can do to improve it, I’ll give you an example of one of these exercises that was conducted recently, I had no idea about any of this, by the way, there was a proposal for relative value credit system. And we chose two reviewers, one from inside the unit and another from the fast trading unit, who are also working on credit market making. And there was just a lot of tooing and froing, on that research, idea, challenging thing, but also making suggestions. And all of this is done, by the way, only on in sample data, we’re kind of keeping a dataset that we haven’t tested any of these ideas on. And then the researcher went back took all of these things on board tested his original idea, and they kind of modified one on the out of sample data. And these suggestions had an improvement on the Sharpe pressure point nine, which for a kind of system that is more or less similar as a pretty significant improvement. And obviously, that still may be out of sample is still back test. But it just goes to show that getting good research reviewers involved in the process and getting them to really challenge things and come up with ideas on how to improve things. And then monitoring all of this can lead to these really positive outcomes.
Corey Hoffstein 32:53
And a conversation you and I had preparing for this interview, one of the things you said was, quote, sometimes the most successful strategies and models don’t back test very well. In the same vein of that idea. There’s all these emergent phenomena that simply can’t be back tested, like the recent rise of Reddits, Wall Street bets or COVID. ‘s impact on the market. How do you go about creating confidence in your researchers and your reviewers for signals and strategies where back tests may be lacking or entirely impossible,
Russell Korgaonkar 33:26
that’s a really good point and a real challenge for many quantity, they kind of back tested like a comfort blanket, they can do that model and they check on the back test, it all goes up. And it just feels so kind of reassuring as it’s going up in the past, it must, therefore continue to go. And obviously, as we all know that reality is completely different to that. And We have certainly found over the years in the last few years in particular, there are certain regime changes going on in the marketplace that make you believe that certain strategies are unlikely to work going forward. Because conditions are likely to be different going forward. Sometimes, you have to design strategies, with a view to a different world in the future to the world in the past and the world in the future, for example, may not feature interest rates that fell from double digits to zero. Okay, it’s definitely not gonna feature them whether they continue near zero, they go up, but it’s going to be different from that perspective. And when thinking about models, and the kind of likely reaction of markets you need to make a few judgment calls and use a back your intuition. I think it’s something that requires a little bit of a change of mindset, from the back test, fixated mindset, their content to stick to, I think it helps to, again, not work in isolation for those types of things, but working groups where you take a few different opinions. If you can find someone who really has an understanding of the particular phenomenon that you’re trying to exploit because it’s an area of specialism for them, then all of that leads to greater conviction. That’s definitely a different approach. But to my mind, it’s a really potentially very valuable one.
Corey Hoffstein 35:13
Let’s talk about some of the research that’s been published lately. That’s got your name on it. You’ve co authored a few pieces, sort of circling this idea of dynamic risk targeting, which is marrying together these different ideas of diversification, leverage target volatility and dynamic risk overlays that you’ve put together, can you walk me through the core philosophies at play in this research?
Russell Korgaonkar 35:39
One thing I should start with is that risk management has been something that has been a real focus for us inside HL for at least ever since I’ve started. And I know for a fact for three decades or so the AHL has been in existence. And you would imagine that it really has to be for an Systematic Investment firm. To survive that long, it’s got to have a good focus on risk management. I think there’s two key elements to risk management. The first is constructing portfolios that themselves have sensible and well observed risk limits, knowing how to monitor those things, and doing a good job of saying, today’s portfolio positions I hold right now today are within sensible risk limits. So that’s number one portfolio construction. And then the second bit is about dynamic Exposure Control, which is to say, I need to be able to react to changes in market conditions, and recognize the fact that risk levels in the markets change over time, they’re certainly not the same, they change and I need systems to be able to change with them. And that bit for us, I would say is the real distinguishing feature I was all of this should be relatively common. But for us that real distinguishing feature is that ability to change exposure at times rapidly based on changing risk conditions,
Corey Hoffstein 37:06
let’s maybe dive into those risk overlays a little because they do seem to be a really key component in your research and writing as helping you make sure you don’t get caught off sides during market conditions where volatility spikes or correlations crash towards one. And so there’s been a couple of approaches that I’ve seen that you’ve written about, I think there’s been a momentum or trend overlay some volatility switching signals, a correlation signal between bonds and equities. And then there was an equity specific signal that was measuring sort of internal diversification of the equity portfolio, sort of a diversification trigger, I was hoping you could walk us through sort of how you think they’re useful, individually, sort of why each signal has a purpose, and then how you think about them in combination.
Russell Korgaonkar 37:55
The first thing with risk controls, I think, is that you should think of them a little bit like alpha signals in that they also diversify. And it’s better to diversify. So rather than having one particular risk signal, having a few different ones that ideally themselves, capture different types of risk, the more you can create that diversified blend, the better. And one thing we all know from risk events, and we’re probably 1415 months on from the last big risk event, the next one always looks different. There’s no point fitting into the last risk event, because the next one always looks different. So we try to come up with risk signals that capture types of risk that we’ve kind of familiar with that we’ve seen before, we’re likely to see again, as opposed to very specific ones. So the controls that you mentioned, momentum is one of them. And that focuses on left tail events from a single market perspective. And it’s fairly easy in the case of most markets to look at the pattern of their return. So you’d look at monthly returns of most asset classes through time. And you’ll find that they’ve got a left tail negatively skewed, and they might not be negatively skewed by their own. In the case of daily returns, they certainly are by the time it gets to monthly or longer. And the reason for that is one of the similar reasons that momentum works in the first place. During market sell offs, different mechanics come into play, people become forced sellers, there becomes more pressure on the market and you get continuations, you get negative trends. That’s the first one. And it’s fairly intuitive and obvious that we just use some trend signals that we’ve used in our hedge fund programs, but just focusing on the left tail, volatility, spiking, switching whatever you want to call it is another one, that it’s a feature of markets that happens very consistently through time that market volatility levels change the distribution of returns. As heteroskedastic, if we want to use the status expression for it, and that means that the volume levels of markets change through time, but the important bit is they change with a degree of predictability. So the persistency of all levels is really a really important feature of markets. And it’s a persistent feature of markets. And there’s never guarantees, but it’s more likely, if you’re in a period of low vol that tomorrow is going to be low vol. and vice versa for high bar. And that, by the way, applies at the market level, it applies across asset classes, it applies to the market as a whole. And then finally, this idea of diversification breakdown, and the bond equity correlation is actually similar to the equity diversification idea that during periods of stress, you get a change in the correlation conditions between markets and asset classes. And this was an area that we actually really leaned on our colleagues at the man Oxford institutes, we were kind of fortuitous, with timing, but they’d been working on these econometric models to help with these kind of regime change identifiers for correlation and diversification. And that just helps with those periods. Like you say that if you’re running some kind of balanced portfolio, you rely on diversification, but at times that diversification breaks down. And so that risk signal is there to try to help identify those occasions. And as I said, None of these things are perfect in their own right. They pick up different types of risk, different characteristics. But what we find is similar to Alpha signals, that if you put a collection of these things together, the net result is a better risk managed system than any single one by itself.
Corey Hoffstein 41:42
I’m going to ask you, it’s probably an impossible question. It’s a more philosophical question, which is that markets are ultimately reflexive and that there’s sort of a I see a game theory problem that if everyone starts to adopt these types of dynamic risk controls, which may make complete sense in isolation for the individual, it might actually have destabilizing effects on the market as a whole. If everyone’s systematically selling at the first sign of a drop in diversification. Well, that becomes a self fulfilling prophecy, because everyone’s just selling everything. And therefore, diversification decreases further. Curious as to your thoughts there. I mean, there’s been this argument lingering for the last 10 years that vol targeting funds, whether it’s CTAs, or risk parity, or insurance products are all sort of contributing to the increased volume of vol, we’re seeing in the markets. Do you think that this is an unfortunate arms race, that could be destabilizing markets? Or do you think the evidence isn’t really there?
Russell Korgaonkar 42:47
So I would go back to one of the things that I said, first of all, that trading strategies are really there to exploit inefficiencies. And the funny thing is that markets, to my mind, are as inefficient now as they were 20 years ago, it’s just the inefficiencies might move around a bit. And it is certainly true that the more capital chases an inefficiency, the less you’re likely to serve it. And it’s also true, therefore, that the too much of trading in the same direction is likely to exacerbate market moves. We have to accept, therefore, that as systematic investors, we’re not alone, we work in an ecosystem with others who trade models is not hugely dissimilar to our own. So the important thing, from my perspective, running capital with those kinds of signals embedded is that we monitor it carefully. And that means that we look at market price behavior, before we execute in the period after we execute, we look at order book imbalance, we look at all of the clues that might lead us to the conclusion that there’s too much pressure being placed by certain types of strategies. And therefore we’re not blind about that, it would be wrong to say that cannot possibly ever be the case, because there’s so much capital in the world doing other stuff. Having said all of that, with the benefit of all of these signals, as you said it is impossible to give a definitive answer, but you cannot have previous includes this point in time, it’s safe to say that the Aum and the trading volumes that would cause pressure on the market aren’t doing so in a way that’s overpowering markets. So none of those events do I think that huge moves have been caused because of CTAs or risk parity. I think huge moves have been caused in which CTA is in risk parity maybe involved and part of the flow, but do I think those moves have been strongly exacerbated or cause we’ve seen no evidence of that.
Corey Hoffstein 44:46
So a lot of this research that you’ve put out I think a couple of the articles came out pre March 2020, some post March 2020. And any researcher will tell you that there is a huge degree of difference between research on Paper and research and practice, what lessons did you take away from managing assets through the 2020 environment.
Russell Korgaonkar 45:09
I’ll give you a little kind of personal story about COVID. And markets in February and March of 2020. So, as you know, the real game changer for markets was around the 23rd of February, which was a Sunday, when some news came out that hundreds of cases have been reported in North of Italy. And I thought to myself, God, that is going to move the markets really significantly tomorrow, I really thought it could be a double digit kind of move. And the market sold off about 3%. On the Monday, the 24th of Feb, but they kept selling that week, the new story at that time was such an obvious train wreck in motion, it was obvious what was going to happen. I mean, maybe you could have said it was obvious before that, that once you’d had a few cases outside China. But clearly, when you had this big cluster, that was going to be a problem. And the direction of news flow was awful. And I went skiing on Friday of that week. And I think the market sold off seven or 8%. And our models being systematic, we’re doing what our models should do in that situation. And they were reacting to what was obviously a kind of rising risk and reacting to that accordingly. And the reason I remember going skiing is because it was a chance for me to actually kind of step away from the day to day of running a fund. And also just get a sense of the mood amongst people. And I just remember taking the flight back on the Sunday, they didn’t use it all of that week are just getting worse and worse about this virus. And no one quite knew how severe it was. But it was obviously pretty bad. And it was kind of spreading, and it was all over the place. And people kind of wearing masks for the first time, a very odd doesn’t feel very odd now or used to but back then it was really quite odd. But you could just get the sense of fear in the marketplace, I felt relatively calm, I suppose that we kind of built our models, not for COVID, obviously not for dramatic market conditions is that but they were kind of doing the right thing. They were recognizing this as a risk event, they were taking down exposure. And I got in on Monday and started talking to different colleagues and getting their opinions. And by the way, there were plenty that said, look, it’s over done, it was 10% down or whatever it was that week, there’s over done, because the long term impact on valuation of short term closures to economy is actually negligible. All these very good answers. However, the one thing that you cannot allow for more you have to is the power of emotion and the power of fear. And it was a fear driven market throughout the first half of March. And the power of that to move markets is exceptional. And when markets start to move in that way they can really keep going. And therefore you have to be careful that risk management is important for those type of events. And particularly when you’re running strategies with leverage, and you’re running strategies across asset classes, you just have to be prepared for those exceptional events. So I think the lessons learned really were just another reinforcement. I mean, those of us who’ve lived through 2001, and 2007, and 2008, you’ve seen different kind of 2010 2012, I suppose you’ve seen these different stress events in markets. And this was a totally different one. But it was another kind of reminder, I suppose that you have to be prepared for that left tail. And that, to me was the big lesson learned.
Corey Hoffstein 48:29
One of the common critiques against systematic strategies is that they’re very opaque in nature. And I know this causes problems with a lot of folks on the due diligence side. And when you talk about opaque systematic strategies, and then you start to add in many potentially complicated layers. So in this example, we’re talking about creating diversification, the application of leverage a number of dynamic risk overlays, and we haven’t even started talking about Alpha signals yet, it can lead to what is potentially a very difficult due diligence process. If you were to sort of turn around and go on the other side of the table, and ask the questions. What do you think the most important due diligence questions someone could ask you is?
Russell Korgaonkar 49:15
It’s probably a two part question. But I think the most important question is What could go wrong? Really? What can go wrong with your strategy? Then? The second bit is Why do you think it works in the first place? What can go right and what can go wrong? By the way, that’s not just due diligence of external investors for funds? I think it’s also a question that internally I would ask of people who are proposing a new strategy. Okay, why do you think that thing is going to work in the future? Not just because it’s worked in the past, but why is it gonna work in the future? And what could go wrong in the future? And those are the questions that every investment manager should have thought carefully about and they should be prepared to answer. Those would be the ones that I would put in my DD pack.
Corey Hoffstein 49:57
I’ve been asking everyone this season the same Final question of each episode. And it’s been interesting because I’ve taken so long to record this season that the answers have vary depending on the external environment because the question has been, it seems like we’re finally getting out of the woods from the COVID situation. More and more people I talked to are vaccinated, more and more places seem to be opening up, we’re starting to get rid of the mask restrictions. Though, the Delta variant not to timestamp this episode, too much seems to be rearing its ugly head. And some of those mask restrictions may be coming back in place. But as the world is beginning to normalize a bit, what are you most looking forward to?
Russell Korgaonkar 50:37
First off, we are planning to go to the Greek islands on Sunday, and we’ve had such a miserable summer in London, mainly because the weather’s been horrible. Getting out with the family. My wife’s half Greek, and she’s got family over there. So I think the main thing for me is just getting to see people again, we’ve been pretty fortunate here that things have kind of felt a lot more normal. And they have, I was lucky enough to go to some of the football matches or soccer as you guys may refer to euros that we had at Wembley and just tremendous to be with people again, fingers crossed, we go in the right direction. And that happens more and more going forward.
Corey Hoffstein 51:15
I’ll keep my fingers crossed for the trip. Russell, this has been fantastic. I can’t thank you enough for joining me.
Russell Korgaonkar 51:21
Thank you very much. Great, really appreciate it.
Corey Hoffstein 51:28
If you’re enjoying the season, please consider heading over to your favorite podcast platform and leaving us a rating or review and sharing us with friends or on social media. It helps new people find us and helps us grow. Finally, if you’d like to learn more about newfound research, our investment mandates mutual funds or associated ETFs. Please visit think newfound.com. And now welcome back to my ongoing conversation with Harley Bassman one of the series I’ve really enjoyed going back to read are your annual stocking stuffers, which is a list of end of year trade ideas. I’m curious looking back, which of those trades were maybe some of your favorites and why?
Harley Bassman 52:14
There’s a number of them. And they all involve long dated options, options that were five to 10 years next three options that you really can’t easily do as a civilian as a non professional. And this is one of the reasons why I joined simplify is that they with me found a way to take these interesting professional is the is the by the way is international swap Dealers Association is the contract that all Wall Street people use. It’s your standard legal contract. So everyone’s dealing on a fair playing field, a way to take is the products and make them available to civilians. This is pure genius. This is breakthrough stuff. And what’s amazing about it is that like all great ideas, like the post it note, like who dreamed with that how obvious is that this will be here is also rather obvious why we do this before the trade I published, which really, I was able to put it into a structured note for retail. But the structure I did was in the December 2012 stocking stuffers, it was published for going ahead into 2013. And the trade was to go and buy the 1800 strike call on the s&p Sell the 900 strike put on the s&p for 10 years at zero cost. And I did this trade personally, actually, it hasn’t expired yet. And that was a trade taking advantage of a number of structural things that you can go to my website, convexity maven.com And you can find it there. That was a terrific idea. There’s buying very long dated options on DOLLAR YEN. And this trade still exists today. And hopefully I will get in trouble. But I suspect that we might offer this kind of risk someday in a simplified ETF but because of the nature of the relationship between the interest rate differential and the volatility term surface, it turns out that you could buy a 10 year option wait seven years and the price goes up. So you can be long convexity long the option limited loss, unlimited gain with positive carry. It does sound crazy, but it does exist. And I’ve written about that one of my stocking stuffers. I also wrote about options on the SX five E. This is the European Dow and that also similar idea it wasn’t quite long convexity positive carry but Up to the massive spread between the dividend yield of European stocks and the interest rate which is negative. Well, if you start compounding a negative rate, you get some pretty crazy stuff. And if you use longer dated instruments to borrow money in a negative rate and then compound that, well it looks pretty good. My best ideas and my most popular ones.