Douglas Greenig - The Fundamentals of Trend Following

00:00:00 [Douglas Greenig]
See, what's really happening with trend following is narratives take hold in markets. Let's say the U.S. economy is going into recession or China will boom when it reopens. So a view takes hold, a narrative takes hold, and prices tend to move and get pushed along in the same direction. But when narratives change, there is usually a burst of volatility. And what this means is that by rapidly scaling your positions, that process allows trend followers to take profits when there's a burst of volatility.
00:00:49 [Vincent Weber]
Hello everyone, I'm Vincent Weber and this is Resonanz Spotlight, the podcast where we explore investment strategies, learn the story behind them and meet the experts who create them. No strategy has been declared obsolete as often as trend following, only to make a strong comeback when least expected. In today's episode, we'll dive into the reason behind the effectiveness of trend following, aiming to understand why it works and how it provides such a distinct return profile. Our exploration will span from the fundamentals of volatility to portfolio diversification and extend to niche commodity markets such as California carbon emissions. Today, I'm excited to welcome Douglas Greenig, CEO and CIO of Florin Court Capital, a leading systematic asset manager specializing in trend following.
Before establishing Flooring Court, Doug served as the chief risk officer at Man AHL and led its portfolio management group. Prior to that, he had several senior trading roles at Fortress Investment Group, RBS Greenwich Capital, and Goldman Sachs. Doug, let's start with an overview. Could you tell us about Flooring Court, its core activities, and its current position in the market?
00:02:10 [Douglas Greenig]
Florin Court Capital is a special kind of CTA. We've been running the floor and court capital program since 2017. And what we do is we apply systematic trend following models to a very broad range of alternative and exotic markets, not normally traded by typical CTAs or trend followers. For example, we trade French electricity. We trade steel rebar in China. We trade Malaysian palm oil, California carbon emissions, South African maize, about 500 macro markets in total across all the major sectors. And what the markets have in Ghana, you see, is they're all operationally difficult. They're markets where there tend to be barriers to entry. And that means there are greater opportunities. They're less well arbitraged.
And we trade trend in a reasonably traditional and pure way on this vast set of alternative markets around the world, around the clock.
00:03:28 [Vincent Weber]
Okay, thank you. I just took a look at your CV. You were not born a trend follower. You started as a mortgage trader. And how did you become a trend follower?
00:03:37 [Douglas Greenig]
Well, life is funny. You take these paths and you take these turns at different points and you end up in places where you never expected to arrive. I started out as a student of economics and I studied economics at Princeton. And actually, I enjoyed the subject a lot and I won a big prize for my senior thesis. And it later played a role when Fisher Black hired me to be his assistant at Goldman Sachs. I then took a little twist and went into pure mathematics. I got my PhD in mathematics at the University of California, Berkeley. And that was a long time ago. I mean, I can't believe the years I'm quoting here. I got my PhD in 1993, okay? And, that's a long way back, like raptors and T-Rexes ruled the earth.
It was a different climate. It was a great, great, great math department then. There were people like Kurt McMullen around, Steve Snell. And it was a wonderful privilege to study mathematics. My field was differential topology. And when I finished with my degree in that year, 1993, I joined Goldman Sachs. As I said, Fisher Black hired me to be his assistant, sort of. He had been excited by some of the work I had done years before in real business cycle theory and applying the generalized method of moments to rebuild the foundations of macroeconomics. He wrote about that later in his final book called Exploring General Equilibrium. A lot of it was about this work that I had done. Anyhow, so there I was at Goldman Sachs, and one thing led to another.
And before long, I was working on and eventually running the bond arbitrage desk at Goldman Sachs in New York. And in those days, we did a lot of derivatives, macro trading, yield curve trading, things of this nature. And so now we're talking about the mid to late 90s. It was a wonderful experience. Goldman, of course, is a legendary place. So many talented people.
00:05:58 [Vincent Weber]
Would you call that being a discretionary trader? Oh, yeah. Of course.
00:06:03 [Douglas Greenig]
We used models, and models informed our trading, but it was not systematic in the sense that you needed to stop and ask yourself why there was this anomaly on a yield curve. A lot of times, if you see a mispricing in the bond market, It actually has to do with the repo market and the financing market. And there are reasons for the double old 10-year or something trading at some price that's sort of off of the generic curve. Anyhow, it was a great experience, and I worked with some amazing people. But it was very quantitative.
00:06:44 [Vincent Weber]
Okay, right. And in terms of, if we look to this, your current philosophy, what were your most defining experiences? over the course of your career that influenced you most?
00:06:59 [Douglas Greenig]
Well, one habit that I have is I really want to understand why things are the way they are. For example, when we talk about trend following, why does it seem to work? What's really behind it? And a lot of people will just sort of take the received wisdom. a received algorithm, a received methodology, and keep applying it. But don't ask themselves a lot of questions about why it works. They don't do a lot of experimenting or thinking. You'd be surprised at how little thinking quants usually do. it certainly surprised me. But that's not my mental habit. So one of the things I discovered as I began to really delve into trend following carefully and that was during my AHL years, was how important volatility is to the process.
See, what's really happening with trend following, for the most part, is narratives take hold in markets. Okay, ideas. Let's say the U.S. economy is going into recession, or China will boom when it reopens, or as is happening now, the view that China will continue to suffer as their economy softens. So a view takes hold, a narrative takes hold, and prices tend to move and get pushed along in the same direction. But when narratives change, there is usually a burst of volatility. Not always, but there usually is a burst of volatility around narrative changes.
And what this means is that by rapidly scaling your positions with using short-term volatility, shrinking when volatility drops, expanding when volatility drops, that process allows trend followers to take profits, in effect, around those turning points when there's a burst of volatility.
00:09:15 [Vincent Weber]
Okay, thank you. That's very interesting. Before we dive deeper into that, just one more question. You experienced many great firms, both like a company, also the academics there. And I know it might sound cheesy, but let's talk a bit about culture. You've been in Goldman, Berkeley department. So how do you view flooring courts culture now? And how did you shape it? And how does it differ from the one you previously experienced?
00:09:45 [Douglas Greenig]
I think a lot of people like to talk about culture. But so much of it is just a bunch of BS nonsense. But it's real and it's important. Seriously, it shouldn't be just a bunch of corporate speak and gobble the cook. So I'll give you my views on it. I'd like to think that the culture that I've created, that we've created at Flooring Court, reflects some of the best features of the various places I've been at, as well as reflecting to some degree. What I hope my character is. So let's start with this. Hard work. Nobody ever gets anything great done without working really, really hard. In the old days, when I was at Goldman Sachs, there was a culture of hard, hard work. And it started at the top.
OK, you can have a situation where for long, it's not a stable situation where the troops in the trenches are working super hard. And the management team is kind of swatting around Europe, taking long vacations and getting updated at long intervals. People expect the management team to lead from the front and to work really, really hard and be really, really engaged. And that starts with me. This starts with my other partners. There needs to be a culture of hard work. And by the way, if the management team is working hard, The odds are very, very good that everybody at each level sees that and it becomes part of the culture. So hard work is a very important thing. The next thing is you've got to be really careful about hiring people.
My wife is a psychologist, and she could tell you about all the interesting, complicated quirks, neuroses, issues, personality disorders that people can have. And you can have some incredibly talented people that are really pretty disruptive in the workplace because they're dealing with all sorts of complicated emotional issues, like the need for attention, the need for excessive credit or dark triad traits. I mean, there's some people with incredible talents who personify things like narcissism, Machiavellianism, and so forth. You need to hire very, very carefully. You want talent, but just as important as talent is you want basically well-balanced people who communicate well and are not bringing dysfunctional interpersonal styles into the workplace.
00:12:30 [Vincent Weber]
Okay, right. Excellent. Now let's talk a bit about your journey to alternative markets. I find it very intriguing because, For most, trend following has been associated with futures market. I get the term CTA. So how did that happen? What inspired the shift from traditional to exploring alternative markets?
00:12:52 [Douglas Greenig]
Every firm in the quant space or in the CTA space has some element that they regard as their special sauce. And part of the recipe at my former employer, Man AHL, was to add as many diversifying markets as they could. So the first alternative market CTA that I am personally aware of was the Evolution Fund at Man AHL. And it was an outstanding, it is an outstanding product, widely regarded as one of the best CTA programs in the world. And so I'm proud to have been associated with it for a period of time.
Now, I had the opportunity after I left, man, to get out a clean sheet of paper and say, if I can start with a brand new infrastructure, if I can start from scratch, how would I do things and make things as efficient as possible and make it as focused as possible? the basic concept is this. It comes from, if you will, I think it was Ron Kahn and Richard Grinold had some Principle of active portfolio management, which was risk adjusted performance equals breadth times skill. OK, if you can add new diversified markets to your trend following program. That increases breadth. If you can improve the models that you're using. That obviously helps with skill and if the markets are less efficient.
Whatever skills you have may work even better in these new alternative markets that are, again, I'm probably referring to Robert Frost, the road less traveled, right? You may find more opportunities on that road. And so the idea is not that complicated. Let's not mystify it. You're adding new markets that are less efficient and uncorrelated with the old ones. Of course, you're going to add a huge amount of performance, even if the markets weren't better markets, just through diversification, breadth in that equation. I would say that what we use one measurement, we count the number of independent bets in our portfolio versus the number of independent bets in the standard CTA portfolio. And we're a multiple, and we've been a multiple.
four or five times more independent bets in our portfolio the way we measure it using a kind of principal components analysis. There are other ways to measure it. Look at pairwise correlations and stuff. It's just a lot lower. We have a lot more diversification. And you would expect on that basis, just on that basis, for our Sharpe ratio to probably be about double a standard CTAs. And it is, or a little more than that. And then we have also found that the markets we trade trend a little bit better than standard CTA markets.
00:16:09 [Vincent Weber]
Okay, thank you. I think, yeah, the benefit was almost obvious, but there are definitely challenges to getting there. And can you expand a bit on the nitty-gritty part, how you had to get your hand dirty to be able to trade all those markets?
00:16:24 [Douglas Greenig]
In a sense, that's the right focus point. It's not that glamorous. But by the way, in life, a lot of the best businesses, involve a lot of gritty, not so pretty hard work to set the stage for that profitability. So in many cases, the data are less available. And there are operational hurdles. These are often not futures markets. Many of them are over-the-counter markets. A lot of them don't look like a future. A credit default swap doesn't look like a future. And so you have to transform the markets into something that looks more like a future, that you can trade with the systems that we have.
So whether we're joining some exotic exchange in order to be able to trade freight, whether we're learning about the inner dealer brokers and sources of liquidity in an emissions market, whether we're dealing with a complicated tax issue in some emerging market, futures market. These sorts of things are the sort of bread and butter of what we do.
00:17:37 [Vincent Weber]
Maybe let's not talk about the actual strategies. Maybe just to start at a high level. So in terms of return, profit, or profile, or benefit to investors, what are you trying to achieve with your strategies?
00:17:52 [Douglas Greenig]
Well, our main program runs at Tenvol. And unlike many systematic shops, We actually deliver our advertised vol, or at least we have done so historically. We're very close to 10 on any kind of reasonable look back. And as I had mentioned, we have a sharp ratio in the neighborhood of one, a little over one. So we're generating about 10% returns at 10 vol, and it's pretty much uncorrelated with everything except the CTA index. The CTA index, we still have some positive correlation because you pick up momentum from big macro trends in there. But we have a pretty big tracking error to it, by which I mean typical developed market CTAs have a sharp ratio of 0.3. We have a sharp ratio of over one.
So our curve goes up and the other one kind of goes up at a much slower rate, is more choppy to the side. Because of our diversification, We look less cortotic. Our drawdowns are less alarming. One of the worst months we had was in March of 2023. There was a big repricing in global fixed income and stuff due to the potential for a credit crisis. You were called Silicon Valley Bank and all that stuff. And some of our peers lost 12 percent. OK, at similar vol levels. and we lost a neighborhood of five. So if you just look at our kurtosis statistics, they look good. We don't quite get pummeled the way less diversified programs do. So we want the positive skew.
We want to be very market neutral in the sense that we don't have a long beta bias, a kind of long treasury bias. We don't want to have a bias in the program.
00:19:53 [Vincent Weber]
Okay. Thank you. Let's talk a bit about risk management. I see you also have a strong background there. So is there something beyond vol targeting? What is there beyond vol targeting?
00:20:05 [Douglas Greenig]
Well, one of the difficulties here is you're actually managing a portfolio. When I spoke of vol targeting, we were talking about one market, a narrative, right? The narrative changing. There's a burst of vol. Vol doubles. You cut the position in half. the trend turns. At first, you lose a little bit of money. Then the trend following system locks on to the new trend. And because you cut the risk, you locked in your profits from the prior trend to a significant degree. Those sort of stylized facts. But now we're talking about an entire portfolio of this and trying to hit a volatility target for that portfolio. Now things are getting tricky because the correlations among the different assets, they're changing.
Think about how much the correlation between stocks and bonds in the U.S., just to choose a non-alternative market, how much that's changed. In normal circumstances, strong economic growth is good for stocks and a little bit bad for bonds. But at other times when monetary policy is in play, you'll have a situation where the Fed sounds more hawkish. It's bad for bonds and that's bad for stocks. Right. So the correlation could be positive or negative between stocks and bonds, depending on whether you're talking about the monetary policy function in the second case or fixing the monetary policy function. And you're just talking about economic growth. These correlations are not terribly stable. And you can imagine when you're looking over 500, we trade about 500 macro assets. All these correlations keep changing.
So we followed an approach to maximize diversification. Because remember what I said in the beginning, more diversification plus better trends equals better pizza. So we want our allocation to really exploit. diversification, to maximize diversification. So we need to take those correlations into account. So we use a tree structure. And essentially the idea is you have twigs, you have branches, you have limbs, big branches, you have the trunk of the tree. And the trunk of the tree is the portfolio. A big branch might be power markets. A smaller branch might be natural gas markets. And then you get down to individual markets. And what we do is we're constantly comparing the delivered vol along a twig, a branch, big branch, a trunk, to the delivered vol to the expected vol.
And then nudging our position up or down. so that you're trying to converge in the direction, okay, of delivering the vol that you want.
00:23:22 [Vincent Weber]
Very insightful, Doug. One last question I have to ask you, that the buzzword of the season, AI, so it's everywhere. What's your take on it? Does it play a role for your business at all? Is it incremental or a game changer?
00:23:37 [Douglas Greenig]
Well, a term has emerged in connection with ESG. I think it's called greenwashing. And it's where people try to dress up what they're doing and pretend that they're more ESG than they really are. And there's probably some AI washing going on where people want to sound like they're involved in the space more than they really are. The thing I would say is AI is really an extension in many cases of traditional statistical methods to larger data sets. It's not always quite as exotic as people think. A neural network is a form of a nonlinear regression estimated using methods like backpropagation. And actually, it's a pretty cool thing. And neural nets can do many interesting things.
So with respect to what I regard as real AI, deeper AI, I think one of the more interesting lines of attack and what I'm interested in is learning how to separate causation and correlation. Okay? We've all learned that statistical correlation is not the same thing as causation. smokers tend to have stained fingers and they also tend to have certain health problems. But it's not the stained fingers that cause the health problems. It is the tobacco consumption that causes the stained fingers and separately causes a range of health problems through the introduction of carcinogens and free radicals and things like that. But there are statistical methods.
that have been developed in the last few years that rely on the notion of conditional independence to help you sort through this stuff and determine what's causation and what's correlation, or at least give you some ideas about that. And the goal is to get a directed acyclic graph, or in some cases, a cyclic graph. of causation. There's an exciting book written a number of years ago by a computer scientist, Professor Pearl, on this very subject. And this is something that interests me, because this is the exact opposite of what many quants are doing in the AI space, which is to say they're fitting very, very complicated loose models that pick up spurious correlations and often perform poorly out of sample, and we all know about data mining.
I'm interested in using the most sophisticated techniques to figure out what the real causal structure of macroeconomic variables is. Now, we have, at present, a systematic non-trend macro fund using a lot of alternative markets and using a lot of fundamental variables. And it has done well in recent years. And that is one place where having a more deeper understanding of the causal structure of macroeconomic relations is quite relevant. For example, there is the monetarist point of view that the cause of inflation 24 months down the line, is an expansion of broad monetary aggregates like the Vizia M4. And that it's not the stuff that most people talk about.
That the stuff that most people talk about are not causes, but are more like the stained yellow fingers for the smokers, if you see what I'm saying. And so I'm interested in certain applications of AI. that are away from the kind of crude data mining approach where you're really understanding the deeper structure of markets and economics. And so we are working on some things like this, particularly in connection with our non-trend systematic macro fund.
00:28:18 [Vincent Weber]
Wow, that's really interesting. By the way, Doug, Thank you for sharing your insight with us today. And I will say it's been fascinating to hear about your innovative approach. And to our listeners, thank you for tuning in to Resonanz Spotlight.