“I did end up in 1984 meeting another trader by the name Poul Tudor Jones and I allocated some capital to him, I met in ’82 John [W.] Henry and allocated some capital to him….and we were the first outside investor for Transtrend in a fund that they launched when they launched when they launched their company in 1992”
On today’s show I am talking to Mike Dever, the Founder and CEO of Brandywine Asset Management. Mike is a Divergent Thinking Systematic Investment Manager and Author of the best selling book “Jackass Investing”. Mike has been trading for almost 4 decades and shares a wealth of insight to how he has stayed successful and how his trading approach and research has stayed current in an ever changing financial landscape.
Thank you for visiting, now let’s get to the interview with Mike Dever.
In This Episode, You’ll Learn:
- The Evolution of Mike’s Trading
- How Mike almost got in to business with Richard Donchian
- How they developed a systematic trading model in the 1980s
- How Mike met and invested with Paul Tudor Jones, John W. Henry and was the first outside investor in Transtrend
- How Mike went from being a discretionary trader to become fully systematic when launching the Benchmark Program, and Why
- About the experience of running the Brandywine Benchmark Program in in the ‘90s
- How Mike discovered the importance of Portfolio Allocation in the 1980s
- The Story of Developing the Predictive Diversification Portfolio Allocation Model
- The Entrepreneurial Storm that Kept Mike away from Trading and Brandywine for Years
- The story of developing the Brandywine Symphony Program
“The question that was being asked, by everybody that was doing portfolio modeling, then and continuing today, was wrong. They were asking, ‘how do we achieve the most optimal performance.’ They got that. The question that they should have been asking was, ‘How do I get the most predictable performance?’ Nothing else matters.”
- How many people it takes to run a system with more than 1,000 strategy/market combinations
- Exploring the track record of Brandywine Symphony Program
- Why one should avoid making decisions based on single return drivers
- 5 Themes in Brandywine’s Asset Allocation Strategy: Fundamental, Sentiment, Event, Arbitrage and Alpha Hedge
- How Marginal Cost of Production Works as a Return Driver
- Are the strategies from the 1980’s just as effective today?
- What is directional arbitrage?
Resources & Links Mentioned in this Episode:
“If you’ve got optimal returns that you don’t know are predictable going into the future, who cares?”
Jackass Investing by Mike Dever
Learn about Dick Donchian the Grandfather of Trend Following
Learn about Poul Tudor Jones of Robin Hood Foundation
Learn about John W. Henry
Sponsored by Swiss Financial Services and Saxo Bank:
Connect with Brandywine:
Visit the Website: www.brandywine.com
Call Brandywine: +1 630 361 1000
E-Mail Mike Directly: Mike@brandywine.com
Niels: You are listening to Top Traders Unplugged, Episode number 003, with Mike Dever, founder and CEO of Brandywine Asset Management. This episode is sponsored by Swiss Financial Services.
Introduction: Imagine spending an hour with the world’s greatest traders. Imagine learning from their experiences, their successes and their failures. Imagine no more. Welcome to Top Traders Unplugged, the place where you can learn from the best hedge fund managers in the world, so you can take your manager due diligence, or investment career to the next level. Here is your host veteran hedge fund manager, Niels Kaastrup-Larsen.
Niels: Welcome to another episode of "Top Traders Unplugged." Thanks so much for tuning in today. I really do appreciate it. On today's show, I'm talking to Mike Dever, the Founder and CEO of Brandywine Asset Management. Mike has been trading for almost four decades and shares a wealth of insight to how he has stayed successful and how his trading approach and research has stayed current in an ever-changing financial landscape.
For those of you who are new to the show, I just want to let you know that you can find all of the show notes, including a full transcript of today's on the toptradersunplugged.com website. Now, let's get started with pod one of my conversation. I hope you will enjoy it.
Mike, before we jump into more of the specific topics and questions that we will cover today, I just want to start out by acknowledging you and thank you for being on the podcast today. I still remember when I visited your offices about 17 years ago. Because it left such an impression on me because of its uniqueness. So, of course, I'm interested to hear whether your offices are still located in this wonderful 17th century gristmill that you renovated so beautifully?
Mike: They are. We're now going on close to 20 years in here.
Niels: Fantastic, and speaking on uniqueness, I think it's fair to say that you do things a little bit different to many of the other systematic investment managers. And this will certainly shine through our talk today. So, I think our audience is in for a real treat and I think we should just jump right in to today's topics, if that's okay with you.
Mike: Excellent, thank you.
Niels: Mike, your journey has been long and distinguished, spanning over nearly 40 years of trading, as well as being a bestselling author of the book "Jackass Investing." And not forgetting, an Internet pioneer.
And it's my impression that these different experiences play an important role in what you do today from a trading point of view. So, perhaps you could take us back to the beginning and tell us this fascinating story and of course, a bit of background on Brandywine as a firm and its evolution over the years.
Mike: Certainly. I founded Brandywine in 1982 and I had been trading for a few years prior to that, just myself on a discretionary basis. And throughout the 1980s, continued to do more of a discretionary trading, but with some support from computers and other research to kind of position my trades.
The other thing that was, I think, key to me and it was just a nice experience through the '80s, was that I had started some funds with some other managers as well. One of the early ones that I came in contact with, I sought him out, actually, was Dick Donchian who a lot of people acknowledge, I guess, as the grandfather of trend following trading.
Mike: He was at Shearson at the time and I went up and met with Dick and started putting a fund together that Dick would manage. I would be the pool operator, he'd be the CTA on it. And we got well down that path before Shearson put an end to that. They didn't want some young punk, me, running a fund with one of their brokers on it. So, unfortunately, we weren't able to put that together.
But my hope was to complement my own trading with funds that would be managed by some other traders. And really, more of it was from a business standpoint, to help smooth out the revenues to the firm.
So, I did end up, in 1984, meeting another trader by the name of Paul Tudor Jones and I allocated some capital to him. Met in '82 John Henry and allocated some capital to him. So, I did have a couple other funds that were out there managed by some other managers.
And so, it was a nice experience through the '80s with my own trading, the research I continued to do, the ability to be on the front of the stage with some of the early pioneers in the hedge fund industry with Paul, with John. And I carried that into the 1990's as well, when I set up a fund and we were the first outside investor for TRANSTREND, on a fund that they launched when they launched their company in 1992.
Niels: I didn't realize that, wow.
Mike: So, it's been kind of a nice experience and seeing what some of these other people are doing as well. What I did through the '80s, I started realizing, with the discretionary trading which was around the clock. That in order to get more predictable performance, instead of what I refer to as the "ultimate black box," which is a brain. And not really knowing day to day exactly how that brain was going to react to certain market conditions and environments. I really wanted to systematize the approach.
So, we sat down in the late '80s, put together a research project that over the next few years became rather intensive and extensive. And developed a systematic trading model that was very different at the time. And even today, I look at it and I realize that there were so many things we did then, which is 25 years ago, now, that people still are not doing today and I believe give us a real nice edge.
But we systematized the approaches and we launched our brand-new Benchmark program in 1991.
Niels: Yeah. Would you say even at that time, you were influenced by what you had seen from these great traders that you were allocating money to?
Mike: It was nice having that experience and exposure. But what I realized was that they were doing things very differently than the way I thought of doing things. Paul Jones was very much a discretionary trader. And as often as he tried to systematize his approaches, I don't think he quite captured what it was that he was doing on a discretionary basis. He had, still has, I think, an innate intuition for markets. And above all else, he just had an enormous discipline for being able to manage risk in his portfolio.
Obviously, John Henry was one of the early pioneers in trend following. And it's a fairly straightforward approach; it's a sound return driver, but it is only one return driver. So, I could pretty much have a good idea what his positions were without even looking at my brokerage statements.
And TRANSTREND, I came into contact with TRANSTREND when I was doing the research for our Benchmark program. And realized that they also were involved in the belief of the broad strategy and market diversification.
Now, we take it a little bit differently than they do. They tend to be, I believe, more technical. I don't know today exactly what they do. But at that time...and we have more fundamental basis strategies in our portfolio. But we still shared that philosophy that essentially, there is a free lunch, and that's portfolio diversification.
Niels: And the Benchmark program, was that systematic and completely without discretion? Or did you keep sort of your discretionary trading in there?
Mike: No, we didn't. It was fully systematic. But what we did was look at a lot of types of trading that I did through the '80s and try to capture that in a systematic fashion. So, for example, we had a number of strategies that were based on market reaction to news events or reports. Because I traded a lot of that type of event reaction in my discretionary trading. So, we systematized the approach, captured it in the best way we possibly could and included that in the Benchmark program. So, there were a number of strategies in Benchmark that were based on sort of that intuitive or discretionary feel. But instead of relying on the black box of the brain to interpret everything correctly each day, it followed a rigid set of rules to be able to repeat the process every time it reoccurred.
Niels: And so, in that sense, just to be clear, were you convinced, even at that time early on, that if you were going to be successful from an investment management point of view, you had to be systematic? Was that something that just was obvious at that time? Or was it more of an operational issue where you say "Okay, it's actually easier if I can do it systematically."
Because the systematic side of things is partly, I think, an operational issue, but it's also an emotional issue, in my opinion.
Mike: Absolutely. The one thing you get from systematic...the one main thing I think would benefit the majority of traders out there is the discipline that it brings to the trading. You can sit there and say as much as you want as a discretionary trader "I'm going to cut my losses at these levels, I've got a bad trade, I'm going to get out when it does this."
But not everybody does that. And you may do that for a while, but at some point, there's that constant battle where you're sitting there saying "But it's a great position. I've got a great position."
And I've been reading some books about the financial crisis. And one the things that you see there that's the same as what you had with long-term capital, is just that utter breakdown in risk management discipline. And it's easy for a person to sit and rationalize why "You know what? The risk management wasn't designed for this scenario." Or "This was a great trade 10% loss ago. It's even better now."
So, I'm certainly not going to get out of it today and just ride it into the ground. So, the one thing, if there's nothing else that you get from systematic trading, is the disciplined risk management approach to trading.
Niels: Absolutely. And so, the Brandywine Benchmark program that ran until 1998 I think. Tell me a little bit about sort of your experience during that time? And then, you moved onto other things, as I remember.
Mike: Sure. So, Benchmark was...when we went out with Benchmark Program, it was really interesting because we funded it with $1 million in 1991. And we had a lot of confidence that what we were going to get performance-wise was going to be fairly similar to what we tested it to perform it like.
And the reason was because of the methodologies that we developed not only in our back-testing of the trading strategies, but in the portfolio allocation model and I'll talk about that for a second.
When we went out in the late 1980's knowing that we were going to have dozens of trading strategies. We were going to be trading across dozens, eventually well over 100 markets in the portfolio. The first thing we needed to know was how we were going to allocate across those strategies, markets in the portfolio.
So, we went out and looked at some of the standard, conventional, accepted, I guess, methods for portfolio modeling, such as mean variance modeling, the Markowitz-type modeling that everybody was doing. Looking at efficient frontier analysis.
And we hired some of the brightest people out there to help us create portfolios based on those methodologies. And one of the first results I got back from one of those researchers was the results of how to allocate across--it was about 15 different strategies, maybe 30 different markets.
And the results came back with a high concentration of one of the strategies trading in one of those markets, a few other strategy market combinations. And then, pretty much else, everything was all set. There were no allocations.
So, when I called him on that, I said "Okay, obviously these results aren't right. I can't do this. It's not going to work going forward." His response to me was "That's the perfect answer." And I started realizing that what they considered "perfect" was a mathematical elegance. A level of mathematical perfection that had nothing at all to do with what I was trying to accomplish, which was to create a trading program where I could look at its back-tested performance and have a high level of confidence that that was going to occur again in the future.
So, I sat down for the next six months to a year and we had a number of people working on this. But it ended up really coming down to more of an intuitive result. And what I realized was that the question that was being asked by everybody that was doing portfolio modeling then and continuing to today was wrong. They were asking "How do we achieve the most optimal performance?"
And they got that. But the question was that they should have been asking was "How do I get the most predictable performance?" Nothing else matters. If you've got optimal returns that you don't know are predictable going into the future, who cares?
And so, what you found was that people were fixing their results. They would get the answers back, the perfect result like I got from my one researcher. And then, they would modify it. They would say "Well, we won't allocate more than 5% to any strategy market combination." Or "We'll add a little bit of these other things, even though they don't really get much of an allocation because we know they diversify the portfolio and they add value."
My response to that has always been "If you get results back that you know are wrong, don't tweak the results. That's the same as putting earrings on a pig. What you've got to do is go back and fix the original model."
So, that's when we came up with the predictive diversification portfolio allocation model. And we use it today.
Niels: Yeah. We're certainly going to talk a lot more about that. But just briefly, Mike, if you would just take us through the next iteration of your career after 1998? Because I think, also, maybe that's partly influencing when you returned to managing money a few years ago?
Mike: Yeah, definitely. At the end of 1996, I started a company called Spree.com. Spree was an early pioneer in e-commerce. And we referred to it as an e-commerce community. Because Spree allowed people to come in and set up their storefronts, sell products off of those storefronts. Spree managed all the back-end of that for them.
And pretty quickly, by mid '98, Spree had grown to be the seventh most trafficked e-commerce company in the world. So, it became a tail wagging the dog. Where I thought Spree was going to be a side project. When I launched Spree, I had a couple of dozen employees in Brandywine. By mid '98, late '98, we had over 80 employees in Spree. So, it now was three times as big as Brandywine was.
And we started getting acquisition interest from a number of companies that were looking at buying Spree. So, I really ended up focusing a lot of my attention on Spree. I brought in some additional other people to manage Brandywine.
And it really was...it was a great experience and a sad experience at the same time. Because Brandywine was and still is, my first love and my first primary interest. And I was not able to stay running that at the same time that I was trying to hitch a ride on this Spree. It was a roller coaster, it was crazy.
So, I needed to be there on the Spree site. And I guess I overestimated the amount of sort of corporate continuity and corporate memory there would be with me stepping away from Brandywine. Because it radically changed over the next two years.
And the research group that I put in place was replaced by a researcher doing equity research, another one doing futures research, because we had some equity programs as well. And it became more of a conventional-type firm in the way it did its research.
So, Spree was a great company, a great experience. And I got out of Spree at the end of 1999, early 2000 when we had venture capital money coming in. I set up a company called Mind Drivers which was essentially taking on the other early stage technology businesses that had grown up around Spree. And I ran that through the 2000s.
And in 2007, sold one of those companies. And started the path back to doing liquid investment trading again.
So, really, there was about almost a decade period where I still had a foot in the door with the trading side, but it wasn't the immersive experience that I had had prior to that or subsequent to that.
Niels: Sure. And just out of curiosity, for your own sort of part, did you continue any systematic trading during these years? Or did you simply walk away from trading while you were doing the Internet side?
Mike: We had a mix. We had one fund that when I got back into Brandywine, we were managing. It was our largest product at the time. The futures trading had essentially dissolved at the time and that's really my fault. I promoted Brandywine as being a benign dictatorship and when they saw that I stepped away, our investors stepped away as well.
So, that was a bit of the corporate continuity issue that I had with that. And we resolved today, bringing in a partner and doing some other things.
But I had one fund at that point that was doing mutual fund arbitrage and we continued to trade that for a few years. And then expanded on that by running another fund that was trading that as well as the venture capital, early-stage investing deals I was involved in.
And that product in 2004 added some futures in a sub account that traded for about two years. But we really realized that unless we were going to be fully committed and focused on one type of trading, it didn't make sense.
So, we refocused that fund entirely on venture capital. And that actually remains open today as a venture capital product. But today, when we got back into trading, everything we're doing now is fully-focused on our globally-diversified liquid assets trading.
Niels: Yes. And of course, 2011 comes and in July, you launch your current program, the Symphony Program?
Mike: Yes. Symphony, what was interesting about this process is when I made the decision to get back into trading on a full-time basis and commit to beginning the Brandywine Symphony Program, I went back to the strategies that we had developed in the '90s. And there were about maybe three dozen that were in the Benchmark Program.
And pulled out, evaluated all the ones and pulling out the ones that I thought were still relevant today. There were some that became irrelevant. One, for example, we had that was looking at freely-traded interest rates. Fed-controlled versus freely-traded. That in the 2009, '10 period when we were starting to redevelop the program, we realized there just wasn't the environment for that. You didn't have that freedom of interest rates in the environment. So, there were strategies like that that we didn't even look at updating the testing on. But there are about two dozen that still remain relevant today. And we updated the testing on those. And the great news about it--I like to say it's the good and the bad.
And the good side and the great news was that those strategies continued to perform, as expected, from 1999 through the 2009 period when we first started updating, to walk forward on those. As we had not only back-tested to perform through the '70s and the '80s, but they actually performed with real money in the 1990s.
So, the strategies remain valid. The bad news, as I like to point out, if I had just simply continued to run Brandywine in that fashion, I would have owned a baseball team today. So, I don't have the Boston Red Sox, but I do have those strategies and they're of enormous value.
So, we went out, we raised $10 million to have non-proprietary capital trading the initial seed account. Started trading that in July of 2011, which was coincident with the publishing of my book "Jackass Investing."
Niels: Yeah. And today, can people come to you both through managed accounts and fund vehicles? How do people come to you today?
Mike: Either way. So, a managed account is $5 million nominal account size minimum. And that's targeting about an 8% standardized deviation with a low double-digit return.
So, if somebody wants to take a $5 million account and is willing to accept higher volatility or risk in their portfolio, they could notionally fund it. And we could have as little as maybe one-third cash to the $5 million investment.
Then we also have two funds. One is our standard Brandywine Symphony fund. That trades at the standard 8% annualized standard deviation leverage. And we would refer to it as a cash-efficient product, which is essentially the same as notionally funding a managed account that trades at three times that standard leverage. And that's our Brandywine Symphony Preferred Fund. And those are able to be invested in at $100,000 minimums.
Niels: Great. And how much does the program run or manage in total today?
Mike: $30 million today.
Niels: Okay, great. Mike, I want to shift gear on you and just touch upon a little bit about the organization. Because you, as well as I think many other firms, I think you run a very lean infrastructure today. And I guess a lot of that is based on all the wonderful things we can do today with technology.
But how do you view sort of what to keep internally in the firm, in our business and what can be outsourced? How have you balanced those two things? Mike: There are two reasons that Brandywine today can be a lot smaller personnel-wise than it was in the 1990s. The one is what you're pointing out is there's so much good technology now available to do a lot of the same things that we did with people in the past.
The second is that Brandywine already had such a great base of research that we could pull on. And so, we were able to probably get the same effect as far as diversification portfolio in the research with one-fifth, one-tenth of the people that we had in the research group back then.
On the things that we can outsource today, in the 1990s, we built our own front to back research trading account reconciliation account management program. Today, we have what we call Cadence, which is our trading research and order generation system. And then, now, we've licensed the DMACs, the books to do all the account management side.
And then you go through, for example, CQG or PatSystem or trading technology to do trade execution. So, we don't have the same need for the staffing of the trading desk that you had in the past as well.
So, we're able to operate with four full-times and some part-time people today and I think we can maintain a pretty lean operation growing over the next year or two.
Niels: And in terms of growth and so on and so forth, looking at the markets and the way you trade the markets, is there anything to be said about an optimal size for the program? Is there anything where you would say "We've designed it to be 'X' amount on the management and probably not further than that?"
Mike: There are certainly markets that we trade in the portfolio today that have lower liquidity than others; we trade in the livestock markets. Some of the other foodstuffs, agriculturals that have lower liquidity in them.
The way we look at it is what asset levels will we be able to maintain, say, 100% optimal diversification and how will that change as assets increase? And roughly, because we're so broadly diversified across trading strategies in markets and the portfolio, we can probably get close to that half-billion, third-billion dollar range before we start allocating sub-optimally to the markets and the portfolio.
And then what will happen is that next tier going from half-billion to, say, $2 billion. We may drop from being 100% optimally allocated to 80% optimally allocated in the portfolio. And then it tends to sort of level off a bit after that. Because it drops a little.
But for the most part, the markets that you remain able to be optimally allocated to, they're very liquid. So, you can go up in the billions quite easily. And you just continue to maintain the same position size, this maximal position size as in those less liquid markets, which lowers their effect on the portfolio as your AUM grows. But we believe that for the first half-billion or so, we're pretty good and able to be reasonably allocated.
But that's really a function of the fact that we've got so many strategies and markets in the portfolio. If we were trading the narrower portfolio, we'd hit that threshold much sooner.
Niels: Sure, absolutely. Now, another area that I want to cover and talk a little bit about is sort of just track record from a very high level. The Symphony Program, in my opinion, has a great track record, starting in 2011, has certainly been an interesting time to launch a program. And I've obviously seen your numbers and they look great. So, obviously, I invite people to look more into that.
But I want to ask you, with such a long experience in this industry. And that's just really performance after 2009 in the CTA space has, in many respects, been quite different to performance prior to 2009.
And of course, you're probably one of a very few in terms of the exception to this because your performance has been pretty strong in the last few years. While many large managers who have been around for two or three decades have had somewhat more of a challenge.
And even some of them, seeing the draw-downs expand quite dramatically...what is your sense and what is your two pennies about performance in the last few years compared to the much longer period that sort of CTAs and trend following, in particular, has been known for being quite a robust investment case?
Mike: Right. I can certainly talk in reasonable detail on what we do to the extent that we want to reveal things, but I can only surmise what might be going on in the rest of the people that are out there registered to CTAs.
And with Brandywine, our approach is return driver-based. So, we're looking at sound, logical return drivers that we can build trading strategies on and diversify across those in a portfolio. So, we don't really limit ourselves at all to any style of trading.
If we think there's opportunity in trading against investor sentiment in the S&P, we'll develop strategies that capture that. If we think there are things that I can do...like the event trading ideas as a discretionary trader in the 1980s, we'll develop strategies based on return drivers that will do that.
So, we've got, now, dozens of different return drivers that are underlying the trading strategies in our portfolios. So, we end up with returns coming from a lot of different sources. And that, from our standpoint, means that we kind of don't care about any specific environment that's out there. For us, the last three years are the same as the prior three years. Contrary to that, I think what I've seen happen--and it's not just the guys that are out there and women that are registered as CTAs. This is the investment industry, in general. It’s that people tend to latch on to a certain style that was successful. And they fine tune maybe a little too much their approach to being suited to that style.
And when the markets change, the environment changes a little bit, that style falls out of a favor. They lose money.
So, the example we've had in the futures industry--if you can call it an industry. And I really kind of don't refer to it so much as an industry because everybody is independent. But there is a group of people that trade futures markets using trend following strategies, because that was one of the return drivers that they looked at and said "Wow. Over the last couple of decades, this makes money." And no different than equity managers who sit there and say "Wow, this value thing makes money. I'm going to be a value investor."
It's just one return driver. There are dozens more that they can employ. So, they start out by limiting themselves to one return driver. And then, they compound the problem, from what I've seen, by fine-tuning that return driver to give them the best possible performance on that past data.
And we don't have any return drivers that when it's properly constructed into a trading strategy, consistently make money over multiple-year timeframes. They make money over 50 years, but in any decade period, almost, I would say any trading strategy that we have has decade-long periods where it loses money.
And the only way that you can avoid that in back testing, is to fine tune the parameters until you get to a point where you don't have those periods. I just read an article. It was interesting, but troubling, this past week. And somebody was talking about, this is somebody that's written a lot of articles in different magazines about trading strategies and strategy development.
It was talking about a strategy that they thought was really good. But over a four-year period, it had one year where it lost money the entire year. And it gave as much as 20% of the overall profits they made over that four-year period. So, obviously, nobody could actually trade that in real time.
And I'm looking at that going "That's just unbelievable. I don't have any strategy that is that good in my portfolio."
And so, I think what happened is Brandywine, everything we do, everything in our research is focused on predictability of performance first and foremost. If you don't meet that threshold. That you have a high degree of likelihood that the future performance is going to match the part performance. Even if it's bad, I don't care. I just need the predictability first and foremost. Then we can sit down and figure out how to combine those things into a portfolio, whether they should be included in a portfolio or not. The rest of the industry, every article I've read, every person I've talked to, other CTAs, equity managers. It's across the board, academics. Across the board, it's trying to find some sort of optimization, optimal performance on a trading strategy or an idea.
And then, the question is "Okay. I know it's perfect here and it won't happen in the future. And they come up with some sort of a formula for how much they'll discount those returns and haircut the performance of that."
And again, it's back to putting earrings on a pig. If you get results that you look at and you know they're wrong and you've got a haircut to tell you what you think you're going to get going forward, don't even try. Start over at the beginning again, create something that's predictable. And don't fool yourself. You're going to develop strategies and they're going to have 10-year losing periods. Doesn't mean that the return driver is not valid. It just means that that's what happens.
Niels: Yeah. Well, in a sense, it's just like looking at the equity markets. They can certainly go 10 years without making any money; it doesn't mean that people don't like investing in equities.
Mike: That's right. And when I talk about it in the book and in my first chapter is that's because in short periods, less than 20 years in the U.S. equity markets and I think you could apply this globally, too, but I did all my research on U.S. equity markets. The dominant return driver is investor sentiment. It has nothing to do with the fundamentals of the underlying companies.
That means that you can have a company that doubles its earnings over a 10-year period. But if investor sentiment turns negative over that period, you can have a decline in that stock price. Or the overall index.
So, essentially, you're saying that those returns are dominated by a single return driver, which is investor sentiment. And if you've got a single return driver driving your returns, you're going to have decade-long periods where it doesn't perform. That's just what happens.
And you can fix it. You can add some filters and fine tune the parameters to make it look like you wouldn't have lost money had you done those things. But now, you've just slashed the predictability performance. So, the results you're looking at, historically, have virtually no meaning going forward.
Niels: Yeah. And I think this is also what I alluded to in the beginning where I said that I think it's fair to say that you do things differently and that I really appreciate that. So, I want to talk about, if my understanding is correct, you kind of have these sort of five..I don't know whether you define them as return drivers, but you certainly have these five themes in the model or in the program. Fundamental, Sentiment, Event, Arbitrage and Alpha Hedge.
Can you talk a little bit about why these and then we're going to dive into maybe each one of them and try and get people to understand what you mean by it?
Mike: Okay. I'll start off by saying those are totally arbitrary classifications that we came up with. You'll understand a little bit as we describe them more. But we have, in the portfolio, dozens of individual return drivers.
And for purposes of illustration or discussion with investors, we categorize those into those five strategy types. And because they have certain characteristics that are similar. But within each of those strategy types, the trading strategies, with the exception of Alpha hedge, have no relationship to each other as far as performance. So, the return streams are totally uncorrelated. But they're based on something similar.
So, the idea is, we don't really allocate based on those strategy types. We allocate based on each of the underlying return drivers in the portfolio. I can talk about different types, but--
Niels: Well, give us an example of a return driver and let's take it from there and see where it goes.
Mike: Okay. Well, we talked about sentiment. And sentiment being the dominant return driver for the stock market in periods of less than 20 years. Over 20 years, you start getting a bigger contribution from actual corporate earnings growth as the dominant return driver.
But it's really until you're out in the mid-20's, 30-year period where that really starts dominating the portfolio and sentiment takes a back seat. So, if you look at it and say that sentiment is a dominant return driver for equities in the short term, you can develop trading strategies based on that return driver, that concept.
And so, one of the things that we do--and I actually reveal an actual trading strategy on the companion website for my book. It's a sentiment-based strategy that we developed more recently, because it wasn't even available to be developed in the '90s. That's looking at money flows in and out of the aggressive and inverse U.S. equity and U.S. bond now, we've extended that, ETFs.
So, if you see a lot of money flowing at the triple leverage long and out of the double leverage short U.S. equity ETFs, it's telling you that the sentiment--people are voting with their money, the sentiment is getting bullish. And this strategy goes in and looks for extremes in that to short the market. It'll be very selective. It might make a few trades per year based on this approach in any given market. And those positions will be held for a week, maybe a couple of weeks at the most. It really, just until that return driver which is the excess sentiment, is wrung out of the market.
So, we're not looking for a trend to develop or anything. We're just saying, we're doing everything we can to do as purely as possible, capture that return driver.
Niels: And if we take that as an example, so you see that obviously, cash is going towards equity, ETFs or whatever it might be. But how do you then implement a model? You're now looking to go short, that market. What kind of indicators do you then use in order to trigger that?
Because this is the sentiment. It means cash is going in one direction. But it doesn't obviously necessarily link directly to the price of whatever underlying instrument you want to trade.
Mike: Right. But remember, what we're trying to do is in as pure a way possible, capture that return driver. So, we could change the strategy. And instead of simply shorting it when the return driver indicates, which is what we do, we could say "We're going to wait until the market actually breaks down." "We're going to wait until there's a key reversal day." We could wait for some other technical indicator.
What we've just done by doing that is turned it a little more now into being, say, a trend strategy. We're waiting for the short-term trend to turn before we do something. And we're not now purely capturing the return driver. We're capturing a combination of return drivers.
So, this is where I talk about predictability of performance. We have a high level of confidence that by keeping the degrees of freedom as small as possible and having that focus just on that single return driver, which is market sentiment, defined in this case by ETF money flows. That the results we're getting, historically, have a high degree of probability. They're going to recur in the future.
And that's all we care about. And then, we look at those returns, those historical returns, and we say "Okay, are we comfortable with those?" And comfortable doesn't mean that they're great returns. It just means that, yeah, it looks like it captures some positive money over time.
And it may be that in a five-year period, it made money, but it had draw-downs that were twice as big as the money it made. I don't care. It's a valid concept that's producing positive returns.
And there will be periods, we know, that it'll go for five years, it'll go for 10 years and it won't make money. As long as the concept is still valid, it's included in our portfolio. Niels: Interesting. And before we talk about the overall, how many return drivers and combinations and so on and so forth, do you have a sort of a similar example? Again, without giving away details you're not comfortable with. But say, a fundamental strategy, what might that be?
Mike: Sure. So, if you look at trend followers, a lot of these CTAs will have trend following strategies. That'll be short markets and see those markets trading down to really low levels. And instinctually, maybe intuitively, they're sitting there going "I don't really want to be short. Nat gas at $2.50, I don't want to be short. Orange juice at $0.60."
But they remain short. That's what trend following does. And I don't disagree with that. If the valid strategy is to remain short because the trend is short, you stay with it. But at the same time, that unease is pointing out something else. That there's a fundamental bias that may start providing some support at certain levels at certain markets.
So, in the 1990s, we referred to it as "marginal cost of production." It's looking at commodity markets that have reasonably stable marginal costs of production that can be determined.
And when those markets are nearing, approaching, going under that marginal cost of production, we trade them from the long side only.
Mike: Now, it doesn't mean that the portfolio might not be short from other strategies. But that particular strategy is only going long. And it's taking short-term, long-only trades in markets that sell off like that. And it may accumulate multiple positions. That may exit some of those positions.
And really, the effect of it is that over an average campaign, which lasts maybe six months for one of these strategies, will be in and out of the market a dozen or more times. The average trade length is around 10 days. So, it'll buy, it'll sell, it'll buy some more, it'll buy some more, it'll sell some more. And each of those trades is a week or two in length. And the whole campaign of multiple trades might last six months.
It's an infrequent strategy; there's times when markets are trading near their marginal cost of production. There's periods like the late 2000s where markets were substantially above their marginal cost of production and nothing happened.
Our latest trade in this was in Nat Gas. In January of 2012, we started putting some positions on, we got out of that a little bit and got back in, I think it was March or April. For the next few months, we're trading from the long side only as Nat Gas dropped from under $3 to under $2.
And then, has it finally rallied back out of that range? That strategy stopped trading, we were out of it. Picked up maybe 50 basis points on the portfolio from that strategy over that campaign period.
Niels: Sure. Do you have an example of an event-based strategy?
Mike: Yeah, so event strategies; I've seen people try to do event strategies in the past and they were very complex. And when I traded in the 1980s, I was looking at a lot of things going into, say...when I started trading, it was 1979. And '79 into the early '80s, the M1 money supply was the report that people focused on. It really moved the markets.
When you got into the '80s a little more, the trade balance became a big number and everybody was looking at that. And what I would look at was, okay, what are the expectations for the trade number? What are the markets doing going into it? How do the markets react after? Immediately following it, then through the rest of the day.
And I'd start coming up with trade decisions based on that sort of plethora of information. When we started systematizing the approaches, we're always trying to break it down to what is the key return driver we're trying to capture? And we've come up with some pretty nice elegant little strategies that are looking at markets, action and reaction to these reports. To come up with trade decisions.
So, one that we had last year. I remember it was one where we bought deferred Euro/Dollar contracts; I think I talked about this in our monthly report based on the employment report. And we held that trade for a little over a month. And it turned out to be just perfectly timed counter-turn trade.
Which is, often, what can happen with these event strategies. You'll have, at that point, a downtrend going in the Euro/Dollar contract. And the employment report came out and that pretty much marked the low for, I think, still to today in that deferred contract.
So, it's looking at the market reactions of events for telling you how, essentially, you could classify this as a sentiment strategy because it's telling you how the short-term sentiment has shifted based on that report release.
Niels: I have two questions on this topic. One is, this is the trick to get into a trade, but how do you get out? What tends to trigger your exit for a trade like this?
Mike: So, each strategy has its own entry and exit triggers. And often, remember what we're trying to do is we're trying to capture the returns from a specific return driver. So, sometimes, there's a tail on it. Sometimes, there's market action that indicates that that influence is over. Sometimes, there's a subsequent report or something that'll trigger that difference. So, in the case of the first one we talked about was sentiment strategies. We're really just trying to capture that extreme sentiment. When that's wrung out of the market, we want to be out. So, you'll see that the money flows and the buy is going on. And at some point, that kind of comes back again and it subsides. That overall sentiment and enthusiasm that you had for being bullish on stocks.
And when that subsides, we're out. The market might still be selling off. And it might be selling off for another month, two months, a year. We don't care. Because the return driver we were capturing there was purely sentiment. We want to capture that in as pure a fashion as possible.
Niels: The other question I had was really more relating to the event strategy. Did you find, in today's world, where news gets around much quicker and more people have access compared to 10, 15, 20 years ago when you came up with some of these ideas. Have some of these strategies changed their effectiveness? Or do you find that they actually work just as well today as they did when you first started using them?
Mike: Right. Well, first of all, I had the exact same questions 20 years ago. Information is moving much more rapidly today. People are able to process things and get information more quickly.
Because we look at the 1990s, 1990, for example, to 1970. The exact same scenario that you have today from the 2010s to 1990. It's continued.
We haven't had, since I've been involved in training, a period where you've had slower information flow. It's always been increasing.
So, what we find is that, as I mentioned, in the early 1980s, late 1970s, you had M1 money supplies, like the key indicator. Well, the way we develop model isn't to say "Okay, specifically, this and this happens." It's looking to determine first and foremost, I guess, whether or not that indicator is still valid.
So, today, I could incorporate strategies using M1 money supply. And I can virtually guarantee you there would be no trading in them. Even though it's the same strategy that we're using for employment report, it's been fairly active the last few years.
Another strategy we do have in there is looking at CPI. I don't think the CPI strategy has made one trade since we re-launched the brand new Symphony trading in July, 2011.
So, the strategy, the concept is still valid. And it can be incorporated in the portfolio and included in the portfolio. It doesn't mean that it'll remain active. It's valid, but it may not be active. And that's up to the strategy to determine whether or not it should be. Same case with the ETF money flows. We could plug that in there, but if starts turning out that they're not a good measure of sentiment and indication of sentiment, the strategy will stop making trading decisions, based on that information.
Niels: You also have a theme called "Directional Arbitrage Strategy." How does that fit into this?
Mike: Directional Arbitrage is using what would be considered arbitrage techniques to take directional trades in markets. So, we're not actually specifically arbitraging...well, there are some opportunities there. But in the directional arbitrage, we're not specifically going after, say, an arbitrage between one market to another market. But we may be taking a directional position in one market based on some price action in another market.
If you look at it, a specific sense, with people using the carry trade in the currency markets, that's a form of an arbitrage that gives you an indication that there may be directional biases in certain markets based on interest rate differentials. So, there are some opportunities there. There are opportunities across yield curves. There are opportunities across just the entire price curve of commodities. Whether in backwardation or contango. That can yield directional opportunities based on the differences between contract months or specific market prices.
Niels: Sure. And the final one, Alpha Hedge strategy?
Mike: So, Alpha Hedge are momentum strategies. A lot of people would refer to them as trend following. We don't for one specific reason. For us, Alpha Hedge, it's an integral part of the portfolio that allows us to do a lot of the things we do in the portfolio that are non-directionally triggered.
So, if we end up with a fundamental strategy that's along Nat Gas for example and the Nat Gas market just continues to sell off, well, the Alpha Hedge Strategies will be short and they'll be offsetting that risk. And as long as each individual strategy has its own relevant return driver and a positive return expectancy over a long time--and I'm talking decades--then we feel comfortable including those in a portfolio.
In the short term, though, you have periods where you may just be fundamentally off on a direction in a market. The Alpha Hedge will never be off on a directional trend. It'll always be on the direction of the trend.
Now, what's interesting. And this goes back to a lot of talk I hear about trend followers having a difficult period and underperforming and losing money over the last few years and maybe that has to do with quantitative easing or a number of things. Our Alpha Hedge strategies have been profitable since we started trading in July of 2011. If we did nothing but trade trend following, essentially, we'd have a nice, positive return going. It's been enhanced from the other strategies in the portfolio. Which is the whole intent of broad strategy diversification.
But that's why I look at the performance in the industry over the last few years and sort of the explanations of why the industry hasn't performed as well as it did in the prior couple decades. And to me, it has less to do with the environment and probably more to do with how people have fine-tuned their strategies to have optimized on the past data, rather than allowing them to be pure and raw and sometimes poor performers over future periods.
It looks to me like a lot of the industry has probably just gotten a little too fine-tuned and curve fit on the past data. And then going into this period, ended up with negative performance.
Niels: Yeah. It's interesting. I think, certainly from my view, I do think the trends are a little bit different in the last few years. I do think they're a bit shorter, I do think that they're a bit more erratic. And I do think that the manipulation or government intervention has had its influence. And I think it's great that your trend following strategies are immune to this and continues to perform.
But certainly, from where I see it, I do think it's a bit different to what we saw prior to 2008. But I think more importantly, on that subject just to give my own view here, I think that a big part of whether you made money or whether you didn't make money in the last few years has really been the sector weight.
Because if you're a fully-diversified manager, there has been quite a lot of the sectors that really hasn't given you that many good opportunities. Whilst there's been one or two sectors, namely the equity markets and the bonds, where the good action has been.
So, I do think that there is some influence being in terms of how diversified are some of these managers. Because, of course, some managers have done well in the last few years. But when I see the names and I think about what I believe they do, I think they are pretty heavily exposed in fixed income and equities. But that's just my view sitting from the outside.
Mike: Maybe. Definitely, things are different, there's always different. And I guess the question is, are they differently different? And so, I look at our performance in momentum-based strategies. And our whole philosophies are to maintain balance across the portfolio, across the strategies and the markets. So, we've got 120, 130 markets that we're trading in.
And some of our Alpha Hedge strategies have lost money consistently over the last three years. But as a group, as a strategy type, they've made money. And they've done it in a reasonably diversified fashion. They've made some money in the currencies, they've made some money in interest rates and stock index, as you point out. But there's been some great trends in the metals. Gold has great trend up and great trend down during that period. Some of the agriculture markets have had just tremendous trends. That trend followers could have captured had they been, first, I guess, balanced across all those sectors and not dominating one sector or another.
And had they had, I guess, a very broad sort of parameter structure that they were using so that they weren't fine-tuned to the specific types of trends that we had maybe in the prior few years.
Niels: I think the other thing maybe, to just add to that. And that is in terms of these trend following strategies that you had in Alpha Hedge...my guess is that you might also be more on the longer term side compared to many CTAs who probably are more in what we would call the "medium term" in terms of how the models work.
Because I do believe, certainly, in the long run, I think long-term trend following is probably more profitable than medium-term trend following. The problem is that it also comes with a certain level of volatility and draw-down that many investors don't really like. And therefore, the pure trend followers have sort of shied away from being too long-term, even though they know that that's probably the right thing to do.
I would guess that you probably trade your Alpha strategies a little bit more on the longer-term side. I don't know whether that's a correct observation or not.
Mike: It's actually a real mix. I mean, we do have some that trade with holding periods that would be, say, a losing trade that's many months and a winning trade that could be a year or more.
But we also have rather short-term. That other winning trade might be two months long. For trend following, which is pretty short-term.
I think a lot of the problem does get to the fact that people don't like the trades, as you pointed out, which is, I think, a great comment. People don't like to trade some of the more difficult strategies of trade.
And I think one of the things that have helped make trend following traders successful over the last few decades is that it's not an easy strategy to trade. You've got to sit there and ride through multiple years, sometimes, of just terrible performance. And then wait for that quarter where it all comes back and you make money.
I think by maybe trying to just focus on trend following and smoothing returns in that style of trading, people have started curve fitting their strategies a little too much on the past data to make it look like they had solved the problem. When in reality, all they did was lower their predictability of performance.
There's nothing worse than doing that because then going forward, you don't know what you're going to get. And I think that's what surprised maybe a lot of people who I don't--
Ending: Ready to learn more about the world’s Top Traders? Go to TOPTRADERSUNPLUGGED.COM and signup to receive the full transcripts of the first ten episodes of the show, and visit the show notes where you can find useful links to other amazing resources. Thanks for listening and we'll see you on the next episode of Top Traders Unplugged.
Become An Insider
Subscribe for free and be the first to receive new and exclusive interviews with the world's top traders. As an insider we'll also send insightful bonus content direct to your inbox.Free Instant Access »
You might also like:
Date posted: 02 Jun 20141 comment