Niels
Marc, thank you so much for being with us today. I really appreciate you taking the time.
Marc
Thank you, Niels.
Niels
Now as I was preparing for our conversation today, I noticed a couple of interesting things that I'm sure we'll have a chance to discuss but just to give you some of my initial observations, it strikes me that you have quite a large number of different strategies that you offer both in the alpha strategy space, but also some alternative beta strategies, which to me suggests that you have diversified your business into more of a solution oriented firm rather than being more of a standard type of alternative investment company. The other thing that I noticed was that you began your career working for some of the very large institutions in the world, yet you chose the entrepreneurial path. I can imagine that that's quite a change in itself, and I'd love to hear more about that. Then I also noticed that you developed this risk aversion index, which I find very interesting and would certainly like to spend a bit of time discussing these things. I'm excited about all of these topics that we can talk about, but before we go into that, and before we find out where your company is today, I would really love for you to take us all the way back to the beginning of your story and tell us about what led you to take this path in life, and really feel free to go back as far as you want, Marc.
Marc
Sure, it's actually a fairly interesting path that led me to where we are today. Before I get started, again, thank you Niels for giving us this opportunity. All the points that you mentioned are definitely something that we can get into, and each one has its own particular story. Starting with the background: originally I'm from Beirut, Lebanon. I graduated from Cal Tech in 1992. At Cal Tech, I was studying neural networks and decision support systems. I was very immersed in that field, and definitely planning on getting my Ph.D. in the field and then continuing my research there. In my junior year at Cal Tech, I received a grant from NASA's Jet Propulsion Laboratory which is one of the 8 NASA centers in the US that Cal Tech runs. The grant was to basically study how you can position tanks in a battlefield to optimize your chances of winning based on the new mathematical technique called scenario analysis. The person who was doing the leading research in the field of scenario analysis was a Professor at the University of Wisconsin in Madison.
So, I traveled over a summer from Cal Tech to Wisconsin, to Madison, to do the research for them. We started getting to know each other, and he finds out I'm from Lebanon and says something like, you know you just came from a battlefield; do you really want to spend the whole term looking at tanks in a battlefield? Why don't you apply the grant to the optimization problem of something like stocks, bonds, and currencies. Mathematically it's the same project. Whether you are using green tanks and blue tanks and yellow tanks or stocks, bonds, and commodities or FX, it's really the same sort of mathematical work that needs to be done. So I said fine.
At that point, I really didn't know what a stock or and bond was. I never spent any time looking at any of this stuff. I remember that I had to go out and get a book that actually just gave me the definitions of these things. We didn't have Google back then. So I did my research and ended up writing a paper that was published at the University and went back to Cal Tech to finish my senior year. In my senior year I was very excited because I was finishing at the top of my class at Cal Tech and I had the highest job offer from Oracle to go and work there for a couple of years before going back and getting my PhD. Back then Larry Austin was still signing the offer letters, so I still have that letter from him. I was very excited. I had the highest job offer from anyone at Cal Tech, which I think at the time (this is 1991), was about like $45,000 or something. Sometime in the spring of my junior year I got a call from somebody calling from New York and said this is so and so, and I'm calling from Salomon Brothers and we'd like to invite you over for an interview. I was shocked at first, but it turns out that they had read my paper and wanted to interview me. So it's fine - a free trip to New York.
I went to New York with absolutely no intention...or no understanding of what the job was. Anyway to make a long story short, I fell in love with the whole concept of trading and money management, at least the way I understood it from a couple of days of interviewing. At the end of those couple of days, the MD pulled me into the office and says something like, we normally don't do this, but everyone likes you, and we're going to make you an offer on the spot which was quite a bit more than what Oracle had offered. So that was the beginning of how I got into this whole thing. Quickly, very shortly after I got to Salomon because I was hired in research, I realized that I would much rather be on the trading side than the research side, so I left after about a year and changed to joint an ex-Salomon trader that had started a hedge fund back in the early 1990s.
He hired me basically to develop trading systems for him and trade currency options. So I did that until about the end of 1994, the beginning of 1995, when I was hired by UBS to start and run the global group in exotic derivatives on foreign exchange, which essentially were every derivative that goes beyond a simple call or a put. I did that until 1997 when UBS and SBC merged at which point I was promoted to run the combined banks, prop trading groups in London and the Americas. Prior to that, in setting up the global exotic derivatives group, I had set it up in New York initially, then moved to London, set it up there, then moved to Tokyo and set it up there, globalized it and ran it from Tokyo until I was promoted to run the prop trading groups. I moved then back to London, and that was around 1998 if you remember the turmoil in the markets in 1998, and the bank had lost a lot of money which really cemented my decision to leave. I was thinking about leaving before that, but seeing the effect on my own compensation and limits and things like that from things that were happening in other parts of the bank. Basically, it cemented my view that I think on Wall Street you have maybe 5% to 10% top producers that pay for probably 80% or 90% of mediocre or below average people. If you're vain enough to think of yourself as that top 10% or 5% then really, by working on Wall Street, you're subsidizing everybody else. So I decided to leave and start my own firm. That was the beginning of me into the hedge fund role.
Niels
Did you know what you were going to do at that stage? Did you know from day one this is exactly how I'm going to do it, or was it more like let me...
Marc
Absolutely. Prior to doing that I had run quite a substantial global prop trading group at UBS. I had about 18 traders, each one a senior in their own right, that reported to me. In my own trading at UBS...actually starting from my days at the hedge fund in 1993, 1994 time, going to the period where I ran exotic derivatives at UBS, which also included a large portion of prop trading, which really was the reason why they promoted me to run the prop trading groups afterwards.
Throughout this whole process I had developed and used systematic models to trade in the market. I've always had the philosophy that if you look at basically any successful discretionary trader, they don't wake up randomly one day and decide to put on a position because they had a dream...or...maybe some do, but usually not the successful ones. Usually it's a thought process that leads you to put on a trade. You look for certain things to happen, either technically or fundamentally, when these conditions are met you put a trade on. When you put that trade on, you know ahead of time how much you are willing to risk on that trade, what position size you need to have for that risk, and where the stop on it should be, and where you take profit on it. When you list all these things one after the other, it's no different than the algorithms that go into creating a trading model. So I think with a lot of the very successful traders, if they had the mathematical and computer background they can very easily turn a lot of that thought process into an algorithm, and once you have it as an algorithm, then it's very easy to program it. At that point a computer is a much more efficient tool to execute your own view for a variety of reasons, stretching from taking the emotions out of it, to efficiency of execution, to giving you the flexibility of trading many more markets that you cannot follow yourself, but a computer can do it in a much easier way and so on. So I always had a very systematic approach from my days at UBS, so on day one, when I left and went on my own. I had a suite of ideas that I knew I wanted to start with and work on. Also when I started I had a partner at the time and he also came with his own ideas and basically our product was a combination of a lot of brainstorming through the background that I brought and the background that he brought.
Niels
Let me just ask you a spur of the moment question here. When you explain why people, if they had the knowledge of programming or putting together systems rather than trying to just do it as a discretionary trader, even though they might use internal rules in their own mind, which obviously is a very logical way of doing things. Why do you think it is so difficult for systematic traders to explain to investors that they shouldn't be worried about the fact that we use computers to do this and in fact that isn't a black box?
Marc
Well, that's a very good question. I don't think it's difficult for a systematic trader to explain their strategy. I think there are a lot of systematic traders that deliberately try to be very opaque about what they do for a variety of reasons: one of them is they don't want to give away the "secret sauce" of what they are doing, and I think to a large extend in the old school trend following world, primarily it's to keep how simple it is from getting out and making people realize why on earth am I paying $2 and $20 for that?
Niels
That's from the manager's side, Marc. On the other hand, I would argue that investors, they have a difficulty with accepting these systems, and they try to make them, in some ways, more complicated by calling them a black box. So I think that you are right about the manager side. That's probably one reason, but I think also the investor side in a bit to blame here, that they don't embrace the fact that technology... if you board a plane, you know that the pilots not going to sit there for 12 hours flying the plane manually. He's going to use an autopilot, and I think most people would prefer he uses an autopilot.
Marc
I think that is changing over time. I think that there is a wider acceptance of quantitative strategies now than there were before. To use your example of the plane, passengers know that the pilot is relying on a lot of analytics and tools to fly that plane, but the most important thing is that they know there is a pilot; that if something goes wrong the pilot can turn off whatever machines are there and just fly the plane. I think with systematic strategies and the way that a majority of them are explained and portrayed, that people are fond of saying we're 100% systematic, and when you say 100% of anything it tends to make people nervous.
Systematic works maybe 95% of the time, but sometimes you have some events that can happen that can hit the world, that can be either endogenous type of events or exogenous type events that are unforeseen. Now again, 99% of the time you want to follow your system because that's basically the reason you build them. The world is usually not that different. However, the way that we describe our strategy is we say we are about 95% systematic and 5% discretionary. The reason that we do so is, again, to use your analogy of the plane and the pilot, investors like to know that, yes you might have the highest tech equipment, but they want a pilot in charge. The example that I give there usually is that look, we're systematic 95+% of the time, but if I'm sitting in my office and my models are screaming to buy stocks and I see a plane heading into a building, I might choose to turn off that model at that point. It's just a common sense approach. The other point that I think confused a lot of investors... let's take an example. If you look at traditional trend following, the Turtle approach to trend following: slap a couple of moving averages together - a short-term one and a long term one, apply it to 50 of your favorite markets across 4 or 5 different sectors, and you've got yourself a hedge fund. Historically, from the 1970s until early 2000, that's pretty much what all CTAs did. You can check that by looking at the correlation and the return of CTAs during that period to the very simple model that I just gave you. It's about 80% and it's a very consistent 80%.
When you take a trend follower... it's a very simple tool that I just gave you. A lot of people might want to get into the space, but they don't really want to question why trends happen, they just know that this exists and that if you do this you are going to make money. They start a fund, and they go out and start trying to get investors. Investors, also at the same time, they can sit there and say well, this black box, they don't even know how it works, but the explanation that the manager gives at that point is look, we know that trends happen in the world and our models take advantage of trends, and they make money when there are trends, and historically we've made a lot of money, and so on. In that kind of process from the manager side as well as from the client side, there is a lot of uncertainty. So the client will say, why do trends happen? How do I know if they happen before they are going to happen again?
When you're having that conversation with a manager that is very orthodox in their approach, look, I don't know why trends happen, I don't care why they happen, they happen, and that's how we take advantage of them, which is really the way a lot of people answered these questions. It creates lack of understanding and people... if you're not comfortable with an investment, it's going to be very difficult for you to get to invest. What ended up happening for a lot of CTA investors is that at the initial pitch they passed, then they saw CTAs making a lot of money. They'd go through a period where they would make 40%, 50% for a year or two. So they get dragged into investing, kicking and screaming, at which point their investing at the top in these strategies, which, by definition, are very cyclical, only to see that investment plummet in value because they pretty much bought at the top of the cycle. Given that they didn't understand exactly what they were buying, they don't have a conviction to hold as it's coming down. They basically get out at the bottom, and we saw this cycle happen so many times with investors who would always buy the trend followers at the top and sell it at the bottom. My cynical side says that, over time, investors as a group have lost money investing in CTAs just because of the timing.
Using that same approach, the way I would explain it, is that first of all looking at CTAs, even though we trade 50, 60 different markets, for any decent sized CTA our bread and butter comes from the financials. You're talking global fixed income. You're talking currencies. You're talking equities. Commodities are a nice diversifier, but that's pretty much what it is, a diversifier. If you look at any economic region: US, Europe, Japan, Asia, every one of those regions has its own business cycle. It has a certain either growth or declining economy and the business cycle around, let's say it's a growing economy, so you have an upwards sloping line, and the business cycle oscillates around that line.
In each one of those economic regions, the job of the central bank is to minimize the amplitude of that business cycle relative to the slope of the line. If a central bank is hugely successful, then they keep that economy on that slope of the growth - there is no volatility. Obviously, no one is that successful. What happens is as the business cycle starts going quite a bit above that slope of the line, central banks come in and try to increase interest rates to slow things down. When you are at the bottom, below your growth trend, they come and lower interest rates to essentially bring you back up. When you look at the way a central bank changes interest rate policy, it's not plus 50, minus 25, plus 75, minus 10, plus 5, minus 10, plus 5, that's not how they do it. They go +, +, +, +, stop. They stop for a while. Then they go -, -, -, stop. Then they stop for a while.
Now think of what happens whenever a central bank embarks on these cycles. Every time that they start with a plus, there is a wave that starts and propagates through all the financial markets. Changes in interest rate policies have effects on currency prices; they have effects on stock prices; they effect on bond prices. Now you start at the beginning of a trend. When they come again and they do another 25 or 50 and so on, now your trend is getting momentum, then it goes again and now the trend is getting more momentum, and then they stop and that's when your trend levels off and, in a perfect world, that's when your position reverses and you take it on the other side. Unfortunately, that would have been too easy if that's exactly how things work. What happens is for every market, every market moves for two reasons. It moves for alpha reasons, and it moves for beta reason. Alpha reason being conditions that are only relating to that market, and beta reason are the overall macro condition or interest rate policy and so on. What I described in the action of the central bank, that's really macro; that's beta. If in the absence of any alpha reason, then we'll get perfect trends and everything will work fine, except that things are not that easy. Every market again has its own alpha reason.
You get the best trends when both the beta and the alpha are pointing in the right direction. Let's think of what would make a perfect trend. Let's say you're looking at crude oil. We know that in a very hot economy an expanding economy there is a natural pressure on crude oil prices to go up as there is more business activity, more manufacturing, so that's an upwards pressure on crude oil prices. Let's say that in the middle of that happening you get a giant explosion in Saudi Arabia that causes crude oil to spike for a local reason - for supply reasons...then you get a real turbo boost to that move that was happening in crude oil and you get the perfect trend at that point, because you have both alpha and beta kicking in the same direction. When you get really bad trends, is when they are completely opposite, where the alpha and the beta are fighting with each other and you get these choppy markets where you are constantly getting stopped out, and back into position, and stopped out.
For a lot of people trend following is a very opaque strategy, but really it's not. It's not alchemy, the concept for it is rooted in simple economics and finance, and I think once, in my experience at least, once I've explained it to investors, they generally agree and then they get it. Now the next thing that I saw is that given how simple this strategy is it's really not a 2 and 20 strategy. People have gotten away with too much over the years charging $2 and $20 for this. What happened, in the early 2000s, I spent a lot of time doing research on how to extract more alpha from long term trend following and I kept coming to the conclusion that the most efficient way to get more alpha from trend following is to essentially not charge $2 and $20. I wrote a paper in early 2000s how using very simple off the shelf trend following techniques like moving averages, or breakouts, or things like that, and some other very simple portfolio techniques, you can build a very reasonable and high expectancy trend following strategy, and that if you don't charge $2 and $20 for it, you end up having results that beat 80%, 90% of the trend followers out there. I wrote the paper in the early 2000s, like I said, and published it and in 2004, we were approached by the endowment of a very large Ivy League university that agreed with our premise and asked us to start a fund for them based on that. Since 2004 we have had a fund that's basically a very well diversified trend following strategy that charges a flat 1% management fee and no incentive fee and since we've launched it, it has had 70%, 80% very stable correlation to the trend following indices, and that ivy league endowment is still the biggest investor there.
Niels
That's a very interesting observation, and I think, as I mentioned in the early part, it's one of the things that I'd love to discuss with you because you were certainly one of the early adopters on this. But here's my question about it, to some extent I feel that managers will have this alternative beta, or this CTA replicator to give investors very similar returns to these CTA indices, but at a lower price, and that's fine. There's nothing wrong with that. However, I feel that the CTA indices themselves, because the underlying CTAs have changed, and therefore I'm not so sure that these replicators today are very... I say it a little bit, but I'm not so sure that they track the CTA indices as well as they used to, simply for the fact that I think CTAs have changed. I think that's another thing that you have made observations about, and that is that managers today have migrated from being pure trend followers maybe to doing other types of strategies, in particular in the last few years, to compensate for maybe lack of trends in the usual sense. I don't know whether you think that's an issue, and maybe do these alternative beta strategies or CTA replicators, do they themselves actually need to be evolving and maybe you do them differently than you 10 years ago when you started?
Marc
Look, absolutely. That's a very astute observation. What happened is if you look at trend follower correlations from the dawn of trend following, until the early 2000s, or early to mid-2000s, they were a very homogeneous group. You'd be hard pressed to find CTAs with less than 70% correlation, and in some cases it was more than 80% and solid. The reason for that is that pretty much all of them followed the same kind of Turtle methodology of long term trend following, with all the good and the bad that that strategy entailed. Starting in mid-2000s, there were a couple of changes that some CTAs started doing that led to what I call the split, in the CTA space, between what I call the old school and the new age CTAs.
The old school is the ones that stuck to the Turtle approach. The new age guys changed things in a couple of ways. One of them is fairly technical, I'm not sure how much we want to go into it, but essentially it has something to do with the fact that basically, historically CTAs sized their position in any given market in a relationship that is inversely proportional to the volatility of the market that you traded, meaning that you want to take the same risk per trade, you want to risk the same amount in your crude oil trade as you did in your Euro dollar trade, but these two markets have vastly different volatilities, so you can't take the same dollar position size in them. What you do is you adjust it for volatility and then you take a vol adjusted position. What happened is the old school way of doing this is you had your position, you stuck to your position until that trend reversed and then you exited and you reversed and do the opposite side with the new position size at that point. One of the biggest criticisms for CTAs, historically, has been that CTAs have high volatility, large profit giveback, big potential drawdowns.
What eventually became the new school guys started thinking about the problem, they correctly realized that the volatility at the beginning of a trend is fairly low. As that trend matures and starts showing signs of weakness and reversion, volatility goes up significantly. Essentially they said, look, it doesn't make sense that we are taking a position at the beginning of a trend when volatility is low and therefore we would have relatively large position size, and holding this same position as the volatility in this trend itself is changing, which is causing us to have the higher vol on the portfolio. So they said, what if we do more frequent sampling of the volatility and adjust our positions accordingly? The thinking there is that you'll be taking profit on your position as the trend is developing and then by the time volatility spikes, and the trend reverses, you go through that reversal with a very small position and therefore minimize the effect on your portfolio, which is great except that it leaves one gap in the thinking. It automatically assumes that an increase in volatility is a precursor to a trend reversal and not a trend continuation.
If that sharp rise in volatility actually is a precursor to a trend continuation, then the risk to them is that they are going to basically sit out that trade, which probably will be at a time when CTAs are the most needed which is when volatility is really high, but then going even higher. An example of that is what happened in 2008. In 2008, if you remember, volatilities spiked quite a bit going into the summer... the end of the summer. By then a lot of these new school CTAs had pretty much exited a lot of their position because of the volatility spike. Then what happened is we had Lehman Brothers, and then really volatility exploded at that point. What happened is all these guys ended up missing a once in a 10, 20 year opportunity in the CTA space, and more importantly they did not deliver the protection to their investors that is part and parcel of what investors expect from a trend following investment.
I won't go into which manager is new school vs. old school, but that's one way to look at it. The other big change that they've made is that they, again, correctly realized that trend following has very distinct characteristics that work in certain market conditions, which are really in many ways negatively correlated to a lot of other strategies, which is really, when you think about it, is why the pitch that we make to investors of why they should go into trend following strategies, because it's low correlation or negative correlation at some point. But these guys said, well, if we include in what we're doing, a negative correlation strategy, we're going to improve our risk-adjusted return. To give you a very simple example, if you take a simple Turtle model - a moving average based Turtle model, you add to it something as simple as a long S&P position - maybe not purely long S&P but long risk, you can improve the risk-adjusted return of what you are doing, except that you'll be really deviating from investors expectation of what you can do in different market conditions. At that point when volatility spikes in the market, and risk aversion spikes, you're not going to be delivering the protection that they thought they were getting with you. What we know is the two large changes that we saw in CTAs starting in the mid 2000s is that you had one group that stayed true to the Turtle approach. Those are the guys that did really well in 2008, and then you had another group that did two significant changes, which is the position size adjustment based on volatility, as well as adding totally different strategies that are long risk with negative correlation profile to trend following into their own fund, and these are the guys that have done very well in the last 4 or 5 years.
The question is, are investors investing in CTAs as just an absolute return strategy, in which case they would go with the new age guys, but in which case I would argue that if you are investing in CTAs as an absolute return strategy then really you should go look elsewhere because there are other hedge fund strategies that do better on an absolute return basis. Or, are investors investing in CTAs as a portfolio tool that works very well against the rest of what they have in the portfolio? In which case CTAs really have to stick to what investors expectation is of their performance in various market conditions. To give you a practical example, when we noticed (it's a long answer to tie back to your question) whether CTAs are tracking or not or whether trackers should change or not. When we noticed this a few years ago, we went back to the ivy league endowment and presented the numbers and said look, this is what we think is happening, we think that you can have better risk-adjusted numbers from us, overall, if we add a long risk strategy to what we do, which we already had because in our other fund Conquest Macro, we have a very successful long risk strategy. It's a very easy thing for us to just turn it on for you in this one, and then the fund will switch from being a Turtle approach to being a new age approach. It's really not rocket science. It wasn't rocket science to do Turtles, and it's still not rocket science to do the new age approach. The answer from the endowment was a resounding no. They wanted us to stick to the Turtle approach precisely because they used us as a portfolio tool, and they said they have plenty of long risk in their portfolio. They think they are a better judge of what long risk strategy to be in than us with a genetic long risk strategy, so they wanted to retain that flexibility in what they do, but they wanted us to continue to deliver the characteristics of what their expectation is from traditional trend following, which is fine.
You mentioned that we have a lot of products. The reason we have a lot of products is... look, the way I think of portfolio construction is it's a two-step process. A portfolio is made of individual models. I think that we are very good at creating different models that do different things in various market conditions. Maybe it's from having three young children and spending a lot of time with Legos, but I think of these models as Lego pieces. You can take those Lego pieces and put them together to make a plane, then you can take it apart and put them together in a different way to make a submarine, or a car, or a house, as long as you have good solid Lego pieces you can build them to create any profile that you want, which is exactly what we do on a portfolio basis. I think we are very good at making these Lego pieces. They're very solid. They're good quality pieces, but the construction of them is really in how we want the portfolio to look: what characteristics, all that other stuff. If I think of our generic Turtle approach as one Lego piece - a big Lego piece, then I think of our long risk as another Lego piece. Each one individually can give you a very different risk profile. By putting them together now, we can go from a Turtle based trend following strategy to a new age absolute return trend following strategy.
Niels
Let me just say to the listeners who are listening, and may not know what we refer to when we say a Turtle strategy, if you go to episode 13 and 14, that's actually with Jerry Parker and he's probably the most successful Turtle, so if you want to hear the whole Turtle story go to episode 13 and 14.
I want to ask you, Marc, to go back and then take us from where you started realizing this and then how Conquest evolved as you were adding these strategies, what was your thinking behind adding this particular one as a separate offering, and so on and so forth, and before we leave your story as a whole, I think we need to go from the inception of Conquest to where we are today, before we dive into more of the specifics. I'd like for you to tell us that story, how the product range evolved in the last 10 years.
Marc
Absolutely. Look, all that stuff that we spoke about so far that was just almost like a side project for us. That has never been our bread and butter. That's something that started out of an intellectual exercise, which we had the luxury and privilege of being the first one to test it and bring it to market, and now you can see how many different people are adopting that approach.
Historically our bread and butter has been Conquest Macro. When I first started, I'd always wanted to have... let me backtrack for a second. If you look at pretty much over 90% of investment strategies, they all tend to do really well in low risk, low volatility environments, and they get slaughtered in high vol environments. Periods like 1998, anytime there was shock in the market, you see a huge, very universal suffering from pretty much every investment strategy. The reason for that is, in a very, very simplified way, active investments (whether it's a hedge fund type and so on) make money by buying the risky asset and selling the less risky asset against it and benefiting from it. Think of this as in the stock market being long, kind of high bidder versus short, low bidder. In fixed income, you can do sort of the borrow short, lend long. There are a variety of ways of expressing this in pretty much every single market.
While those strategies, whether you do it in fixed income or whether you do it in equities and so on, while they have low correlation, most of the time when the environment is not very risk averse, their correlation goes to 1 on the downside, whenever risk aversion rises. The reason for that is that each one of those strategies that is benefiting from buying the risky asset and selling the less risky asset needs one very crucial thing for it to work, which is liquidity. Liquidity is like the oxygen for these strategies. When there is oxygen each one is doing its different thing, low correlation, everyone is happy. However, when risk aversion rises in the markets one of the first casualties of rising risk aversion is liquidity. Liquidity dries up significantly, so what happens is that suddenly all these strategies that look to be uncorrelated, where in reality they did have one common risk factor which is liquidity, but they start all losing money at the same time. That's when investors start scratching their head saying, what happened? We thought we were diversified, why are they all losing money at the same time? The reason is because they pretty much all long a lot of liquidity, and when that disappears they lose at the same time. We've seen that happen time, and time again.
My thinking about Conquest Macro, from the beginning, was that I wanted a product that would do well in a risk averse period, because it's needed. Like I said, over 90% of strategies out there don't have that profile. Historically, the way people hedged some of that rising risk aversion is they could have allocated to short sellers. The problem with short sellers is that stocks go up over time, and it's pretty much a negative expectancy strategy. Very few short sellers survive for decent parts of time. Markets can go up, as we saw in the last few years, for a significant period of time. I mean Keynes has never been more right than now. Markets can go much longer and be much more irrational than it can be solvent. The problem with short sellers is when you have that type of strategy, given that markets can go for long stretches of time, let's say going up in this case. Investors find it very difficult to hold onto a short seller over that period of time. Just every month, losing, and losing, and losing. So they end up redeeming out of their hedge, probably just at the time where they needed it the most. So short selling is not sort of an ideal hedge to portfolios.
The second strategy that people allocated to, to hedge some of the risk aversion risk, was pure long volatility strategies - just buy vol. That was great in the 80s and 90s when, I believe vols were mispriced. That was pretty much my career until I went off... I traded volatility, and it was mispriced. However, in the 90s and early 2000s we made significant headways in, not just understanding volatility, and the jump in vol, and second derivatives, and so on, but we also with computers and all these models... whatever edge that we had in the models that we had built to give us a better mousetrap for measuring volatility, by the end of the 90s early 2000 everyone had them. As a result, I think volatility became much more correctly priced. If you assume that volatility is correctly priced, and your strategy is based on just buying that, then really you have zero expectancy, but if you factor in transaction costs and things like that you end up with a negative expectancy. Again, when you look at investment in pure long vol strategy (I don't know if any of them still exist) but they had that characteristic of bleeding, losing, losing, losing over significant periods of time and then, generally, investors redeemed before vol spiked and they made a lot of money.
The third one that people allocated to was trend following - long term trend following. There is a big misconception out there is that investors believe that long term trend following is a long volatility strategy - it's not. I actually wrote a paper on that. The paper was published and it won the best paper award by Institutional Investor. I think it was published in the Journal of Alternative Investment. Long term trend followers are long the second derivative of volatility. If a volatility event happens, and that event is sufficiently large, then it can kick into place trends that can go for quite a distance and over time for trend followers to really benefit from them - think of what happened in 2008. An example where trend following made money, but it could have just as easily have lost money, let's say, go back to September 11, 2001, when that tragedy happened, trend followers made money, but the reason that they made money is that when September 11th happened, obviously stocks went down and bonds went up, just to simplify things, however the markets, the trends in stocks and bonds, had turned a few months before into short stock and long bonds, so when you got that exogenous event, which was September 11, it basically, again, using our example of alpha and beta. The beta was correctly in place, and that was an alpha event that really pushed everything in one direction. If September 11, 2001 had happened a year before, when long-term trend followers were long stocks and short bonds, they would have gotten destroyed.
The point that I am trying to make is that long term trend following is not long short term vol, but it can be very long term vol, or the secondary derivative of vol. However, that doesn't hold true in another piece of research that we put out and was published in a book by Risk books in London in 2008. What we proved is that when you come down in the trend following spectrum, from the long term trend following to the very short term trend following, starting one month and under, essentially trend following strategies that are very short in duration start acting very much like pure long volatility strategies. I'm sort of simplifying a little bit, but the whole paper is available, I think you might be able to find it online or through Risk books in their book. Looking at what's available for an investor as protection against short-term risk in volatility spiking, or risk aversion, or so on, really there wasn't any good product out there. That was the idea why, with Conquest Macro, we wanted to have a product that essentially had two mandates. The first one, and very important one, is to be an absolute return strategy, because if you are not a good absolute return strategy, investors will not hold you long enough to fulfill your promise, basically. The second mandate was to try and generate the bulk of that absolute return in what we call periods of risk aversion, which are periods of high volatility and high volatility of volatility.
Our expectation for the fund is to return somewhere between 5% and 10% in risk seeking periods - in low vol environment periods, and return over 30% in risk averse periods. In our actual trading, since we started, we have annualized over 30% in those risk averse periods - we have checked that box. It's on the risk seeking side where we've had to do a lot of work to improve our strategies. Depending on when you look at our track record, and so we've had many improvements that have happened over the years, I would say that our actual track record is probably flat to slightly positive in risk seeking periods, which by itself is still a very significant improvement over short sellers and long vol strategies, but looking at our track record since we've made a lot of the improvement in our risk seeking performance, that's tracking closer to 5% to 10% annualized. If I think of our return, based on our risk index and our analysis of the risk environment, we model the world to be roughly about 70% of the time risk seeking - low vol, and about 30% of the time risk averse. Just a quick back of the envelope calculation, if the 70% of the time that you are risk seeking you are going to make 5% to 10%, call it 7% and change average. You are going to make roughly 5% in that period. Then in the 30% of the time that you are going to make 30% from your risk averse performance you are going to make another 9%ish, so I think we're at 14%, 15% type strategy over time, but with a very, very important portfolio benefit in the way that we deliver those returns.
Niels
True. Here's a question. That, obviously, is very, very interesting that trying to design a program that not only gives investors the bulk of the return when they most need it but actually also can make returns when they don't really need it, but obviously it's an absolute return strategy and therefore it's nice to have an absolute return through those periods. In a sense, you could say it's the best of both worlds. With traditional trend following my observation is that it can't deliver both. There's simply going to be a period where it will lose money. Of course, the question is then, how much will you lose, and so on, and so forth. It's kind of universally accepted that trend following can't be the best of both worlds. It sounds to me like a very tall order to try and do both. How do you achieve that?
Marc
It's a lot of very hard work and a lot of trial and error. In a way, 15 years on, I think that we have the best product that we've ever had and I think it's both an ability to design individual models that over time will deliver exactly what... not just the return but the risk profile that you expect. I don't understand how anyone can promise a certain return profile because really returns are a function of the market it gives you and no one really know ahead of time what the market will give you. Our biggest effect comes on risk control. If you build up a certain risk profile, that risk profile is going to be associated with different return and different market conditions. We start with that concept and what we do is we have, I think our risk index - we have one of the biggest benefits that we've reaped in the portfolio was when we came up with the idea of the risk index. Again, at that time very few people thought of risk indices or even had a risk index.
Niels
Why did you want to develop that?
Marc
It was very simple. In trying to have the best product that responded very well in periods of higher volatility, we wanted to have a quantifiable way to go and measure what we defined as high volatility versus low volatility.
Niels
Was it more to visualize to people, or was it something that you use in the program?
Marc
No, no, no, we use it on actual programs. The premise before we started building the risk index was that different strategies behave differently in different risk environments. If we had a proper way to measure the risk environment through a risk index, we can use that to affect our risk allocation methodology across the different strategies within our portfolio and therefore have a better portfolio - a simple premise. The first step was to go out... because historically when people looked at risk, they looked at it in individual pockets, meaning either through the prism of the VIX, but really VIX is only equities, or through effects on volatility for people who traded, effects on options, or swap spreads for people who did more fixed income, but people were not looking at a much more comprehensive view of risk. What we basically observed is that one of the benefits of having a risk index is it really pushes your tentacles across all the different parts of the capital markets. On a daily basis, you're able to measure the temperature of pretty much all the different areas from liquidity risks to credit risk... the way we define it is we said what are the different risks that people look at? You have liquidity risk. You have credit risk. You have emerging market risk. You have equity risk. You have foreign exchange risk. When you have a much more comprehensive view of these risks, it gives you a much better way of assessing market vulnerabilities let's say, because a lot of the time, depending on where the risk aversion ends up coming from, you start seeing signs of that in that particular dark corner of the market way before anybody else starts feeling it or seeing it. So it gives you some preparation time to go and think about what you want to do and how you want to do it.
Once we built our risk index, it turned out that our intuition was spot on, which is that there is very strong statistical evidence that different strategies consistently do very different things in very different risk environments and that there was a certain level of autocorrelation in the risk index that allowed you to use some of that information. In periods of switching from one risk environment to the other, there is the expected noise in the data around those points, which is not something insurmountable. You can very easily filter out that noise, but once you got into the body of the risk environment, your probability of staying in it was much higher than the probability of reversing, until you had an event that caused it to reverse, and again you go back to that noise and so on. What we found is that using an asset allocation strategy greatly benefits our return because it allowed us to put our resources where they had the highest expectancy given the risk environment.
Given that we thought we had a very good mouse trap for figuring when the risk environment... what it is, what it's telling us, and when it's changing. As a matter of fact, we put these changes in our portfolio in March of 2005, and measuring the effect of those changes for the next few years, it ended up improving our risk-adjusted returns by about 30% - just that one change alone. Again, once we were able to see the benefit from allocating to different strategies that have different risk profiles and different correlation profiles, using our risk index, it spurred us into going out and creating a lot more strategies - sort of more Lego pieces, let's say, that do different things in different risk environments. When we first came up with a risk index, we had three sub-strategies within Conquest Macro: we had a long vol strategy, we had a short term trend following strategy, and we had a non-trend and a counter-trend strategy. Using the risk index, what we found is that, in risk averse periods our best performer was our long vol component, which makes perfect sense. However what we also found out is that in risk seeking periods short term trend following ends up doing significantly better than long vol, because even in very long volatility environments you still have some short term trends that you can take advantage of. What we also found, and this again caused me to scratch my head a little bit at first, when I put in our risk bucket that has in it both counter-trend and non-trend strategies, on the surface it showed no effect on the risk environment on that sub-strategy, which puzzled me for a little bit until we dug down into the individual model level. Each one of those sub-strategies has many different models. What we found is that pretty much 1/2 of the models were effected in one way, and 1/2 of them were effected in another way, and the net effect was no change in that risk bucket. This made sense, because if you look at our long vol sub-strategy, individual models that have correlations, let's say between teens to 50, 60 or so; if you look at the short term trend following risk bucket, again correlation is probably 40, 50, 60 and so on; in both of these are all positive correlations, therefore you really saw a clear effect from the change in the risk environment. When you look at our non-trend and counter-trend strategy it has both strategies that do really well in risk seeking periods, and strategies that do well in risk averse periods, therefore it has, as components, negatively correlated strategies to each other, therefore the effect from the risk environment was neutralized by having the component being negatively correlated to each other.
As I said, we made the changes to go from static risk allocation to dynamic risk allocation based on the risk environment. We improved our returns by about 30%. However, the way I viewed that improvement was more like the sort of improvement you get from cost cutting. It was just a better rearrangement of the deck chairs. I still wanted a component that actually added positive returns in risk seeking periods. That's where we went out and built our force of strategy which is what we call long risk, which is a strategy that essentially has about 80% correlation to hedge funds; that uses only futures and FX; that as a standalone could be a very good product by itself, because, gain it gives you a sharp of about 1, 80% correlation to hedge funds using only futures and FX, which as a standalone strategy would qualify it as a CTA, so you can be technically invested like a hedge fund, but getting a tax treatment of the CTA, which is, by the way, in our product offering we also offer it separately as a standalone.
It was hugely helpful to the Conquest Macro portfolio, because now we had one more risk bucket that we can allocate to in the risk seeking period that really would give us pure, simple, absolute return on the positive side. The rest of the evolution of the models within Conquest Macro is historically we started off with only price based models. After about 7, 8, 9 years we progressed into an area that we call Quant Macro, which is models that take a combination of two things: fundamental data, as well as technical data. That, again, allowed us to have more leeway on how to allocate within the different risk environments and so on.
Niels
So you have these four strategies within the Macro program, are you able to visualize and talk just briefly about how each of them implement what they do, just to make it simplified a little bit...
Ending
Ready to learn more about the world's top traders? Go to TOPTRADERSUNPLUGGED.COM and sign up 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.