| Main |
2. Leda Braga Systematica Investments Systematica Investments CEO Leda Braga has always enjoyed solving thorny problems. As a quantitative analyst on the derivatives research team at J.P. Morgan in London during the mid- to late ’90s, the Brazilian-born Braga built sophisticated models to price exotic interest rate products and hybrid instruments. In 2001 she was recruited by BlueCrest Capital Management co-founders Michael Platt and William Reeves to create pricing tools and analytics while looking for new trading opportunities. It was there that Braga, a self-described geek who has a Ph.D. in mechanical engineering from Imperial College London, came up with the idea for BlueTrend, a systematic, trend-following hedge fund launched by BlueCrest in April 2004. The fund, which trades futures in more than 150 markets around the world, has delivered an annualized return of 12 percent and has had only one down year, 2013.
Late last year BlueCrest spun out Braga’s group of 100 traders, researchers and technologists as Systematica. Although BlueCrest retains a significant economic interest in the new firm and its existing funds, BlueTrend and BlueMatrix, 49-year-old Braga is majority owner of the Geneva-based operation. Innovation will be important for her as she plots the course for $9.2 billion-in-assets Systematica, which also has offices in Jersey, London, New York and Singapore. Her team must continue to refine its quantitative investment models, as well as design lower-fee products with greater transparency to meet the demands of current and future hedge fund investors.
Institutional Investor’s Alpha: When you were growing up, did you have any intention of going into finance?
Braga: No, not at all. I started a career in academia. I was a lecturer at Imperial College between 1990 and 1993. It was going well. I had research contracts, I was supervising Ph.D. students, and I was teaching a lot. The problem was that it was the Thatcher years in the U.K. You could get funding, but it was hard work, and at the same time there was this whole City of London phenomenon happening around us. Then you really ask what it is to be an engineer. If you are a pure scientist, you’re trying to answer questions and improve knowledge and develop a framework for knowledge of a particular aspect of nature or phenomenon, but if you’re an engineer, you’re trying to solve people’s problems. I kind of thought there was this whole movement around derivatives that required people with better numerical skills, so I applied to some of the banks, and J.P. Morgan moved very quickly.
“When we started in the hedge fund world, it wasn’t obvious that constantly trying to improve your model was wanted or required.” — Leda Braga What was it like as a member of the derivatives research team at J.P. Morgan during the ’90s?
Derivatives research was a small team. There were about 12 of us for all asset classes, for all offices. We were in charge of modeling decisions, and those modeling decisions could be very complex, like in the case of exotic claims that require heavy models and equations in bits and pieces, but they could be simpler things as well.
I did derivatives research for many years. It was very good grounds for understanding the whole financial system, and I learned a lot. There was this feeling, though, that derivatives research people, we were very, very technical, but in the financial world it pays to be a little more of a generalist. We were the geeks. We never went anywhere. We just did more and more and more derivatives research.
When you joined BlueCrest in 2001 and startedto research systematic trading, how innovative were the strategies being deployed by other managers?
Renaissance Technologies was the big name that we looked up to, and we still do, so we knew that there were these guys that deployed science in their investment process, for sure. They were deemed to have 90 scientists in a room on Long Island somewhere, so these guys were obviously innovating. It was clear from the output — the Medallion Fund was already famous — but it wasn’t clear how they were doing it, and they were very secretive about it. That stood out as a center for innovation — and we’ve always looked up to those guys.
I think that the core avenue of innovation is this: The world has gravitated to a place where all data is captured and the processing power is there to process most of it. If you’re going to make investment decisions, why not deploy the technology that you’ve got to analyze the data to the best of your ability? Investment management is information management, isn’t it? Nowadays when you look at how we make decisions on everything in life, it’s always studying the data on the matter, whatever the data is, because the data on the matter will be available.
How different was BlueTrend from other trend-following systems at the time you launched it?
The idea of trend following has been around for a long time. It’s difficult to say how different BlueTrend was because I don’t have access to what others were doing. From one angle it was very similar. The correlations were high. The correlations in the trend-following space are high, generally around 80 percent. Having said that, correlation in the ’80s doesn’t mean that you’re going to get returns that look like each other. To evaluate the correlations you need a long time window, and the correlation is only one measure. But it doesn’t actually say that the outcome is the same. So we weren’t really looking too much at what others were doing.
We had momentum signals, we had an asset allocation algorithm, and we had execution assumptions. We knew we could put research effort into all three avenues and we could improve what we had on day one. We felt if we put our minds to it and if we tried different things, we’d find improvement.
Is the idea of improving — always getting better, always tinkering with the system — important when it comes to systematic trading?
Yes, I think it is, and I think we owe some of that to J.P. Morgan because J.P. Morgan is a very intellectually driven organization. They really listened to the quants. In other organizations the quant is the geek on the side. He’s not paid that much, and you don’t listen to him that much, and he writes us a model that occasionally we use. J.P. Morgan wasn’t like that at all. They used to pay attention to what was said, and I think that is partly why they didn’t get themselves in a pickle over the credit problems [during the financial crisis], because the quants did say, “Look, there are limitations to this,” and the J.P. Morgan crowd listened.
When we started BlueTrend, I used to go on the road and present it to investors and say: “Look, we’ve got this model and it’s trend-following. So you know what the trend-following piece is, and you know it works, and we can’t quite explain it. The academics dispute it. We’ve got all this methodology. This is how we developed it, and this is the kind of result that you get.”
Then I would say to them, “But we’ve hired three new quants, and we are doing some research, and we’re going to update all these models.” The reaction from investors wasn’t necessarily positive at all. They’d go: “Hmm, you’re going to change it? Why are you going to change it? Are you not happy with what you’ve got?” I was really taken aback by that because it came so naturally to me to say that we had a model but we were improving it, and I realized that in the algorithmic trading world people derived a lot of confidence from the fact that you would say: “I never change it. I did it and it works, and it’s always worked, and it’s always going to work.”
Fast-forward ten years to now, clients go in and they ask me what the research agenda looks like. Clients have come to expect that you need to adapt and evolve and get better. That has been a big shift in the industry.
Why has that happened?
I think apart from anything, every year the markets throw another year’s worth of data at you. That data might contain patterns of behavior or movement that weren’t in the data set before that will highlight a weakness of your model, that highlight a strength, maybe, and you’re obliged to evaluate it. When we started in the hedge fund world, it wasn’t obvious that constantly trying to improve your model was wanted or required, but over time it has become the norm.
Can you talk a little bit about BlueMatrix?
BlueMatrix is an equity quant market-neutral program. What does it do? It trades a lot of stocks. It trades 4,500 names worldwide, so it’s a global program. It listens to a lot of sources of alpha, so it uses a variety of data sets. It uses price, but it uses volume. It uses directors’ dealings. It uses financial statements. It uses analyst revisions. It uses data on news and events. It uses data on the supply chain, how companies relate to each other.
Each one of these data sets will typically have one or two investment theses attached to them. You could say that if analysts are upping their revisions on a company, we expect the price to go up. That’s an investment thesis, so you can say that if a company is upstream on the supply chain from another company, then if the company that it supplies has a price spike, then you expect the supplier to follow that spike within a certain time. The algorithm listens to all these alpha sources, and it tries to combine them into a portfolio and assembles that portfolio while making sure that the portfolio is neutral to the market. The program is a residuals trader. It’s trading relationships between stocks.
What do you do to encourage innovation at Systematica?
According to the business school thinking, it’s useful to classify innovation into two avenues: radical innovation and incremental innovation. Both are difficult to implement. It takes people willing to be free thinkers and say something that half the people might mock them for. Nobody on the Systematica team purports to know the outcomes and to know the truth. We all listen to each other. Everybody thinks they can challenge everybody else.
Incremental innovation is for us somewhat easier to implement because if you think about it, the programs that we run, they’re live. They attract investment. They will go through tough patches. They will display behavior that we like or that we don’t like, and so to try to address that, to fix that, to study that, is a very natural thing. If you think about the BlueMatrix program, it’s got a lot of sources of alpha in it, but the guys on the team keep dreaming up other sources of alpha. “Oh, now the mobile phone companies sell location data of mobile phone users.” You can actually track if people are going to shopping malls more often, and that should predict the return of the consumer goods company. People will come up with ideas. “Look, this data shows this and that. This is a plausible thesis, don’t you think? Shall we investigate it?” So incremental innovation is good there.
In terms of more-radical innovation, where you do something very differently, it’s a bit tougher, right? In the BlueTrend program, for example, we started with one trend-following system, and we were lucky and the run was good. By about 2010 there was enough experience in the research team that they felt that they were ready to try something different. It was a bit like if you are BMW and you make a very successful 3 Series car, do you want to be the engineer that comes along and says, “We should scrap the 3 Series because it’s not good enough, and I’m going to come up with . . .”? It’s always safer to do it incrementally, isn’t it?
At the time, what we did with the guys [on our research team] — which I think proved to be a success — was I said to them: “Look. We are paid to generate returns. You are not paid to find the absolute truth. So how about you guys articulate a different system for trend following? If we like it and if it passes the tests, we’ll allocate to it and BlueTrend can have two trend-following systems inside it.”
We did that and they articulated a new model and we allocated initially 3 percent of the assets to it, and so 97 percent of the assets were running on the original model. Over time we got more comfortable. We got more interested in this new model, and we allocated more and more to it, and then, finally, in 2014 there was a research project to merge the features from both into one brand-new module. That was a way to not shut down the guys or to not dismiss the guys that wanted to do something more radically different.
Was there a big difference culturally between the systematic and discretionary businesses at BlueCrest?
Culturally, the systematic division was always quite different from the discretionary side. The discretionary trader in general is an individualist because he needs to make decisions by himself, he needs to bear the downside by himself, he needs to have a tremendous amount of self-confidence. “Often wrong, never in doubt,” as a friend of mine likes to describe it. Perhaps that personality is not conducive to teamwork.
The systematic side is pretty much the opposite. We look at trade opportunities as a whole. The guys who are good at data analysis are perhaps too mathematical to execute the trades, and the technologies and opportunities are too specialized for traders to code it themselves. So you need technologists for real, you need traders for real, you need researchers for real. These three groups need to rely on each other. You cannot be a one-man band in a systematic shop. We liked to think that we were a team and the discretionary side was not. That, combined with the fact that investors were on a crusade for lower fees and greater transparency, sort of dictated that we went our own ways and split up the company. If you are a systematic shop, you can respond positively to these pressures. If you are a discretionary shop, these pressures can cause difficulty. Letting us go our own way just gave us the freedom to respond to these pressures and turn them into a bit of a competitive advantage.
Do you see innovation playing an important role when it comes to things like fees and designing products for investors?
Definitely. Look, it’s amazing. You’d think that there could be a pension fund manual out there, but there isn’t really. I was part of the board of the CERN Pension Fund in Geneva, the research institution, and I’ve seen how some of our clients think. Different people have different priorities. They have different targets. They are dealing with different liabilities and different inflation levels, and the trustees of the board can be sensitive to different things. To be able to, in a cost-effective way, shape your offering to cater to the client is important.
Are there new technologies or technological innovations that would enable you to push your model to do more?
Yes. One of the things that we have decided as a team is we want to be value-added managers. So you sign up for our funds, you get your returns, but you get some other things. We’re looking to share research information through technology with some of the investors, and that’s one initiative right there. The investment in data and technology is by far the biggest bill we have.
Do you think the hedge fund industry will experience radical innovation, or is it going to be incremental?
I think the evolution of an industry segment always looks incremental because you’re looking at a large group of people doing something and you’re looking at a distance. But then some of the sheep in that herd were a bit radical in order to drive the incremental change.
We try to be radical. Innovation is the development of something new that applies itself to something commercial or to something that is useful. What is the thing that inhibits innovation? To dream low — dreaming low is not good, right? You’ve got to dream high. You’ve got to try to solve the problem as it is. “Look, guys. We need to develop something that works at all times, that can be done for a very low fee and that will be transparent.”
One thing that I believe in is not telling people what to do or how to do it but just telling them the goal. Let’s just share the goal and let you guys choose the route. I think we all do that in our company. Give people time to get it wrong. Don’t give them too much grief because if you’re going to try lots of things, then lots of things are going to go wrong. Just because there is a problem, it doesn’t mean that you will find a solution next week. You need to battle on and try things. You need a good, friendly, open atmosphere where people cooperate and discuss and they’re not afraid to suggest and to try.
