Beyond human bias: The systematic advantage

Lukasz de Pourbaix (LDP): You've been with Fidelity for 20 years, which is a long time. In that time, you've been running a variety of systematic strategies which is quite unique, given Fidelity is very much known as a fundamental equity shop.

Tell us a little bit about that journey.

Matt Jones (MJ): It is a long time and it's nice to talk about because systematic equities and what I’ve done at Fidelity for 20 years hasn't really changed. We have been doing it probably about 22 years in total. It has been a long journey. My early days started off here in Sydney in a fundamental equity team, like Bankers Trust in international equities there, essentially as a quant within a fundamental research team. A lot of learnings began in those early days of portfolio construction and trying to capture alpha from a from a fundamental research team. That took me to JP Morgan for about 18 months - two years and finally to Fidelity about 20 years ago in London. I spent the first 13-14, years in London, and six now back in my hometown.

LDP: The systematic business has grown from strength to strength. I think it's roughly around USD$22 billion at the moment, which is incredible, and it is testament to the growth and interest in systematic strategies more broadly.

MJ: I think the increase in growth, especially more recently, is around things like customisation, repeatability, a bit of a stronger process around risk controls and portfolio construction, which you didn't see a lot more in the early days. Then, it was all about a lot of things, about taking big bets everywhere and big risks and tail risks and more. More recently, over the last five - 10 years, some of those tail risks have been so big that people are starting to open their eyes to something a bit more risk control, a bit more disciplined, a bit more repeatable.

LDP: We're hearing that from a lot of clients. They are looking for strategies that, to your point, have transparency i.e., it's very clear as to the investment process and outcomes.

MJ: A bit more stable and persistent. I try to say systematic equities, especially the way we do it at Fidelity, is like that blue chip stock that back in the day you could buy, lock in your bottom drawer, pull out 20 years later and it hasn't blown up and it's still ticking along.

LDP: In my mind, when I think of systematic, it can range from a quantitative black box model through to an approach which integrates fundamental research, such as our process. How would you describe systematic investing and in terms of your role, how do we position systematic investing at Fidelity?

MJ: I think that's the difficult bit. Systematic is like quant and in a way, it can be very broad and diverse in the way it's implemented and the different approaches they use. Like you just said, it can be very high frequency, black box algorithmic trading, like some of the big hedge funds of the world. On the other end of the spectrum, which I think is where we sit at Fidelity, it's using systematic processes and strong portfolio construction, but their pure input into our process is a fundamental analyst writing research. It's very much forward looking compared to a lot more quantitative strategies, which are looking at a bit more backward-looking data.

LDP: It's an interesting point, because the forward-looking bit is certainly a critical element, because if you cast your mind back to say 2008 and the Global Financial Crisis, in that period we did see some pure, what I'd call quant strategies, when you had that big market dislocation really come unstuck. But if you had that fundamental insight, it provides more pragmatism to the process.

MJ: We saw that because in running strategies, even long/short strategies, and working in Europe during the Global Financial Crisis (one of the biggest events I've seen in my career and hopefully we'll never see again) you do learn a lot of things from that and it was really nice to see that by running a more systematic process, a bit more quantitative, but using the deep fundamental inputs, we actually navigated that better than our quantitative peers at the time because we weren't chasing rapidly dislocating risk models. We weren't chasing momentum. We held onto that long term view and that's very much worked until recently and the last two - three years with macro shocks from Trump trade wars. I'm not seeing the rotations and the volatility as some more of our quantitative peers in the way we implement our process here.

LDP: How do you marry the systematic quantitative element with the fundamental? How do those two things interact from a practical, everyday perspective?

MJ: It's quite neat and it's quite simple. I think we're very lucky here at Fidelity because the company has been around since 1969 and our big parent company, FMR in Boston, has been around a lot longer. We've got this long, deep history of analysts writing research notes, doing the deep fundamental research, and that buy or strong buy, or sell or strong sell rating comes back as a number and straight away, we can start to quantify that and build it into a process and speak of that history. It's not just recent. We've got 30-40 years of this history of research notes, so we can capture that long term history, analyse it, see when it works best, and work out how to incorporate it in portfolios. In our process, an analyst researches companies like Commonwealth Bank, ANZ or BHP, and they speak to all the analysts around the world to come up with a simple buy to sell rating that turns into a number. We capture that in our portfolio construction and investment process to make sure we focus on what we call the higher conviction names, the names they love the most. Essentially, we can look at things like their model portfolios to narrow it down to two or three names they might really like in their sector. We take those numbers simply and put it into a filtering process to focus on what are the best ideas at Fidelity and then wrap that in a strong portfolio construction. When I talk about portfolio construction, it sounds technical and a bit scary at times, but it's not. It's just building a portfolio in our respect that doesn't have any big bets. Like I was saying earlier, I'm not a value manager and I'm not a growth manager. I'm an active alpha, stock picking manager and that's what we focus on.

LDP: So, the way I could describe it is a fundamental research process then a systematic portfolio construction process, which really takes away some of those biases that sometimes, consciously or unconsciously, occur as a portfolio constructor?

MJ: Exactly. It's more about the unconscious, conscious ones. A lot of the times, we say we will make sure there's no unintended biases in our portfolio. What that means is we want to make sure that we're only betting purely on our stock picking capabilities at Fidelity. We don't want to see anything else creep in, like a value bet or a growth bet or a large cap bet or a small cap bet. We just want to purely focus on that stock picking risk, because we're a stock picking house, and that way you minimise any other big risks and big rotations or big kind of dislocations in your portfolio.

LDP: Given you've been doing this for more than 20 years, in terms of your career particularly on the fundamental research side, I'd be interested to get your thoughts on what apart from the analyst ratings, gives you a sense of conviction and quality?

MJ: I think the beauty of our strategy and the beauty of Fidelity is that everything comes down to the fundamental rating. It's quite simple. It's quite transparent. It's not over engineered in any way. An analyst has a stock on a buy or sell rating. What's really important is the change between those ratings. If an analyst goes from a buy to a sell, for example, or a sell to a buy, that's probably one of the strongest signals at Fidelity to either get in or out of that name or increase or even reduce your bets in that name. It's a very strong signal. It doesn't fire often, because our analysts take a three-to-five-year view. So, this isn't changing every day and that's probably one of the most important signals at Fidelity in my career, of looking at things in a more quantitative sense. I think the biggest thing we can do, and obviously the fundamental to do this exceptionally well, is reading the research note. A number doesn't capture everything. When you read the research note, you get a feeling from the analyst on how much conviction is behind this name, whether there's something you need to reach out to them to challenge them on just to make sure that you have their best ideas in the fund. I think that's a very critical, important step on top of this process that we run that a lot of quant funds or purely systematic processes don't do. It's really good to understand in the world when you think of things like AI like natural language processing, what does that mean? It reads the words, and it tells you what it thinks is going on. We do that on a lot of our research notes to to help speed things up. It is important to get that conviction behind the research note. Just, not just what the number says is on the page.

LDP: Touching on your comment around AI, your model and the process evolves over time and AI is a good example where it's very much at the centre of a lot of conversations across industries. You mentioned that you're starting to incorporate some of those things. If you're to think ahead, where do you think it may go in terms of working alongside what you do from a day-to-day perspective?

MJ: I think the key word is alongside. I think working alongside AI is where you'll see the most critical and incremental improvements and efficiencies in everything that we do, especially in the research team. I'm a believer, so I don't believe AI will take over a really great quality research team. Will AI take over an average research house, yes, but I think it's the next step beyond that where you add the incremental value that I won't be able to achieve. Like all quant models and all AI, they're not 100% accurate, so you want to improve on that accuracy. That's where the human being, the analyst, will step in. AI will take up a lot of the grunt work. You know, an analysis picks up a new sector, instead of taking three months to come up on speed on the sector, AI could help them come up to speed on the sector in a matter of a month or two, which means they're then adding the value on top of that a lot quicker. I think that's where it'll be most critical and that's where we'll use it a lot more in part of our process as well.

LDP: I think it's an interesting point, because there's a lot of noise around AI. Having that quality and if you've got an edge, whether it's in investments or whichever field, that's still going to be of high value. From a Fidelity perspective, that research pedigree is certainly our edge. It's quite well known that the critical bit on AI is the data that you're feeding into this. If you've got the data that no one else has, which we essentially do, arguably the best research house in the world, with that history of research data, if no one else can get access to that (which is what a quant loves, access to data that no one else can have access to) and we can build models off that to help the research team perform even better. Then, it's a no brainer. It's a win, win situation.

You're based in Sydney, Australia and as part of this global organisation, you collaborate with our teams around the world. How does that practically work from your perspective?

MJ: It's quite interesting, going back to my early career and studying in the industry, I think around 1998 at Bankers Trust in Chifley Tower, in an international equity team run under Bankers Trust New York, globally, outside of Sydney. To me, it's not unique and this is quite easy, but what's interesting over the many years is seeing how much more efficient it's become. I think back to watching the analyst in the old days, picking up the phone, trying to get contact with CEOs across the rest of the world. At Fidelity, that's incredibly easy. We've got a presence in most countries. Infrastructure and technology has evolved even in my 20 years at Fidelity infinitely where I can put trades down on the desk, share them with my colleagues in London. We can see the impact immediately on all the funds we run then we can send trades from anywhere in the world. I have sent trades from Mexico, even back in my early days in London. So, if I could do it back then yes, Sydney is an easy place.

LDP: I could imagine 20 years ago and beyond, would have been very different - sending the pigeons to carry messages. From a global perspective, you manage a range of different strategies for different client bases within an Australian market. We are running a systematic strategy in the Australian market, so I think one of the interesting things with that is that with a systematic approach, it's very malleable in terms of whether you're running a long only equity strategy, whether it's a long-short, market neutral strategy. You can adapt it for different types of strategies.

MJ: Other than the renewed focus on portfolio construction, a bit more discipline around how you build portfolios and risk. Always one of the biggest selling points is that adaptation to solve clients’ problems and solutions. Do you want to exclude certain sectors? I don't have a sector view or as it would/could if client wants that we can incorporate in, but it's very much solving a client problem. What do you want an exposure to? We can give you an exposure to that, whether it be indexed enhanced (like the SMAs), or full geared, long-short market neutral funds, bit higher alpha, higher risk, from infinite between those two spectrums.

LDP: Yes, the scalability of the process is quite amazing in the systematic approach. Feels like it's the era of the systematic investment style at the moment.