Trading Insights: Systematic investing, with Aspect Capital’s co-founder Martin Lueck

So what we're trying to do here is to capitalise on trends which appear occasionally and unpredictably in some of that very broad range of market sets. We can't predict where they're going to develop, we can't predict when they're going to develop, but by careful risk management you can jump on opportunities wherever and whenever they occur. Hi, I'm Eloise Goulder, Head of the Data Assets and Alpha Group at J.P. Morgan, and today I'm so delighted to be sitting down here with Martin Lueck, who is Co-Founder and Research Director at Aspect Capital, a global systematic investment manager. So, Marty, thank you so much for joining us here today. Absolute pleasure, Eloise. Thank you very much for the opportunity. And in fact, this marks our second conversation with your firm, with Aspect Capital, on this podcast series. It was about 18 months ago that I sat down here in this very same room with your Co-Founder, Anthony Todd, for a really insightful conversation. I really enjoyed that podcast and hopefully I get the last word. Absolutely. So, I know, Marty, that you really are a pioneer in the systematic investment management industry. You are the L in AHL from MAN AHL. So perhaps before we talk about the origins of Aspect Capital, could you tell me a little bit about your career before then? Sure. So, it was in many ways a happy accident. I studied physics at Oxford with my dear school friend, Mike Adam. And when I left Oxford, I went to work for Nomura, but spent my lunches with Michael, who was then working for his father in a small commodity broking business. And Mike had bought an early Hewlett-Packard workstation, and on his father's suggestion, was testing chart strategies. And Mike's dad said, why don't you see if there's anything in this, if you test it statistically? And I was absolutely hooked. So, left my job at Nomura, joined Mike, working for his father, and we met a third former physicist, David Harding, who was from Cambridge, and also fascinated by the potential of applying systematic methods to financial markets. We were discovered by the MAN group, who were interested in taking our product and distributing it through their network. That, though, was a distinctly retail business. So, the vision of Aspect was to bring that product, bring that approach to institutional investors. So, what really was and is the investment philosophy at Aspect? Over the years that we have investigated these systematic models, it became clear that the most valuable component here is the strategy commonly known as trend following. Through trading the futures markets, we're indifferent to whether a market is rising or falling. And through the breadth of highly liquid futures markets available to us, we can get enormous breadth across a range of asset classes. Literally hundreds of markets that gives us the opportunity to trade from bull markets in equities and fixed income to bear markets in grains and foodstuffs, depending on what's going on at any given time. And in the retail context, there was a lot of opacity to this. And we felt to have a meaningful allocation in institutional investors' portfolios where it really would bring valuable diversification, it had to be made transparent. We had to educate and develop partnerships with our investors so that they would understand it, because the return dynamics are very different from holding traditional assets. So, what we're trying to do here is to capitalize on trends, which appear occasionally and unpredictably in some of that very broad range of market sets. We can't predict where they're going to develop. We can't predict when they're going to develop. But by careful risk management, you can jump on opportunities wherever and whenever they occur. And that lends itself to the ability to capitalize on trends, both rising and falling, that may occur in financial markets. And that utility of the strategy is very different from everything else that institutional investors are likely to have in their portfolio. So, trend is clearly at the heart of what you do. And I know that trend as a feature has been around in markets for hundreds of years. Why do you believe that trend is so significant and is so persistent? Well, empirically, when Mike and I started looking at those systematic models, there's one really prominent driver, and that is momentum. That is the autocorrelation in markets. It's a persistent feature of markets, but it's episodic, and it's very unpredictable. And why does it exist? Well, I think it is inherently a property of human-intermediated markets. It's a manifestation of human behavioral biases, of FOMO, the fear of missing out, of anchoring, of a really distinct loss aversion. And if you read the works of Daniel Kahneman and Amos Tversky, they bring science to that understanding of how much more we hate losing money than making an equivalent amount of money. It's human nature to sit on a losing trade. You put it in the bottom drawer and cross your fingers that it won't go away, and if you start to make money, you'll take profits. Those are exactly the two behaviors which are the worst possible way of trading, and that's one of the key drivers of trends. Another feature is information dissemination. And even if all information was available to all market participants at the same time, which, of course, it isn't, different market participants will react to that information at different rates. The efficient market hypothesis is, in my opinion, deeply flawed. So trends are an inherent property of markets. Another recent example is the return of inflation in 2022. The central banks were absolutely in denial. All of the signs of inflation were brushed off as transitory until they weren't, and then we all wound up playing catch-up, and that drove one of the most successful trends we've seen in fixed income for a very long time. It's fascinating how deep-rooted an attribute trend has been over the years, and presumably you see trend as distinct from something like value or relative value in the sense that they are inefficiencies, arbitrage opportunities, whereas trend is an attribute within markets. Yes, so we talk about having divergent or convergent strategies, and trend is a potentially profitable strategy that is able to capitalise on divergent market opportunities, whereas something like relative value or a relative momentum strategy is a convergent strategy. Yes, and I'm aware that trend hasn't performed so well year-to-date. What do you think the reasons are for this relative underperformance in trend? Trend generally has two weak points. The first is when you have an absence of trend. If you go through a quiescent period where there isn't much happening, then the model will be looking for trends but not committing very much. More painfully, though, the second Achilles heel of the strategy is sharp reversals. When you've had a really strong trend that the strategy has been able to capitalise on, and those suffer a sharp reversal. The model takes time to assimilate new information and new condition, and it will go through a period of losses as it reduces its position. So coming into the end of last year, we were very much riding that American exceptionalism trade, and that was very heavily represented in some of the positions in the portfolio. Come some of the ructions and policy decisions of this year, that all gets thrown into doubt, if not reverse, which means that those well-established positions are suddenly thrown into question. The risk management of the portfolio will reduce them very quickly, and then over time, the range of different timescale filters, these will actually scale down the position, and if it persists in the new direction, build a position in the opposite direction. That's very clear that there should be these two distinct areas where trend performs less well. One, the absence of the trend, and two, the reversal in the trend. And as you say, the unwind in the US exceptionalism trade, it has really been a reversal in trend. And presumably, the breadth of your programmes across multiple assets and multiple regions is really very powerful, because it may be that certain trends arise that none of us are even aware of, or that the popular press hasn't started writing about yet, yet they're starting to work, and you have this ability to capture them. I think that's absolutely spot on. We wrote a paper recently which captured these thoughts. It talks about unpredictability alpha. So people ascribe trend as a strategy that offers crisis alpha, and I think that's just too narrow. We think that its applicability is much broader, related to sort of global unpredictability. So if you think about the world that we live in today, with tectonic plates moving in every different asset class, whether that's the geopolitical unrest now, climate change which is fomenting demographic shifts, real shifts in dispersion in the economic environment, we're going to see opportunities across all asset classes of dispersion in currency markets, in equity and fixed income. You look at what's going on in the US and there's pressure to reduce rates, whereas in China they're at a different stage in the cycle. Across commodity markets, the shift towards energy transition, food insecurity, these are all important trends, and this strategy allows us to take advantage of opportunities in all of those markets without having to have an a priori thesis of where it's going to develop. So I'll point you to the cocoa market, which we have studiously tracked and traded for many, many years, and it's gone for at least 20 years without contributing much. But we adamantly continue to have a risk allocation to this market, because you can't predict where these opportunities are going to present themselves. And halfway through 2023, we find that we're starting to build a meaningfully important position in this market. In 2024, that turned into the most successful position in the portfolio. So just as you said, it's often very hard to know what the next opportunity is going to be, and sometimes trends develop hidden in plain sight. It is impossible, in my experience, to time this strategy. If you think we're going to jump in as soon as we start to see signs of a trend developing, chances are you've already missed the substantial belly of the trend. And I'd offer one more example, which was in 2014, we saw the shale oil revolution apply pressure to energy markets and crude oil. It had been sitting around $100, $110 a barrel for a couple of years, and then started to slide, and it kept dropping and dropping and dropping down to below $30 a barrel. So again, that was another really successful trend. Well, I think both of those examples, the cocoa and the crude oil, also show the strength of your conviction in sticking with the trend, even when for a period of time for those particular asset classes, it wasn't necessarily paying off. And almost the necessity within your organisation to remain unbiased and not to take a discretionary view that a certain asset class might no longer be in favour. Because at some point in time, it will reflect itself in the trend and you will reap the rewards of that. But I'd imagine that that can be very difficult at times, particularly when multiple trends don't work for a period of time. You mentioned earlier that you do have some non-trend strategies, some convergent strategies. Could you speak to those, how significant they are within your portfolio and at points at which the trend isn't working so well? Yeah, the origins of the firm were in trend, because we were absolutely convinced for all of the attributes that we've spoken about, but it became clear quite quickly that the return stream that it gives you is very, very different from, say, equities or bonds. Equities tend to go up, up, up and then have a short, sharp, painful period and then pick themselves up. You have more positive days trading equity markets than you do losing days and that we as humans are behaviourally comfortable with. By contrast, the return profile of a trend following strategy is actually to have more losing days than it does profitable days. So by the early 2000s, we realised we had to complement it with a range of those non-trend strategies to provide buffers in periods of sharp reversals or periods of no trend. Trend is in many ways the active ingredient. It's the medicine in an investor's portfolios. And number one, you have to finish the course and give it time to work. And number two, often to make it palatable, you have to sugarcoat it. And I'm not trying to be glib because these are in their own rights very useful strategies, relative value or seasonality, and these are complementary to trend. And the different programmes that we'll offer our investors, often in customised solutions, will have a different ad mixture. So some clients are very comfortable with the pure trend, the uncut active ingredient, and others are more comfortable with much smaller allocation to trend and higher allocations to some of the more mean reverting or convergent strategies that we've spoken about, a more absolute return profile. I love that analogy. This is the active ingredient that's so critical. And presumably from the investor's perspective, it depends what else they own as well. If this is their standalone strategy, they're likely to want the sugarcoat, they're likely to want the other strategies. But if this is more of a hedge for a time at which traditional equity markets, let's say, will underperform, then perhaps they don't need it. That's absolutely correct. And indeed, that happens at a bigger scale across the investment industry. So we've seen consultants work with institutional investors to develop their risk mitigation portfolios or crisis risk offset portfolios, where what they're doing is embedding trend, which is referred to as the second responder, alongside the first responder, which may be a long duration or an option based strategy that gives you much faster responsiveness to a crisis. Trend gives you that convexity when a market, principally an equity market, goes into a bear phase of say two to three months or longer. And you may also have an allocation to say, systematic global macro or discretionary global macro to try and provide an absolute return boost. Absolutely. Well, you clearly utilise a huge amount of data, given that you're trading multiple programmes cross asset. And also, you clearly have multiple models in place, including not least your trend models. So Marty, we always ask our guests, what do you deem to be more important, the data or the model? So what's your take on this one? Oh, I love this question, Eloise. And I love listening to all of the different answers that you get. And I'm going to come down very heavily in favour of the model, because I think you can see cloud patterns in data. If you just look at data and make random inferences from it, my point is that you have to have a model, it's just as in physics, and you construct a model of how the world is. And then you use data to test or to prove or disprove your model. Now, I'm not implying that finance is a hard science. But I do believe as a quantitative investment manager, you have to be able to articulate what the model's doing, and then use data to support or disprove that model. Of course, they're intertwined, data is going to lead you to a formulation of the model. But I adamantly would lean into the model as being the crucial piece. Fascinating. I think that's very compelling that actually you have to have a hypothesis, you have to have a starting assumption about the way the world works, the way financial markets work, and then to test that hypothesis, and your hypothesis is reflected in your model, you test it using the data, we don't often get that answer. More often than not, we hear that data is the more important piece, albeit I would say all of our guests acknowledge both are important. You have to have an underlying model. And if we talk about the application of machine learning in financial markets, we're very much of that school of thought that we have to have an underlying premise. Of course, machine learning is a tool and it allows us to find relationships in data that might be hidden to the naked eye. But I'm not going to trade much, if any, of that model, if I can't articulate a narrative of what it's finding. Well, that lack of interoperability comes up so often with machine learning and issues using it within finance. Could we turn to LLMs and AI tools as a whole? It's obviously a very fast-moving space. I'm sure there are many areas where you are cautious, given the need to have that starting point hypothesis. But could you explain how much you are looking into these newer tools and how useful you think they could be? It's really interesting. It's a great space. It's not new, though. The application of machine learning and neural networks to financial models, people have been trying this space for a very long time, candidly, without a lot of success. Ironically, about five, six years ago, there was a big sort of burst of hype. If you weren't investing big time in machine learning models, you were going to be left behind in the quant finance world. Internally, people's excitement was in inverse proportion to how much they actually understood about the space. Folks that we had on the research team who had done their PhDs in machine learning were going, steady on here, let's not get over our skis, because these aren't physical systems. Financial markets are not like learning to fly a plane or the game of Go. You can latch on to what the predominant set of drivers is right now, and it can change in a heartbeat. That's the first thing to bear in mind. We love it. It's a fascinating space. We approach it very carefully. We insist on having interpretability. We do very constrained data sets to understand what we're getting out of it, and then gradually relax those constraints to see what we can find. I think that it's much more now a tool that we apply across the research space. We're looking at signal generation, but then it may also be used in volatility forecasting or portfolio construction or execution methodology. It's a great technique to add to the quiver, and provided you're very conservative about not over-relying on it and making sure you understand what it's finding it, if you ask a question of a machine learning model, it will give you an answer, even if there wasn't an answer there. It's a useful tool. It's not a panacea. MS. LLMs can sound so convincing, and I think that's also part of the worry, is that they can give you this false answer when it shouldn't exist and sound so convincing at the same time. MR. Absolutely, with all of the footnotes as well to pay into papers that just don't exist. MS. Yes. Could we turn to the human? We've obviously spent a lot of time on your processes, the critical significance of your models, your hypotheses, but could you speak to really where you feel human judgment is most critical? MR. Well, we make a big thing about being systematic. I think that's crucially important, both to the repeatability, the testability, and once again, the utility to an investor so that we can model what it is we're trying to do. But having said all of that, quantitative finance is still a people business, and I think we recognized that early on. As Antony mentioned, when you met, one of the very first things when we established Aspect was a culture document. These were the attributes and behaviors that we were looking for and we expected from people, and that's manifested itself in how we do research. We expect it to be a collegiate and academic exercise. We're looking for people that are interested in cross-fertilization of ideas. So, focusing on that breadth and collegiate and collective research effort means that we value diversity, diversity of experience and thought. Bring those together into a room, I think you get a better program. MS. So, Marty, we always ask our guests what's next, and are there any areas that you're really delving into at this stage that you want to share? MR. Well, I often get asked the question, well, after almost 40 years of doing this, surely I've figured out trends by now. But no, there are often some amazingly new insights that the team has. So, within the research team, there's proliferation of data series. I know you've covered this a lot. This is an area of your particular expertise. There are always new data sets that we can harvest to find some nuance. We've talked about LLMs and machine learning techniques for parsing that data, what is useful to us and what's just noise. All of that is an infinite playing field that we're super excited about. I think the real skill of what we do is in distilling an agenda down so that you focus on the most meaningful projects. Apart from what's going on in the research space, I think the investor landscape is also interesting. I think about a period of TINA, of equity markets being unrivaled. The only rival to public equities has been private equities and private assets. So, portfolios are pretty undiversified and pretty illiquid now. I think that the liquidity characteristics of what we do are going to become increasingly valuable once again. It's kind of like salt on your food. You don't know you need it until you don't have it. The other feature is that the flexibility of this systematic approach means that we can customize portfolios and make them much more useful to our investors. There is a resurgence of interest in this idea of portable alpha. You can actually very efficiently take existing portfolio betas, whether that's exposure to equity markets or bond markets or other assets within an portfolio and overlay on top of that a diversifying component such as trend or more absolute return type programs. Well, Marty, this has been such a fascinating conversation. I think everything from learning about the origins of Aspect Capital through the origins of AHL and Man AHL to hearing your perspective on why trend is such a significant market attribute and the drivers of that through to the significance of the model, the hypothesis, the explainability, all the way through to the culture, the collaborative culture and the importance of sharing knowledge and ideas across your organization. Absolutely fascinating to hear all of this. So, thank you so much for taking the time. Eloise, thank you. It's been really my pleasure. I very much enjoy our conversation. I hope we can do it again. Well, likewise. And thank you also to our audience for tuning in to this bi-weekly podcast series from our group. If you'd like to learn more about Aspect Capital, then please do go to their website in the show notes. Otherwise, if you have questions or if you'd like to get in touch with us, then please do go to our website at jpmorgan.com forward slash market-data-intelligence, where you can always contact us via the form. And with that, we'll close. Thank you. Thanks for listening to Market Matters. If you've enjoyed this conversation, we hope you'll review, rate and subscribe to J.P. Morgan's Making Sense to stay on top of the latest industry news and trends. Available on Apple Podcasts, Spotify and YouTube. The views expressed in this podcast may not necessarily reflect the views of J.P. Morgan Chase & Co and its affiliates, together J.P. Morgan. They are not the product of J.P. Morgan's research department and do not constitute a recommendation, advice or an offer or a solicitation to buy or sell any security or financial instrument. This podcast is intended for institutional and professional investors only and is not intended for retail investor use. It is provided for information purposes only. Referenced products and services in this podcast may not be suitable for you and may not be available in all jurisdictions. J.P. Morgan may make markets and trade as principal in securities and other asset classes and financial products that may have been discussed. For additional disclaimers and regulatory disclosures, please visit www.jpmorgan.com forward slash disclosures forward slash sales and trading disclaimer. For the avoidance of doubt, opinions expressed by any external speakers are the personal views of those speakers and do not represent the views of J.P. Morgan. Copyright 2025 J.P. Morgan Chase & Company. All rights reserved.