Strategie Quantitative con Nicolas Mirjolet di Quantica Capital (EP.103 Special)

In questa puntata speciale Nicola commenta l'ultimo paper di Quantica Capital insieme a chi l'ascritto. Nicola Mirciole, co-Eddo Fresserci della stessa quantica. Il paper sentitola "If you can beat it, stack it" e se il titolo non ti dice già tutto, te lo dirà lui. L'intervisa è stata registrada nel studio di Mr. Tip. Grazie Gioccio per i pressi to. Ti abbiamo nascosto qualche regalino in giro, ma non ti dremo mai dove. Ultima cosa, l'intervista in inglese, se lo montata io, può ascoltarla anche tu, quindi non avere paura. Se invece l'inglese di mettere ancora un po' d'ansia, può rimediare con lo sponsor della puntata. Prepli. Il migliore modo per fare lezioni, uno con un insegnante madrelingua. So Nicola, welcome to big to fail. So, you recently wrote a super interesting paper on a topic that we try to introduce many times on this podcast, which is portable alpha. But before going there, can you introduce yourself to our audience? How do you come to specialise in systematic trend following? And what drew you to quantica capital? Yeah, thanks for having me on the podcast Nicola. Pleasure to be with you. Yeah, I've been a quant for a bit more 20 years now. I came here to Zurich a bit more than 20 years ago. Did some student projects that touched a bit on finance was interested, managed to get an internship and ended up at the family office. They had discovered quant finance, systematic investing, and got the opportunity to launch a hedge fund. That is garbage, trash, equity, long short hedge fund. That was an incredible adventure, incredible learning for me, starting really from scratch, and going out, trying to raise money, delivering the performance that we believe would be able to deliver. Basically going to the whole experience of building a hedge fund. I did that for eight years until basically the learning curve was starting to flatten and we hit some glass ceiling, if I may say. And there I got the opportunity to join a bigger group and that led me eventually to join a quantica capital in 2020 shortly before the pandemic. And I have been in this role of leading the research at quantica now for the past six years. Now it's at the other end of the spectrum. I started on the short term conversion type of strategies, betting on single stock minuversion in a way. Now I'm on the other side of the spectrum where we are basically trying to capitalize on the divergence of macro asset classes, long term divergence. So it's actually quite an interesting evolution. For 20 years from short term, minuversion is too long a term divergence strategies, but all systematic, always systematic. Now we are here because you wrote a paper that is titled "If You Can Beat It, Stack It." So what was the light bulb moment that made you want to write it? It's a long story. I mean basically we run a quantica capital is an independent business, our clients are institutional allocators from the US to Japan and they have specific allocation requirements. That whole possible alpha topic came up is something that came up the last two years, maybe I would say originating from the US and spreading a bit across the regions and the type of allocators. So that's what triggered the initial work that we did in this space and realizing that thinking first what was driving this interest and number one is for sure the very strong equity performance that we have seen over the past decade. Actually I was surprised myself. I mean everyone knows that equities have been doing really well, but if you look at US equities they have been annualizing at an incredible I think the S&P is around 15% annualized over 10 years, 15% annualized over 10 years. So it has been a pretty good strategy like just buying whole equities, it's difficult to beat, if you take if you take tech equities is even even better. So that's a reality and just generally global equities I think are 12% per annum and even in Switzerland. So in that environment it's very costly actually if you want to build a more resilient portfolio, if you want to just hedge your portfolio, find potential hedges for a crisis that may come up, what do you do? I mean in the traditional sense you would sell part of the equity exposure to fund your diversifier and that today means opportunity cost. If we continue to analyze at that level, I mean the one knows right, but there is an element of that that has been leading to the question how can I still add diversification without reducing my equity exposure. And then if you try to formulate that problem more mathematically it basically means okay I don't touch my equity exposure. So the only way to add diversification is actually to layer it on top. So you stack it, but if you stack it on top usually you would that implies that you increase your risk. Now in our view that's easy, that's basically leverage right, but in our view the interesting problem is if you add the constraint that your risk should stay equal and then it becomes very interesting. It's like the question is how much of a diversifier can I lay on top my equity exposure without changing my risk profile. I want to keep my standalone equity risk profile. That was the starting point of this paper and the paper basically is trying to provide a framework and an answer to this problem by making a few simplifying assumptions to actually come up with a clean solution and a cleaner representation of what is actually a good diversifier for my equity portfolio that doesn't come with the opportunity cost. And that's a pretty powerful concept in today's market because again equity markets do really well. Yeah because I read I don't know who said it but like the issue with our diversification is that it's a process of addition by subtraction and so portable alpha is trying to solve this. In the paper you argue that the cost of diversification is less appropriate of the diversifier itself and more about how it's funded. So can you unpack this decision? So basically when you so the cost of that of diversification obviously increases in a way with how much returns you get from holding what's the risk premium you earn from holding equities right and at the moment as we have seen like if you as equities you get 15% per annum. There is no diversified in this world. Liquid diversifier that will give you 15% per annum risk premium. Like the trendfilling risk premium what we do trend following historically has generated between 4% on average for the industry we have produced around 6% per annum over the last 20 years. That's pretty good actually when you consider as well with what type of correlation to equity market that comes with it right. Trendfilling performs in any type of market environment historically there's a good track record for that. So every time tried to fund your diversification by selling equity again if you combine 15% analyzed source of return with a 6% annualized source of return and you do 9010 8020 whatever it's gonna lower your return. It's gonna increase your risk adjusted return so the unit of return you extract per unit of invested risk is gonna increase but your absolute return is gonna get lower and that's a cost. That's an opportunity cost that you would not get if equities were returning zero then you don't need possible alpha to improve your return. The main point is that of portable alpha is that why can you keep the risk constant when you combine equity with your diversified simply because it's a function of correlation primarily. Using diversified is how they correlate with your with the equity portion that you want to diversify and the main result of the paper is very simple is that the more and it's intuitive right? The more negatively correlated the diversifier is to your equity exposure the more risk capacity you will free up. So if you combine a very negatively correlated asset with equities you can layer a lot of it on top of equities without increasing your risk because it is diversifying by definition. The issue is that the more negatively correlated an asset is to equities the more costly tends to be and the best example is long volatility you can always buy volatility to hedge your equity exposure but if you just buy for instance a vix future and you just buy 100% notion of that we know that in less than a year you have lost 100% of your investment. So it's costly in terms of expected return exactly so the diversifier will basically allow you if it's lowly correlated or negatively correlated with equities to add a portion of it on top of your equity exposure without increasing the risk. Now the key and as we are trying to highlight when it comes to okay what is a good diversifier? It's actually the tradeoff between the risk premium that the diversifier gives you and that should be positive ideally and how ideally negatively correlated that it is to equity and that's kind of in a way you would think that's incompatible like you get free diversification and you get compensated for that in a way but that's that's why trend following you so interesting. Maybe before going to trend following because I think that in the paper and I really wish that everyone goes and read it because there is a really elegant formula that you created to explain the capacity of this overlay that you put on. Do you think that there are other common mistakes that invest or make when they think about how much of a diversifier they can hold? I mean you already mentioned that the correlation is the key but the correlation is the key and what we show in the paper is that we are trying to, we tried to come up with an analytical solution to the problem of what is the maximum weight I can lay down to my equity exposure of a diversify and that's a function we show it's a function of correlation. The relationship tells you that the more correlated diversify is to equities, the less you can stack it on top of your equity exposure because you don't just free up enough capacity. So just a very simple example. Assuming you have US equity as your benchmark and you add your P-neiquity that still will be correlated, let's say it's 0.9. So you expect those two markets to be very closely connected together. So if you add it your risk is going to automatically increase that's leverage. So there is no, it's not a good diversify of course. Then if you take a Treasury's for instance so Treasury's historically have been the natural portfolio diversify a 60/40 portfolio why? Because Treasury's tend to be negatively correlated in risk of environment. So it's a very easy and natural hedge to use over the past 10 years has been less good because of high inflation, high inflation. We know that in high inflation environment, equities and bonds tend to correlate positively. If it correlates positively that's more akin to again more leverage. So dangerous. The danger of these kind of overlay solutions is obviously that you create a overexposure. That is that you lay on top and ask it that we'll go down at the same time as when equities go down. Where trend following is different is that we know historically that trend following tends to become more negatively correlated to equities in crisis. In times of crisis it's one of the few strategies that has this nice behavior. Meaning that when the crisis hits so when equity markets correct we know that the longer the crisis goes the more diversifying trend following becomes. That means that however you size your trend following exposure based on the long term and that's what we do in the paper. Basically we assume we make a very simplifying assumption that correlations are long term constant which is wrong obviously but it's just a purpose of coming up with a simplified model and you can see that with trend following you tend to actually underestimate the capacity you can lay on top because that capacity tends to go in times of crisis. So that's a nice thing to do because I back to your question what is something that you can go wrong is obviously adding an asset that's going to correct. That goes down at the same time simultaneously than when your equities go down. You want to absolutely avoid that. You can't eliminate that risk. I mean it's part of every portfolio construction but the whole point is to choose an instrument that complements your equity exposure in a structurally robust way over the long term and that means understanding how correlation characteristics of that assets behave relatively to equities over the long term and trend following has luckily a more than 40-year track record of delivering pretty good returns in the worst crisis like GFC. The GFC 2008 was the prime example. Yeah because someone might say no you run a trend following a shop so how would you respond if someone would raise the kind of conflict of interest in writing a paper where a trend following comes out so good. I mean of course we are biased right. We are trend followers by conviction. We have been a trend follower for 21 years. We truly believe in trend following and we have always been I would say a high conviction trend flow because we haven't done anything else in 20 years and we have always been invested quite heavily in in what we do. I mean we started actually a Quantica as an Opelate solution so it's not something new to us. What is new is that investors are looking into this a lot more actively. So I would say in our case the idea was to demystify in a way what it means and what is a good way to build a portable alpha solution that doesn't go against you at the worst point in time. And so I think trend following is certainly not the only solution that's what we actually showcase in the paper. There are other assets that fulfill those criteria of a good portable alpha candidate and we mentioned treasuries we mentioned gold as examples and there are many others actually. There are not so many actually but there are a few. What is a good candidate? I think ultimately what we want to show is like okay you want to build a portable alpha solution. You want to you're interested in this. What are the candidates? Let's let's ignore trend because yeah you may argue if we are trend flow is maybe where that's okay but then consider fixed income or gold. What is the commonality between these assets? They are long term uncorrelated to equities. Maybe not in a short time like we mentioned fixed income has been more correlated and the higher the inflation the more correlated it is but long term we see that the correlation is basically not not very large and when you have risk of events fixed income still tends to be to be diverse. So that's one one part the correlation structure and the other part is the risk premium. So what is the risk premium that you earn from these diversifiers and why is it positive? Do you understand why there's a positive risk premium? There's two sides of trend following. I think the the correlation structure of trend following you get it by construction unless you screw up the portfolio construction if you design it properly this is by design. The question that is open for debate and really remain a debate is why do you earn a positive risk premium and that if you don't believe in it you shouldn't actually do trend flowing because then you will not generate any benefit. For treasuries from gold it's the same but ultimately a good portable alpha candidate is one that has a structurally low correlation to equities over the long term and that has as well long term a positive expected risk premium. I think from from my perspective the risk premium side you can predict it. Today over the next 10 years we cannot say I cannot say which one of those candidates will perform better and actually for trend following fixed income gold or any other candidate we don't know but what we know I mean we have more conviction in the correlation aspect. I think it's more likely that these assets remain independent and we believe long term that they will deliver the risk premium. So when you have that I think the best is just to combine these instruments together maybe equal risk and that's actually a quite attractive structure that we have done for some clients as well so we do have as well we have run for instance portable alpha on gold rather than equities. Why? Because gold, trends and equities are mutually independent over the long term. It's all about that. Actually investing doesn't need to be too complicated in the end. If you understand the correlation picture if you understand what is the source of the inefficiency that you are earning over the long term and if you accept that you cannot predict the returns that one asset will deliver over say three to five years then just mix them. If you have a strong hypothesis and if the correlations characteristics are robust it will be a good portfolio. And then I guess like the third element that you did mention is the volatility of the assets but I guess that like the way that I explained to me is like well once I introduce leverage I kind of can scale the volatility of each asset where I want to be because normally like the issue with stocks and bonds is that the bonds are like really low volatility and if they if you take them unlevered then the combination is not good for the investors because it drugs down the expected return of the portfolio. Exactly so one thing that we haven't touched on is obviously to so the classic way of diversifying to sell equity to fund your diversifier that's straightforward but it comes with a downside if for instance you want to allocate two bonds which has much lower volatility you are you need to sell even more equity in terms of your notion exposure to just build enough bond exposure. That's the other benefit of these stacking structures. To implement it you can't buy cash securities so you need to buy capital efficient instruments and that is now in in our case its futures futures is a very efficient instrument and that allows you basically because it's a margin instrument so to to build a hundred dollar risk exposure you only need typically a fraction of the capital which means that to build your hundred dollar equity exposure you can either buy an ETF you buy an ETF you need a hundred dollar you consume them and that's it and and and you have no capital left. In the case of Portable Alpha you have to buy your equity exposure through a future or a swap for that purpose or any any equivalent structure and let's say to buy a hundred dollar global equity exposure you need let's simplify you just need 10% so 10 dollar. You are left with 90 dollar that you can still invest and that's where we said before taking trend following or any other diversifying asset but the interesting thing is if we go back to what we said initially if you invest the amount in your diversifier that is that maximum scaling capacity that does increase your risk then interestingly you could in some cases and that's that's kind of structure we have been interested in building is like you you kind of can allocate another hundred dollar exposure without without actually spending an additional dollar of capital right and and that's the that's kind of the beauty of it if it's done well. Because again the danger is. is if this additional $100 is leverage, that's what you don't want, right? You want a diversifier that reliably creates that additional risk capacity without increasing your initial equity exposure, right? - So there is a point that we touched in our prequel that I really want to ask you again, because I think it's really interesting. Typically, retain investors sees turbulent times as not great for trend following. But you said that short-term crashes are like the tariff turn-through or the beginning of the war in Iran is what creates the opportunity for the long-term trends. So can you elaborate on this? - The positioning of the trend follow always reflects the prevailing macro narrative, because you follow trends. And so for trend following to make money, you need a stable macro narrative. If it's changing all the time, there is no trend to build on. Now the most stable that macro narrative is, the more you expose yourself to shocks. The biggest risk when you run trend following is that something hits the market that no one expects. And markets are always pricing some form of macro equilibrium. There's always a narrative price in the markets. And trend-filling strive when that narrative is just the strongest and the most stable. Now, Liberation Day, so the unexpected terrorist announcement or oil price shock of March this year, they are the prime definition of shocks. If you understand how trend-filling is built by definition, the one day after this shock happens, usually it's negative for a trend-following because all the markets will reverse. And the macro narrative valid a day ago is completely off. So that's only natural. Actually, that's one way to explain the risk premier that you earn with trend-filling. Why is it that trend-filling produces your long-term, uncorrelated return to equity, it's too fixing, come to actually any traditional portfolio risk factor. And on top of that, you annualize at 4, 5, 6% per annum. That shouldn't be. You shouldn't be able to get protection and not pay for it. And you're actually being rewarded. Why? Because you're actually taking a significant risk is that as soon as you detect trends, you build up positions and you will build up risk proportionately to the strengths of those trends. And it's intuitive as well to think that the stronger these trends are, the longer they last, the more likely there will be some reversal. And usually we know the more stretched markets are in one direction, the more abrupt and shocking the reversal is. So this is what liberation day was. It was pretty insane actually, the moves that we witnessed. Yet the negative returns of trend-filling were actually not really surprising. And then there are two scenarios from there. Either that reversal creates a new type of crisis that is entrenched and then trend following will adjust because you follow the trends. And a shock like that, the one thing that it does, it triggers a transition to something new. Usually you don't go back to the macro narrative that was valid before the crisis. And that is where the opportunity usually is for trend. Because the bigger the shock, the bigger the likelihood that we're gonna move to some sort of new macro narrative, new market equilibrium if you want, or cause asset picture, that will be radically different. And then the second part of this is that it takes time to figure it out. Markets don't reprise that overnight. It can take a long time. And we have seen that in 2021, with the inflation shock, it took the market 18 months actually to correctly price in the thermal rates, like US rates went from zero to five percent that is for a 10 year. It took almost two years. And that's a fantastic opportunity for trend-filling. Because trend-following is actually striving on capturing that opportunity, that mist pricing of markets for a per-range peer of time. That's why Liberation Day may have seen like the end of trend-following. That's what we heard here and there is like, well, in a world where the guy in charge is changing his mind every week, it cannot be a great environment for trend. So trend is definitely not a strategy you wanna be exposed. History tells you exactly the opposite. If you go back in time and we published actually in the weeks following Liberation Day a research paper titled "From tariffs to trends" where you just wanted to pass on the message, those shocks historically, they produced the biggest opportunities and fast-forward to today actually the second half of 2025 was one of the strongest half-year records for the strategy in over three decades. Who would have seen this coming? Well, actually, trend-follows would have seen this come because the history repeats. And just one thing that we can say about Iran war, so it was less extreme in terms of market reaction because trend-follows were already positioned on the long side. I mean, basically the being-long energy helped tremendously diversifying the maybe some of the losses that were experienced on the equity side. But here again, we go again, like this oil shock is gonna create some new cross-asset picture that we probably don't even know for sure where it's how it's gonna look like. We see interest rates have two big impacts actually of this oil shock so far is that we have seen rates really going up again in a way we haven't seen since that 2022. We see obviously the role of commodities continuing to be absolutely critical. There have been a lot of opportunities in commodities for trends over the past six years, actually, since COVID basically, and that continues to be the case this year. So we'll see. So I have a personal question, because the paper is focusing, I think, on 21 years. So do you think that is enough data to provide a reliable test? - Well, it's a very good question. You never have enough data. - Yeah. - So when we write the papers, the idea is to illustrate concepts. The way we write these papers, the topics always come out of client conversations. We have been writing one paper every three months for the past six years. We always think about it from our side. How can we make a quantitative piece? So we just not just talk high level, but make it quantitative. So you substantially, actually, you have use by proper analysis. Now, in terms of the look back, there is a trade-off here. If we just talk about the last three years, you can arguably criticize us of just cherry picking the period. If we go beyond 21, I actually didn't remember it was 21 years, but usually we don't go back further than the year 2000. Simply because what we tend to see is that if you go back to the '90s, and we have data set going back to the 1960s, because when you look at futures markets, they emerge in the 1960s, like in the US, you had agricultural futures. But what you always see is that the results just get better and better and better. What you can then easily be accused of is just to make your results look better. Because the markets, when way less efficient, maybe as well, yes, we know that trend-filling, worked phenomenally well in the '80s, for those who had access to the market, I mean, the data back then wasn't that easily accessible. So that's the reason why we tend to not do it, and as well, the universe was small as you did with all the problems, is that, you know, you would have to select a small, as a subset of markets and so on. So technically speaking, not easy, that said, you know, there's a lot of parallels actually between today and the 1970s, like in terms of the inflationary environment. We believe actually that the results being able to analyze the behavior, the cost asset behavior in the 1970s is actually interesting in terms of evaluating the potential scenarios that you may face as well in the future today. And one thing that we said after COVID, we used 40 years of data actually for that paper to make that case what you see over those 40 years that every time the opportunity set in one asset class shrinks for trend falling, the opportunity set and the other asset classes expand. And that's the beauty of it. And the road that we wrote it back in 2020 because we came out of a decade of declining interest rates and back then, the main question mark that investors had is, look, we know how you make money. You make money on declining rates. But what will happen when rates increase? That's another criticism. Wasn't another criticism of trend following actually. It's like, it's great just because you have been just writing the declining yield trend, the QE environment of the 2010 decade, right? The structure was really bad because if you think about it, that strategy is an overlay on the cash rate, the more the cash rate was going down, the less absolutely turn you add. And that's true. And so it's true. But that's true. That's true. That's true. Obviously, it was not an easy environment. It was actually the most challenging environment for trend following. If you take the period 2015 to 2018, that was pretty tough. But if you look at the post GFC decade until COVID, a big chunk of the returns, it's true, where by fixed income. And so investors naturally get worried if, let's say, 80% of your returns are driven just by one trade. With trend following, that's always the case. Actually, there's always just a few trends that drive your PNL. Now, fast forward to today, and that's really fascinating. And we are just about to-- that next paper will be exactly about starting on that observation. If you look at the return attribution, since the start of 2020 for the current decade, you can actually explain almost the entirety of trend following returns through commodity contribution. In the previous decade, that was negative. Comolities had a zero to negative contribution over 10 years. It reflects just one of the fundamental principles that you should follow when running trend following and when investing in trend following as well, is that you cannot agree or ignore where the trend opportunities will happen tomorrow. But what you know is that, if you're diversified enough there's always something going on somewhere and when the asset class has been or when the sector that has been your biggest performer in the recent past when that opportunity shrinks right somewhere else where new opportunities arrive and it's interesting to see that history repeats over and over again but it's a hard sell right to tell like in 2020 say yeah we believe that if interest rates rise maybe the opportunity is gonna show up in commodities and funny enough we have seen exactly that last six six years so we'll see if it's 20 if this decade will remain the decade of commodities but like last year and again this here commodities are central piece to trend flowing and it's actually if you think going forward what are the risks to traditional portfolios I mean you hear about stackflation risks right you hear about the potential comeback of inflation potential recession down the line you know the good thing about trend flowing is that you don't really care about what scenario is gonna the one thing we don't need to worry about is actually which scenario is gonna happen because we know and that's really based on how the strategy is built and just the track record the long-term track record that whatever is gonna happen the strategy is gonna adjust the positioning to that macro environment because it's agnostic to the correlation structure of the market that's its strengths so and trend flowing becomes all the more valuable that the positioning is not a strategic portfolio positioning that you would be able to take like being long commodity like being long strategic long energy over the long term is probably not that a good trade but tactically when inflation comes back you want to be long how do you get in and out trend flowing is a great strategy for that but then what's the one thing that you wish retail investor would take away from the paper that you suspect that will miss the most the motivation to write this paper was not necessarily just a portable alpha subject for a long time I was thinking how do you show the value quantify the value of correlations everyone can quantify the value of returns right you know everyone looks at returns but correlations yes you're uncorrelated great but what does it mean for my portfolio that's where portable alpha is actually a nice tool because if you basically you just transfer your standalone view on on trend following into your portfolio context and there suddenly trend following shines relative to other strategy suddenly the strategy that has three times lower sharp ratio than the then points in the equity long short strategy suddenly looks as good as equity long short because you take into account a correlation effect so that's one thing is that one thing is not stopping at just the standalone metrics and look at what's in it for me in my portfolio and trend looks always better and has a great track right actually improving the portfolio characteristics the second thing and I think that is valuable for any type of portfolio is that it's a great complement to your equity holding look at portable alpha well constructed portable alpha again I would be extremely careful about understanding what is being done like always investing in what you understand and and why it will deliver what it is supposed to deliver and again that's why trend following is a very reliable building block in combination with equities and then it's a great it's just a better way to build your long equity exposure you can buy an MSG world ETF great you do nothing wrong by doing that and or you can buy like what we call we call it enhanced equity because it is enhanced equity because if you do that well combining a hundred percent equity exposure with an exposure to trends gives you an additional rate excess return which is equal to the risk premium you earn to trendfully without increasing the risk without increasing the risk that means that you have a beta one allocation and you may have an allocation to MSG world you may buy into active equity managers you may do different you may implement your equity exposures in different ways but this is nothing else than just an equity building block which we believe and we invest as well personally that way it's just one piece of your equity allocation it's a great piece and and it's something that the the enhanced the enhanced return they are generated through that system at in a way through a systematic macro overlay which is friendful and well you call out thank you so much this is a message that we are really passionate about and I was really glad that you were able you accept that you come here to talk about it because yeah one thing is to hearing from me one thing is to hear for someone that does it and leave it and wait it so thank you thank you so much for the time thank you Nicola it was a great pleasure speaking to you today and yeah thank you. [Music]