Michele Cancelli - How QIS Are Transforming Investor Portfolios

00:00:07 [Michele Cancelli]
And this, I think, has been what has changed completely the way people are using QIS. Instead of sitting in the alternative piece of the portfolio, enter the core portfolio, be maybe the platform providing equity agent solution, equity carry solution, race agent solution, so on and so forth. And that is actually where QIS can be amazing.
00:00:30 [Vincent Weber]
Hey everyone, I'm Vincent Weber and this is Resonanz Spotlight. the podcast where we dive into investment strategies, explore the forces shaping the financial world, and most importantly, get to know the experts behind it all. Today, we are talking about quantitative investment strategies or QIS platforms. These are systematic, data-driven investment approaches that are reshaping how institutions and investors think about risk, return, and portfolio construction. And I've got the perfect guest to break it all down. Joining me today is Michele Cancelli, Global Head of QIS Structuring and Trading at Citigroup. Michele has been deeply involved in the evolution of Citi's QIS platform and he's seen firsthand how systematic strategies have grown into a core part of institutional investing. Michele, thanks for being here and welcome to the show.
00:01:21 [Michele Cancelli]
Thanks a lot.
00:01:23 [Vincent Weber]
All right. So, Michele, before we get into the nuts and bolts of QIS platform, let's start with you. So, what got you into this field? Were you always drawn to quant finance or was it something you fell into along the way?
00:01:38 [Michele Cancelli]
Thanks, Vincent. So, I would say I already appreciated the quantitative fields, mathematics, physics, mainly when I was there. younger. And when looking into what to do in life, specifically coming from a really small village in Italy, I was looking for something dynamic that would allow you for having like a potential big growth in your career path and meeting those international people and having a overall, basically, dynamic career experience. So finance seems to be back in time the case. Maybe nowadays it's actually technology, but back in time was finance the way to go. So that was basically what brought me there.
And specifically related to the field of trying to do something more related to quantitative models was to look for something where you can apply, call it, Applied research to your job, so searching for new information in the market and trying to deploy it in some way in strategies or models.
00:02:51 [Vincent Weber]
Thank you. So I love really hearing how people find their way into this world. So now LeadingCity is a QS platform. I imagine some of your past experience really influenced how you think about the business. So what are some of the big lessons from earlier in your career that have shaped your approach to running QIS today?
00:03:12 [Michele Cancelli]
Yeah. So, I mean, I used to be, I started at J.P. Morgan, then moved to Goldman Sachs, where I was leading the equity volatility for development team, together with my back-in-time boss, Brentin André. He's also my boss nowadays, by the way. That was the leader of the QIS platform over there. So really, I think my current role really shaped eight years ago when I joined CP. And the early stage of, I would say, my career there was around deciding how to differentiate at the very beginning versus versus the street and some already existing platforms specifically thinking on how to play on our strength so my strength has been always volatility products specific inequities and then cross-assets. So first thing first, you start from what you really know well.
And this is why also the very beginning, the platform of Citi was tilted towards volatility products. Secondly, it was trying to provide also differentiating. and alternative to people as well to go for other providers and not to simply come up with me-too products and me-too concepts somewhat. And obviously, at the same time, onboarding some of the good concepts that you learn through your career and also be followed by the right people, as indeed it happened also with my boss. So we brought here some logic and some basically Some business logic, I would say, that made sense. For instance, the moment we started here, we created a full cross-asset team integrated in one pod, creating structuring under the same hat, somewhat designing the full ecosystem really well, even before starting working on it.
00:05:07 [Vincent Weber]
Okay, right. So let's set the stage for our listeners. So if someone's hearing about QIS platform for the first time, How would you describe them? What do they do and how do they fit into the broader investment landscape?
00:05:20 [Michele Cancelli]
Yeah, actually, given the popularity of QIS lately, I've been asked this many times. I would say I'll describe it as such. So to me, QIS is no more than the aggregation of simple transactions in one place now in one index okay an index ultimately could be seen as an application that is simply doing a lot of things for you each in each single action being something really really simple then that rolls up into an app or something that actually is really simple to use for the end user and investor so really let's say that qis start from data basic data typically option data single stock data swap data raw data And simple ideas of maybe I want to buy a call option every day on those strikes or sell a put option on those strikes.
And instead of having someone executing these every day on screen for the investor, you simply say, what about I wrap this up into an index? What they say, multiple now decades ago, people created by when creating equity benchmark, S&P, Euro stocks, ultimately those are just indices aggregating. the purchase of single stocks into one index qis is this exactly the same simply on slightly more sophisticated transactions or maybe more tilted to our options for instance and then basically the popularity of it comes from i would say there is i mean the way the the the fact that is really easy to use as a product because you are able to wrap extremely labor intensive
00:07:09 [Vincent Weber]
simple yet and hopefully strategies into one single transaction what do you think is what the forest is okay so essentially it's about using data and technology to be investment strategy that follow clear rules or systematic strategy how does it work actually behind the scene i'm trying to i mean we put for a listener to to get an idea of how
00:07:39 [Michele Cancelli]
the factory looks like from the inside if you could walk us through the process from from designing a strategy to executing it yeah maybe actually indeed going to the production cycle of one strategy allows you to understand really well the ecosystem so let's assume we want to develop a new strategy in a new field i don't know would be a option on on treasury futures let's say first thing first you need to have the data because you need to do some research so you need to start with data you cannot just start from from a blank piece of paper data first so this is the most important item that everybody needs to dedicate time to clean to have it available in the right format be able to also use it afterwards then you're going to
have some dedicated people that we call pro developers which are ultimately you can see them as you're going to have we run the business as a cost asset of and volatility development and across a set of linear development rolling up into a global head of development and then under those people you have the specific specialist for each single asset class so you're gonna have the linear commodity specialist the volatility effect specialist so on and so forth so it's like like a bit of a tree now so you say this especially for the specific field will start using the data to test potential intuition. So how do you come up with an idea? It's up to you.
Meaning that where I used to be a post-developer back in time on equity volatility, you can come up with ideas simply because you're staring at a Bloomberg screen and you see patterns, for instance, around intraday delta hedging, and you start thinking, hmm, every day at Canzela, the close these things get. get additional acceleration to the close due to potentially some people trading on the back of it could be you're reading research academically so you're speaking to colleagues many times you have a lot of interesting ideas be simply by understanding what people do on the floor and what clients do on the floor and also ultimately as well academia and most importantly i think of all is speaking to clients because ultimately the starting point is What is the need you are trying to address?
You're not trying to create something in a vacuum. So speaking to clients, you find that maybe multiple people have the same problem and you try to find a solution. So you then, with the data, try to test your idea. Example, I believe that selling treasury official options provides a positive carry. once someone put it on and Delta Edge it and create and try to monetize this basically implied manualized premium. So I sell options, I collect something and I need to pay out something else. So hopefully I would make money by doing this. That's my assumption. You test it and then you need to go to the full rigorous testing. Typically for a flagship strategy, you might do, I don't know, hundreds of tests of do I pick calls, puts?
which maturity which 10 order so which channel which which which which which underlying to start how do i also size those positions because assume i want to focus on call options above 30 delta sure How do I size them? Do I put equal notion, equal V, gamma, gamma weight, whatever. So you need to test all of these and then be also truthful to yourself about your testing. Try to minimize the temptation of optimizing too much. Because in my simple example, again, I want to create a strategy that is doing carry on treasury option. So I sell those options, I collect the premium, and I delta edge them, and I pay up. This is my strategy.
Then by definition, maybe the 3% strike, one and a half more maturity is going to have the best sharp ratio. So you are tempted to say, hey, that's the one. By definition then, either because you don't have enough capacity or because you're simply data mining, maybe that specific path in the last 10 years was the one that was never hit. So it's the best strategy of all. So you need to be truthful, which is obviously. part of the process that is basically peer review, my colleague, the committee review of your strategy, and trading review, and ultimately client review, because you propose this to clients and you see the feedback. So assume basically, literally, month of work on this, you come up with one flagship implementation of a strategy.
I found all the parameters, that's the strategy set. And let's compute it. So I have a full 10-year back to solve the strategy. From this point onwards, the strategy will be live and alive by its own light. Then a whole step is there is a new team called Origination that instead will work with our sales force to bring those ideas to investors. basically bringing those ideas, some marketing material and some specific land discussions where you say, hey, this was my idea. What do you think about it? Does it fit your portfolio? Does it suit in terms of risk you're trying to capture? Hopefully the answer is yes. By the way, this phase is also long.
It could be that I see all the time you go with flagship strategies to clients the same way you go. When you are in a restaurant, you eat at the menu. No, you need to know what is our suggested means, what is our suggested desserts, and what is our suggested drinks. Obviously, then people can add and change parameters and ask him to remove something or to add something else. So the origination process is also really dynamic. But assume that then you end up with a specific implementation. You get a trade, basically. And then the implementation of the trade pass through. the official implementation of the strategy by our independent calculation team that is completely separated to the business so that you don't have conflict of interest.
It's also regulated by multiple regulators and rules, specific rules and internal governance committees. And once you have the official index up and running, then you can trade with our trading desk, which will offer this exposure in many formats. It could be funded format, derivatives format. option format to the end investor and the praline desk is basically executing you can see that it's executing of vr for clients the strategy okay so and that's this is the security stuff because obviously then you would have posted services about this the competition to clients you can have modification of this strategy additional other strategies and continuous development because once you develop something I would say the typical life cycle of a strategy is in the range of some years.
You don't, I would say, if I think about what was the flagship way of doing volatility carry 10 years ago, has nothing to do, inequities let's say, has nothing to do with what is now a flagship way to do volatility carry. Nothing to do. None of the old strategies are actually considered at par. implementation nowadays in any bank. And the reason is it's a bit like evolutionary technology, that you don't buy a 10-year-old phone and think it's good enough for nowadays standards. Exactly the same concept.
00:15:35 [Vincent Weber]
Right. You briefly touched upon customization, and that seems to be a big topic when people talk about QIS investor. And I mean, really, the investor has such different needs. You have pension, hedge fund, your private client, So how do you go about customizing and be able to fit all these different objectives? So is it just like tuning a few parameters in the model or is there more to it? I think there are two ways to see this.
00:16:11 [Michele Cancelli]
One is sometimes we just act as execution of ideas directly coming. And sometimes we don't even need to know what the idea is behind, meaning you have maybe clients can trade a portfolio of, I don't know, 200 instruments, futures, options, whatever. They give orders and they create their own team without even us having to inquire what is their investment objective to start. This is what would be called like managing the series and we have a platform called Atlas to do this. In that case, it's none of our IP to be deployed there. This obviously is a part of the business. I don't think it's the majority.
Typically, on the customization, as I said before, you need to have your flagships, in my view, simply because you want to make sure you have a stronger, you need to have some some idea of what you're doing. And you say, if you want to do vocabulary or if you want to do edging, this is in specific asset class, this is our two solutions. This is the way we will do it. And this is important because you're supposed to be truthful to your style of development, the style of the platform. Everybody has some different ways to implement things and you want to show, basically, what is your style. Because maybe your style is not for everyone and maybe some Some people might like particularly the way you develop strategies.
But then customization then comes, as I said, on an ad hoc basis with, I would say, nowadays, I cannot tell you the exact percentage, but I would say at least 50% of the customers, they end up in specific variations. Now, the variations are maybe modification of the flagship to fit specific portfolio needs. where I think we have the most entertaining discussion and when indeed we are somewhat the expert providing the solution in typical execution. So it's also part of the evolution of the business in the sense that up until, call it 2019, pre-COVID, this business was labeled as risk premium. You might have heard this term in the past. banks were competing more against hedge funds or the pitch was more the Arialman account, don't invest in hedge funds and pay 220.
You can do it in QIS, it's cheaper and you can get the same. Now, this pitch was a bit short-lived in the sense that by definition, then you end up putting in a portfolio 50 strategists trying to be a multi-strat. But because you are not a fiduciary, You cannot actually change every week the strategies, the balance, the risk management. It becomes an extremely complicated process. And also, by the way, that's not the work of a non-fiduciary. So fiduciary, as managers, hedge fund, they have indeed people dedicated, portfolio managers dedicated to risk management. While banks, by definition, they are broker-dealers. So we sit in between. While instead where we are the expert is on execution. By definition, banks, they have amazing pipes for executing everything.
I'm thinking about the previous question about when we set up QIS here. We, of course, also leveraged on our amazing rates franchise and fixed income franchise for execution, access to liquidity, so on and so forth. That is where actually we can be entertaining infinite discussion around when to execute, how to execute. What time, how many times, size, capacity, so on and so forth. So that is typically the input from us that doesn't really change because we say that's the capacity that we can do that what we suggest to do. And then instead, from a client perspective, it's more the design of the risk. I want to be more skewed into downside pools. I want to be more skewed into something longer gamma, shorter digga, whatever.
Then becomes more of a fine tuning of the risk, which indeed is more a fiduciary decision. but it's more sitting with clients.
00:20:35 [Vincent Weber]
So what's really been driving the demand for QIS investing post-COVID year? You mentioned, is it the need for exclusive surveys?
00:20:45 [Michele Cancelli]
Yeah, it's been, I think, the change in paradigm that we saw in the way the investor saw QIS. So I give you our own experience because, as I said, Nowadays, we are a leading platform in this business. Obviously, at the beginning, we had to build. So, 10 years ago, we had to start and build the franchise. We experienced this. One of our differentiating points was to say, use QIS to do portfolio overlays. and don't use it as a, I will say, cheaper multi-struct hedge fund because that's not what banks are good at. But instead of sitting in the alternative piece of the portfolio, enter the core portfolio, be maybe the platform providing equity agent solution, equity carry solution, raise agent solution, so on and so forth.
And that is actually where QIS can be amazing because If you, let's say, even if you're an edge fund, you're a multi-struct, you have multiple already pods, all made of really competent individuals. But it might be that at the top level, you have risk coming at you that you want to systematically edge. You want to be longer gamma on the downside in a minus 10% scenario of S&P, example. So that is where a portfolio overlay, systematically implemented, can really help you. And this, I think, has been what has changed completely the way people are using QIS. And this is why, for instance, the biggest thought that grew in the last three years, let's say, exponentially, has been the usage of HFAN or QIS.
In the past, I would say pre-COVID, you couldn't have a QIS discussion with any HFAN, pretty much. After, when people understood what this is, it's actually the average cheaper way to implement portfolio delays at more higher level of risk management or a cheaper way to implement a specific strategy then this ultimately makes a lot of sense and also with the increase of usage of more complicated products option shorted adoptions i think about i know our dispersion index that has been really really successful is managing and thousands and thousands of options. By definition, doing this in-house implies having maybe two dedicated traders. So it makes a lot of sense to source it. The moment you go to short-rated options, zero-day, one-day, two-days, with intraday hedging, this becomes a proper full-time job of someone.
And maybe it does not make sense for everyone to have a dedicated PM just doing a one-day S&P option. and delta aging them throughout the day. But you can actually, at the right price, you might be really, could be economical, really economical to source it. This has been basically the could be changing paradigm. Then as having said that, let's not confuse QIS as a passive benchmark tracker. So we're not speaking about strategies which are, excuse me, my French, dumb. They're not. They're actually quite sophisticated. There is a lot in it. We bring a lot of intellectual property. So the intuition should be simple, but then implementation maybe is particularly sophisticated. There's a lot of research and studies behind.
But it's different from saying invest in my strategy because it will deliver to you for sure three sharper and it's going to be competing with the best multistruct out there. is more different. The proposition is really different. It's more to say, use it as a portfolio overlay.
00:24:48 [Vincent Weber]
Thank you. Very interesting. So, of course, with any investment approach, there are challenges. So, what are some of the biggest misconceptions or concerns your investors have about QS? So, I would say, I would say,
00:25:08 [Michele Cancelli]
The first concern typically comes with transparency. So people which are not used to the product, they might say, okay, I buy exposure to a specific Bloomberg ticker. What is inside? How do I know that I'm in control of the risk? But the answer to this is typically, let's look into this and let's have specific risk reports because by definition, you can have full look through on everything you're trading. And we provide this really proactively to clients. And so this is the work I was mentioning before of origination. You start with an idea and then you give a look through analysis, portfolio analysis on specific scenarios because clients could have specific risk scenario either for regulatory reasons or for non-portfolio management reasons that they really do care about. So you might want to do this.
This, I would say, addresses the first concern. I will not call it a concern, really, because maybe the second concern could be, I mean, you can call it a concern, let's say it could be cost, but cost, again, same stuff, is fully transparent. Everything is there. So there are no hidden fees. Everything is disclosed. So as long as you're comfortable of the charges at single instrument level and overall charges, as I said, people are smart. They can simply make the math of how much would cause them to do in-house versus not sourcing. Comparing banks, this has been really, I would say, easy as well for investors to address. So yeah, those are, I would say, maybe the two main concerns, but it's part of our job to actually address them in the cleanest possible fashion.
Because as I said, everything is transparent. Those are not meant to be black boxes. They're meant to be extremely transparent, fully replicable indices, basically.
00:27:04 [Vincent Weber]
Great, thank you. So looking ahead, where do you see QRIS platform evolving in the next 5 to 10 years? Are there any trends or innovation you're particularly excited about?
00:27:18 [Michele Cancelli]
So, yeah, look, Ozzy, I'm biased. So it's somewhat in it. So by definition, I'm bullish. But I was actually asking myself last year. We experience, the industry experience a large growth. I would say probably the wallet has been growing for the last four years in a row systematically across the street. We specifically have an exponential growth for our platform. And I was almost wondering why for the industry. What is the reason of this? Because this has been obviously us, industry, and us, CTO, for doing a really good job. to invest in solutions, but also sometimes you feel like it's almost happening to you. There's even an over-demand of this. And you're asking why. First explanation I give to you is technology.
Meaning, if I think about how a flow trader or an exotic trader in various divisions or the traditional trading desk are run nowadays versus 10 years ago. I don't think the difference is that big. So the way they see that is, the way they act, the way they execute is not that different. They will also interact with clients. If I think about QIS 10 years ago, maybe 12 years ago, let's say, was Excel-based, was a few tabs of Excel, data, some VLOOKUP, index match, whatever, whatever, and you compute an index, it was maybe trading two rolling futures. If I think about QIS today, but even today. five years ago, you're speaking about hiring data scientists that are managing non-sequential databases. Those people then have been interacting with extremely sophisticated product developers.
Excel is not even open anymore. So if I think about that, I say technology, because we are able to deliver a product that is simply not matched, has not been matched by the evolution of the rest of the flow that i observed this is number one and and that's my first explanation and the second explanation is that the domestication of of the product itself as i said multiple times in in while originating transactions We used to have the feedback of, ah, this is too expensive. This is a black box. This I don't understand. This is a back test that it will never work after. But this, really initial comments, which are obvious because people should be skeptical, no? It's a bit like with crypto, no? First comment is, it's worthless and useless.
Then maybe you go through it and you understand maybe that there is something behind. So what happened is that simply people went to spend the time and now the QIS product is adopted throughout the full industry because we see family offices, asset managers. So we have nine accounts that were always there. So insurance, pension, asset managers were there all the time. But hedge funds, as I said, family offices, retail, different products obviously for retail, but you have the full spectrum. I don't think there is an investor that does not know the concept at least of QIS. And this has been, of course, part of it had been the good job that the industry did. On the other side, I think it's also the people seeing something in it. And this is what is driving this.
So from here, if I look at the future, you say, okay, what would be the drivers? Number one, technology. Technology and I would say sophisticated execution of complex strategies. Is this going away? Don't think so. if any will continue to proceed in this way think about again zero day option one day option they used to be non-existent five years ago now they are also existing with different classes point number one so that's not going away so that trend will continue and the second trend is as i said adoption adoption by investors of the concept so people spending time understanding the technology behind and understanding that actually it's not the black box at all and it's not that expensive neither.
So those two elements are not going away because the moment people did the investment, in theory, it will continue to reinforce itself. And it's no longer the business concentrated in few ports. It used to be maybe concentrated in some specific regions, some specific client types, up until, let's say, five to seven years ago nowadays your hand as a client's handle client type totally different client needs really balanced classrooms as well like if i look at our platform is pretty much equally weighted between like edging and carry pretty much literally not not exposed to a specific risk factor you say okay if the mode of x goes to 30 everybody's down not really it could be someone is up a lot so it's it's a platform the way it should look like.
00:32:47 [Vincent Weber]
Okay, great. Thanks. Guido, this has been a great conversation. So before we wrap up, I want to take a step back and ask a more personal question. Over the course of your career, what's one big lesson you've learned, whether about investing, leadership, or even just navigating this industry that you think our listeners should take away?
00:33:09 [Michele Cancelli]
I would say the first one, obviously, which is quite maybe naive, is learn from the mistakes and don't think you're always right. In the sense that obviously, specifically when you're successful, you take momentum as a team, you might be complacent and thinking we're going to make everything right all the time. So be always hypercritical about what you're doing. Look at what other people, competitors, don't be dismissive. It's really simple, it's really easy to fall into the trap of thinking you're the best. You're winning all. Everybody else doesn't understand anything. So instead, think that maybe in the small bank somewhere, someone actually, there's going to be some extremely talented individual you want to speak to.
And actually, if you hear that someone in another really small competitor is succeeding in a region, don't think, I know, for sure it's not true. Think actually, wow, maybe we missed out something. Be humble in the sense, and it's not only, as I said, not only learning from your mistakes, but actually think that there is always something to learn from everyone, clients, but also competitors. The second item is instead, I would say, more related to team building and also creation of a brand in the sense that I think it's always a struggle to try to differentiate yourself, but You need to try to create, I think I learned today, to try to create something that people can speak about, people know, people can differentiate for something else.
Put some rules, some, let's say, frame around what you're doing and offer this frame to investors and offer this frame also to people internally to work for you. And by definition, this might not work for everyone, meaning that you might invest not liking you might invest loving you. in the amount of people loving working for you and people that don't like to work for you. But it's okay. As long as you are able to continue to grow and differentiate yourself and have a specific style and also be true to yourself. Like don't keep deviating your style because you heard that someone else is doing something else and you try to cover all the battles. So sometimes you lose some battles. Focus more on the medium to long term of the overall, I would say, effort.
00:35:34 [Vincent Weber]
that's i would say maybe are the two main lessons learned throughout michele this has been an absolute pleasure thanks so much for taking the time to join us today and to everyone listening Thank you for tuning in to Resonanz Spotlight. If you found today's episode insightful, make sure to check out resonancecapital.com. We've got a wealth of resources on QIS, including articles, white paper, and index data to help you dive deeper into this space. Stay curious, stay informed, and join me next time as we continue exploring the world of investment strategy and the people shaping them. I'm Vincent Weber, and this has been Resonanz Spotlight. Until next time, take care.