The CIO Agenda: The Rise of Quant Credit

Paul Kamenski and Robert Lam join The CIO Agenda to discuss the alpha opportunities on offer from a systematic approach to credit.

Please accept statistics-cookies to listen this podcast.

Quant credit remains a relatively niche investment area compared to quant equity

The attraction of quant credit is clear. Fixed income is still dominated by large, slow-moving, buy-and-hold money and unlike equity markets it is not yet heavily trawled by quant managers. Consequently, it is a market rife with inefficiencies and quant strategies exist to harvest the opportunities that such inefficiencies throw off.

Quantitative investment strategies are increasingly using the growing amount of data generated by the fixed income and credit markets to deliver uncorrelated returns and eliminate the human biases to which discretionary investors are susceptible.

Paul Kamenski and Robert Lam, Co-Heads of Credit at Man Numeric, join Sandy Rattray and The CIO Agenda to discuss why a systematic approach to credit is moving from the niche to the mainstream and the opportunities for alpha on offer.

 

Recording date: 9 April 2021

 

Episode Transcript

Note: This transcription was generated using a combination of speech recognition software and human transcribers and may contain errors. As a part of this process, this transcript has also been edited for clarity.

Sandy Rattray (00:05):
Hello and welcome. I'm Sandy Rattray CIO at Man group. Welcome to The CIO Agenda.

Today, I'm joined by Paul Kamenski and Rob Lam, who are co-heads of credit at Man Numeric and we'll be talking about the rise of quant credit.

So let's kick off – Paul, systematic strategies have grown enormously in many asset classes, but fixed income. And specifically credit is only really getting going now. So what's taken so long?

Paul Kamenski (00:34):
I think things have really taken a long time because it's a real challenge. Two things that stick out to me particularly are the overall market structure, as well as some of the data challenges. So just to touch a bit on the market structure, corporate credit is an OTC market without that sort of depth of liquidity that we've seen in other asset classes. And that means that really trying to understand liquidity, trying to understand transaction costs is a real challenge and a hurdle, especially for systematic strategies where that's a key input to the process that can really dictate what comes out in terms of the overall portfolio.

On the data side, there are real challenges there as well. I think if you're interested in a particular equity or a particular company it's very clear which equity to buy, which stock to buy, not so much on the corporate bond side, you have an entire capital structure, oftentimes with multiple line items, different subordination, different durations, and less clear issuer curves to try to identify pricing even within that same capital structure. So at this point early pioneers, I think, have really tackled those challenges. And it's an exciting time when there really is some track record now for some of those earliest adopters.

Sandy Rattray (01:40):
And Rob, we've heard some of the challenges in systematic credit, but what's the exciting bit, what's the value proposition, where's the opportunity for investors?

Robert Lam (01:50):
So here's a stat that I think highlights how different a fully systematic approach is. 99% of high yield assets are managed using traditional discretionary processes. And when I think about that inherently, when the market composition is so one-sided, inefficiencies and distortions are naturally created. And so we believe that there's a really large opportunity to go against the grain here. So not only do we believe that differentiated investment processes can lead to differentiated outcomes, but also there is an opportunity to improve at each step of the investment process that people are using to invest into these corporate bonds. So everything from the underlying data, the forecasting of returns, risk modelling portfolio construction, right down to what the model picks in terms of the final CUSIP that is being bought and sold and bake that fully into a systematic process. But I think you're right, Sandy, it is hard and it's hard to exploit these inefficiencies in the credit market because of some of those complexities in some of the systematic processes steps, right? And includes everything from, you know, being able to capture those key drivers of credit spreads, to modelling the cost of execution and the probability of fill, which are actually quite unique to credit. So taking a step back, I would say the challenges definitely exist. They exist in the data. They exist in the quantitative modelling techniques, but at the same time can be very fruitful to tell.

Sandy Rattray (03:18):
So in a way, the fact that there are these challenges, that's the source of the opportunity. Well, let's talk a bit aboutsort of where we go from here. And Paul you've been involved in quant equity for many years and Numeric as an investment engine has been investing in through quantitative strategies and equities for over 30 years. So what do you think the trajectory for quant credit is if we've got 30 years of experience of quant equity, is it going to look similar?

Paul Kamenski (03:47):
I think in some ways, yes. And I think in some ways, no, I think the trajectory is certainly clear. We saw a strong evolution away from voice execution and block execution on the equity side towards a much more efficient market where liquidity was more readily available. We're already seeing that in the corporate bond market, the protocols and venues have been in place for years. And if anything, especially through the work from home life that we've had for a year plus through much of the market. And I think generally seeing how that's evolved in terms of liquidity, if anything, we've seen a really strong pickup in terms of execution and liquidity, having this positive feedback loop of quantitative or systematic strategies, oftentimes needing that price discovery, needing that access to liquidity, to be successful in a more meaningful way. I mean, a very significant way.

Paul Kamenski (04:32):
And then once you have that sort of momentum in terms of having more and more liquidity, I think it feeds positively on itself and supports even more growth in this space. That to me is a clear thing that will be similar. I think what's different is it's going to happen much faster. We already know so much from the development in terms of staying at the front edge, the bleeding edge of systematic strategies and other asset classes that I think the pace of adoption, the pace of advancement will just be that much faster with credit. And I think we all know it's hard to try to stay at the forefront. It's a never-ending work, but I think that it, whereas things that might've taken five or 10 years on, for example, the equity side might take only one or two within credit. So certainly I think that the pace will be a bit different.

Sandy Rattray (05:12):
So slow start, but maybe a little bit more rapid from here. Clearly the case for running active strategies in fixed income and specifically credit is that there are inefficiencies that we can take advantage of. If the market was efficient, then why would you need an active manager? So maybe Rob, you can just give us a little bit of a colour from your experience of what are the sorts of inefficiencies that we're looking for?

Robert Lam (05:37):
Inefficiencies in the credit markets can, can certainly come from a variety of factors, right? They certainly stem from the fact that there's a very archaic market structure, how you trade corporate bonds as Paul was highlighting earlier through phone calls and IB chats, it's the predominant way of risk discovery, which is quite strange and different than other asset classes. Also, the investor base is very different as I highlighted earlier. So ultimately, I think what this means is that you have humans making human decisions and exhibiting human biases, which are inevitable in human processes. So I think a lot of the analysis that we focus on is around how do you exploit those inefficiencies? You know, whether that's momentum exploiting anchoring biases or informed investors signals exploiting investor segmentation, or even value signals exploding over or under market reaction. So I think there's a lot of different ways that you can exploit these inefficiencies that really come up in credit markets. I'm sure we'll talk about this later on in the podcast, but we think that there is even more interest and more differentiated ways to come up with some orthogonal alphas around alternative data, which is a relatively untouched area within the corporate credit universe. So being able to capture KPIs and consumer behaviours and competitive dynamics, I think a lot of these quantitative techniques are just simply out of reach for many discretionary credit investors.

Sandy Rattray (06:59):
And let's talk a bit more about the human side of the corporate bond market. At a certain level it does seem as though it's a much more complicated market. As I think Paul mentioned a few companies out there have multiple lines of equities, but most of them only have one line whilst most bond issues have many bonds and they may well have different terms associated with them, quite hard to code up into a structure. So do you think Rob that we can ever expect fully systematic strategies in credit or is it likely to have a component of sort of human intervention?

Robert Lam (07:37):
I think there is an opportunity and we certainly favour the approach of being fully systematic, but I think that the question really targets the root of one of the most challenging parts of a fully systematic process in corporate bonds is that there are so many different nuances between each of the corporate bonds, whether that's more definitional or characteristic differences around durations and coupons and the impact that that might have in the economics of the bond, or even if it's less characteristic based, but more so trading based. I E how do you define the transaction cost that you would pay to for that corporate bond? Or what is the probability of fill given that runs are not actionable, I think those lead to significant challenges, but if you're able to tackle those challenges, I think that that's going to be one of the keys to a fully systematic strategy.

Robert Lam (08:28):
And I know there are many different flavours of systematic credit strategies given that this is early days and the definitions haven't been fully defined, but some other people might favour a wish list-based approach where your model actually outputs a whole slew of bonds that you would like to buy and sell. And then you would go out and best practice trying to accomplish or try and fill that list within a reasonable cost. So inherently that approach would lead to a model versus live and certainly exposes you to things like adverse selection. So we do think that it is worth it. It is rather difficult to be able to try and capture all of those dynamics, but it enables a fully systematic process in this space.

Sandy Rattray (09:08):
So let's now start talk about alpha a bit more and factors, and Paul, certainly you may have had the similar experience, but my experience of factors in equities has been that 10 years ago, really only quants talked about factors. Today, discretionary managers talk about factors in almost the same way as systematic managers do. There are ETFs out there. There are all sorts of different ways of buying and selling equity factors really simply and cheaply. In credit it's not the same. So do the factors work in credit and people aren't talking about them, or do we have to think about different factors in different effects in credit to equity?

Paul Kamenski (09:51):
I think we're certainly biased, but I I'm decidedly convinced that they certainly work within the credit landscape within the credit asset class. I think the key themes, if you think about concepts of value, trying to buy cheap securities that applies across asset classes. And we've certainly seen that across Man Group and seeing that in research through academia and so forth, that absolutely still applies within the credit asset class. I think it is new though, to talk about credit through a factor lens. I think that that's decidedly different than how it's been approached with historically a decidedly discretionary, more relative value approach and strong quality and defensiveness characteristics, but not through the lens of factors per se. So I think taking a step back there is the whole spectrum of, as you were highlighting more simple, raw factor approaches all the way through to more highly refined approaches.

Paul Kamenski (10:42):
And I think that that spectrum will exist into some extent already exists within the credit landscape. And as Rob was highlighting, those buckets are that spectrum. We continue to become more clear as the years progress. But what we found is that again, a lot of those core key concepts, when you're thinking about valuation and trying to identify cheap securities, certainly when you do have quality or defensiveness characteristics, there are many different concepts there to try to pick up conservative management approaches where we think that there are attractive opportunities there, but also concepts of trends and underlying business trends, concepts of momentum. I think that's a relatively new concept generally within the credit landscape. I think there's decidedly that asymmetry within credit returns where you're not going to trade tighter than a zero OAS. You're not going to trade with a yield lower than a treasury.

Paul Kamenski (11:31):
With that capped upside I think there's always been some hesitation with within momentum conceptually, but what we found is that actually broadening that concept, not focusing just on security trends within credit or, or equities, but even moving much more beyond that, identifying other alternative data sets to really capture forward-looking trends. Now-casting data is really an exciting prospect where we're seeing a lot of value. So I, a lot of the same pillars apply across the asset classes, but a key point is that really within each of those pillars, we really have found that there's quite a differentiated approach that's needed. So for example, within valuation, we find that the cheap equity does not imply a cheap bond and vice versa. So I think there are certainly nuances to take into consideration as well.

Sandy Rattray (12:13):
If we look at individual bonds, then how much of a single bond performance can be explained by the credit factors that you've evolved and how much of it ends up being unexplained is, I don't know if there's a good answer to that.

Paul Kamenski (12:37):
It's a tough question. What we found is that generally attribution's always difficult. Certainly when you're trying to attribute to a certain factor, you always run into the unexplained portion or the double counting, or sometimes even triple counting from an attribution point of view. But we have found that actually there is a significant portion of that overall return that is explained by our alpha signals. And I think maybe even more so than that, one of the things we'd like to focus on, or one of the ways I think about this too, is really trying to understand the consistency of some of these signals in their efficacy, especially from a monitor unicity point of view. And I think one key point that we found is that, although there are raw factors that can explain some portion of the overall credit return, their hit rate tends to be very low. They tend to be extraordinarily volatile. I think what we find is as you strip away a lot of risk factors and control for a lot of concepts of rating duration sector, so forth or volatility, what you can find is that you get a much more consistent excess return profile for some of these alpha signals, once you really go through more of that highly refined approach.

Sandy Rattray (13:38):
Okay, let's actually go out on that a little bit further and get into portfolio construction. And maybe Rob, you can take a stab at this. When we look at building portfolios and equities, there's a large number of tools that you can buy off the shelf to build the portfolios and you know, many competitors, and they've been around a long time and they've got optimisers and the data is generally pretty clean and robust. There are differences between the different providers but what's it like in credit? Can you just buy an off the shelf optimiser and a set of portfolio construction tools, or do you need to build them yourself?

Robert Lam (14:15):
Given how new systematic credit is and how different I think the asset class is or how unique it is relative to other asset classes, we very much so favour building our own tools in house, across the entire investment process. So you highlighted portfolio construction is one. I think that's very applicable. There are a lot of different levers that you have to pull in different techniques that really help stabilise the portfolio. And especially within corporate credit when the transaction costs are so high. And when there is a unique market structure that leads you to also have to things like probability of fill, you really need to pull on those levers in order to stabilise, you know, whatever optimizer technique that you'd like to use. So those don't come off the shelf, those need to be built in house, but I'd also extend that to things like risk modelling as well, where there are certainly off the shelf risk models.

Robert Lam (15:10):
You know, we've seen some that get heavily standardised on the equity side, but those certainly are not standardised on the credit side. And there are advantages in some way that the credit market offers or the credit asset class offers that you can really leverage within things like risk modelling to be able to better capture the downside and better be more on top of the market changes that can change quite drastically within corporate bonds. So ultimately I think given how new the asset classes, I would say that, you know, the real systematic people or the handful of systematic competitors that I would put on the list should be considering building these things in house.

Sandy Rattray (15:46):
You touched a little bit on costs of trading in there as well. And I suppose we compare again, back to equities. So equities are pretty volatile and yet the bid offer spreads and just the cost of trading has enormously tightened over time and they become much cheaper and so costs while significant, are a relatively more minor factor for many equity investors today. That's not true in credit. So why is credit so expensive to trade and what can you do about it?

Robert Lam (16:19):
Yeah, it's certainly a difficult problem to tackle. And I recall, you know, over 10 years ago when I was a single main corporate bond and CDs trader, I was sitting there at the trading desk, making markets clients would call in or, or IB chat me, you know, looking for liquidity where they're actually not sure where the market is. And I would try and connect a buyer and seller or think about a price that I would take down that risk. But ultimately at the end of every day, I would think about how odd is that workflow in order to understand where a corporate bond trades. So given that workflow, it's actually quite difficult to be able to quantitatively interpret that process and be able to put that into your model. So I think some of the questions that you really have to answer in order to come up with a good understanding of transaction costs within corporate bonds is actually three pronged.

Robert Lam (17:08):
So the first thing that I think you'd have to answer is what is that headline intraday inside market for a particular bond. And I compare this to the sticker price of a car, and obviously no one likes to pay the sticker price of a brand new car. So then the second question that you actually have to answer is what is that dealer behaviour for them to actually improve or fade from their runs, which are not firm inactionable, and that's also quite different than the equity markets. And then the third question that I think you have to answer is that what is that additional non-linear cost that you're going to have to pay as you request for more notional and you scale up the size of your strategy as well. So it's quite different. It's a multidimensional problem given kind of the liquidity profile and the volatility profile of the different bonds.

Robert Lam (17:57):
But when you put that altogether, I think you can have a much clearer sense of transaction costs and liquidity. And those are not easy questions to answer. And I think you kind of have to have an ensemble approach to different data sources in order to be able to really answer those. And I can run from things like the institutional runs that you might be getting the intraday quotes, the axes, the inventory realised print RFP data. You know, those are very large data sets and rather difficult to manipulate given the noise in them. But those are very important parts of it in order to capture all the different pieces of this part of the puzzle.

Sandy Rattray (18:31):
Okay. One of the challenges I think for credit investors generally, but particularly Kwan credit investors is that many of your issues are not public companies. So you don't have the data, the financial data easily available on them. So Paul, can you give a little bit more colour on the challenges in data and especially the private company problem?

Paul Kamenski (18:52):
So start actually by highlighting the price discovery and liquidity piece is decidedly one of the most challenging, but I know that we already talked about that. So focusing more on the private company side of things, that's certainly the case. So a significant fraction, depending on if you're looking at high yield or investment grade, there's more so in high yield, a significant fraction of the standard benchmarks for example, US high yield or US investment grade corporate credit are from private companies. That is to say they have publicly traded debt. They are benchmark eligible names, but they do not have publicly traded equity. So when you try to identify where their financial statement data is, or try to source that information, you certainly can't go to the standard vendors that have that you can't go to Edgar to download the 10Qs and 10Ks. One of the really interesting efforts is to try to tap into that full breadth of information for all those issuers, where those issuers do still provide financial statement information, but within data rooms. And so it's a decidedly harder, more time intensive, more labour intensive process to extract that information. And it's something that we think really can be an edge.

Sandy Rattray (19:56):
And just to sort of drill on all of that just a little bit more. So you talked about 10Qs and 10Ks, which obviously applied to the US market. If you're looking outside the US, is the data challenge hardware again.

Paul Kamenski (20:09):
There I'd say it's certainly on the financial statement side, we see a similar effect, but what is decidedly different is again, getting back to liquidity. So I think a key challenge, especially when you're looking at trading corporate bonds outside the US is trading anything that's not captured by trace. So I think there's a real advantage, a real benefit to having information from trace, being able to effectively have that tape in very short order with high fidelity information, seeing what traded, how much traded and at what price, even with the caps that are in place within the trace data, you really can get a good feel for where the market is, where things have traded, what liquidity looks like. There is no perfect source like that outside the United States, there are regulated versions of this through MiFiD, there are various vendors that have their own sort of give, to get methods, to try to capture information. But liquidity generally is a partial information world outside of US corporates.

Sandy Rattray (21:02):
Okay, now let's move to the second of my final three topics, I think, which is ESG. So we have this huge growth in ESG investing in equities is the same thing going to happen in credit?

Robert Lam (21:14):
Absolutely. I think it's very topical ESG in credit has gone through such an evolution, right, in the very early days as equity saw it was restriction list-based and then it moved into ESG and alpha source and can it produce consistent returns and actually add value to the portfolio. And then the different ways of defining what is ESG? Do we want to add on carbon output to make it a little bit more carbon efficient? And then these days climate is very topical as well. So I think overall ESG is certainly a topic and a growing area of interest within credit. I think that given the flows and given how much ESG is being baked into new mandates, particularly out of Europe, I think that's going to dictate the cost of capital for many credit companies within the yield and investment grade market.

Robert Lam (22:07):
Even taking that a step further, not only is it going to dictate the cost of capital, but sometimes I actually think it's going to even limit their ability to even access the capital markets at all. So they're going to have to find more expensive forms of capital and kind of manage their liabilities through that way. So I hope this doesn't come off as more greenwashing cause I'm sure there's a plethora of papers that really highlight this topic, but we truly do believe that this is going to be a large shift, a large focus over the coming years.

Sandy Rattray (22:37):
And it's interesting Rob, that you mentioned cost of capital so early in your answer, because when you talk to equity investors about ESG, they do talk about cost of capital, but they think of it as something which will take many years to start to impact. So you had it fairly early on in your answer. I suppose you're saying that because you think that the investment community will start to reflect ESG into their decisions and it'll be very directly reflected in the spreads that companies have to pay to issue debt.

Robert Lam (23:06):
Whether portfolio managers will add that bond or that issuer into their portfolio, which is the most direct form of impacting your cost of capital and dictating what coupon that these companies are going to have to pay to borrow money, but even taking it a step further than that, it shows up in the data as well. So if you look at ESG data in the corporate credit universe, whether that's high yield or investment grade, there is a raw tilt to that. And that's happened over many years, actually. So not just recently where companies with higher ESG scores actually quite drastically have lower yields and lower coupons than companies with really poor ESG scores. So the market is telling us this, and they've actually told us this many years ago, and I think it's just really starting to bubble up and people are starting to see that more.

Sandy Rattray (23:54):
So now let's just move to our final topic of discussion, the future of quant credit. And I've got two questions here, so we'll split them up, but firstly what do you think the future of systematic fixed income investing sort of holds? What's it going to look like in five or 10 years’ time?

Paul Kamenski (24:17):
So I think perhaps the biggest change that, that I'm excited about certainly and hopeful for is to change that number that Rob referenced before where 99% of high yield assets, for example, are managed by discretionary approaches. No one knows what that steady state number could or should be. But I think directionally five, 10 years from now that number in terms of the percentage or market share that that systematic approaches have, will be decidedly higher. I think also we will see a real breadth of offerings where certainly we've seen more smart beta type approaches have a strong hold within the equity markets all the way through, again, highly refined approaches. I think that's not really well distinguished right now in the systematic credit space. And I think that will continue to evolve and will continue to segment itself more to really provide a full spectrum of offerings.

Paul Kamenski (25:04):
And I think there's just going to be a lot more familiarity. I think in the early days you have discretionary managers that are sort of the dominant player in the marketplace. And so the whole credit market was familiar with trying to understand and appreciate those types of approaches. And I think we're already getting to the point where a lot of prospects, allocators, clients, investors generally are starting to appreciate systematic approaches within the corporate bond landscape. And I think that's just going to continue. So that backdrop, along with that positive feedback loop of more and more quantitative strategies entering the market and having a positive feedback loop that we see on the execution side, hopefully as we've seen in the equity space, meaning that there'll be better liquidity, tighter spreads, lower TCs [trading costs], I think will be advantageous generally. So very excited about the next five to 10 years.

Sandy Rattray (25:50):
And then Rob, let's just take a little bit into what might happen with alpha over that period. Typically when portfolio managers tell me, “gosh, it was terribly difficult.” I usually think that's great that it was terribly difficult because alpha tends to be more present and things which are difficult. If they're easy, then lots of people will be seeking the same return and the alpha will go down. So we've talked quite a bit about how essentially systematic credit will become easier. Does that mean we should expect the alpha to decrease over time?

Robert Lam (26:23):
The markets are reflexive, right? We see this in many different asset classes and just by the sheer fact, I think that we know that there's an inefficiency and that we're talking about it today actually kind of decays the alpha to some extent. And I believe that quantitative credit strategies in credit will continue to grow. So it will be natural to see some moderation in alpha over time. But with all that being said, I still think that we're in the early days and given all the difficulties, given all the challenges that we'll see, it's going to take some time for the market to catch up. So with all that being said, I think that that's why it's also so important to continue to innovate on the research side. My base case is that the underlying insights in our strategies today will actually look drastically different in five years. And it'll actually look drastically different again in the subsequent five years. So I do expect that moderation over time, but from our perspective, it's just about continuing to push to that forefront of where do you see the inefficiencies at that point in time and how do you exploit them and express them in your model.

Sandy Rattray (27:27):
Thank you, Paul. And thank you Rob for taking us through all the intricacies of systematic credit and the exciting journey which we think the markets and these strategies will go through. You can follow The CIO Agenda on Apple podcasts on Spotify or wherever else you receive your podcasts to receive our future episodes. Thank you and goodbye.