Trading Insights: QIS developments and the use of LLMs

[Music]

Eloise Goulder: Hi, I'm Eloise Goulder, head of the Data Assets & Alpha Group here at J.P. Morgan, and today I'm delighted to be joined by Deepak Maharaj, who is head of Equities and Cross-asset QIS structuring, to talk all things QIS and systematic strategies. So Deepak, thank you so much for joining us here today.

Deepak Maharaj: Thank you for having me.

Eloise Goulder: Deepak, could you start by introducing yourself and explaining where you really sit within the business here at J.P. Morgan?

Deepak Maharaj: Sure. So I lead the equity and cross-asset QIS product team. And we sit within the broader market sales and trading division and cover the space of quantitative investment strategies, which is commonly known as QIS. So QIS are essentially systematic strategies that aim to provide clients with exposure to common investment themes such as market beta, risk premia, alpha, or hedging type strategies. It's an area that's grown in popularity over the last few years and in particular, the equity QIS space.

Eloise Goulder: That's fascinating. And I'm looking forward to hearing about the rationale for clients using these products and why they've grown in popularity over the years. And it's worth noting that we have previously featured several other members of your team, Deepak, on this podcast series, so we will show links to those episodes in the show notes. But Deepak, could you start by giving the background to the QIS business as a whole?

Deepak Maharaj: Sure. So, when you think about the equity QIS products, I'd say they're perhaps the longest standing family of strategies, and it's probably why there's a widespread QIS ecosystem today. So if you think back to the seminal research papers of people like Fama and French or Jegadeesh and Titman, those were the original systematic equity strategies that were created. On equity factors in particular, it's fair to say they fell out of favor in the last quant winter from 2018 to 2020, but they have staged a resurgence in the last few years, and at least from a performance perspective, that has meant investors have become more interested in this space. And then in terms of exciting new opportunities within this space, the broader use of equity risk factors within client portfolios, so this is either as a tool for hedging or for tactical trading. There's the advent of intraday momentum strategies on single stocks, so tick-by-tick data. There's trend following applied to equity factors, which I know we've spoken about trend following in the past on this podcast, but it's an exciting development to see them applied to equity factors. And then finally, the use of LLM in strategy construction, as no discussion is complete without the use of LLM.

Eloise Goulder: Absolutely. Well, Deepak, you mentioned there four areas, the equity risk factors, which are perhaps the mainstay and the genesis of the equity QIS business, and then intraday momentum, trend following applied to equity factors. And then of course, use of LLMs in strategy development. So Deepak, could you dive into a bit more detail on some of those?

Deepak Maharaj: Yeah, sure. So if you think about the intraday strategies as an example it's fair to say they've been around for a number of years, but applied to futures. So this development is leveraging the intraday technology, but applied to single names. And that is allowing investors to capture various microstructure effects that may happen in single securities, such as intraday momentum, mean reversion, or even relatively new.  The amount of compute and resources that you would need to be able to effectively run such strategies and that's only becoming possible in recent years with the improvements in technology and infrastructure. The second area, trend following applied to equity factors, that is definitely a new area of research and development, and we see that as being a complement to existing trend following mandates. It's been shown that equity factors do exhibit trending behaviour, and that can be captured in a systematic trend following algorithm. And we see that as very much diversifying existing trend following products. Because if you think about traditional CTA-type products, the majority of them tend to be on things like equity futures or bond futures or market beta instruments, whereas equity factors are inherently long short in nature, so you're really accessing a different spectrum of risk and return. Finally, with the advent of LLMs and the incredible power that you have in these algorithms, there was a lot of interest in whether they could be used in systematic strategies. And we found a way to incorporate them into an existing family of thematic indices. That family is called Quest, and it stands for Quantitatively Selected Themes. And that family essentially uses so-called big data in the form of news articles to find companies which are associated with a given theme. And the LLM enhancement that we developed was essentially to refine the words that we are using to search for company theme associations. I'd say it's a continuation of our efforts in the space of incorporating new technologies such as AI into product development, and makes for a richer offering for investors.

Eloise Goulder: It's so interesting that you're now using GenAI in your investment strategies via this use of LLMs. Deepak, how transformative is this versus your prior work on NLP, natural language processing, to assess sentiment, which I know you've been doing for years?

Deepak Maharaj: Sure. So I think it's true that the application of LLM in our Quest product was essentially an evolution of an existing model, which was using NLP. And we did some tests internally when we were developing the LLM product to see how effective it was relative to the old model. And we indeed found that it was much better at identifying specific keywords and themes compared to the previous model. So in this small area of the strategy ecosystem, LLMs have been shown to be more effective. Having said that, given it's such a new technology, it's difficult to say at this point like how this process will evolve, and how widespread the use of LLM in systematic strategy construction will be. One of the challenges, with LLMs is that the results can be random. That is the antithesis of systematic strategies which are designed to be repetitive, rules-based and replicable.

Eloise Goulder: Yeah, it's such an interesting point. And Deepak, you mentioned at the start that equity QIS has really been a growth area within our business, so I'm intrigued as to which clients are utilising QIS and whether the breadth of those clients has increased over time.

Deepak Maharaj: Yeah, it's an interesting point because we've seen quite a wide spectrum of clients accessing QIS strategies and in particular, equities QIS strategies. If you break down by client type, on the hedge fund side, we see a lot of them interested in things like equity factor products where they're using this either to take exposure tactically to specific factors or to hedge factor loadings within their portfolio. It's interesting when we talk about equity factor products they are becoming increasingly commoditised. And if you think back, say, maybe 30 years ago when sector investing was considered an exotic product, it's clear that equity factors are heading in this direction of becoming more commoditised. Especially the more common ones such as value, quality, momentum, low volatility, and so on. And we see investors increasingly using these as tools either for tactical trading or for hedging in their asset allocation. With regards to the thematic products, we also see interest from hedge funds and asset managers on the thematic side where this is essentially a market access product for thematic investments. And thematic investments have grown in popularity over the last few years with the emerging technology, decarbonisation, and digitalization of the economy. And Quest is essentially a systematic way to create these themes, and also allowing you to measure thematic exposures within your portfolio. So it serves two purposes here. And then for the sort of broader spectrum of clients' intraday is perhaps an interesting case because essentially it's accessing a new market segment which was previously only available on futures, but now you can have an intraday strategy on some of the mega caps that everybody talks about and potentially capture more alpha.

Eloise Goulder: Well, Deepak, you've discussed all of these products that you're working on, and you've discussed the fact that this equity QIS business is a growth area with more and more clients wanting to engage with it. So what's next?

Deepak Maharaj: It’s the new data sets and leveraging those new data sets because there is so much data out there. Whether we're looking at intraday data, which is tick-by-tick on specific stocks, or things like social media data, sentiments, credit card, et cetera, there's just a vast amount of data out there, and one area which we're exploring is the application of macroeconomic data to equity strategies. And I think that's a relatively under-explored area because it's hard to access good quality time series-based macro data. And fortunately, we now have this available with our partnership with Macrosynergies, so we're very excited to analyse what risk premia or alpha that could be made available from this dataset. The other area which we're quite excited to leverage is social media or sentiment data. And again, it's relatively new and I know there's been a lot of work done internally to clean this data up and make it in a form that could be consumable by systematic investment strategies. So that's something we can't wait to get started on.

Eloise Goulder: Well, there's clearly a huge amount to be working on, Deepak, so thank you very much for taking the time to speak with us today.

Deepak Maharaj: Well, thank you for having me.

Eloise Goulder: Thank you also to our listeners for tuning into this bi-weekly podcast series from our group. If you'd like to learn more about Deepak's work and the QIS team, then please do reach out to your J.P. Morgan sales representative. Otherwise, if you have feedback or if you'd like to get in touch, then please do go to our website at jpmorgan.com/market-data-intelligence where you can reach out via the Contact Us form. And with that, we'll close. Thank you.

Voiceover: Thanks for listening to Market Matters. If you’ve enjoyed this conversation, we hope you’ll review, rate, and subscribe to J.P. Morgan’s Making Sense to stay on top of the latest industry news and trends – available on Apple Podcasts, Spotify, Google Podcasts, and YouTube. The views expressed in this podcast may not necessarily reflect the views of JPMorgan Chase & Co, and its affiliates, together J.P. Morgan, and do not constitute research or recommendation advice or an offer or a solicitation to buy or sell any security or financial instrument. They are not issued by Research but are a solicitation under CFTC Rule 1.71. Referenced products and services in this podcast may not be suitable for you, and may not be available in all jurisdictions. J.P. Morgan may make markets and trade as principal in securities and other asset classes and financial products that may have been discussed. The FICC market structure publications, or to one, newsletters, mentioned in this podcast are available for J.P. Morgan clients. Please contact your J.P. Morgan sales representative should you wish to receive these. For additional disclaimers and regulatory disclosures, please visit www.jpmorgan.com/disclosures

© 2024 JPMorgan Chase & Company. All rights reserved.

[End of episode]