What’s next for quant investing

Andrew Francis:

Welcome. My name is Andrew Francis and I'm the Chief Executive of RQI Investors. RQI Investors is a Sydney based quantitative equity manager. We manage $30 billion across Australian global and emerging market equities. We have two main strategies, our value strategies, which we're very well known for and have been managing Australian global and emerging market value strategies for the past 17 years. And our diversified Alpha strategies, again, we're managing Australian and global equities, and we've been managing those strategies for approximately five years.

I'd like to welcome Dr. Joanna Nash. Joanna's the head of portfolio management and a senior portfolio manager at RQI Investors. Jo, I know you've done quite a lot of work about looking at can quants continue to deliver. I just thought I'd frame that by saying quantitative investing is a very broad term, and there's lots of different strategies out there. And even ourselves, we manage a value strategy and a core strategy, our diversified alpha strategy. But lots of ... There's hedge funds, there's fixed income funds, but what we're talking about here is quant equities in general. So I guess, do you think quants have had a really good run? Do you think that can continue? And what are some of the reasons why quantitative managers have done well?

Dr. Joanna Nash:

So you're definitely right, Andrew. We have seen some really strong performance coming from quants, and I think it's only made it more stark when we've compared it to some of the fundamental managers, which have been a little bit more challenged. Now, I think this is where we see the benefit of quant investing. And so what I mean by that is quants look at a whole range of insights when they're thinking about a company. So they'll look at quality, they'll look at the momentum of the company, they'll look at relative valuation, growth estimates, sentiment, et cetera. And so then that all gets combined to give us an idea of the company. And this is sometimes a little bit different to fundamental managers, which may be looking at just one or two aspects of a company. And so therefore, if what we tend to see is, what we hope is that each of those different components will be a little bit additive and give us an overall consistent performance in our portfolios.

Now, what we've seen recently is actually probably a little bit abnormal. What we've seen is all those components have all been working at the same time. And so that's given us this outsized performance that we've seen in quant. And again, compared to fundamental managers, even more stark. Now, do I expect that to continue? I wish it would, very much so, but I think what we'll see is we'll come back to a more normal level where we get a period of time where one component may not be as strong, other components aren't, they still combine together to give us that consistent performance over time. So yes, I do think quants will continue to deliver probably at a slightly lower level than they have been, but yes, I still think they will.

Andrew Francis:

I think you made a good point there. And like I mean where you're saying combining all the insights, because I guess if I looked at the market and the real advantage of quantitative investing, well, one of the advantages is the breadth and also the portfolio construction and the risk control. And do you think that has been one of the primary factors that has led to quantitative investing doing well over the last few years?

Dr. Joanna Nash:

Look, and I think that is a really good point. So quants make a lot of little bets rather than a small number of larger bets, which is again, what the fundamental managers tend to do. And so in times like we've seen in the last couple of years where we've had some sort of, I guess, macroeconomic factors or political factors coming in play, what you'll see is that one component of the market may perform very strongly for a little while and then the next component performs very strongly. And so if you've got larger bets, then you are at risk of maybe having some performance, but then it sort of reverting very heavily when those different macro conditions come into play.

If you look at a quant when we're taking a large number, sorry, of small bets, what happens is none of those bets are going to have an outsized impact on the portfolio, but rather you're getting that nice risk control across the whole portfolio. So when those macroeconomic conditions come to play or when those political components come in, yes, certain components of your portfolio may be impacted, but they're not going to be as heavily impacted as we see with some of those fundamental managers, which then allows us to again provide that consistent performance throughout those different market cycles.

Andrew Francis:

Great. Thanks, Jo. And I guess we're very well known for our value strategies. And so that does have more, I guess, a overall single factor tilt. But again, that is done pretty well and pretty well versus stock picking managers. So can you tell us a little bit about some of the things that have worked for us in our favour in those strategies?

Dr. Joanna Nash:

Yeah, look, and you're right, we do have those value portfolios, which tend to have that value tilt to it. I think the thing that we benefit from that is we're trying to pick up that long-term value premium in those portfolios, but know that there are these short to medium term alpha opportunities. And this is where our alpha model helps us sort of balance up. We're getting that long-term tilt, but also being able to take advantage of the differences that are happening in the markets in that shorter and medium term. So this is where those different components that we spoke about, looking at quality of the company, looking at relative valuation, looking at sort of momentum and sentiment can come into play to help us balance off and still provide shorter term and medium term alpha opportunities, but also still capture that long term value premium.

Andrew Francis:

And if I was an advisor, like I mean, some of that sounds quite technical, would it be fair to say some of those additional alpha insights are trying to remove stocks that might be cheap for a reason? Is that a fair ...

Dr. Joanna Nash:

No, that is 100%. And that's where we look to tilt a little bit towards what we call quality. So this is looking at companies that have consistent cash flows, that have well managed companies, all those aspects that sort of add to being able to capture that value premium, but do that within the higher quality companies. And so we avoid then companies that happen to be cheap for a reason. We don't want those ones. What we want is companies that we know will have that consistent performance, will be able to provide us that value premium in the long term.

Andrew Francis:

More broadly, not just the value strategies, but more broadly quant, as we said at the start, has had a good run. What do you think are some of the risks in quantitative strategies? And we quite often hear ... Well, on our side quite often here about the risk of crowding, is it a crowded trade? Is there a risk of a quant winter, et cetera, et cetera. What do you think some of the risks are, but also what has changed from sort of ... Because there was a quant winter a little bit over seven years ago where quantitative strategies didn't perform as strongly. What do you think? One, do you think there's risks of crowding? And secondly, do you think that also that quant strategies, what's protecting them from going through another quant winter, so to speak?

Dr. Joanna Nash:

Yeah, no, I think there is a risk of crowding always. And I think what we hear when everyone's giving a lot of attention to quant, you may think, okay, there's too much money piling in there. You're all following exactly the same insights. Are we going to have the reversal or that quant winter that we saw previously? I think there's a couple of reasons why we won't see that. So the quant winter was driven very much by poor value performance. And what happened was those other insights didn't get to offset it enough. So we ended up having everything was sort of neutral, maybe a little bit positive, value is very, very poor. And so we ended up with that underperformance coming through.

Now, what's happened since then, I think two main things have happened. We've got what we call, I guess, more diversified insights, and this is coming to play where we get some of that AI and machine learning coming in. So if you think about a quant, typically they were based on data, right? So it was going to be balance sheet data, maybe some analyst information, maybe some price-based insights. Now we've been able to open up to a whole lot of what we call unstructured data. So this will be things like text-based insights. Maybe it's looking at visual images, maybe it's looking at sort of credit card transactions, all these sort of unstructured data which is harder to deal with. We can now deal with due to things like AI or things like machine learning.

So AI allows us to read all the text. So now we have a whole lot of different sources of data we can look at, which we couldn't look at previously. The machine learning allows us to sort of look at that unstructured data and also look at what we call non-linear relationships. So previously we used to say, well, A plus B equals C. The markets unfortunately aren't just A plus B equals C, they're a lot more complicated. And the machine learning allows us to look at those interactions between different insights and look at that non-linearity to help us be able to look at the data better.

And so what that's meant is, even though you've got a number of quants playing in the markets, they're all looking at data in a slightly different way. There's more data for them to use, and so you've got less chance of that crowding. They're still going to have some aspects where they look at, everyone's going to look at some sort of form of momentum, but the way that they look at that momentum, the way that they analyse it will be different because they've got all these different techniques available to them. So I think that risk of crowding is a lot less than what we did have back during that quant winter.

Andrew Francis:

Great. Thanks, Jo. And I guess importantly for advisors, how do you see quantitative strategies sitting alongside or replacing either fundamental stock picking strategies or index-based strategies?

Dr. Joanna Nash:

Yeah. So I think the advantage with quantum, and you referred to this earlier, was that risk-controlled nature of quant. So quant will deliver you the performance but at a certain level of risk. And so that allows people when they're constructing their portfolios to sort of have a better idea of the risk that they're going to have at that overall portfolio level. It then allows them to sort of use quant potentially as like a core allocation where they can sort of have that sort of risk controlled returns coming through and then combine it with some of those more volatile managers who potentially can't sort of control their levels of risk as much as quants can.

We're also seeing it a lot in people sort of slowly wanting to take on a little bit more risk in their portfolios, but wanting to know how much they're taking on. So rather than just being purely passive, looking at some of those sort of low tracking error or sort of enhanced portfolios, which is where quant comes into play really well, such that they can sort of get a little bit more alpha on top of the market returns, but again, do that in a risk controlled manner.

Andrew Francis:

Great. Thanks, Jo. That's been really insightful. So thank you for your time.

Dr. Joanna Nash:

No problem at all.

Andrew Francis:

I'd like to welcome Dr. David Walsh. David is the head of investments at RQI Investors. David, we just heard from Jo talking about AI and machine learning, and I know that you've written extensively and researched extensively those couple of topics. So maybe if we start with just you talking about how we use AI and machine learning within RQI Investors.

Dr. David Walsh:

Yeah. Thanks, Andrew. It's one of those topics that can be poorly defined if you're not careful. So just to clarify a little bit, machine learning is a broader topic and AI is kind of like part of machine learning. So we use machine learning quite extensively within the business and AI more broadly is kind of used in parts, in different parts of it. So the machine learning or AI parts that we use tend to be around the use of particularly research tools. So we like it in research. We use research tools associated with machine learning and AI to extract better insights, non-linear patterns, things we couldn't currently study, better ways of studying what we currently do. And those tools and techniques that AI machine learning generates are really quite important in terms of the way we explore the ideas and how we think of constructing our research.

Andrew Francis:

And how do you think it gives RQI an edge, or not just RQI, how do you think it gives quantitative investors an edge, those different tools and techniques? And I guess also it's not new, is it? Like this is, everyone's talking about it that it's new, but it's been in our toolkit for a long period of time. And maybe if you can give us a few thoughts about that, but importantly, how it gives us an edge and quantitative investors an edge more broadly.

Dr. David Walsh:

Yeah. Yeah. So it's certainly going to make returns and processes more robust in the way that returns are generated, you would expect, because we're now exploring things when non-linearities can be picked up. So our alpha models now allow ourselves to pick up things which couldn't previously pick up.

Andrew Francis:

I'm just going to interrupt. So when you say, what does non-linearity mean?

Dr. David Walsh:

You can imagine something increases, if you increase the base level by one unit and the underlying exposure increases by two, increases by two, the underlying exposure increases by six. That's a non-linear exposure. It steps up more, the more exposure you get to it. So you can see that more in, for example, an analyst signal when extreme outliers matter a great deal more in terms of their revisions than ones who need consensus. So that sort of thing matters in terms of ... And you can pick it up much more using machine learning type of tools.

So robustness in terms of performance of funds and alpha generation is probably a natural outcome of using these AI techniques. I don't think it's going to find us new ideas. So what we think of as idea generation within RQI is not really about letting the machines tell us what results we find. It's more about what the insight is and how we generate ideas themselves and then the AI machine learning tool lets us express that. So it's exploring new areas, but the insight's not going to change. So more robustness, better evolution and quicker returns, quicker processing its returns. I think all these ideas is pretty important.

And we'll put one point at the end of that. There's a bit of a danger in this, in that if we let ourselves go too far towards letting AI tools generate these sorts of ideas, we need to run the risk of crowding. There's an issue associated with crowding always within quant investors because there's a common set of ideas that are exploited sometimes in a common way. But with AI, if you do it right, you can differentiate your ideas much more by having different approaches, given the broad set of tools that AI gives you.

Andrew Francis:

That's a great backup for what we're just talking about with Joanna. I was saying the risk of crowded trades. And she was saying with more and more of these, I guess, non-linear signals and I guess the broader data sets that the risk of crowding has been reduced, albeit that there is some aspect of it, but probably not as stark as it was say 25 years ago or 20 years ago.

Dr. David Walsh:

Yes. Yeah. And a lot of that's got to do not just with the use of AI in alpha generation, although that's important, and then what the broader set of alpha sources we can use. It's also about the way in which we think of portfolio construction and the way in which we think of alpha construction as well. The AI lets us do some of that that we currently, couldn't do in the past. Again, the insight's really important, making sure we get the idea right, then using the AI machine learning tool to decide how we construct alphas, as in the broad mix of alphas, how we construct portfolios, differentiates us potentially and prevents crowding. But there's a danger that we all end up lining up the same thing if we're not careful. So we need to be quite innovative and insightful in how we apply it.

Andrew Francis:

Do you think the advent of AI has also, I guess, lowered the barriers of entry in quantitative investing? Like if we go back 20 years ago, obviously very, very big quant firms with very big teams of data scientists and developers, et cetera, had a really key competitive advantage. Whereas today, smaller businesses can compete more nimbly with those much bigger businesses. Would you say that's ... Is that fair?

Dr. David Walsh:

Yes, it is. But I think the same thing kind of applies now as it did then in some ways. So in the past, the barrier to entry for these ideas was about the data sets you had and could you build them because they weren't commoditized. The tools you would use like machine learning AI tools, which were now becoming increasingly commoditized, back then were not. And perhaps the talent that you could use, that was often quite smaller group, isolated in a small, fewer group, a smaller group of investors. Now those toolkits, the knowledge about them, the data and the machine learning tools that are available are more commoditized, which means yes, there's a lower barrier to entry and yes, we would expect more smaller players to get into the market, but exploiting machine learning tools with better data, the more commoditized site set doesn't generate your better alpha out of sample.

It's still an in sample problem and you still need to have the right insight. So that's where it comes down to. That's always been the case, still the case now. One downside from that is potentially that we'll see much more volatility for a little while while it suggests because we'll have these players trying to chase alpha, which is largely in sample. They might generate returns slightly out of sample for a while, then compete themselves out by crowding. Again, the crowding issue comes up. We need to make sure that we're going to be using insight that's differentiated and tools are differentiated to avoid that.

Andrew Francis:

So I guess when you're talking there, David, I mean, it's no different in quantitative management to fundamental management, people really matter because it's the people that are generating the insight. Is that still true today, would you say, and even more so?

Dr. David Walsh:

Absolutely. Even more so, that's right. And not just in the idea generation for research, that's something we know about that people matter much more in terms of compared to the data or the code than the same way it does in a quantitative investor, it does in a fundamental investor. There's no question about that, but it's not just in the researchers. So there's quite a move at the moment in terms of the tools available for coding that tend to make people think that coding will become much more commoditized, that a lot of the people who generated code in the past will become squeezed out by Claude code or something like that, generating the code for them.

In fact, it sharpens up the need for good quality developers and people who understand the research they can do and the development they can do. So in some sense, both the researchers and the application, the development of those researchers is sharpened, the insight and the skills are required and made better by these tools, not commoditized.

Andrew Francis:

So I guess looking forward three to five years and thinking about where do you expect the biggest advancements in quant investing will come from, or is it a bit too hard to say or looking in your crystal ball?

Dr. David Walsh:

Yeah, it is hard to say specifically what'll happen. We can say more generally things like coding will become much more commoditized, that the breadth of data available to be analysed will become more commoditized. The insights potentially will need to change, that'll be part of the research process that you need to think about. But definitely the coding, the structure of the market, the need for people, both researchers, developers, understanding portfolio construction traders will need to change the way they think of the world. So that's definitely going to be a change we'll see over the next three to five years is the way in which these things are applied and the need for people in that process.

The second thing I'd say is the trading process currently, and with the advent of Agentic AI, it hasn't really found its way into the trading processes yet, the way the dealers work or the way that traders work, the way that brokers work. Now that'll be more of a barrier to entry because there's a lot of institutional operations there and a lot of history. But I think that's an evolution we'll see over time is more agentic AI based trading systems, potentially on behalf of fund managers, quant fund managers who will implement these trading based on some kind of agentic process, which is not just an algo the way that it's currently set up, but much more advanced than that. I can see that definitely happening.

Andrew Francis:

And just define agentic.

Dr. David Walsh:

Agentic. Agentic is the idea that you can both give an AI the ability to analyse and the agency to act on behalf of what it analyses. So it can go off and find data, make decisions and implement them that you've told it to do without you getting involved.

Andrew Francis:

Well, I guess going back to the previous questions is, I mean, with agentic AI, then is there a need for humans?

Dr. David Walsh:

More need than ever, I would say, but it's going to be more specialised. The people still matter. So agentic AI will implement what we ask it to do and we will build the models and hopefully be more differentiated than others so we can generate those, but writing and building those models in a right way, developing the tools that do that, generating the insights, that's still a human thing. AI is never going to do that. I don't believe it's ever going to do that. So it's still very important. The nature of the work will change quite significantly, but the need for people is probably going to increase rather than decrease, the right people that is.

Andrew Francis:

Well, that's probably a good segue into the next level of the final question. And like having good portfolio construction and risk modelling, is that probably even more important now than say just picking individual AI signals or thematics or that sort of thing? What's your views there? I mean, we put a lot of time and effort into model management and so what's the interaction there with some of the new tools and techniques?

Dr. David Walsh:

So the alpha modelling, that's the building of alpha composites and alpha models. The portfolio construction that follows from that, the way we model risk are all going to be critically important again in the future. In fact, there's a danger we can see that if we have too much common portfolio construction tools, too much common risk modelling, and the industry focuses too much on just AI type modelling, we'll end up aligning our portfolios in the same way and generating crowding that we did pre the GFC. That's an issue which we saw with the GFC. We don't want that to happen again.

So making sure that your portfolio and construction, your alpha modelling and your risk modelling are differentiated, idiosyncratic, the way you do things is different and has your own insight in it. That's critically important. Chasing a new AI idea, it's a tool like any other tool. It's a good tool. It does a lot of things we can't currently do, explores things in interesting ways, but it doesn't generate things that we can replace the current foundational issues we see in our business, which is alpha construction, risk modelling, portfolio construction that'll get us all the way there with a quality set of alpha insights.

Andrew Francis:

Great. Thanks, David. And look, I'll leave on one last question. I did say that was the last question, but one further question, because it does tie into what Jo was talking about before. And I think quant has had a good run. So if you're looking, what do you think some of the biggest risks for quantities if we're looking three to five years out? Quant investing.

Dr. David Walsh:

Quant investing. So the quant-

Andrew Francis:

Inequities, I should say because it's not a heterogeneous group either. It's not a homogenous group.

Dr. David Walsh:

No, right, of course. It's everyone from slow to fast and risk factors to idiosyncratic and whatever. Yeah, exactly. So there's been a lot of common factor performance that we see, which is being driven by common, by more generic quant models chasing those things. The danger is that becomes more prevalent, that we squeeze out too much of the alpha that's available and we bet more on more on risk factors, and that tends to lower the overall tone of the industry and make it more difficult to generate alpha. So the danger there is the industry's going to evolve more towards a simple risk premier idea. And this idea's come up several times in the past, never proved to be true, but there's a danger we can head in that direction if we're not careful. We don't differentiate ourselves properly.

So we get back to that point you made about portfolio construction, about alpha and about risk. That's where you differentiate yourselves, right? I think quantitative investing is systematic application of good ideas. It's not AI ideas, it's not data-driven ideas per se, it's just good ideas applied systematically. And if you apply those good ideas systematically, then the question is, how do you get the good ideas? How do you then apply them? That's where your edge is going to be. So quantitative investment management is in those two fields.

Andrew Francis:

Thank you, David, and thank you for sharing your insights. That was excellent. And most importantly, thank you to the audience for your time and listening to us today and for those that are listening that are investors with us, we're very, very grateful. So thank you.