Turning data into dollars: Macquarie on AI, insights, and smarter investments

Note: This interview was recorded Wednesday 23 October 2024

Billions of people worldwide now have internet access. Every action we take online generates new data, and businesses across every industry want to use that data to understand customer behaviour, shifting major trends, and how best to manage inventory and human resources.

According to Benjamin Leung, Macquarie Asset Management's Head of Systematic Investing, that data can also be used to make better investment decisions.

How big is the opportunity? Well, humans now create data at such an incredible rate that we’ve had to invent new words like zettabyte to measure it. What is a zettabyte?

Like the number of galaxies in the universe, it’s hard to wrap your head around. Technically, it equals a trillion gigabytes. Numerically, it looks like this - 1,000,000,000,000,000,000,000 bytes

To put that in some sort of context, a petabyte is the equivalent of 11,000 4k movies, which would take 2.5 years of nonstop binge-watching to get through.

That’s petabytes, though. One zettabyte is equivalent to 1 million petabytes. One estimate suggests there are currently 44 zettabytes of data in the entire digital world. The mind boggles.

Given the scope of the opportunity, more and more data providers are popping up, allowing Leung and his team to create increasingly robust investment models - which can ultimately lead to better performance. 

“We are seeing increasing numbers of small niche data providers,” says Leung.

He adds that the proliferation is a function of recognition of the value of data in the investment decision-making process, as well as the amount of capital going into the AI space.

“Both of those things have created a proliferation of data vendors, and that's been really enriching to our research process”, says Leung.

The mountains of data now available have allowed Leung and his team to build models that interpret earnings calls and evaluate the sentiment they can generate. These models can provide powerful analyses that would have previously taken hours or even days for analysts to conduct (if they even got the time at all during earnings season).

In this episode of The Pitch, Leung shares more about the novel uses of data and AI in his investment process and the technology that has him most excited for the future of systematic investing. 

Edited Transcript

Chris Conway: How is AI changing the systematic investing game?

Benjamin Leung: AI. It's very topical at the moment. I think the answer is it is and it isn't. Nothing is absolute in this world, and I'll explain why.

As a systematic investor, we've always been experimenting with AI and machine learning techniques and some of the technology that's enabled things like large language models and ChatGPT. But it's fair to say that the aspiration of having a generative AI to give me a stock portfolio is still quite a distance away.

Where it is making an impact, really, is the opportunity to tremendously increase productivity as an analyst. Never before in history has a population had such a powerful tool in reach? You just need to look online to see how many creative uses; people using ChatGPT-4 or things like it.

Ultimately, whether it is a game-changing moment for you will depend on how you utilise it as an individual.

Anecdotally, I've found enormous benefits from using generative AI. One of the things that I use it for, and I've heard other experts talk about this, is really using it as a challenger tool. I caution that we shouldn't ask it for investment advice, but it's actually a fantastic tool to soundboard or test your ideas against.

If you wanted to buy a stock, you could talk to ChatGPT and ask, "Why wouldn't you buy a stock?" or "Pretend you're Warren Buffett, tell me why..." or "What else should I think about?" And that's really been insightful for me as an individual, and it also helps you break out of the behavioural biases that I've spoken about.

And I really encourage people to utilise the tool and be comfortable with it because, as you're saying, it's only a matter of time before it becomes an essential part of our existence.

Chris Conway: Ben, you analyse an incredible amount of data as we've talked about. What are some of the major trends you've seen over recent years when doing that analysis?

Benjamin Leung: We are seeing increasing numbers of small niche data providers. That is a function of a couple of things.

One, I think it's a general appreciation of the value of data as part of the decision-making process. Secondly, probably in terms of the amount of VC capital that's going into AI space.

Both of those things have created a proliferation of data vendors, and that's been really enriching to our research process.

And for us, we've been seeing activity in the non-structured and alternative side of things in terms of ESG information and that's created a lot of ideas that we are exploring at this point.

Chris Conway: Ben, just a quick follow-up, if I may. Non-structured, some people out there won't understand that concept. Can you just quickly talk us through that?

Benjamin Leung: So non-structured data is basically referring to information that is not organised for a computer to understand. A computer deals pretty well with a spreadsheet, with a table, et cetera, but give them a broker report, give them a speech, and they often struggle. And I guess that's where LLMs ( large language models), are really going to be useful.

But the ability to access those transcripts now has become a potential source of value add, and we're seeing market participants enter that space.

Chris Conway: Thanks, Ben. What are some of the novel uses of AI when it comes to analysing data?

Benjamin Leung: We’ve been experimenting with machine learning for a long time. We use machine learning techniques to help resample our data to increase the robustness and, I guess, the strength of our models.

We use large language models to evaluate sentiment in transcripts and things like that. And we use various forms of machine learning to enhance our sentiment detection, just to mention a few.

And, of course, the example I gave before around using AI as a challenger, there's a lot of innovation by the analysts themselves as to how they enhance their research process.

They come across a research problem that they can't get their head around. They would ask ChatGPT or GitHub to help them understand how they could formulate and translate that into a mathematical formula. And that's where we're seeing usage there.

Chris Conway: Ben, what is one set of data that you would love to be able to quantify, but perhaps the technology doesn't quite allow for it just yet?

Benjamin Leung: If I dream big, I'm really excited about the prospect of quantum computing. I'm not an expert, but based on what I understand, I think it's changing the paradigm away from the existing way in which we analyse data based on ones and zeroes and make it more analogous to how the world works in a more analogue way.

The ability for it to consume huge amounts of information and make predictions, I understand, is enormous and the industry and the world is making breakthroughs in that technology even as we speak. So, if I had to put something on the table, I'd be very excited to see how quantum computing plays out.

Chris Conway: Ben, it's an incredibly exciting space. Thank you so much for taking us through it. Thanks for sitting down with Livewire.

Benjamin Leung: Thank you, Chris.

ETF

Macquarie Core Australian Equity Active ETF (MQAE)

Australian Shares

View

ETF

Macquarie Core Global Equity Active ETF (MQEG)

Global Shares

View

Discover Macquarie’s Core Equity Active ETFs

Macquarie Systematic Investments has spent over 30 years evolving a robust systematic process that has delivered long-term, consistent returns across a range of strategies for leading super funds and institutions, and now bring two core active ETFs to the Australian market.

Explore more at etf.macquarie.com.