The APB Quant Investment Forum – Sorting signal from noise
Asian Private Banker (APB) co-hosted a lively and thought-provoking lunchtime roundtable discussion alongside its partner Robeco in a private room at the Artemis Grill restaurant in Singapore on November 4, 2025.
The Quant Investment Forum brought together Weili Zhou, Robeco’s deputy chief investment officer & head of quant investing & research, with the chief investment officers and senior fund selection specialists at six of Asia’s leading private banks and wealth management companies.
The forum put a sharply focused spotlight on the twin aims of quant investing systematically: to achieve different return profiles and to more effectively manage market volatility for private wealth clients in Asia Pacific.
Given heightened topicality by investors’ ever more energetic search for alpha and diversification at a time when equity markets face unprecedentedly high valuations and concentration risks, the roundtable also explored the role and impact of the rapid, arguably game-changing, advances in the technologies now being utilised by quant solutions providers.
The stakes are already high, and they are only increasing as the barriers to entry and success in systematic investment both increase and get more costly. As Zhou of Robeco put it succinctly: “If you’re just average as a quant manager now, you’ll have a very tough time in future. In the meantime, the best fundamental managers will have to become more quant-like. And the best quant managers of the future will need to be even more technologically savvy, data-smart and well-resourced than they are now. The very best will possibly have to go as far as developing their own large language models similar to those of Google or OpenAI.”
With this in mind, the forum also sought to examine how private bank fund selectors can best evaluate the competing claims of rival quant providers, and how they can build portfolios and products with combinations of fundamental and systematic strategies.
It sought, in short, to separate the signal from the noise.
It was by no means all plain sailing for the quant evangelists at the roundtable. Gatekeepers raised numerous challenges to quant orthodoxy and, inter alia, questioned the failure of previous models to cope with rapidly changing market conditions, the performance history of quant over fundamental strategies, and the unproven nature of many current models, given the prolonged and continuing 17-year bull market in global equities.
No question was ducked, and no challenge went unanswered.
Here are the edited highlights of the discussion.
Participants:
Jean Chia, managing director, global chief investment officer, Bank of Singapore
Dimitrios Papacostas, director, Citi Investment Management, Asia Pacific
John Ng, head of funds investments, DBS Bank
Hugh Chung, chief investment officer, Endowus
Steve Brice, global chief investment officer, Standard Chartered Bank
Yves Bogni, head mandate investment team, UBS Wealth Management
Dawn Foo, head of wholesale distribution Asia ex-Japan, Robeco
Lawrence Hanson, head of Asia Pacific distribution, Robeco
Rob Huisman, client portfolio manager – quant, Robeco
Weili Zhou, deputy CIO & head of quant investing & research, Robeco
APB: For the purposes of our discussion, the definition of quantitative investment is the use of mathematical, statistical and modelling techniques with the aim of achieving excess returns. With that in mind, how do you individually, and how does your bank, asset or wealth management company approach quantitative investment?
Weili Zhou, Robeco: I joined Robeco around 20 years ago and was, at that time, heavily influenced by my fundamental-minded colleagues. The role of quant research and investment back then was to support and facilitate the fundamental investment process. I was taught about valuations and price movements and that quickly evolved into us producing very sophisticated ranking models and algorithms. From not very much when I started, our quant investment has built into a Euro100 billion business and our adoption of and preference for systematic investing is clearly at a totally different level now versus then.
I’m also super excited that a lot of alternative data, machine learning and NLP AI [natural language processing artificial intelligence] are tangibly making their way into mainstream investment – not just in laboratories and testing, but also in real-use and daily investment processes. This is significantly boosting output and helping us to make better decisions, increase returns and provide better risk management within holistic portfolio management.
Hugh Chung, Endowus: Before joining this digital wealth platform three years ago, I was the CIO of a single family office for about four years and, before that, spent much of my career you might say competing with quant as a fundamental equity investor in hedge funds in Hong Kong. So, as both a bottom-up risk-taker and as an allocator within both a family office and a digital platform, I have been able to see different aspects of the pros and cons of many and various quant and fundamental approaches.
At Endowus, we make extensive use of quant and systematic strategies. Historically such strategies have not been so readily available to retail investors. But now firms like Robeco are increasingly entering the wealth space and we believe this is a good thing. In our core flagship portfolios, we’ve already embedded different systematic strategies. And I’d say quant has now become an integral part of the client’s journey.
APB: What will drive further growth?
Hugh Chung, Endowus: A big part of it will be educating our client base on what these strategies actually mean and how different signals and factors within strategies change during different regimes and at different times. The short answer is that we use quant strategies a lot and only expect this to grow. Many members of our fund selection team have backgrounds in quant investment. We want to provide the last mile in introducing good, proven, accredited systematic strategies to our clients, both standalone and as part of portfolios.
APB: Can you tell us roughly what the percentage of quant or systematic funds currently is on your platform right now?
Hugh Chung, Endowus: I would say around 20 percent of the assets under advisory on our platform are quant-related.
Lawrence Hanson, Robeco: I’m responsible for Robeco’s commercial businesses across Asia Pacific out of our network of offices in Hong Kong, Singapore, Tokyo, Sydney and Shanghai. While Robeco is widely recognized for its leadership in sustainability, what’s less known—particularly within the wealth management community—is that the majority of our regional business is actually driven by quantitative strategies.
What’s interesting to me, when we look at the adoption of quant, it’s very much a mainstay for institutional asset owners in the region, while it’s at the relatively early stage of take-up in the private wealth space. Its growth potential is significant and exciting. And I agree with Hugh that education will be one of the most important keys to success here.
Jean Chia, Bank of Singapore: I take care of our strategy research teams and am in charge of the bank’s discretionary portfolio management teams. Have we used quant strategies? Yes we have. We have quant funds on our product shelf of course, but clients need to be aware that such methodologies may not be all-weather and performance can lag during periods of dislocation.
APB: How about today?
Jean Chia, Bank of Singapore: We have started to deploy data very broadly in a process we call “robust optimisation” which we have fully incorporated into our asset allocation methodology. Our data teams are not financially trained – they’re more a combination of physicists and engineers. Our foundational premise is that uncertainty in markets and policy is going to prevail. The models our teams are focused on need to optimise the most desirable asset class combinations from and under the almost infinite number of market conditions that can eventuate from this premise.
Steve Brice, Standard Chartered Bank: As the CIO of the bank, from a CIO perspective, we use quant in different ways. We incorporate both outside views (quantitative) and inside views (qualitative) into our decision-making process. First, from a strategic perspective, we use it as a “smart anchor” for our asset allocation decisions. On the tactical side, we have four different quant models that we use, just to provide different perspectives. The main challenge has been that, post-Covid, a lot of the signals were irrelevant so we’ve not relied purely on quant – there’s always a need for balance with a strong qualitative input.
We take a similar approach for equities selection. We have a quant ranking tool that serves as an overlay once we have made our qualitative selections. And in the bond space we use quant to filter for bonds that we believe will not default. So we use quant to widen the variety of options that our front-line bankers and relationship managers can talk to their clients about.
Dimitrios Papacostas, Citi: When I look at how we use quant within our discretionary portfolio management (DPM) activities, I think it’s fair to say we’ve made great progress over the years. At the same time, we’re probably not currently firing on all cylinders. We have certain quant models that we use for idea generation. And we do have a number of quant portfolios that our clients tend to like. What we’ve noticed overall is that in markets that are reasonably efficient, clients like quant for a combination of cost reasons and the orderly and low-touch nature of its implementation. We run US large cap and international equity quant portfolios. But it’s an area in which I would like to dig deeper and pursue more solutions.
John Ng, DBS: We have a handful of specific “quant funds” on the platform. Not that many, as we have not seen a very significant demand for such dedicated funds. However, we know of several funds which are not specifically marketed as “quant funds” but do use some level of systematic and quantitative processes in their portfolio management. Going forward we expect there to be more added.
APB: How is Robeco partnering with global private banks and wealth managers for mutual benefit in the quant space?
Weili Zhou, Robeco: We’re currently working closely and in partnerships with private bank partners from Australia to Africa on innovative private wealth quant solutions.
A good recent example is a partnership with a large Swiss wealth manager to help it completely reconstruct its approach to risk management within its holistic portfolio management business. That meant helping the bank to redesign its core portfolio to give it a tracking error range of two-to-three percent. Another stipulation was that this portfolio’s performance and constituents had to be more accountable and explicable to their end-clients. They also – naturally – wanted strong alpha.
We’ve run this as a sub-advisory mandate and we’ve had to structure it to cope with the more frequent inflows and withdrawals of a private wealth – as opposed to institutional – mandate. We started the year with funding of around EUR300 million and, thanks to its strong, stable and explainable performance – and its more predictable alpha – we have now grown the mandate to over EUR1 billion.
APB: Do you believe such partnerships could be replicable in Asia?
Weili Zhou, Robeco: Very much so. We have been working on similar partnerships with other private banks, wealth managers and single-family offices where we have co-created quant solutions and vehicles ranging from US equity to fixed income ETFs. We will be extending these kinds of partnerships into Asia starting in early 2026.
APB: How do end-clients (ultra high net worth investors and family offices) typically respond to the concept of quant investment when it is first raised to them?
John Ng, DBS: When it comes to funds and strategies, I don’t think that most clients differentiate between quant and fundamental. That’s largely because even the most traditional of asset managers who started off as fundamental managers are now using a certain level of quant in their processes. Only a few clients who are very familiar with investment strategies do specifically ask for dedicated hedge fund or systematic quant strategies.
Jean Chia, Bank of Singapore: If you want clients to sit up and listen right now you have to mention AI. I think AI’s impact on investment and how it might adapt and augment quant investment – rather than the quant versus fundamental discussion – is more top of mind at the moment.
APB: Do clients get lost in the reeds sometimes, given the technical, mathematical and statistical complexities of quant investment?
Weili Zhou, Robeco: Quant investing might sound very mathematical but, at the end of the day, it’s data driven. In fundamental investing you take market data and financial statement data and then provide a human analysis that overlays this. Quant does the same but in a more disciplined and more extensively researched way. We look at any number of valuation metrics but then choose the one that we think is the best.
Some inputs and metrics that go into quant models might be the same as they were 20 years ago but the way of assessing these inputs and metrics and investing in these signals is now very different. Quant can take into consideration much more data and much broader coverage – across say 5,000 stocks versus 500 – and test investment signals that conceptually work, that intuitively work and then, crucially, that prove they can work over time. These are the algorithms that have brought us to such long-term and consistent success.
APB: How do you address Jean’s point about the obsolescence of some of the older quant models and strategies over the past 20 years?
Weili Zhou, Robeco: Compared to 20 years ago, two things have changed dramatically. One is the explosion of available data. 20 years ago, the volatility and return of a stock combined with the financial statement of the company were what gave us most of our insights and inputs. Now we can analyse, for example, what the CEO of a company says and what investors tweet, while at the same time looking at the results of this company’s employment engagement survey. This gives us a totally different order of magnitude of data that we can use to understand, analyse and differentiate between individual companies and their investment merits.
Second, computing power is doubling or tripling every two to three years. Given that the access to cloud, that means there is now almost unlimited computational power available to us to digest, interpret and exploit the signals given off by this ever-expanding quantum of data. In combination, this gives a considerable edge to quant investors.
APB: Is there also a downside to this in that the deluge of available data can sometimes see the signal obscured by the noise?
Weili Zhou, Robeco: The explosion of available data can indeed be a curse as well as a blessing. But, crucially, that’s why we add an all-important human intuition overlay. We have to know which signals should be listened to, prioritised and acknowledged and how much weight to give to each one. That’s where intuition alongside our team’s combined economic, political and social know-how and experience come in and why they need to co-exist alongside pure quant in any successful strategy.
APB: As global equity markets approach peak concentration and valuation levels, what role can quant solutions play for private banks and wealth investors seeking both alpha and diversification?
Weili Zhou, Robeco: In the past five years the average return from global equity markets has been double-digit. Downturns have been predicted during that period but have so far not materialised. There will be one – but we don’t know what will cause it and we don’t know when it will come. So it’s very important to have a cushion and a hedge in your total allocation. Investors will want to continue to invest in the alpha they like – naturally to continue to benefit from positive momentum – but at the same time avoid some of the big names and holdings that can quickly go against them.
Achieving this in a controlled way by systematically reducing tracking error and associated risks while maintaining alpha – with predictable portfolio characteristics – related to core holdings is a key reason why enhanced indexing quant strategies are now so well loved and supported by both institutional and private wealth investors alike.
APB: Can you give us a definition of enhanced indexing?
Weili Zhou, Robeco: Enhanced indexing combines the pluses of passive investing – broad market exposure, liquidity, and diversification – while aiming to outperform the market. It takes a benchmark as a starting point and then aims to generate returns above that benchmark while systematically keeping the relative risk at the portfolio, country and sector levels – indicatively in a one-to-two percent tracking error range. We target a passive-plus return in the long run with enough alpha to both cover fees and ensure investors are better off than if they invested in a purely passive strategy.
APB: Are the CIOs and fund selectors among us persuaded by this?
Steve Brice, Standard Chartered Bank: We are, as I’ve mentioned, big believers in quant. But there are a few general challenges. First, how does any fund manager pitch a fund? They give the overall story of what they do, and then they give examples and stories to support their pitch. And quant isn’t always as great at giving those stories. Partly because there’s no human involved!
Second, you might have a quant model that’s been tested in a bull market, but does it work under different regimes? Is it sustainable? If you’re saying yes it is, then why isn’t it being competed away? If it’s always outperforming in every market condition, then surely other providers will quickly replicate it and any competitive advantage will quickly disappear.
Third, quant providers might say their models work over time. But if you have two or three years of underperformance, people will start to lose confidence. You have to find a model that tries to outperform consistently and certainly one that doesn’t underperform for more than one year consecutively or people start losing faith in it.
Jean Chia, Bank of Singapore: What is so unique about Robeco’s approach? Quant may be the ‘how’, but it’s the ‘why’ that I would like to know more about. What’s your particular philosophy around quantitative investment that’s so special?
John Ng, DBS: Echoing Jean’s point, I’m sure that all Robeco’s funds – including the ones we have on our platform – have a certain level of quant behind them, even if they are not called “quant funds”. But are there any strategies – other than those of the obvious systematic hedge funds – that are purely quant? I get that you apply quant factors using big data and AI in all of your strategies, but how do we make the distinction between the fundamental and quant elements within these strategies?
Dimitrios Papacostas, Citi: For any and every kind of quant strategy, there’s a fundamental point that its parameters are still designed by people. People need to select the factors, decide if their data is tidy and decide on algorithms, such as when to use Random Forest or XGBoost. People ultimately make the rules and constraints. Doesn’t that mean that even at its most technical, no quant strategy can ever be 100% quant? Where are the borders between the model and what people have chosen to put into and overlay the model as guidelines to how it works?
APB: How do you respond to the points and challenges that Steve, Jean, John and Dimitrios have articulated, Weili?
Weili Zhou, Robeco: How do we differentiate ourselves? Well, pure quant investing, very importantly, has to be entirely evidence-based. Everybody can have investment ideas. I see thousands of them over my desk every month. But for any to be incorporated into a robust quant strategy, we have to be able to prove they can actually work for the broadest number of securities, for the longest period of time and over the widest range of market conditions.
Second, pure quant investing has to take an A-to-Z systematic and disciplined approach. There should be very little human discretion about it or interference in it. Portfolio construction and the timing, size and generation of buy or sell orders should also be extremely disciplined and systematically implemented. We have to take human greed and fear entirely out of the equation.
Hugh Chung, Endowus: I would agree with Weili. For me – having come from a fundamental background and now utilising a lot of quant strategies, the biggest difference in someone using quant filters for fundamental investment – versus someone using a totally systematic strategy – is basically being able to follow through with a factor from the beginning to the end. So if I get a signal as a fundamental analyst, I can choose to use it or not. For a quant strategy, it’s something that has to be followed – unless, of course, there is a discretionary overlay.
Weili Zhou, Robeco: I’d add that if a quant manager cannot explain what’s behind every signal and idea, then you’re talking to a bad quant manager! At the end of the day, we as quants need to be able to explain and account for every basis point coming out of a portfolio. Although the whole process has to be evidence-based and systematic, we are the designers of the model and while there might be 40, 50 or 60 signals in my model, if I’m giving you 100 basis points this month, I need to be able to look you in the eye and explain which signal was responsible and why.
APB: How do you address Jean and Steve’s points about the frailties inherent in certain past quant models and the vulnerability of quants in cases of sustained underperformance?
Weili Zhou, Robeco: Quants may have got a bad reputation in the past – and attracted a certain negative stigma – because they’ve responded badly to a meltdown and blamed some kind of generic “black box” malfunction. These were subpar quants. Today I need to be able to account for and explain every element of performance – up or down – in an entirely transparent way using and communicating the use of all available evidence.
Hugh Chung, Endowus: To follow up on Steve’s question about the sustainability of alpha generation for quant strategies. If systematic strategies’ market share among global asset managers’ total AUM continues to grow, and active or fundamental strategies’ market share falls, is there still room for continued alpha generation by similar systematic strategies by or is it ultimately destined to decline going forward?
Weili Zhou, Robeco: The alpha decay will be slower for the more differentiated and ultimately better resourced systematic strategies. The ability of the latter to, for example, buy unique data sets, invest in greater computing power or connect four or even five datasets to get an enhanced interaction effect for prediction will mean the best quants will retain and even extend their edge.
APB: Can you give us a tangible example of how Robeco is doing this?
Weili Zhou, Robeco: Certainly. We invest in listening to audio from CEOs when they are making earnings and other key announcements. We continuously expand the types of datasets in our process. You can imagine how big the database that houses this audio data needs to be. And how deep your pockets need to be to tap into and effectively use it. And to develop the machine learning infrastructure around it required to engineer it into a proven quant strategy. This is what I would term true differentiation.
APB: How do or should fund selectors sort the wheat from the chaff among the providers of quant solutions in today’s markets?
Weili Zhou, Robeco: I would check on three key considerations when differentiating between quant providers. First is infrastructure. You’ve got to go into fine details. This means examining, for example, the number of CPUs [central processing units] and GPUs [graphics processing units] a quant manager uses, the number of terabytes in its database and the precise size of its annual data budget.
It also means scrutinising how much time managers have spent developing the effective predictability of machine learning and how much output they can derive from the explainability of their machine learning. Can that quant manager break down and attribute the linear and non-linear action and feature parts of their machine learning? If they cannot, then they haven’t done a proper quant job.
Second is stress testing for a manager’s quant models’ performance in different markets and through different economic and market cycles. Third is the breadth of the manager’s coverage – a top quant should have models for emerging markets as well as developed markets, for fixed income as well as equities and so on.
Jean Chia, Bank of Singapore: Given that data on emerging markets tend to be less readily available and less granular than for developed markets, how did quant managers cope with, for example, the real estate market issues of recent years in China? Did your models predict and navigate them well?
Weili Zhou, Robeco: We were indeed able to avoid the worst excesses of China’s real estate market problems. We have made a significant investment in data in China and, more important, have learned which signals to look for and which to ignore. There are large reserves of high-quality data in China obtainable from government and regulatory sources – obviously in Chinese – that we gather and analyse on a proprietary basis.
John Ng, DBS: At the end of day you sort the wheat from the chaff by performance, risk control and volatility. Whether it’s for quant or fundamental strategies.
APB: If a fund selector asks you to pit quant against fundamental, how do you respond?
Weili Zhou, Robeco: I’m not here to advocate for quant over fundamental! I’d never advise any selector to ditch fundamental in favour of quant. At Robeco we have very strong and solid, often complementary fundamental equity and fixed income strategies and capabilities.
We’ve recently spent a lot of time analysing quant versus fundamental performance data from the past 20 years in global and emerging market equities. One key finding and conclusion was that the correlation between the two was low. It’s the same when we compare in-house performance of Robeco’s global equity fundamental and quant strategies: the performance of both was strong but the correlation between the two strategies was negative. The key reason is hundreds of different decisions around investment processes. At the end of the day, my fundamental colleagues might end up with 40-50 stocks, we might end up with 400-500.
APB: What’s your key conclusion?
Weili Zhou, Robeco: That the diversification benefits embedded in quant strategies are demonstrable, powerful and of particular relevance in today’s heavily concentrated market conditions.
