Goldman Sachs Asset Management: Hedging macro uncertainty with quantitative precision
Goldman Sachs Asset Management: Hedging macro uncertainty with quantitative precision
Goldman Sachs Asset Management: Hedging macro uncertainty with quantitative precision
In an article, originally published by Fund Selector Asia, Alison Lau, Head of Public Equity Client Portfolio Management in Asia Pacific Ex-Japan, explains how the Quantitative Investment Strategies team's informational edge can unlock alpha opportunities in European equities.
As markets fragment and data explodes, data and technology-driven platforms built for scale are well positioned to consistently create an information advantage to uncover insights before the rest of the market does.
Longevity and success in systematic investing are not simply a result of years in operation. They require accumulated intellectual property, research archives and lived experiences across market cycles. Even then, outperformance is not a given.
In practical terms, making insights count relies on being able to synthesise more data, quicker, and to extract economically intuitive signals before they are fully priced.
This is the focus of Goldman Sachs’ Quantitative Investment Strategies (QIS) team, which has a track record of data-driven investing of over 35 years. During that time, the goal has been to refine a single idea: that durable alpha comes from processing information better, faster and more consistently than the rest of the market, uncovering opportunities before they become mainstream.
It helps that this is part of an organisational mindset. The QIS team is able to consume hundreds of datasets daily based on the support of an annual firm-wide data and technology budget. “We are able to access more data thanks to the broader resources of Goldman Sachs,” said Alison Lau, managing director, Head of Public Equity Client Portfolio Management in Asia Pacific Ex-Japan.
But access alone is not enough. The edge lies in converting raw information into systematic, repeatable insights.
With investing today becoming more complex as data volumes explode and news cycles accelerate, managers able to analyse information at scale – and interpret it with discipline – are better placed to generate consistent alpha, added Lau. “This competitive edge drives our investment process.”
A legacy that leads to an ‘information edge’
The QIS platform offers an institutional memory that few quant entrants can replicate.
New researchers can leverage the endeavours of prior portfolio managers and data scientists, with access to a deep research library and documented evidence of what worked – and didn’t work – in different regimes.
This is important in periods of stress, Lau explained. During the liquidity shock of March 2020, for example, the team observed not just drawdowns, but the breakdown of correlations and co-movements across strategies. When subsequent liquidity crunches hit, portfolios were structured with a clearer understanding of how signals behave under pressure. For allocators, that translates into resilience as well as return.
More broadly, Lau describes the QIS philosophy as “fundamental thinking, done systematically”. Every signal is grounded in economic intuition.
Making AI matter
Over the past decade, advancements in AI have enabled the team to process entire earnings transcripts, regulatory filings, sell-side research and news flows in one pass, capturing nuance and tone at scale.
Take sentiment, she explained, as a cornerstone of discretionary analysis. Around 2010, the approach to capture sentiment around a company relied largely on counting positive or negative words. With the introduction of transformer-based language models, context has become measurable.
The result is a sentiment framework that mirrors what a fundamental analyst would do, but applied consistently across thousands of companies and continually updated.
Beyond sentiment, leveraging AI to process a large quantity of various text documents could help discover themes and map corporate networks. For example, when AI-related spending accelerated in late 2023, Lau said the models identified not only semiconductor leaders but also secondary beneficiaries – such as companies supplying ventilation systems for data centres, copper producers tied to power infrastructure, and other linked firms whose exposure was not yet fully recognised.
By systematically identifying these interconnections, portfolios can be positioned ahead of slower price discovery.
Europe: ripe for a systematic approach
Putting the QIS approach into practise, European equities present an appealing allocation – especially given they are generally still underrepresented in global portfolios after years of US concentration.
Notably, the region is structurally fragmented: ~50 countries, more than 20 exchanges, and multiple languages and regulatory regimes. Further, valuations of European equities trade at a discount, dividend yields are higher and fiscal shifts – particularly in infrastructure, defence and energy – point to supportive macro dynamics. Currency trends and diversification considerations strengthen the case.
For discretionary managers, that complexity raises research costs. By contrast, for a systematic platform with scalable data ingestion and multilingual NLP capability, it creates opportunity.
Information in Europe often diffuses more slowly and economic linkages are less obvious, said Lau, pointing to an example of a German auto parts supplier, which may be influenced as much by French consumer confidence or Chinese demand as by domestic conditions. This makes quantitative network models well suited to identifying such non-linear relationships, she added.
Against that backdrop, inefficiencies created by fragmentation and uneven coverage provide fertile ground for systematic alpha.