A “human-machine team” approach to generate investment insights is important in a rapidly evolving and data driven world, Ahmed Talhaoui, head of BlackRock Systematic Group for Emea and Apac, told a media briefing in Hong Kong last week.
However, “systematic investing is not a black box”. Instead, it combines human knowledge, experience and creativity with the efficiency of machines.
BlackRock’s systematic investing platform, which started in 1985 in San Francisco, now manages over $200bn for institutional, wealth, and retail clients globally. The team has members in San Francisco, London, Singapore, Hong Kong, Tokyo and Sydney, and the platform has seen significant growth with 90% of AUM outperforming benchmark or peer group median over 5 years, according to Talhaoui.
The objective is to deliver consistent and differentiated alpha by making smaller, diversified bets across a wide range of securities. This approach seeks to balance risk, return and cost objectives which entails relying on a research-drive process to maintain diversification across sectors, countries and execute trading strategies that minimise costs.
“This approach amplifies human decision-making by using a data-driven method, applying the scientific method, and having disciplined portfolio construction,” said Talhaoui.
Indeed, “this unique blend of technology and human expertise helps overcome behavioural biases and cognitive errors, and ultimately aims to deliver more robust portfolios for clients,” said Anthony Kruger, Apac head of platform strategy for BlackRock Systematic Group, told the briefing.
BlackRock’s systematic investing team leverages Large Language Models (LLMs) to enhance its investment process. These models allow the team to process and analyse vast amounts of data efficiently.
By using natural language processing (NLP), the team can read and interpret millions of job reports, broker reports, social media posts, earnings call transcripts, and newswires in multiple languages.
“This capability helps the team build scale within their operations and ensures that they capture more nuanced insights that might be lost in translation,” said Kruger.
Moreover, the asset manager’s systematic investing uses customised LLM tailored for financial markets, enhancing the breadth and relevance of investment insights. By fine-tuning on 30GB of financial text, the BlackRock model forecasted post-earnings stock returns more accurately than generic GPT models, they claim.
Of course, there is no guarantee that a positive investment outcome will be achieved, and diversification and asset allocation may not fully protect investors from market risk.
But, in a world where technology and data are paramount, systematic investing could stand out by leveraging both to create a dynamic and adaptive investment strategy.
“Although, systematic investing relies heavily on data and algorithms, human input remains crucial,” Talhaoui stressed. Creativity and judgment are essential for generating innovative investment ideas and identifying emerging trends. “This human element complements the pure quantitative approach,” he said.