Giuseppe Paleologo on Quant Investing, Factor Models, and AI at Hedge Funds

Defining Quantitative Investing

• 🎯 Quantitative investing is broadly defined by the number of independent or quasi-independent bets an investor makes, requiring the ability to scale a method across a large number of opportunities.
• 💡 While many investors use quantitative overlays, true quant investing is differentiated from discretionary or value investing by its systematic approach to a large cross-section of assets with low expected returns in each.
• 🔑 A key characteristic of a factor is that it must be pervasive, affecting a wide range of securities, and persistent over time, not just a short-term anomaly.

The Role of Quantitative Research

• 📊 The global head of quantitative research provides centralized quantitative services, including developing factor models for equities, managing firm-level and PM-level hedging, and offering portfolio advisory services.
• 🧠 Idea generation for factors involves identifying attributes that are pervasive, persistent, and interpretable, distinguishing them from mere themes like AI or specific political events.
• 🛠️ Quantitative researchers help Portfolio Managers (PMs) construct better portfolios, understand performance, manage risk, and sometimes act as therapists.

Factor Identification and Alpha

• 🔍 Isolating factors requires techniques to separate their impact from other correlated or overlapping factors, ensuring that the identified return is truly from the intended source.
• 🚀 While well-known factors like value and momentum can provide priced returns, true alpha is ideally a return with no associated risk, which is rare; often, it's an exploited factor with unique characteristics.
• 📈 The capacity of factors can be exhausted, leading to diminished returns or increased crashes, while others, like medium-term momentum, may still offer positive, albeit smaller, Sharpe ratios.

Quant Investing in the Modern Era

• 🤖 Generative AI and LLMs are currently being used to enhance productivity, with potential future applications in more complex investment strategies, especially in data-rich environments.
• 🌐 Large firms with scale, historical data, and proprietary datasets may have an advantage in leveraging AI, but the rapid pace of change makes it challenging to adapt.
• ⚠️ Regime changes in markets are difficult to detect and act upon effectively; it can be easier to detect changes in a portfolio manager's behavior and adapt to that.
• 📉 Original factors like size have become less explanatory on their own, often being combinations of other factors, while others have seen their capacity exhausted or exhibit different characteristics like increased crashes.