Richard Falk Wallace, Co Founder & CEO of ARCANA, at AI for Financial Services NYC – Risk Modelling & Performance at the Age of AI / ML

Current practices in risk and crowding analysis

Top funds are increasingly focused on understanding unique insights and idiosyncratic returns, separate from macroeconomic factors. This approach applies to both equity long-short multi-manager funds and long-only investors. The emphasis is on decomposing systematic components of return from residual return, with a keen eye on crowding among different types of investors.

The cutting edge: advanced risk insights and crowding

Advanced funds are developing sophisticated methods to understand public stock risks, using factor exposures and crowding signals from various markets. This includes detailed analysis of crowding at a granular level, examining exposures to specific investor types, such as hedge funds and multi-managers. This focus on crowding helps refine investment decisions and mitigate risks.

AI's role in enhancing stock selection

AI and machine learning (ML) offer significant potential in refining stock selection. AI can process vast amounts of unstructured data, such as investment memos, research notes, and meeting transcripts, translating them into structured data sets. This helps in expanding systematic understanding and narrowing the scope of discretionary decisions, thus enhancing the overall investment process.

Future prospects: systematic vs. idiosyncratic risks

The future of equity stock selection lies in separating systematic and idiosyncratic risks more effectively. AI can help identify and model these risks, integrating insights from unstructured data and fundamental processes. This systematic approach can improve stock picking performance by providing a clearer understanding of the risks involved.

Unstructured data and AI integration

Integrating unstructured data into the investment process involves capturing fundamental process data and understanding regime shifts and thematic factors. AI can help translate chaotic information into a systematic framework, aiding in stock selection and risk management. This integration enhances the ability to identify systematic pieces and synthesize insights for better decision-making.

Tooling and system integration

Developing user-friendly tools and systems is crucial for integrating AI insights into the investment process. These tools should facilitate easy access to insights for CIOs, portfolio managers, risk managers, and analysts. A blended approach combining AI-driven insights with traditional stock picking methods can significantly enhance public market equity management.

Practical applications and current use

Currently, few funds use AI in a systematic way for stock selection. AI is mainly used for efficient search and document analysis. However, the hope is to develop frameworks that incorporate vast amounts of investment data into coherent systems, distinguishing between alpha and beta insights. This requires robust infrastructure to separate and utilize investment research effectively.