June 2024

JP Morgan Published on Morgan Markets

Homepage - J.P. Morgan Markets

JPM Macro, Quantitative & Derivatives Conference Summary

Maro Kolanovic, Dubravko Lakos-Bujas and Arun Jain

J.P. Morgan hosted its 31st Macro Quantitative and Derivatives Conference on June 18th at the NY headquarters. The conference was attended by ~550 investors representing ~260 institutions from across the world.

The keynote, stand-alone presentations, panel discussions and live investor surveys covered various aspects of Macro & Quant Investing with focus on five main areas:

1) Big-picture macro – keynote discussion compared the current regimes with past and highlighted the strategies that could thrive better; investor survey on macro conditions mostly bullish, presentation on EM opportunities: friendshoring, demographics;

2) Trend-following strategies – facts, fiction, performance and risk factor diversification;

3) Quantitative Investment Solutions – panel discussion on Systematic Fixed Income and on multi-asset QIS;

4) Tail/Crash Risk –The Risks of Chasing Sharpe and Underestimating Tail Risks, Optimizing for Crash Risk in Multi-Asset Portfolios;

5) AI in Investment Process – Evolving Intelligence: The History and Future of AI in Quantitative Investing, Using ML to Select Hedge Funds, Panel Discussion & Investor Survey on AI / Alt Data Usage.

April 2024

JP Morgan Published on Morgan Markets

Homepage - J.P. Morgan Markets

Prompt Engineering & LLMs - Pioneering Alpha Discovery

– Ayub Hanif, William Summer, Khuram Chaudhry

This research note summarizes JPM’s February 2024 DeepFin Investor Tutorial ‘Prompt Engineering and LLMs: Pioneering Alpha Discovery in Finance ’ held in-person, in London. This event took a deep dive into Prompt Engineering and Large Language Models and explore the cutting-edge role of prompt engineering in enhancing LLMs for advanced financial analysis + alpha discovery.

February 2024

JP Morgan published on Morgan Markets

Homepage - J.P. Morgan Markets

The Evolution of Quant Factor Investing

– Ayub Hanif and Khuram Chaudhry

In this research note, JPM revisits traditional quantitative equity factor investing using Machine Learning (ML) to answer the question: which factors are most important in explaining the cross-section of stock returns? Adopting a Machine Learning based approach, we can build relative dynamic importance strategies. We build a Random Forest to calculate feature importance. Our backtesting shows that Machine Learning selected factors outperform a traditional Long-only and Long / Short approach. The results are greater and use fewer factors.

JP Morgan published on Morgan Markets

Homepage - J.P. Morgan Markets

– Berowne Hlavaty

J.P. Morgan is actively applying Machine Learning techniques to analyze investment trends and patterns across all asset classes. A short list of the Big Data and AI Strategies team reports are featured on the QDS Big Data & AI portal on J.P. MorganMarkets. Natural Language Processing (NLP) continues to be an area of rapid advance, which we are applying across the research franchise. This coincides with increased adoption of NLP tools by our clients, driven by improving theoretical approaches and tools from academia and industry, especially ChatGPT and other Generative AI models.