"Causal Inference for Traders: Moving Beyond Correlation in Financial Markets"
In modern markets, most trading models are still built on fragile correlations that unravel the moment regimes shift. This book is for systematic traders, quantitative researchers, data scientists, and risk managers who want to move beyond black-box prediction and towards principled decision-making. It bridges the gap between academic causal inference and the messy realities of financial data, showing how to reason about "what would have happened" under alternative trading rules, signals, or policies.
The book develops a complete toolkit for causal analysis in markets: from returns, microstructure, and backtest hygiene, through probability, estimation, and machine learning foundations, to formal causal frameworks with DAGs, potential outcomes, and identification rules. Readers learn how to define estimands like ATE and CATE in P&L terms; deploy matching, weighting, and doubly robust methods; and exploit quasi-experiments, DiD, RDD, IV, and synthetic control in time-series and panels. The final chapters convert effects into tradable policies via offline evaluation, policy learning, causal reinforcement learning, and robust, governed deployment.
The text assumes comfort with basic statistics, linear algebra, and programming, but it is self-contained in its treatment of causal concepts. Throughout, financial examples and implementation-oriented discussions emphasize realistic workflows and failure modes, making this a practical field guide rather than a purely theoretical monograph.
