The Evolution of Quantitative Equity Investing: Half a Century of Alpha Research and Institutional Adaptation

Abstract

This survey interprets the evolution of quantitative equity investing as a history of changing constraints rather than a succession of increasingly complicated forecasting models. Portfolio theory made diversification calculable; characteristic‐based investing converted value, momentum, profitability, and low‐risk effects into scalable processes; alternative data shifted competition toward information engineering; machine learning expanded the admissible class of return functions; and large language models and agentic systems are beginning to reorganize research itself. Yet each advance has exposed a new implementation boundary. A predictor must survive point‐in‐time reconstruction, model selection, portfolio translation, turnover, transaction costs, shorting frictions, capacity, latency, regime change, and institutional governance before it becomes alpha. The survey develops alpha decay as the mechanism linking these generations and gives particular attention to the limited transfer of machine‐learning research into production. It argues that artificial intelligence will accelerate both discovery and commoditization. Sustainable advantage is therefore unlikely to reside in a generic model class. It is more likely to arise from proprietary information transformation, economic structure, cost‐aware portfolio construction, execution, organizational memory, and tightly controlled human–machine systems.