All Days Are Not Created Equal: Understanding Momentum by Learning to Weight Past Returns
Beckmeyer, Heiner; Wiedemann, Timo
Abstract
By flexibly weighting the information contained in past realized returns, we construct a momentum strategy that outperforms and subsumes the performance of traditional stock momentum. The strategy performs well in crises and continues to work in the most recent decades. We show that the way past returns are weighted is in line with the strategy exploiting an underreaction to the information contained in realized returns, but also investigate alternative behavioral and risk-based explanations. We find that the response to earnings announcements, market-wide jumps and large individual returns realized in the formation period are most informative about future stock returns.
Keywords
Momentum, Machine learning, Big data, Anomalies