Post-Fundamentals Drift in Stock Prices: A Machine-Learning Approach

Deviations of accounting fundamentals from their preceding moving averages forecast drifts in stock prices. Comprehensive machine-learning measures based on such deviations yield annualized alphas that exceed 18% (8%) for equal- (value-) weighted portfolios. The return predictability goes beyond momentum, 52-week highs, profitability, and other prominent anomalies. The profitability applies strongly to the long-leg and survives value-weighting and excluding microcaps. We provide evidence that the predictability arises because investors underreact to deviations from prevailing fundamental anchors.