| The price momentum anomaly is one of the premier market irregularities in the study of financial markets. Momentum trading strategies are well documented for consistently delivering returns in excess of expected market returns, although the driving factors behind this irregularity lack explanatory consensus many decades later. Different streams of financial theory have endeavored to attribute drivers of the price momentum effect to behavioral biases, market liquidity, reactions to earnings, stability price premiums, fundamental value models, and a wide range of technical patterns and measures.;This inductive quantitative multivariate statistical study assessed 24 potential factors driving the price momentum anomaly by using multiple tests of ANOVA and multiple discriminant analysis (MDA) to compare short-term momentum results against asset pricing model outcomes and well established characteristics of the momentum anomaly. In an efficient market, predictable classification patterns of stock performance should not emerge from any of the momentum segments tested. Analyzing seven distinct categories of momentum across eight industry sectors using 7,208 stocks in repeated measures identified that the strongest discriminant power was found among six of the 24 financial variables tested: RSI, Log of Beta, Log of ATR, Log of Price, Log of Dividend Yield, and Log of EPS -- each at p < 0.001. Additionally, the technical financial variables were found to explain the variance in price momentum significantly better than the behavioral and fundamental variables tested in this short-term study.;In all cases tested across the seven different momentum segments and eight different industry sectors, the null hypotheses were rejected in favor of the alternative hypotheses that predictable aspects of momentum were found at statistically significant levels. The momentum classification accuracy of the MDA tests exceeded chance probabilities by more than three times across all sectors, but showed the best results in Basic Materials, Services, Technology, and Utilities where the probability of classification into the correct momentum segments exceeded chance by more than four times. By revealing predictable elements of market momentum, these results significantly challenge principles of the Efficient Market Hypothesis, whereby no predictable patterns that support excess market returns should exist. |
