The primary breakthrough of the Daniel, Hirshleifer, and Sun (DHS) model lies in its ability to subsume the explanatory power of traditional risk-based frameworks by isolating two distinct psychological biases: short-horizon underreaction and long-horizon overreaction. While the Fama-French five-factor model and the Hou-Xue-Zhang q-factor model rely on firm characteristics like investment and profitability as proxies for risk, the DHS model demonstrates that these are often manifestations of systematic mispricing. By augmenting the market factor with a short-horizon factor (PEAD) and a long-horizon financing factor (FIN), the researchers achieved a significant reduction in the Gibbons-Ross-Shanken (GRS) statistic, which measures the extent to which a model fails to explain a set of anomaly returns. In comparative tests against 34 prominent anomalies, the DHS model yielded a GRS statistic of 1.93, notably lower than the 3.01 produced by the Fama-French five-factor model, indicating a superior fit for empirical data.
The mechanism driving the short-horizon factor is rooted in limited investor attention and the gradual diffusion of public information. The PEAD factor, constructed from earnings surprise deciles, captures the tendency of prices to drift in the direction of an earnings shock for up to several months. Quantitatively, the long-short PEAD strategy has historically delivered an average monthly return of approximately 0.65 percent. This suggests that market participants consistently underreact to public signals, a phenomenon that persists despite the rise of algorithmic trading. The persistence of this anomaly suggests that the cost of processing information or the psychological friction of updating beliefs prevents immediate price adjustment, allowing for systematic exploitation by short-term traders.
Conversely, the long-horizon FIN factor addresses overreaction, a bias often linked to investor overconfidence. It is built on the premise that managers exploit mispricing by issuing equity when shares are overvalued and repurchasing them when undervalued. This factor captures the long-term reversal of these trends, typically playing out over three to five years. The FIN factor effectively replaces the value and investment factors found in other models, suggesting that what was previously interpreted as a risk premium for distressed or conservative firms is actually the correction of previous over-optimism or over-pessimism. The DHS model posits that investors overreact to private information—such as a firm's growth prospects—leading to the long-term mispricing that corporate financing decisions eventually signal.
From a historical perspective, the DHS model represents the culmination of behavioral finance theories first proposed in the late 1990s. Earlier research by Daniel, Hirshleifer, and Subrahmanyam in 1998 laid the theoretical groundwork by suggesting that investor overconfidence and biased self-attribution lead to the patterns observed in momentum and value. The 2020 model provides the empirical bridge, transforming these psychological insights into a tradable factor framework. The transition from the 1970s Capital Asset Pricing Model (CAPM) to the 1993 Fama-French three-factor model was driven by the discovery of the size and value effects. The DHS model suggests a similar paradigm shift is necessary, moving away from the assumption that all factors must represent non-diversifiable risk.
For institutional investors and portfolio managers, the practical implications are profound. The DHS model implies that alpha found in many smart beta strategies is actually behavioral beta that can be systematically captured. Managers can improve the Sharpe ratio of their portfolios by tilting toward the PEAD and FIN factors. Specifically, the DHS three-factor model has shown a squared Sharpe ratio that is nearly double that of the Fama-French three-factor model in certain sample periods. This suggests that a significant portion of the cross-sectional variation in stock returns is predictable based on the timing of corporate financing decisions and the market's reaction to earnings news. By distinguishing between these temporal horizons, analysts can better identify whether a stock's performance is a result of fundamental risk or a predictable correction of investor sentiment.