A1249
Title: Investor attention interaction and asset pricing: Evidence from co-search behavior
Authors: Jieying Zhang - Tohoku University (Japan) [presenting]
Yasumasa Matsuda - Tohoku University (Japan)
Stanley Iat-Meng Ko - Tohoku University (Japan)
Runyu Dai - Tohoku University (Japan)
Chia Chun Lo - Huatai Futures (China)
Abstract: The aim is to investigate the explanatory and predictive power of investor attention interaction (IAI) in asset pricing, using 21.33 billion sequential searches from the U.S. Securities and Exchange Commission's EDGAR system. A time-varying, directional IAI network matrix is constructed based on investors' co-search behavior. Embedding this structure within a spatial arbitrage pricing theory (S-APT) framework, it is shown that the IAI network improves the performance of factor models by reducing pricing errors (alphas) to statistical insignificance. The inclusion of the IAI network reshapes risk exposures and return comovements across firms. It is further demonstrated that the IAI network predicts future firm fundamentals, including return on assets (ROA) and standardized unexpected earnings (SUE), and strong return predictability is exhibited. Extensive robustness checks confirm that these findings are not driven by alternative model assumptions, construction methods, or spurious linkages. Finally, evidence that the effect decays with network distance and that it decays even faster in degree-preserving placebo networks supports an attention-mediated information-diffusion mechanism. Results highlight the importance of incorporating investor attention dynamics into asset pricing models.
