Cosearch Attention and Stock Return Predictability in Supply Chains

The ability to make predictions based on online searches in various contexts is gaining substantial interest in both research and practice. This study investigates a novel application of correlated online searches in predicting stock performance across supply chain partners. If two firms are economically dependent through a supply chain relationship and if information related to both firms diffuses in the market slowly or rapidly, then our ability to predict stock returns increases or decreases, respectively. We use online cosearches of stock as a proxy for the extent of information diffusion across supply chain-related firms. We identify publicly traded supply chain partners using Bloomberg data and construct cosearch networks of supply chain partners based on the weekly coviewing pattern of these firms on Yahoo! Finance. Our analyses show that the cosearch intensity across supply chain partners helps determine cross-return predictability. When investors of a focal stock pay less attention to its supply chain partners, we can use lagged partner returns to predict the future return of the focal stock. When investors' coattention to focal and partner stocks is high, the predictability is low. Our simulated trading strategy using returns of supply chain partners with low coattention generates a significant and positive return above the market returns and performs better than the previously established trading strategy using returns of all supply chain partners.