Essays On Stock Return Predictability

The dissertation consists of two essays. Essay I examines the return predictability by firm level R & D and innovation measures and shows that technology spillover helps to explain the positive innovation-return relation. Essay II propose a novel measure of conditional value premium based on firm's stock split announcement. This measure is shown to have a strong predicting power over value premium both in sample and out of sample. Essay I: I show that technology spillovers are important information phenomena that benefit both other innovators (as emphasized in the Industrial Organization literature) and stock market investors. I find that the premium associated with R & D and patenting activities is largely restricted to firms located in more isolated technology spaces with fewer spillovers. Moreover, there is a strong lead-lag effect among firms engaging in innovative activities: the stock prices of firms in more isolated technology spaces react more slowly to new information than do the stock prices of firms in more competitive technology spaces. Finally, announcement-day returns to patent grants are greater for more technologically important patents (measured by forward citations), but only for firms in more crowded technology spaces. My results indicate that investors are able to value innovative investments by exploiting the information flows associated with greater technology spillovers. Essay II: I propose a novel conditional value premium measure based on the present-value relation that the stock price impact of a firm's public announcement reveals the firm's expected discount rates. Specifically, because most splitting stocks are growth stocks on which, by construction, the value premium has strong influence, the average splitting stock announcement-day returns track closely conditional value premium. I find very similar results using announcements of divested asset acquisitions in which acquirers are usually growth firms. Consistent with risk-based explanations, my conditional value premium measure correlates positively with future GDP growth and helps explain the cross-section of stock returns.