Technological similarity and stock return cross-predictability: Evidence from patents’ big data
Natalya V. Khimich,Ron Bekkerman
DOI: https://doi.org/10.2139/ssrn.3147218
Abstract:We investigate asset pricing implications of innovation links between firms. We identify technologically similar patents and firms via textual analysis of 7.7 million patents. We find that information about technological similarity is impounded into stock prices, although not immediately and fully; stocks of technologically similar firms cross-predict one another’s returns. This effect is distinct from the industry and supply-chain momentum and predictable lead–lag relations between large and small firms. The magnitude of predictability increases with information opacity and investors’ inattention to links as proxied by common analyst coverage of technologically similar pairs, suggesting attention constraints as a potential explanation. ___________________________ *Corresponding author. Drexel University, 3220 Market Street, Philadelphia, PA 19104. Tel: 215-895-2118. Email: nvk24@drexel.edu We thank Eliezer Fich, Fabrizio Ferri, Curtis Hall, David Hirshleifer, David Gefen, Leonid Kogan, Richard Sloan, Mark Vargus, Frank Zhang, conference participants at the 2017 AAA annual meeting, and seminar participants at Drexel University for insightful comments and suggestions. In addition, we thank Olga Donin, Research Associate at the University of Haifa, for building the data infrastructure used in this research.
Economics