Lens or Prism? Patent Citations as a Measure of Knowledge Flows from Public Research

Top Cited Papers
Publisher Website
Google Scholar
Abstract
This paper assesses the validity and accuracy of firms' backward patent citations as a measure of knowledge flows from public research by employing a newly constructed data set that matches patents to survey data at the level of the research and development lab. Using survey-based measures of the dimensions of knowledge flows, we identify sources of systematic measurement error associated with backward citations to both patent and nonpatent references. We find that patent citations reflect the codified knowledge flows from public research, but they appear to miss knowledge flows that are more private and contract based in nature, as well as those used in firm basic research. We also find that firms' patenting and citing strategies affect patent citations, making citations less indicative of knowledge flows. In addition, an illustrative analysis examining the magnitude and direction of measurement error bias suggests that measuring knowledge flows with patent citations can lead to substantial underestimation of the effect of public research on firms' innovative performance. Throughout our analyses we find that nonpatent references (e.g., journals, conferences, etc.), not the more commonly used patent references, are a better measure of knowledge originating from public research. This paper was accepted by Lee Fleming, entrepreneurship and innovation.