Are Patents with Female Inventors Under-Cited? Evidence from Text Estimation

Abstract

Utilizing leading machine learning techniques to analyze the textual content and quality of patents, we demonstrate that patents with female lead inventors are under-cited relative to what would be expected had the lead inventor been male, with the effect amplified for significant (novel and impactful) patents. Male inventors are the greatest contributors to the undercitation of patents with female inventors, followed by female inventors and male examiners, while female patent examiners appear to be even-handed. Marketbased measures of patent value appear to recognize this undercitation of significant femaleinventor patents. The results have potential implications for research conclusions that rely on citation-based assessments of patent quality.

Keywords: C13 Innovation, Gender, Patent, Machine Learning, Big Data, Inference

JEL Classification: J16, J24, J71, O30, C13