This paper by Gerard Hoberg and Gordon M. Phillips explores how firms differentiate themselves from competitors using new time-varying measures of product differentiation based on text-based analysis of product descriptions from 50,673 firm 10-K statements filed with the Securities and Exchange Commission. The authors develop a new set of industries and corresponding measures of industry competition where firms can have distinct sets of competitors, which better explain managerial discussions of high competition, identified rivals, and firm characteristics such as profitability and leverage. They find evidence that firm R&D and advertising are associated with subsequent differentiation from competitors, consistent with theories of endogenous product differentiation. The paper also discusses the advantages of their new text-based classifications over existing static classifications like SIC and NAICS, including the ability to capture within-industry heterogeneity, product and industry changes, and cross-industry relatedness. The authors conclude that their new classifications offer significant econometric benefits, particularly in explaining firm characteristics and improving the accuracy of standard errors.This paper by Gerard Hoberg and Gordon M. Phillips explores how firms differentiate themselves from competitors using new time-varying measures of product differentiation based on text-based analysis of product descriptions from 50,673 firm 10-K statements filed with the Securities and Exchange Commission. The authors develop a new set of industries and corresponding measures of industry competition where firms can have distinct sets of competitors, which better explain managerial discussions of high competition, identified rivals, and firm characteristics such as profitability and leverage. They find evidence that firm R&D and advertising are associated with subsequent differentiation from competitors, consistent with theories of endogenous product differentiation. The paper also discusses the advantages of their new text-based classifications over existing static classifications like SIC and NAICS, including the ability to capture within-industry heterogeneity, product and industry changes, and cross-industry relatedness. The authors conclude that their new classifications offer significant econometric benefits, particularly in explaining firm characteristics and improving the accuracy of standard errors.