Evolution of Financial Studies Over Forty Years: What Can We Learn from Machine Learning?

Abstract

How did the finance research topics evolve in the past forty years? We apply machine learning models of textual analysis on 20,185 abstracts of finance articles published between 1976 and 2015, and identify 38 research topics. We present the fastest growing topics of published and working papers. Our algorithm can be used to categorize the articles without JEL codes. We use citation network to present how topics are related, and cluster the topics in five “territories”. Moreover, we find a strong bibliometric regularity: the number of researchers covering n topics is approximately 1/2^n of those covering just one topic.

Keywords: Textual Analysis, Machine Learning, Network Analysis, Evolution of Financial Studies

JEL Classification: G00, G10, G20, G30, B26