Measuring Corporate Culture Using Machine Learning
92 Pages Posted: 21 Oct 2018 Last revised: 19 Feb 2019
Date Written: February 15, 2019
Using one of the latest machine learning techniques — the word embedding model (Mikolov et al. 2013) — and 217,387 earnings call transcripts, we obtain five corporate cultural values: innovation, integrity, quality, respect, and teamwork (Guiso, Sapienza, and Zingales 2015) for 77,541 firm-year observations over the period 2001–2018. We conduct a large number of tests to validate our measures and demonstrate the advantages of our approach over alternative methods. We show that corporate culture plays an important role in deal incidence and merger pairing, and that post-merger, acquirers’ cultural values are positively associated with their target firms’ cultural values pre-merger.
Keywords: machine learning; word embedding; unsupervised learning; corporate culture; cultural fit; acculturation; mergers and acquisitions
JEL Classification: C45; G34
Suggested Citation: Suggested Citation