Exploring the Schumpeterian Entrepreneur’s Personality: A Natural Language Approach with Neural Networks Estimates

This study examines the personality traits of Schumpeterian entrepreneurs by comparing them with those of non-Schumpeterian entrepreneurs, using a sample of Forbes billionaires and analysing posts from X. The RoBERTa language model, fine-tuned with the Pennebaker and King 1999 Essay I dataset, was employed for natural language processing of X’s posts to evaluate personality traits. Two quasi-experiments with propensity score matching were performed for the treatment and control groups. The two quasi-experiments highlighted Schumpeterians’ second-order stochastic dominance in openness to experience and non-Schumpeterians’ second-order stochastic dominance in neuroticism and extraversion. These findings offer insights into the promotion of entrepreneurship by considering personality differences. However, the generalizability of the findings is restricted by the constrained sample size. Due to data availability restrictions, only 113 entrepreneurs from the Forbes billionaires list, who have an active, verified X account, were included in the sample. Further research with larger and more diverse samples is required to confirm and expand these findings.