Abstract
We use a large and unique corpus of job specifications for C-suite positions to document and explain heterogeneity across firms in the skills demanded for high-level executives. A novel classification algorithm maps the text for each executive search into six separate skill clusters, all of which show considerable variation. Patterns in the \textit{social skills} cluster are particularly striking: it features large growth in the sample, is the relatively most common cluster in CEO searches, and has a large sample variance. We propose a mechanism whereby social skills faciliate the exchange of problems within firms; construct proxies for the need for such coordination; and show these correlate with the presence of social skills language. These correlations do not appear for the other skill clusters. The results suggest that the varied structure of firms induces demand for different executive skills.
(with Joe Fuller, Stephen Hansen and Tejas Ramdas)