Stanford University


Identifying the Multiple Skills in Skill Biased Technical Change



We use an unsupervised machine learning technique, iterated exploratory factor analysis, to characterize occupations by the importance of eight endogenously derived orthogonal skills. These factors have clear interpretations and intuitive relationships to the wage distribution. We measure the relationship of each of these factors to wage and employment growth, directly and as mediated by IT usage. Leadership intensive occupations saw significant increases in both wages and employment. Physically intensive occupations saw significant decreases in occupational wages, and cooperation intensive occupations saw employment growth. The increase in leadership intensive occupational wages and decrease in physically intensive occupational wages is more pronounced for occupations and industries that use IT capital more intensely. We contrast our results for leadership and cooperation skills with those from Deming (2017) on the growing importance of social skills. We provide evidence that wage and employment growth in social skill intensive occupations nests two distinct trends. The first is an increase in wages for leadership-intensive occupations concentrated in occupations and industries with high IT capital intensity. The latter is an increase in employment for cooperation-intensive occupations concentrated in occupations and industries with low IT capital intensity.

Related Research

Stanford University