The Cyber Risk Premium

Abstract

This paper studies how stock markets perceive and price cyber risk. We estimate the ex-ante likelihood for a firm to experience a data breach using logistic LASSO regressions combined with cross-validation. Ranking firms based on this proxy for cyber risk, we find that it influences both investor portfolio choices and stock prices. In particular, institutional investors tend to sell stocks with high cyber risk and buy those with low cyber risk; this tendency is stronger during periods with higher data breach concerns. We show that a one-standard deviation increase in cyber risk is associated with a premium of 3.41% per annum.

Keywords: Cyber Risk, Cybersecurity, Risk Premium, Machine Learning, LASSO, Cross-Validation

JEL Classification: G11, G12, G14, G17