Persistent Leverage in Portfolio Sorts: An Artifact of Measurement Error?

Studies such as Lemmon, Roberts and Zender (2008) demonstrate how stable firms’ capital structures are over time, and raise the question of whether new theories of capital structure are needed to explain these phenomena. In this paper, I show that trade-off theory-based empirical proxies that are observed with error offer an alternative explanation for the persistence in portfolio-leverage levels. Measurement error noise equal to 80% of the cross-sectional variation in the market to book ratio, coupled with slight mismeasurement of other factors, matches simulated data moments to empirical moments. This suggests that unobserved investment opportunities play an important role in explaining leverage ratios.