Scenario-based systematic risk in earnings

Abstract

We develop a scenario-based measure of risk in earnings. We start by forming a range of scenarios, or plausible states of the economy, represented by earnings persistence parameters from cross-sectional earnings prediction regressions. We then combine out-of-sample earnings components with these scenarios to predict earnings. The standard deviation of predicted earnings of all scenarios is our firm-year risk measure, sigma. We demonstrate that $\sigma$ captures systematic risk in earnings and that our measure complements firm characteristics that have been proposed as measures of risk in predicting cross-sectional returns. The differential one-year returns of the interquartile range for $\sigma$ is 1.9\%. We contribute to the literature by developing a priced scenario-based risk measure using only fundamental information.