Can Long-Run Risks Explain the Distress Puzzle?

Abstract

The distress puzzle refers to the empirical regularity that firms with high measures of default likelihood earn anomalously low returns, despite having relatively high CAPM betas. This paper shows it is possible to qualitatively explain this anomaly using a consumption-based asset pricing model related to the literature on long-run risks. Apart from the usual assumptions of time varying, persistent conditional mean and volatility of consumption and dividend growth rates, I assume a natural cointegration between the levels of consumption and aggregate dividends. To model distress, I employ a simple intensity-based model of default. Distressed firms have short expected lifetimes, so they do not covary with the long-run risk factors, and thus earn low expected returns. Healthy firms are long-lived in expectation, and in the presence of the cointegrating relation, their prices do not respond strongly to the innovations in aggregate dividends, which lowers their betas. The model is calibrated to match the standard set of stock market moments and reproduces them well.