Large Sample Estimators of the Stochastic Discount Factor
We propose several large sample estimators of the stochastic discount factor (SDF) for pricing risky assets. Our estimators can utilize not only a set of factors implied by a specific asset pricing model but a set of latent factors estimated by multivariate statistical methods. We suggest a correction for the bias induced by having a finite time series and show how to use the correction in exploiting unbalanced panel of individual stock returns. The estimators perform well in simulations designed to mimic the the U.S. equity markets. A Lasso penalized version of the estimators does a good job of excluding systematic, but unpriced factors. When applied to large cross sections of equity returns, the estimators provide evidence about which factors command a risk premium.
Robert Korajczyk, Soohun Kim
Korajczyk, Robert, and Soohun Kim. 2019. Large Sample Estimators of the Stochastic Discount Factor.