Hansen and Jagannathan (1991) (hereafter HJ) derive restrictions on the volatility of stochastic discount factors that price a given set of returns. This article studies the sampling properties of HJ bounds that use conditioning information. One approach is to multiply the returns by the lagged variables. We also study optimized HJ bounds with conditioning information from Gallant, Hansen, and Tauchen (1990) and based on portfolios derived in Ferson and Siegel (2001). We document striking finite-sample biases in the HJ bounds, where the bounds reject asset-pricing models too often. We provide a useful bias correction. We also evaluate asymptotic standard errors for the bounds from Hansen, Heaton, and Luttmer (1995).

Hansen and Jagannathan (1991) (hereafter HJ) derive restrictions on the volatility of stochastic discount factors that price a given set of returns. This article studies the sampling properties of HJ bounds that use conditioning information. One approach is to multiply the returns by the lagged variables. We also study optimized HJ bounds with conditioning information from Gallant, Hansen, and Tauchen (1990) and based on portfolios derived in Ferson and Siegel (2001). We document striking finite-sample biases in the HJ bounds, where the bounds reject asset-pricing models too often. We provide a useful bias correction. We also evaluate asymptotic standard errors for the bounds from Hansen, Heaton, and Luttmer (1995). Copyright 2003, Oxford University Press.

ABSTRACT Even though stock returns are not highly autocorrelated, there is a spurious regression bias in predictive regressions for stock returns related to the classic studies of Yule (1926) and Granger and Newbold (1974) . Data mining for predictor variables interacts with spurious regression bias. The two effects reinforce each other, because more highly persistent series are more likely to be found significant in the search for predictor variables. Our simulations suggest that many of the regressions in the literature, based on individual predictor variables, may be spurious.