The Cross-Section and Time-Series of Stock and Bond Returns


Speaker


Abstract

 

We show that the cross-section of expected returns of stock portfolios sorted along the book-to-market dimension can be explained using a factor, which we call the KLN factor, that is a linear combination of (contemporaneous) forward rates. It has a correlation of 82% with the Cochrane Piazzesi (2005, CP) factor, itself a linear combination of the same (lagged) forward rates. Since the CP factor is a strong predictor of future excess bond returns, the high correlation between the KLN and the CP factors suggests a tight link between the cross-section of stock returns and bond risk premia. Inspired by this finding, we develop a parsimonious no-arbitrage stochastic discount factor model that can price both the cross-section of stock and bond returns. Featuring only three priced factors, all of which are yields, the model obtains a mean absolute pricing error of 40 basis points per year across 5 maturity-sorted government bond portfolios, 10 book-to-market-sorted stock portfolios, and the aggregate stock market. The CP factor is responsible for pricing the spread between stock portfolios, the dividend yield prices the mean of stock returns, and the level of the term structure prices the cross-section of bond returns. With two additional risk price parameters, the model also replicates the dynamics of bond yields as well as the time-series predictability of stock and bond returns. The value premium seems related to fundamental economic risk. Since the returns on value stocks are high when CP is high, which typically occurs near the end of a recession, value stocks carry high returns exactly when investors need it the least.
 
The Erasmus Finance Seminar is jointly sponsored by ERIM and the Tinbergen Institute.
 
Contact information:
Viorel Roscovan
Email