Assuming no arbitrage and that a polynomial in the time to maturity approximates spot-rate increments, we develop a consistent approach to estimate factor risk premia in the bond market. This approach is attractive, as the underlying risk factors are the time-varying coefficients of the polynomial specification, and the factor loadings correspond to familiar bond risk characteristics, such as duration and convexity. Our two-factor model—with a positive (negative) duration-factor (convexity-factor) risk premium—is easy to estimate and provides straightforward economic intuition for the concave term structure of bond risk premia and the downward-sloping term structure of bond Sharpe ratios. Our model decomposes the conditional factor risk premia into a “status quo” component—the expected factors’ realizations when expected spot-rate increments are zero—and the factors’ forecasts. The status quo factor risk premia provide measures of the state of fixed income markets that both outperform alternative measures and are easily estimated in real time; moreover, the underlying risk factors and their loadings are familiar to and readily interpreted by investors and traders. This paper was accepted by Lukas Schmid, finance. Supplemental Material: The online appendix and data files are available at https://doi.org/10.1287/mnsc.2023.00212 .
Balduzzi et al. (Thu,) studied this question.