Abstract:
How much do term premiums matter for explaining the dynamics of the term structure of interest rates? A lot. We characterize the expected path of nominal and real short-rates as well as inflation using the universe of U.S. surveys of professional forecasters covering more than 500 survey-horizon pairs. We obtain term premiums as the simple difference between observed government bond yields and survey-based expected average short rates. Our term premiums measured directly based on expectations accommodate perceived structural change and learning effects, are consistent with a lower bound on nominal interest rates, and uncover a number of important facts: 1) the bulk of the variation in medium- and long-term bond yields is driven by term premiums, not expected short rates or inflation; 2) term premiums co-move more strongly across maturities than expected short rates or even yields themselves; 3) the term premium, not the term spread or the expected path of future short rates, predicts quarterly real output growth; 4) macroeconomic factors are important drivers of term premiums, with demand shocks playing the most prominent role; and 5) the secular decline of U.S. long-term bond yields over the past thirty years is primarily the result of a decline of expected inflation and term premiums while expected future real rates have fluctuated around 2 percent.
