Subjective Shadow Rate Beliefs at the Zero Lower Bound

Abstract

We investigate the sensitivity of estimates of the "shadow rate", the counterfactual one-period interest rate that would obtain absent the zero lower bound. Using U.S. Treasury forward rates and short-term survey forecasts, we estimate models with different factor structures and assumptions about forecasts. Revisiting the channels of monetary policy from 2008-15, we find robust evidence of a structural break in monetary policy transmission before and after the Great Recession. While most models agree on the effects of calendar-based forward guidance, only models with distorted forecasts attribute non-trivial portions of announcement effects to expectations of short-term rates rather than term premia.