Subjective Shadow Rate Beliefs at the Zero Lower Bound

Abstract

We estimate shadow rates (Black (1995); Wu and Xia (2016)) using forward rates on US Treasuries and forecasts of short term interest rates from the Blue Chip Financial Survey. We estimate a suite of alternative models with different numbers of factors, with and without forecast data, and assuming forecasts are full information rational expectations or that forecasters have distorted beliefs. Regardless of how expectations formation is modeled, our estimates that incorporate forecasts imply: (1) deeper easing by the Federal Reserve during the ZLB period relative to previous estimates, (2) longer real-time forecasts of the duration of the ZLB period, and (3) a structural break in the effect of monetary policy on the macro-economy before and after the Great Recession. Our estimates also imply that the perceived path of policy was much tighter than the Fed’s official projected path, suggesting the FOMC’s statements were not perceived as fully credible during the Great Recession.