A0462
Title: Signaling processing monetary policy surprises
Authors: Sebastian Laumer - University of Richmond (United States) [presenting]
Italo Santos - Boise State University (United States)
Abstract: High-frequency identification has become the standard approach for identifying monetary policy shocks. Recently, however, this method has come under scrutiny. Several studies show that high-frequency instruments are contaminated by either central bank information effects or the Fed-responds-to-news channel. A methodology is developed to quantify and address both sources of contamination simultaneously. First, it is shown that instruments orthogonalized to one form of contamination remain correlated to the other. Second, a forward regression algorithm is implemented to select the optimal set of predictors. The algorithm consistently chooses a mix of economic news and central bank information variables, suggesting that both channels are present in high-frequency monetary policy surprises. Third, using the selected predictors, distinct shock series are estimated for monetary policy, central bank information effects, and the Fed-responds-to-news channel. It is found that monetary policy shocks reduce output and prices while tightening financial conditions. Fed-responds-to-news shocks, by contrast, raise output and prices and ease financial conditions. Finally, central bank information shocks increase output on impact despite rising interest rates, spreads, and excess bond premia.
