Monetary policy surprises: Robust dynamic causal effects

We build on the Local Projection Instrumental Variable (LP-IV) approach to assess the effects of a monetary policy shock on the US economy, focusing on inference. We pay particular attention to the specification of control variables that warrant exogeneity of instruments. We introduce a response parameter that is always identified despite missing controls. Simulation results reveal that the standard LP-IV approach yields severe over-rejections as control requirements are relaxed, in contrast to our procedure which controls size and has good power. Empirical results are robust to expanded credit spreads data and to various competing instruments building on High Frequency Identification.