We measure monetary policy shocks as changes in the Fed funds target rate that surprise bond markets in daily data. These shock series avoid the omitted variable, time-varying parameter, and orthogonalization problem of monthly VARs, and do not impose the expectations hypothesis. We find...