Assessing Monetary Policy Effectiveness in Rich Data Environment

We assess impact of monetary policy actions upon inflation in a country while considering changes in global commodity prices in rich data environment. We apply Factor Augmented Bayesian Structural Vector Autoregression (FABSVAR) methodology of Bernanke et al (2005) upon Pakistan’s monthly data for July 1992-June 2015. Unlike Bernanke et al (2005), we combine variables of similar nature in groups to extract factors. We think putting all sorts of variables in one group impairs the factor extraction. Moreover, rather than working only with the response of variable of interest (inflation in this study) to shocks in factors under consideration (which, consist of different interest rates, monetary aggregates, exchange rates in this study) we propose use of eigenvector to obtain Impulse Response Functions (IRFs) of shock to individual variables in a factor. We find significant and desired impact of monetary policy decisions upon inflation in Pakistan. Administered prices, however, are found to have no response to interest rate changes in the country, which is understandable. By analyzing IRFs of inflation in Pakistan to shocks in interest rate we do not observe any price puzzle. It is simply because we consider the relevant variables omitted in previous studies reporting price puzzle in Pakistan.