Title: Quasi ex-ante inflation forecast uncertainty
Authors: Wojciech Charemza - University of Leicester (United Kingdom)
Carlos Diaz - University of Leicester (United Kingdom)
Svetlana Makarova - University College London (United Kingdom) [presenting]
Abstract: A new measure of inflation uncertainty is proposed, which is, to an extent, net of the effects of the monetary policy decisions. The measure is computed using the parameters of the weighted skew-normal distribution fitted to inflation forecast errors, which can be interpreted as representing monetary policy outcomes. These parameters are then used for recovering a hypothetical distribution free of such outcomes, called the distribution of the pseudo-ex-ante uncertainty. Under some general assumptions, it can be interpreted analogously to that of the distribution of the ex-ante uncertainty in a survey of professional forecasters. The measure we propose, named the uncertainty ratio, is the ratio of variance of this hypothetical distribution to the mean squared error of the original weighted skew-normal distribution fitted to forecast errors. This measure is computed for annual inflation measured monthly for 38 countries. With the use of this ratio, we provide some evidence that the central banks independence contributes positively towards a reduction in forecast uncertainty. We also evaluate ex-post and pseudo-ex-ante forecast uncertainty term structure for U.K. and U.S. using new indicators and compare with earlier results.