Research

Inflation expectations play a crucial role for macroeconomic dynamics and more specifically for monetary environment. However, inflation expectations is an unobservable variable. So, the quality of the correspondent measure in a great extent predetermines its feasibility for macroeconomic analysis. Today, survey-based measures of inflation expectations prevail in macroeconomic analysis. However, the drawbacks and/or unavailability of such measures give a rise to other identification strategies. Extracting inflation expectations from the actual data (e.g. series of interest rate and actual inflation) basing on SVAR identification approach has become a valuable alternative/supplement for measuring inflation expectations. In this paper I show that the existing strategy of inflation expectations identification through SVAR approach is very sensitive to the state of monetary environment. When a monetary environment is unstable (e.g. high and volatile inflation), the assumptions of the baseline approach are not hold, and it produces biased estimations. I emphasize two sources of this bias in estimations and suggest procedure for obtaining unbiased estimates. My identification strategy includes a number of steps. I suggest applying Markov regime-switching framework for extracting an unbiased mean for ex ante real interest rate. Further, I use two-stage SVAR identification strategy. First, I identify an unexpected shock to actual inflation, which is crucial for obtaining a proper measure of inflation expectations. Further, I net the series of ex post interest rate from this ‘noise’. Second, I run a baseline SVAR procedure, for which I use the data adjusted at the first step. Finally I obtain an unbiased and informatively rich series of inflation expectations.