Constructing Core Inflation Index for Russia

The paper studies a possibility to construct a price index (“core inflation” index) for Russia, which is more appropriate for assessing monetary policy than consumer price index. The index is derived by applying to Russian data trimmed mean method, which explores cross-sectional dimension of the data. We find that cross-sectional distributions of individual price changes are significantly long-tailed. In this situation ordinary sample mean is not very efficient and robust estimator of the central tendency. Trimming 50-100% of observations leads to a large gain in efficiency. Moreover, trimmed mean series is shown to be less volatile than official consumer price index.