Cleveland Fed's Median CPI Inflation vs. BLS
Greg Mankiw has an interesting post on the difference between standard CPI inflation reported by the BLS and an alternative measure of inflation calculated by the Cleveland Federal Reserve based on the median CPI. Professor Mankiw points out that:
The average of any data set can be thrown off by a few extreme outliers; the median is a more robust statistic to estimate the central tendency in the data.
Right now, the two measures of inflation are diverging substantially. The standard CPI shows deflation over the past year, but that average is due to a few anomalous sectors, such as energy. If you look at the median CPI, which shows what a more typical price is doing, the inflation rate does not look very unusual (see chart above).
MP: The BLS reported on Wednesday that annual standard CPI inflation (deflation) from May 2008 to May 2009 was -1.30% (see chart), mostly because of a -27.3% decrease in the energy prices and a -14.3% decrease in transportation prices from a year ago (when gas was almost $4 per gallon and oil was about $125 per barrel).
In contrast, the Cleveland Fed reported on the same day that its adjusted, Median CPI increased by +2.4% year-to-year through May 2009 (see chart).
The Cleveland Fed has been studying and reporting median CPI for a long time, here is a paper from 1991 on "Median Price Changes: An Alternative Approach to Measuring Current Monetary Inflation."
Here's an interactive graphing feature from the Cleveland Fed for creating charts of inflation.