Consumer Price Index

Earlier this week the Bureau of Labor Statistics (BLS) published its monthly Consumer Price Index (CPI) stating that for the month of February 2021 it had increased 0.4% and year-over-year 1.7%. Unfortunately, the way the Index is calculated it is ripe for misrepresentation. 

In this post I will simply point out exactly what the BLS discloses with respect to the calculation methodology and how it can be misapplied. Of course, the plasticity of the way the Index is crunched would make any reasonable person not place much reliance on it. The government does not mind the way it is calculated because some payments, like Social Security or some debt interest, are adjusted based on the CPI value. Therefore, they will favor that the metric be kept as low as possible.   

The CPI's limitations are as follows:

• It may not be representative of the various subgroups that comprise the population, such as elderly or poor. This is an inherent problem for all aggregate measures. The individual gets lost in the numbers. 
• It cannot be used to measure regional differences in price levels, nor is it supposed to measure everything that impacts lifestyle. It is like Mark Twain said, knowing the price of everything and the value of nothing. 
• It is an estimate based on surveys. Based on those samples, statistical methods are undertaken to come up with the results. And because the CPI is statistically derived, it is prone to sampling and non-sampling errors. The former relies on a sample of data and not on all purchases. The latter is more difficult to controls because there is always the possibility that people will not adequately disclosed the information being sought.

Against this background, we should take the BLS's CPI data with a grain of salt and not with the seriousness some would like to attach to it. A better measure for the CPI is the metric reported by Shadow Statistics, which is reported a 9.4% CPI increase year-on-year for the month of February.