The article takes an in-depth look at the Consumer Price Index (CPI), which is widely used as a basic measure of inflation. In practice, the author says, when measuring the CPI economists usually use the so-called Laspeyres price index, which does not take into account changes in the structure of consumption resulting from price changes in a given time interval. The problem is that the Laspeyres index can be biased due to commodity substitution, the author says. The article discusses potential sources of the CPI bias and the application of two approaches for calculating the index. The first approach is connected with superlative indices, while the second uses the so-called generalized Fisher price index. The article presents the results of empirical and simulation studies conducted by the author. An empirical analysis for the 2010–2013 period points to the existence of a small and unstable CPI substitution bias, according to the author. The simulation study, in turn, makes it possible to conclude that imputations of prices of new and disappearing goods are crucial for CPI bias calculations, Białek says. Moreover, the sign and value of the correlation between prices and quantities do not generally influence the CPI substitution bias in the broad sense, the author notes.