Regression methods are the second, in addition to probabilistic methods, most popular approach in the quantification of inflationary expectations of the market entities. These methods may be used in practice though they require information both on how those entities perceive the current inflation and on their expectations as to the future price developments as well as gathering adequately long time series. The quantification of the expectations is done in two stages. The first stage consists in the evaluation of parameters in the equation of the relation between the current inflation, measured by the official statistics, and its perception among the survey respondents. The second stage - assuming that the same relation links inflationary expectations expressed quantitatively and qualitative opinions of the respondents about the future inflation - consists in the usage of the aforementioned equation in the calculation of measure of expectations expressed in numbers. The presented procedure requires an assumption on the unbiased nature of perceptions of the current inflation, though does not impose the unbiased nature of the expectations.

The article discusses the theoretical basis of regression methods; a review of the models used in this approach has been made, with particular regard to their advantages and drawbacks. Moreover, the selected models have been used to quantify 3-month inflationary expectations of private individuals in Poland in the years 2001-2004. The results suggest that the inflationary expectations within this time horizon are a good predictor of the future inflation. However, considering the small time sample (that, inter alia, makes testing of the hypothesis of rationality of those expectations impossible) the obtained results should be approached with prudence.