Filtering short term fluctuations in inflation analysis

Many economic time series, specifically inflation, are inherently subject to seasonal fluctuations which obscure the real changes of the series. In this respect, seasonal adjustment is a powerful tool when removing such fluctuations. On the other hand, seasonal adjustment may provide highly volatile series, making it still difficult to interpret the movements of the series. The reason is that seasonal adjustment deals with certain type of movements that are completed on specific seasonal frequencies. However, it is possible that there may be other short term fluctuations occurring at non seasonal frequencies. From this observation and in the context of inflation, an improved methodology aiming to deal with all short term fluctuations that are completed within a year is proposed in this study. The two-step approach combines wavelet filters and band pass filters. This method yields much smoother time series than seasonal adjustment does. Moreover, the filtered series capture the dynamics of the inflation in sub groups well. Hence, this two-step procedure provides a useful tool for improved short term inflation analysis.