The problem of annual inflation rate indicator

Abstract

AbstractIn economic practice, the annual inflation rate is commonly used. As a one‐sided moving average of the annualized inflation rate, it is approximately 6 months behind the annualized and monthly inflation rates and the CPI. The annual rate of inflation is, in fact, a smoothing transformation that removes the seasonal component from the annualized inflation rate, assuming the presence of all seasonal unit roots. In reality, however, the CPIs, and thus the annualized inflation rates, do not contain most of seasonal unit roots, resulting in spurious cycles of annual inflation rates, which pose difficulties in assessing and interpreting their development dynamics. Their practical applications in econometric analyses are limited because the unit root tests too often do not reject the zero hypothesis of the presence of unit roots when annual inflation rates are actually stationary. Parameter estimates of the annual inflation rate models are biased and inaccurate, allowing only incorrect short‐term forecasts. The above facts make use of the annual inflation rate indicator in real economic situations questionable.