Abstract

We have received a number of queries recently from market researchers concerning our methods in order to determine the underlying wage process, with particular regard to the correction of whitening. The whitening estimation methods we have developed have been described in outline in the boxed sections of our Reports on Inflation. In this study, however, we would like to present the estimations so that they can be easily reproduced.

The study briefly describes our three methods designed to adjust the possible distortions of wage data. The whitening of the economy may cause such distortions. Since we do not directly observe the whitening process, we use indirect methods. A model of the wage data generation process is established in each case. There was a break in these processes during the second half of 2006. The time series, adjusted for whitening, yielded by a particular method is the fictitious wage time series, which would have been attained if there is no break in the data-generation process. The three methods are based respectively on wage differences by economic sector, labour force categories and the estimation of wage distribution. The three methods result in a similar estimate of whitening in terms of both its extent and time-pattern. The level of distortion may have been more significant in 2006, and gradually decreased subsequently. The wage indices have still shown a significant increase in their dynamics even after the adjustment for whitening. It is thus not true that the acceleration in the wage statistics published by the Central Statistical Office (KSH) has been entirely the result of the government measures aiming at the whitening of the economy.