A proposal to regenerate time series in the case of classification revisions

The statistics are needed on account of the reasons like analysing the economic and social structure of the countries, pursuing the development in the world and taking new investment decisions. The usage of standard classification in the compilation and dissemination of these statistics leads to produce standart and comparable statistics. Thus, these statistics shall be comparable among countries and years and it will be possible to acquire long time series to be used in the pursuit of the development in the world and interpretation of the economical changes. Nevertheless, the need of revision the classification has been appeared so as to reflect the changes in the economy, techonology and the other areas. This revision studies mostly make it difficult to realize the continuity of time series and result in inconsistency among the statisctis produced after the revisional process and time series composed in the former period.In this study, the data set of annual industry and service statistics covering the years 2003 to 2007 has been used and data of 2007 at the enterprise level which has more than or equal to twenty employees is accepted as a frame and a sample has been drawn from this frame. For the turnover, the number of employment, the production value and the value added with factor costs variables, at the micro level the direct estimation, post-stratification and EBLUB; at the macro level the number of enterprise, the number of employment and turnover based concordance coefficients have been employed to estimate four digit levels of NACE Rev.2 and in a nutshell, the methods with best performances have been applied to the data set of 2003-2007 to regenerate time series.