Change of methodology for the Industrial production index

The Czech Statistical Office (CZSO) changed the methodology for calculation of the Industrial production index with effect from January 2006. To maintain the possibility to analyse development in industry, the CZSO recalculated time series according to the new methodology backwards since 2000.
The News Release “Industry - NACE 10-41 (monthly)” - “October 2005” on results of industrial production in the Czech Republic published on 12 December 2005 contained information that starting with the results for January 2006 the final results for the year 2005 will be published. From internal analyses and a methodological audit of statistics of industry made it resulted that the methodology for the Industrial production index (IPI) used up-to-now is no longer suitable, because the set of products-representatives (hereinafter referred to as the representatives), from which it was arising, has become outdated prematurely. It is the reason for a deviation of some partial results. Thus, the CZSO accelerated its preparation for a change to an alternative methodology for the IPI calculation. The change was to start from 1 January 2007, the year 2005 being the base year in harmony with the tradition. The CZSO has further decided to make an alternative calculation of industrial production indices not only for the year 2005, but also for the preceding years so that IPI times series long enough are available for business cycle analyses.
The up-to-now way of calculation resulted from the monthly measuring of the volume of production of about 1,600 representatives mainly in physical units of measure, i.e. to a great detail. For each representative, an individual index was calculated. From the set of indices, using a weighting scheme (derived from composition of gross value added of industry) the average indices for aggregation of products up to the industry in total were obtained. Consequences of large changes in the composition of industrial production have an impact on indices of aggregations of products (for divisions, subsection and sections) that was not balanced, because the influence of individual indices with values exceeding even thousands of % could have been compensated only a bit by those indices the value of which was falling to zero. The high number of measured products was at the same time limiting a lot the possibility to find out defective representatives and to replace them effectively in time.
Internationally accepted IPI concept enables to use a rather large scale of methods as well as input data used. In accordance with experience from abroad, such alternative of calculation was now chosen, in which input calculation of individual indices starts at a higher level of aggregation. However, on that level, it is not possible to combine different physical units of measure and thus it is necessary to rely on monetary indicators adjusted for inflation. As a suitable basic indicator the following usually surveyed indicator was selected: sales of own goods and services of industrial character in constant prices, which can be used as approximation of the volume of production. It was used in most of industrial branches; only in the case of several branches, characteristic for their production of collective homogeneous products, the calculation included the up-to-now used data on the volume of production in physical units of measure. In the new calculation the current weighting scheme was used, which is based on the structure of gross value added of industry in the year 2000. Since data on sales are currently available only in the breakdown by enterprise by prevailing activity, the change of calculation means at the same time also a shift from the product-oriented method to an enterprise oriented one. Namely due to that reason, the new way of IPI calculation has to be understood as a new, alternative approach rather than as a revision or correction of the results already published.
Methodology of calculation based on enterprise sales will be further improved in the years to come, namely in relation to the announced shift to the base year 2005 in the year 2007 as well as in relation to the shift to the new version of the CZ-NACE classification, which is planned for the years 2008-2009. The target is to better respect the influence of auxiliary activities on changes of enterprise sales, which should characterize changes in prevailing activity.
The change of methodology of the Industrial production index did not cause and will not cause changes of indices of gross domestic product in constant prices.

The results according to the new methodology in the breakdown by division, subsection, section and for industry in total: