Iris Rošker and Laura Šuštar Kožuh, Statistical Office of the Republic of Slovenia - PDF

Description
Conversion of short-term indicators' series on production, turnover and prices for industrial activity according to NACE Rev. 2. at the Statistical Office of the Republic of Slovenia Iris Rošker and Laura

Please download to get full document.

View again

of 11
All materials on our website are shared by users. If you have any questions about copyright issues, please report us to resolve them. We are always happy to assist you.
Information
Category:

Travel/ Places/ Nature

Publish on:

Views: 24 | Pages: 11

Extension: PDF | Download: 0

Share
Transcript
Conversion of short-term indicators' series on production, turnover and prices for industrial activity according to NACE Rev. 2. at the Statistical Office of the Republic of Slovenia Iris Rošker and Laura Šuštar Kožuh, Statistical Office of the Republic of Slovenia Key words: Short-term statistics - NACE Rev. 2 new activity classification backcasting series new base year 2005=100 weights indices industry industrial production industrial turnover indices industrial output price indices - import price indices The STS regulation (Council Regulation (EC) No 1165/2005) with the accompanying amendments and other documents required that the beginning of transition to the newly implemented NACE Rev. 2 activity classification started with the data for January 2009, simultaneously with the switch to the new base year 2005 (or 2006 for some prices). According to the demands of the regulation, the whole series of data, back to at least 2000 or further if possible, had to be provided. There were several possibilities for backcasting of data, e.g. on micro level, on macro level, or a combination of both. For data on production and turnover the model of macro transference was chosen with the application of the translation tables, in which the shares of 'old activities' (according to NACE Rev.1) that appear in 'new activities' (according to NACE Rev. 2) are stated. For backcasting of the industrial production index (IPI) also new weights had to be prepared, while for turnover and other non-weighted variables there was no need for that. In the fields of output price indices for the domestic and non-domestic markets and import price indices, the data were recoded to the new classification - NACE Rev. 2 at the lowest level, i.e. the product level. On the basis of the Prodcom correspondence table all data for all months from January 1998 on (for import price indices from January 2006 on) were recoded to new codes from Prodcom 2008, which was adapted to the new NACE Rev. 2. With recoded data the new indices were calculated according to the new classification. All data were backcast at the same time. Because all data were recoded at the lowest level, backcasting was much easier. We used the same program to calculate indices. We created two databases, one with data according to the old classification and one with data according to the new classification. For output price indices on the domestic market and the non-domestic market and for import price indices the switch to a new base year 2005 was already made in The entire process of the new statistical classification of activities of business entities NACE Rev. 2 coming into force and the notifying of users and public in general was centrally led. A special group of experts from different statistical fields was established with the task to introduce the new classification - NACE Rev. 2 in statistical surveys. 1. First step: preparation of weights according to new classification NACE Rev. 2 and the new base year = 2005 Because of the introduction of the new activity classification and the new base year it was neccessary to prepare the new weights, not only for the calculation of the European aggregates (e.g. for Eurostat) but also for our own national calculations of aggregated activity levels of industrial production indices. The main source for the calculation of new weights for the area of industrial short-term indicators were the double-coded micro data from the structural business statistics (SBS). But because of the lack of some (non-domestic) variables within the SBS statistics, we had to find also additional sources for the remaining variables within the final accounts of the enterprises and within the international trade statistics data. Thus, the sources of the newly calculated industrial STS indicators weights were as follows: Table 1: The sources of the newly calculated industrial STS indicators weights Requested variable Value added at factor cost Turnover total Turnover domestic non- Turnover domestic Turnover nondomestic area Turnover nondomestic non- area Purpose (weight for) Dissemination code Sources PROD (production) V12150 SBS TOVT, TOVV, ORDT, PRON (total turnover nondeflated, turnover deflated or volume of sales, new orders received (total), output prices for industry and services -total) TOVD, ORDD, PRIN (turnover domestic market non-deflated, new orders received domestic market, output prices - domestic market) TOVE, ORDE, PREN (turnover non-domestic market non-deflated, new orders received nondomestic market, output prices non-domestic market) TOVZ, ORDZ, PREZ (turnover non-domestic market non-deflated area, new orders received non-domestic market area, output prices nondomestic market area) TOVX, ORDX, PREX (turnover non-domestic market non-deflated non- area, new orders received non-domestic market non- area, output prices non-domestic market non- area) V12110T V12110D V12110E V12110Z V12110X SBS final accounts final accounts international trade statistics (exports) international trade statistics (exports) Imports total IMPR (import prices total) IMP-T international trade statistics (imports) Imports area IMPZ (import prices -area) IMP-Z international trade statistics (imports) Imports area non- IMPX (import prices non- area) IMP-X international trade statistics (imports) 2. The course of transference of SBS data Because the data from the above listed sources (for 2005, some of them for 2006) were gathered still according to the old activity classification, it was neccessary that at the first stage they were transferred according to the new demands. That was achieved so that the enterprises were provided with the new classification codes from the already translated statistical business register (SBR). Namely, at the Statistical Office of the Republic of Slovenia (SORS) we decided to translate all of the yearly states of SBR back to Thus, when we had all the enterprises translated according to the new classification, their accompanying data were aggregated according to the NACE Rev. 2. For the calculation of weights for the turnover split according to the domestic and the non-domestic markets, the final accounts data of the enterprises were used. By means of the final accounts items 'turnover from the domestic market sales of production and services' (without rentals) and 'turnover from the foreign market sales of production and services' we calculated the shares for each of these two items in the total turnover from final accounts. These shares were finally applied to the already 2 (namely in the first step) calculated values of the total turnover (source: SBS) in order to obtain the three comparable values: total turnover, turnover domestic market, and turnover non-domestic market, of which the final two were estimated on the basis of the abovementioned shares from the final accounts items and the value of the SBS total turnover. A similar method was further applied in order to calculate the euro and non-euro indicators. Here the export data from the international trade statistics were used to divide the shares of turnover among the euro and non-euro zone by having applied the EU-15 list of countries. Consequently the calculation then demonstrated as follows: turnover total = turnover (domestic market) + exports exports = exports ( zone) + exports (non- zone) We used such method with shares so as to gain comparable results from the three different sources (SBS, final accounts and exports statistics), that all have different methodology and thus more or less different results. Othervise, should we have used the absolute values instead of shares of final accounts and exports statistics, we would never have got in accordance with the SBS results. 3. Conversion of data series according to NACE Rev. 2 The expert group in charge for the coordinated switch of the statistical surveys to the new classification in SORS prepared a presentation for statistitians with the aim to smooth the transference, and thus prepared the plan of actions with the schedule. The group also prepared the list of possible methods that could be used with the guidelines when and where each of these methods should be used. Methods of recalculation of the statistical series were to be used in these cases: translation key in situation of relation 1:1 or n:1 between the old an new classifications, micro level recalculation with the transference of activity on the enterprise (or LKAU) level and the new calculation of indices, macro level recalculation with the use of translation tables or matrices, combination of micro and macro level methods, e. g. recalculation using micro level method for the recent years and matrices (macro level) for the preceeding years. Every method has its advantages and disadvantages. While the main advantage of the recalculation of data according to micro method is its accuracy, it has several disadvantages, mainly connected with its exhaustive course from stage to stage: firstly the new activity code for all of the enterprises and their LKAU s for several past years has to be defined, then all the data processing for all past periods has to be repeated, then the data imputation problem arises (preparation of the new imputation model according to the new classification), in case of sampling surveys, the design weights have to be re-calculated, finally we are confronted with the problem concerning the missing data for certain new activities that were out of scope in case of the previous classification, but they are included in the observation within the new classification. The advantages of the macro method or the recalculation of the backcast series using the translation matrices are: it is not neccesary to provide the translation tables for all of the reference periods (i.e. months), but only for the yearly situations, 3 recalculation is done on indices and not on the micro data of enterprises, and thus the whole process is less extensive and exhaustive. The disadvantage of this method is that the direct connection between the indices on the aggregated level and the micro data is threatened. 4. Transference of the industrial production and turnover The macro method was selected for the industrial production and turnover series backcasting. For the periods 2004, 2005, 2006 and 2007 we applied the data on newly recoded industrial groups and new weights to aggregate the results on higher activity levels, while for the years from 2003 back to 1998 we used the translating matrices directly. The linking of both series was done with the use of the chain indices. The reason for such an act was based on the fact that the methodology of the industrial production index calculation was changed in 2004, so that from 2004 on its calculation was based on the value of turnover considering the changes of stocks. In the previous period the index of industrial production was calculated from the volume data and that was the reason why the index of turnover for the period before 2004 was estimated. This estimation was performed by using the data on value of the sold products on the NIP level, which were available in the old methodology data. Other motives for the macro method pick up are: the easier acquirement of data for the activities that were previously out of scope, great extent of the survey (in terms of time and finance) and demografical changes of the enterprises. 5. Preparations made before backcasting Some special conditions had to be fulfilled in order to perform backcasting successfully. At first the enterprises in the statistical business register on 31 December 2007 had to have both activities, the old one and the new one, likewise applied also for all the yearly states of the SBS back to For 2008 the survey sample was chosen according to both activity classifications: according to the old one for the current data publication and according to the new NACE in order to create a data base for the future (2009) indices calculations. Then the translation matrices had to be prepared for and the new weights for 2004, 2005, 2006 and 2007 on the basis of the SBS data were calculated. Using the administrative sources (VAT data base) and some series from other surveys (turnover indices in retail trade and services) we also calculated the series for the activities that were previously out of scope. Creation of the translation matrices On SORS the decision to prepare the translation matrices (for ) was made for the variables: according to the enterprise's turnover, according to the no. of persons working in the enterprises (also available on the level of LKAU), according to the no. of enterprises within the activity, (also available on the level of LKAU). For the industrial activity the translation matrices according to the no. of employees were most suitable. Namely, in Slovenia, due to its small economy, besides the enterprises also the KAU's are observed in short-term industrial statistics, and for them we do not have the information on turnover available from final accounts. Thus, the translation tables (matrices) were made for all the years (starting from 2007 until 2000). In these tables the shares of old activities that occurred also in some new activities are stated. 4 Table 2: Part of the translation matrices Year NACE Rev. 2 NACE Rev. 1 Share Source of indices , , , , , , , , , , , , , , , Sources of indices: 1 monthly industrial statistical survey 2 other statistical surveys or VAT data Figure 1: Flowchart of backcasting for the industrial production index OLD WEIGHTS old classification NEW WEIGHTS new classification SBR YEARLY STATE old + new class. code TRANSLATION MATRICES industrial groups level AGGR. NEW SERIES 2005=100 OLD INDICES SERIES old classification SERIES OF INDICES, prev. out of scope, 2-digit level old class. SERIES OF INDICES industrial groups level new classification, SERIES OF INDICES industrial groups level new classification, 2005=100 5 Figure 2: Backcasting for turnover, new orders, stocks (non-weighted series) SBR YEARLY STATE old + new class. code TRANSLATION MATRICES industrial groups level from 2-digit level classif. upwards SERIES OF INDICES from 2- or 3-digit level old classification upwards SERIES OF INDICES, prev. out of scope, 2-digit level old class. SERIES OF INDICES 2-digit level new class. and all the aggregaets SERIES OF INDICES 2-digit level new class. and all the aggreggates 2005=100 The main differences between the series broken down by the old classification and that broken down by the new classification: Mining and quarrying change of designation from C to B (there are also some differences on the lower level of classification, of minor importance) Manufacturing: - change of designation from D to C, - because of the split of publishing activities and recycling of materials out of scope this area of observation is smaller by about 2.6% (in terms of no. of employees), - activities that are newly included into observation from the area of agriculture, wholesale trade and services comprise a small share and are scattered over the different groups of activities, - a lot of changes are to be found, also rearranged among the variety of groups within the manufacturing activities. Electricity, gas and water supply: - instead of the previous designation E, there is now D - water collection, treatment and supply is now within the division E, which is not relevant for us since it has always been out of scope, - some additional differences are also on the lower classification levels. In order to check the correctness of the backast data we carried out a comparison of the data that were derived from the previously described macro method with the data derived from a micro method recalculation. Namely, we also applied that method to gain some aggregates for 2008 when we observed the population according to both classifications. We introduced the so called 'test data base' with the micro data for parallel calculation. Then we calculated the correlation coefficients which were 6 all between 0,96 and 0,99, and made some visual control of the diagrams, which all verified our hypothesis that the backcast series were reliable. Graph 1: IIP - Industry total: comparison between the NACE Rev. 1 data and macro calculations aggregation using the weights, indices, Slovenia, Jan Sept Quality check jan.06 mar.06 maj.06 jul.06 sep.06 nov.06 jan.07 mar.07 maj.07 jul.07 sep.07 nov.07 jan.08 mar.08 maj.08 jul.08 sep.08 NACE Rev. 1 macro method - aggr. with weights Graph 2: IIP - Industry total: various calculations, indices, Slovenia, Jan Sept jan.06 mar.06 maj.06 jul.06 sep.06 nov.06 jan.07 mar.07 maj.07 jul.07 sep.07 nov.07 jan.08 mar.08 maj.08 jul.08 sep.08 macro method - aggr. with weights macro method - aggr. with matrices micro method 7 Graph 3: IIP - Capital goods industries (B), indices, Slovenia, Jan Sept jan.06 mar.06 maj.06 jul.06 sep.06 nov.06 jan.07 mar.07 maj.07 jul.07 sep.07 nov.07 jan.08 mar.08 maj.08 jul.08 sep.08 macro method - aggr. with weights macro method - aggr. with matrices micro method Graph 4: IIP - Consumer goods industries (C), indices, Slovenia, Jan Sept jan.06 mar.06 maj.06 jul.06 sep.06 nov.06 jan.07 mar.07 maj.07 jul.07 sep.07 nov.07 jan.08 mar.08 maj.08 jul.08 sep.08 8 macro method - aggr with weights macro method - aggr. with matrices micro method Graph 5: IIP Intermediate goods industries (A), indices, Slovenia, Jan Sept jan.06 mar.06 maj.06 jul.06 sep.06 nov.06 jan.07 mar.07 maj.07 jul.07 sep.07 nov.07 jan.08 mar.08 maj.08 jul.08 sep.08 macro method - aggr. with weights macro method - aggr. with matrices micro method 6. Conversion of the output price indices for the domestic and the nondomestic markets and import price indices In 2005 we decided to do some revisions on price indices. At that time in Slovenia we only published data about Output price indices for the domestic market. The first step of the revision was to prepare a new computer program for calculating data. We had two choices. One was to prepare a program to calculate the indices in two stages, firstly to calculate the basic indices (current month compared to December of previous year) on the level of class and then all the indices that are published. Eurostat announced that a new classification would be introduced. The best choice was to fill the data base with all basic data. Our data base was filled with all data needed for calculating the indices. We prepared single files, for every month of every year from January 1998 forward. The single file included data for the selected month about: reporting unit (serial number, register number, name), product (NIP 1 code, price code 2, specification of product, unit of measure), prices (for December of previous year, past month, current month, code for currency and code for country 3, weight). The preparation of several single files was comprehensive work that took a great deal of time and demanded a lot of precision. Parallel with preparation of single files about price data the expansion of the sample of reporting units was in progress. The sample of units that were reporting prices for the domestic market was enlarged from 309 reporting units up to 345. In 2007 Slovenia has introduced 1 NIP National Nomenclature of Industrial Products 2 Price code internal code, that split diferent products that belong under the same NIP code 3 Only for Output prices for non-domestic market and for Import prices 9 the Euro. That consequently meant the obligation to distinct the Output price indices for the domestic market to Euro and noneuro. The need to enlarge the sample was on the spot. So the sample of reporting units was enlarged from 74 units in 2001 when SORS started to collect export producer prices, up to 220 units. The purpose of all that work was to insure better quality of the calculated data and maximum availability of calculated data. The sample of reporting units for Output prices was enlarged. Within Output price indices
Related Search
Similar documents
View more...
We Need Your Support
Thank you for visiting our website and your interest in our free products and services. We are nonprofit website to share and download documents. To the running of this website, we need your help to support us.

Thanks to everyone for your continued support.

No, Thanks