Services - Methodology
Methodology – monthly sales indices in services (covering sections H + I + J + L+ M + N without CZ-NACE 72, 75 and 81.3) in CZ - NACE Rev. 2
The Czech Statistical Office holds monthly statistical survey of sales in services since 1997. Results of this survey were published from the beginning in monthly intervals and these original time series are still available on the CZSO web site. Since 2008 a new international classification of activities has became effective (CZ-NACE Rev. 2) replacing the original one (marked as OKEČ, i.e. CZ-NACE Rev. 1.1). This fact had to be considered also in back conversion of time series. Conversions were not an easy discipline, they were made for individual enterprises and due to limited capacities it was not possible to maintain the original length of time series.
New time series of monthly sales indices cover enterprises with predominating activity in transportation and storage (CZ-NACE section H), accommodation and food service activities (CZ-NACE section I), information and communication (CZ-NACE section J), real estate activities (CZ-NACE section L), professional, scientific and technical activities (CZ-NACE section M without 72 and 75) and administrative and support service activities (CZ-NACE section N without 81.3).
These time series include data for individual months since 2000 (in case of y-o-y indices since 2001). Along with the time series conversion according to CZ-NACE Rev. 2 new methods of estimates were applied and some errors in classification of enterprises according to their predominating activities were corrected. In compliance with the Eurostat´s demand only sales excl. VAT are surveyed. New deflators for conversion of sales into constant prices were created. Concurrently, base year was changed to 2005.
I. Sales index: services
Sales indices serve for short-term development analyses and for possible prediction of the future development of individual industries. Data source is monthly statistical survey SP 1-12 with annexes common for the following divisions: trade, accommodation and food services, transport and storage, information and communication activities and market services.
II. Formation of the selective sample
Reporting units are selected from the Business Register. Sample strata are defined by the size of the unit (number of employee) and principal activity. The sample is taken for each strata separately. The selection includes units with 50+ employees as well as those units whose number in the strata is less than 50. In other groups of enterprises proportional sampling is used related to the scope of sales as reported from the administrative data sources. The probability of including the unit into the sample is directly proportional to the contribution of its sales to total sales of all units in strata. This is to secure that the units reporting above-the-average sales have more chances to be included to the sample than those with low sales, which are not thereby exposed to unnecessary administrative burden. However, the importance of small enterprises is not undervalued as the products of weights of individual enterprises and their sales are within the strata equal.
The criterion for including all units into the sample is applied also during the observed year; basic and selective samples are hence updated regularly with new units, which correspond with this criterion.
III. Description of indicators observed
Aggregate sales from the sale of own products, goods and services in the respective month are observed. Sales are recorded excl. VAT.
Initial questionnaire, which is sent as an annex to the units included for the first time to the sample contains information on principal activity of the reporting unit and review of sales by individual month of the preceding year.
IV. Description of the sales index calculation
For each lowest aggregate of CZ-NACE (e.g. 50) y-o-y sales indices at current prices are calculated for individual months. For these CZ-NACE categories y-o-y price deflators are used for individual months of current year and for base year price deflators are related to individual months and to the year’s average. Absolute values of monthly sales at current prices for the lowest CZ-NACE aggregates are converted into the price level of 2005 using the method of chaining deflators. By the summation of the lowest aggregates of absolute values of sales at constant prices in individual months we get higher sale aggregates from which we calculate y-o-y sales indices at constant prices. The sum of monthly sales of aggregated CZ-NACE categories at current prices for 2005 is equal to the sum of their sale at constant prices.
Sales at constant prices are deflated by price indices excl. VAT related to the base year 2005.
V. Retroactive corrections, revisions
Along with the February data processing, data for January are revised. In required, data revisions are carried out along with the processing of the first month that follows the publication of quarterly data, e.g. with the processing of April data, data for January-March are revised. Final data are published following the processing of all months of the observed year, not later than by June of the following year.
VI. International comparability
The area of short-term statistics in the European Union is regulated by the EU Regulation No. 1158/2005 amending the Regulation 1165/98 on short-terms statistics. Under this Regulation the EU member states are obliged to provide information on turnover in services (at current and constant prices) on quarterly basis. Indicator published by the CZSO is comparable in terms of methodology. The CZSO, in addition to the EU requirements, monitors turnover in time series also on monthly basis. Time series will be published on monthly basis beginning with data for
March 2010.
VII. Seasonal adjustment
The size of sales during a year in some industries is heavily affected by season (e.g. air transport, etc.). In order to find out long-terms trends and to be able to compare month-on-month or quarter-on-quarter development, the CZSO produces seasonally adjusted data, when the part of sales caused by a different number of working days and by seasonal swing is separated from the original “rough“ series, using a mathematical model. Upon seasonal adjustment not only y-o-y but also m-o-m (q-o-q) development can be compared.
Method TRAMO/SEAT recommended by Eurostat is used for seasonal adjustment. Programme application DEMETRA version 2.1 is employed in the CZSO.
VIII. Comparability with other surveys in terms of methodology
In conversion of time series monthly indices were retroactively adjusted to yearly structural statistical surveys in 2000 – 2008 and for 2009 to administrative data or results of quarterly statistical survey in 2009. In 2010, however, the y-o-y indices from monthly and quarterly surveys may slightly differ because monthly data received from enterprises are designed to serve estimates and only after a quarterly and sometimes even annual end of data collection the results from all statistical surveys can be unified.
IX. Types of data released
1) News Releases – Services – quarterly, on the 41st calendar day following the end of reference period.
2) Time series – on the CZSO web site: /csu/czso/slu_ts
a) year-on-year index – share of sales in the observed period and sales of the same period of the previous year.
b) base index – share of sales of the given observed period and average sales in the base year. In new converted time-series under CZ-NACE Rev.2 the year 2005 is taken as the base year.
