Czech Statistical Office (CZSO)

Business Statistics Processing Coordination Unit

Na padesátém 81
100 82 Praha 10
Czech Republic

Structural business statistics (SBS) describes the structure, conduct and performance of economic activities, down to the most detailed activity level (several hundred economic sectors). SBS covers all activities of the business economy with the exception of agricultural activities, public administration and (largely) non-market services such as education and health. Main characteristics (variables) of the SBS data category:

• Business demographic variables (e.g. Number of active enterprises)

• "Output related" variables (e.g. Net turnover, Value added)

• "Input related" variables: labour input (e.g. Number of employees and self-employed persons, Hours worked by employees); goods and services input (e.g. Purchases of goods and services); capital input (e.g. Gross investments)

Business services statistics (BS) collection contains harmonised statistics on business services. From 2008 onwards BS become part of the regular mandatory annual data collection of SBS. The BS’s data requirement includes variable “Turnover” broken down by products and by type of residence of client. 

The annual regional statistics collection includes three characteristics due by NUTS-2 country region and detailed on NACE Rev 2 division level (2-digits).

Statistical Classification of Economic Activities in the European Community (NACE):  NACE Rev.2 is used from 2008 onwards. Key data were double reported in NACE Rev.1.1 and NACE Rev.2 only for 2008. From 2002 to 2007 NACE Rev. 1.1 was used and until 2001 NACE Rev.1

The regional breakdown of the EU Member States is based on the Nomenclature of Territorial Units for Statistics (NUTS). 

The product breakdown is based on the Classification of Products by Activity (CPA) as stated in the Regulation establishing CPA 2008 and its amending Commission Regulation (EU) No 1209/2014 (from reference year 2015 onwards).

Starting reference year 2021 onwards SBS cover the economic activities of market producers within the NACE Rev. 2 Sections B to N, P to R and Divisions S95 and S96.
Until 2007 the SBS coverage was limited to Sections C to K of NACE Rev.1.1 and from the reference year 2008 to 2020 data was available for Sections B to N and Division S95 of NACE Rev.2.

From 2008 reference year the data collection Business services covers NACE Rev 2 codes: J62, N78, J582, J631, M731, M691, M692, M702, M712, M732, M7111, and M7112.
From 2013, as the first reference year, to 2020 information is published on NACE codes K6411, K6419 and K65 and its breakdown.

SBS constitutes an important and integrated part of the new European Business Statistics Regulation N° 2152/2019.

Data requirements, simplifications and technical definitions are defined in Commission Implementing Regulation (EU) 2020/1197.

Statistical unit used for the compilation the SBS series is enterprise. The enterprise almost always corresponds to legal unit.

Target population: All the active enterprises (consisting of legal and natural persons) classified in Sections B to N and P to R and Division S95, S96 of NACE Rev 2 and institutional sectors of financial corporations, non-financial corporations and households (employers and own-account workers). Exceptions are housing associations and non-profit institutions classified into the institutional sector of non-financial corporations.

Frame for the identification of the population: National statistical business register.

Czech Republic

Inclusion/exclusion of the activities of Foreign Affiliates

If a foreign enterprise establishes a branch in the Czech Republic, it has the obligation to register the branch at the Register of Companies, thereby it enters the Statistical Business Register and subsequently the population of the Annual Structural Business Survey carried out by the Czech Statistical Office (CzSO). And the other way round, if an enterprise registered in the Czech Republic establishes a branch abroad, it is requested to report only unconsolidated information (i.e. data not including the activities of foreign affiliates). However, if an enterprise registered in the Czech Republic carries on business abroad, but does not establish an affiliate abroad, the results of these activities are included in the data reported to the CzSO.

2005-2022

Not applicable.

• Number of enterprises and number of local units are expressed in units.

• Monetary data are expressed in millions of €.

• Employment variables are expressed in units.

• Per head values are expressed in thousands of € per head. 

Ratios are expressed in percentages.

2022

Data refers to calendar year. If an enterprise reports the data for a fiscal year, it is recalculated to the respective calendar year.

Starting with reference year 2021 two new regulations currently form the legal basis of SBS:

• Regulation (EU) 2019/2152 repealing 10 legal acts in the field of business statistics (EBS Regulation), and
• Commission Implementing Regulation (EU) 2020/1197 laying down technical specifications and arrangements (general implementing act) pursuant to Regulation (EU) 2019/2152

Year 1995 was the first year for the implementation of the Council Regulation No 58/97 (SBS Regulation).

The Council Regulation No 58/97 has been amended three times: by Council Regulation No 410/98, Commission Regulation No 1614/2002 and European Parliament and Council Regulation No 2056/2002. As a new amendment of the basic Regulation it was decided to recast the Regulation No 58/97 in order to obtain a new "clean" legal text. In 2008 the European Parliament and Council adopted Regulation No 295/2008 and the provisions of this Regulation were applicable from the reference year 2008 to reference year 2020. Regulation No 295/2008 was amended by Commission Regulation (EU) No 446/2014.

Not applicable.

Confidentiality policy is based on the Act No 89/1995 Sb of 20 April 1995 on the State Statistical Service.

For the detection of confidential cells we use so-called “key variables” (Net Turnover (12110) and Number of employees and self-employed persons (16110)) and then we apply the confidential pattern to the other variables.

Flag A - too few enterprises (

Flag O - the share of the individual data of one statistical unit is significantly predominant (information on the percentage of predominancy is considered confidential). 

Flag D - Secondary confidential data in order to protect data flagged with A, B

Confidential SBS data is protected against the risk of disclosure using Tau-Argus SW program.

Data is disseminated nationally.

See Catalogue of Products.

See Catalogue of Products.

All the data provided to Eurostat, unless marked for use as a contribution to European totals only, is provided to users through the regular publications or on request.

Microdata is provided only to the selected public authorities and wholly in conformity with the State Statistical Service Act.

Annual.

News releases are not used for the publication of the SBS data. The Czech Statistical Office prefers other publication formats.

Paper publications and Statistical Yearbook of the Czech Republic available in paper and electronic versions (English versions, see Annex).

Time series of SBS indicators supplemented by related indicators constructed according to the national methodology are published on the website of the Czech Statistical Office.

Online database - see Public database.

Access to microdata is granted only to the selected public authorities and wholly in conformity with the State Statistical Service Act.

All data from the Annual Structural Business Survey not regularly published, respecting the rules for the protection of confidential data, can be provided on request, in the detail corresponding to the SBS requirement, to all users.

Available as part of the paper/electronic publications and time series. This documentation is available in Czech and also in English.

See methodology for services (the methodology is the same for all NACE)

The whole process of data production from questionnaires design, metadata definition, through data collection, imputation methods, primary and secondary data validation, confidentiality treatment and dissemination is in detail described in the internal documentation (Technical Projects and Statistical Metadata System) to which all interested parties have an access. If the Annual Structural Business Survey is the subject of the methodology audit, the documentation is available to auditors. This comprehensive documentation is available in the Czech language.

Basic documents are published on the website of the Czech Statistical Office.

No major problems with relevance, accuracy, timeliness and punctuality, coherence and comparability.

Internal Users:

User: National Accounts Department.
Needs: SBS data plus additional, mainly accounting, data far beyond the scope of SBS. 

User: Prices Statistics Department
Needs: Breakdown of Turnover by product type (CPA) 

User: STS
Needs: Weights of base year 

User: IFATS
Needs: Data referring to the foreign controlled enterprises. 

Other users: Innovation statistics, R&D statistics

External Users:

Users: Czech National Bank, Ministry of Industry and Trade (MIT)
Needs: Microdata from the Annual Structural Business Survey

Ad hoc requests: ministries and other public administration institutions, universities (especially microdata access requirements)

Regular consultations with some of our main users are organised.

All the SBS series provided to Eurostat are complete.

Taking into account the available administrative data, the high response rate, imputation method chosen and the multi-stage validation system the overall quality can be assessed as favourable. The possible source of potential errors is the lack of information related to the smallest units in the population, what might result, among others, in the high misclassification of statistical units in the Statistical Business Register. The potential bias is, however, small with limited effect on the overall accuracy of estimates.

The preliminary results may be biased since at the time of its transmission to Eurostat not all questionnaires nor administrative data are available. However, we consider the potential bias to be relatively low.

Sampling error is monitored regularly and it is small. The coefficients of variation (CVs) are calculated and for the indicators agreed by the SBS and BD WG provided to Eurostat via eDAMIS.

CVs are calculated for indicators:

220101 Number of employees and self-employed persons

250101 Net turnover

250401 Value added

Detailed methodology on Variance estimation is included in the separate file (see Annex).

The potential non-sampling error that is considered to be small may result from:

1. Relatively high degree of misclassification particularly of the smallest units (natural persons and legal persons without employees, which are not VAT payers).
2. Non-response. However, it is not considered to be serious problem.
3. Relatively frequent modifications of the content or/and the extent of the administrative sources.
4. The lack of information related to the smallest units (especially natural persons - not VAT payers - without employees)

The methods used for taking into account the unit non-response:

The imputation procedures dealing with unit and item non-response are based on:

1. the available administrative data (selected indicators from balance sheet and profit and loss statement; approximation of total revenues from the VAT data; number of social insurance policyholders);
2. the data available from the other surveys (especially annual Labour Cost Survey; quarterly business structural survey; PRODCOM);
3. the annual structural business data from the previous reference periods.

The quality of administrative data does not allow to be directly used in the imputation procedures. Therefore, prior to their use in the statistical processing they must be verified by the procedures detecting and correcting the errors, e.g. errors in order of the data (e.g. balance sheet data is reported in thousands, while profit and loss statement data in million) or consistency errors (assets do not equal the liabilities, revenues minus expenses do not equal the profit/loss, etc.). Moreover, the process for the verification of administrative data for a subsequent use, includes also the procedures for the imputation of missing or erroneous administrative data.

After the verification, the administrative data and other relevant data enter the process of the imputation. The very basic principle of the imputation process as such is based on the average share of an estimated variable and a variable available from the administrative or other external sources (hereafter referred as the explanatory variable) calculated from the surveyed units that provided CzSO with correctly completed questionnaire belonging to the same stratum as a non-respondent unit or a unit not included in the sample. For the estimation of the estimated variable the average share is applied to the explanatory variable available for the non-respondent unit or the unit not included in the sample.

In the initial stages of the imputation process, the explanatory variables are indicators available from the administrative or other external sources. In the next stages, a previously imputed variable can become the explanatory variable for the estimation of other variables.

The measures taken for minimising the unit non-response:

1) Average unit non-response is taken into account when designing the sample.

2) The questionnaire consists of several individual parts. Each part of the questionnaire is focused on one or more areas (e.g. cost/expenses, revenues and assets; investment; financial assets and liabilities; breakdown of turnover by products; selected indicators broken down by residence of clients; selected indicators broken down by regions; etc.). The individual parts can be combined so that we can tailor different versions of the questionnaire for different groups of respondents according to their size/importance. As a consequence the biggest enterprises complete the version of the questionnaire consisting of all (or most) parts, while the smaller ones those versions consisting of only selected parts. There are altogether more than 500 different versions of the questionnaire. Quite a high level of personification enables us to lower the burden on especially small respondents and at the same time it contributes to the minimising of the unit non-response as every sampled unit completes only the minimum information necessary for the statistical production.

3) Every respondent can choose one of the three ways of questionnaire completion: 1) web based questionnaire 2) interactive PDF form of the questionnaire or 3) the paper form of the questionnaire.

4) If a unit is selected for the sample of one survey, it cannot be selected for any sample of the other surveys unless it is classified to the stratum where all the units are sampled. This method known as negative coordination of the samples contributes to minimising the burden on especially small units. Moreover, it indirectly helps to increase the response rates of the individual surveys since it reduces the risk of non-response due to overburden of respondents.

5) Every non-respondent unit is contacted successively by phone or e-mail and then repeatedly by registered letter sent in most cases to so called data box. The data box is an electronic storage site, intended for delivery of official documents and for communication with public authority bodies.

Bias:

If characteristics of the non-respondent units significantly differed from those of respondent units, the estimated part of the population could be biased. In addition, a certain bias can be caused by the high misclassification of the smallest units (especially natural persons) in the Statistical Business Register.

The potential bias is, however, small with limited effect on the overall accuracy of estimates.

Deadline for data-collection: 25 April 2023

Deadline for post-collection phase: 31 January 2024

Deadline for the Annual Structural Business Survey termination: February 2024

Dissemination of SBS data: 30 June 2024

No delay

As regards NUTS classification, data is fully comparable.

2005 - 2020, 2021-2022

Number of enterprises in the business register - Business register contains units that are from the administrative point of view active, however, as far as statistics is concerned they are not active (e.g. units that were established, but have not started their activity yet, units that finished their activity but have not been erased from the register yet etc.). In this respect, it is important to underline that business statistics consider a unit to be active only if there is an administrative “trace”, i.e. the unit is VAT payer or statistics have information on the unit from the income tax return or on its employees from the Czech Social Security Administration.

Value added (VA) of National Accounts (NA):

There are numerous methodological and conceptual differences between VA of National Accounts and VA of Business Statistics. The inclusion of the non-observed economy in NA can be given as an example. Generally, however, the cross-domain coherence between SBS and NA data is considered favourable.

Evolution of turnover and persons employed from short term statistics:

Minor inconsistencies in the evolution of turnover - resulting from the fact that when STS data are compiled, the accounting data are considered provisional - can be observed.

The data is fully compatible with Business Demography.

Fully consistent.

Cost and burden are regularly monitored for the whole Annual Structural Business Survey (ASBS) covering requirements far beyond the SBS requirements.

On the basis of the ASBS burden model, it was estimated the burden resulting from the SBS regulation (NACE BTSXO_S94). Number of respondents (irrespective of response) for whole ASBS: 35 836. Average time to complete the questionnaire: 10,08 hours, of which 2,5 hours for items entering the computation of SBS indicators. The burden resulting from the SBS regulation: 89,6 thousand person-hours. 

At present, unfortunately we are not able to make a reliable estimation of the costs resulting from the SBS Regulation using ASBS costs model.

When compiling the final data for the year Y - usually in February Y+2, if needed, data for Y-1 can be revised.

Usually, preliminary and final data are estimated in the same way.The only difference is that at the time when preliminary data are compiled, neither all surveyed data nor all administrative data are available. The data for 2022 is an exception in this regard.The revisions described in the chapter 15. Coherence and comparability were applied to the final data only.

Relative Mean Absolute Revisions (RMAR) of preliminary data versus published final data is calculated by Eurostat to ensure the comparability among countries. Relative Mean Absolute Revisions (RMAR) of preliminary data versus published final data were 0,0626 in case of variable 250101 (Net Turnover) and 0,0190 in case of variable 220101 (Number of employees and self-employed persons).

TYPE OF SOURCE: Sample survey combined with administrative sources

Following sources were used for the compilation of the SBS data:

1. Annual Structural Business Survey - by far the most important statistical source - surveyed are legal persons, market producers, classified in the institutional sector S.11 with a few exceptions, e.g. excluded are enterprises with the principal activity NACE 01, non-profit institutions classified in the sector S.11 or housing associations, on the contrary included are certain types of enterprises classified in the sector S.12.

2. A group of Annual Structural Surveys focused on financial and insurance activities

3. Annual Structural Survey of non-profit institutions, housing cooperatives and other selected institutions

4. Annual Labour Costs Survey - some data referring to number of persons employed, employee benefits expense, wages and salaries and social security costs is estimated exploiting information from the Annual Labour Costs Survey.

5. Quarterly business statistics survey - an auxiliary source used in the imputation.

6. PRODCOM survey                                                                

7. Administrative data:

1. Selected information from the income tax returns
2. VAT data
3. Number of social insurance policyholders and the basis of social security insurance assessment                                 

SURVEY (corresponds to Annual Structural Business Survey):

Type of sample design: stratified

Stratification criteria: Activity, Employment size class, other: institutional-sector, legal form, turnover and assets

A random stratified sampling is combined with an intentional sampling of statistically significant units.

The units classified in the institutional sector 14 are not sampled. The required data are estimated solely on the basis of the available administrative data.

Administrative data:

- Kind of administrative data available: micro data

- Population:

• Income tax data - entire population
• VAT data - VAT payers
• Number of social insurance policyholders and the basis of social security insurance assessment - employers

- Frequency

• Income tax data is provided to CzSO in three batches (T+7 months, T+9 months and T+12 months)
• VAT data is available on a monthly (monthly payers)/quarterly (quarterly payers) basis
• Number of social insurance policyholders and the basis of social security insurance assessment are available monthly

- Administrative data, especially Tax income data, is subject to revisions.

FRAME

The sample of the Annual Structural Business Survey is drawn every year from the Statistical Business Register (SBR).

SBR is updated continually throughout the year. The principal and secondary activities are identified using the top-down method and variables value added, number of employees or turnover.

Annual data collection.

Annual Structural Business Survey named P5-01: surveyed are legal persons, market producers, classified in the institutional sector S.11 with a few exceptions, e.g. excluded are enterprises with the principal activity NACE 01, non-profit institutions classified in the sector S.11 or housing associations, on the contrary included are certain types of enterprises classified in the sector S.12 (Financial Corporations).

Every respondent can choose one of the three ways of questionnaire completion: 1) web-based questionnaire 2) interactive PDF form of the questionnaire or 3) paper form of the questionnaire (currently almost unused).

There are several types of verification checks built-in the web-based and interactive PDF form of the questionnaire classified into the two types: interactive and batch checks and four categories: errors (Occurrence of the phenomenon is not possible), anomalies (Occurrence of the phenomenon is possible, but unlikely. The phenomenon must be commented. If found out by a statistician, it must be always verified by contacting the respective unit), non-response (The respective cell of the form is not filled in, even though according to the available information it should be) and information (Occurrence of the phenomenon is possible, but unlikely. Respondent should comment the phenomenon. If found out by the statistician, it does not have to be always verified. Decision whether it should be verified or not is up to consideration of the statistician responsible for data verification).

The web based questionnaire contains all defined checks, interactive PDF form only selected ones. Notwithstanding the way of the questionnaire completion, every questionnaire is verified by the internal verification application – the off-line version of the on-line web application for the questionnaires' completion.

Part of the data verification process is also their confrontation with available administrative data.

After the initial verification the data is sent to a central database for further processing. In the central database the surveyed data together with administrative and other external data enters the imputation procedures in order to be grossed up to the whole population. Subsequently, it is analysed and checked by the branch responsible experts.

The cycle - collection of the questionnaires, their verification, grossing up to the whole population and analysis and verification of the data concerning the whole population - is repeated several times. Particularly when there are new input data (from the survey or administrative sources).

The partner organizations provide CZSO with the administrative data in the form of extracts from the internal databases.

See 18.3 Data collection.

The method of imputation/grossing-up: Estimates were based on the theory of superpopulation model of regression estimates. Simply speaking for regression estimates, the subject of the random variable is not variable Y_i, but it is the remainder (random error) of the regression model.

Modification of RATIO estimators (or Superpopulation model) was applied in the CZSO and used for estimation of expected values for all unique units with no response or not in the sample in corresponding stratum in the frame. Additivity of estimated values was preserved at all level breakdowns by stratification variables. A regression estimator using model with one or two auxiliary variables (without an intercept) was applied for mass imputation. Weights were calibrated subject to totals of two auxiliary variables for all active units in the corresponding stratum. Given the difference of the indicators measured, seven different weights were calculated. The variable number of units at the end of the year was combined with one of the auxiliary variables:

1. Number of social insurance policyholders
2. Total revenues from tax income statement
3. Total assets
4. Tangible fixed assets
5. Intangible assets
6. Basis of social security insurance assessment

The auxiliary variable to calculate regression coefficients are given below:

1. Number of social insurance policyholders
2. Revenues from tax income statement

Statistical Business register (SBR) data as of the date of sampling (generation of samples) were the source of values for auxiliary variables.

Initial weights were calculated from SBR data both as inverse values of probabilities with which a unit would have been included in the sample in a given stratum. The adjustment of the weights to final values for the calculation of estimates takes into account non-activity of units, non-responses and shifts between CZ-NACE divisions (resulting from differences between the BR and actual information found by the statistical survey). Then weighted least squares method was applied in each stratum. The regression coefficients are then used to impute all surveyed variables for non-respondent units or units not in the sample. Different regression models are applied to different group of variables (e.g. turnover based variables, employees based variables).

For the strata with inadequate number of correctly filled in questionnaires, data from neighbour strata were employed to obtain estimates via regression model (so called “borrowing strength”).

As regards estimates for detailed breakdown, they were computed as an aggregate of unique unit estimated values.

Further relevant information can be found in the chapter 13.3 Non-sampling error.

We applied a correction in case the reference period differs from the calendar year.

No other comments.