Reference metadata describe statistical concepts and methodologies used for the collection and generation of data. They provide information on data quality and, since they are strongly content-oriented, assist users in interpreting the data. Reference metadata, unlike structural metadata, can be decoupled from the data.
North Gate II - Bd du Roi Albert II, 16 - 1000 Bruxelles
1.6. Contact email address
Confidential because of GDPR
1.7. Contact phone number
Confidential because of GDPR
1.8. Contact fax number
Confidential because of GDPR
2.1. Metadata last certified
31 March 2023
2.2. Metadata last posted
31 March 2023
2.3. Metadata last update
31 March 2023
3.1. Data description
The annual Business demography data collection covers variables which explain the characteristics and demography of the business population. The methodology allows for the production of data on enterprise births (and deaths), that is, enterprise creations (cessations) that amount to the creation (dissolution) of a combination of production factors and where no other enterprises are involved (enterprises created or closed solely as a result of e.g. restructuring, merger or break-up are not considered).
A summary of the available indicators is listed below. The data is available at EU, country and regional level, with breakdowns for type of activity, legal form and size class.
For the population of active enterprises: • Number of active enterprises • Number of enterprise births • Number of enterprise survivals up to five years • Number of enterprise deaths • Related variables on employment: 'employees' and 'persons employed' (employees and self-employed persons)
For the population of active employer enterprises: • Number of enterprises having at least one employee • Number of enterprises having the first employee • Number of enterprises having no employees anymore • Number of enterprise survivals up to five years • Related variables on employment: 'employees' and 'persons employed' (employees and self-employed persons)
For high-growth enterprises, the following indicators are available at EU and country level: • Number of high-growth enterprises (growth by 10% or more) • Number of employees of high-growth enterprises • Number of young high-growth enterprises (up to five years old high-growth enterprises) • Number of employees of young high-growth enterprise
3.2. Classification system
From 2008 onwards NACE Rev.2 classification (Statistical Classification of Economic Activities in the European Community) is used for all indicators.
The Regional breakdown of national business demography data at NUTS1, NUTS2 and NUTS3 level is based on the Nomenclature of Territorial Units for Statistics (NUTS).
3.3. Coverage - sector
Starting with reference year 2021, BD data cover the economic activities of market producers within the NACE Rev. 2 Sections B to N, P to R and Divisions S95 and S96. The total economy is presented as Industry, construction and services (code BTSXO_S94).
For the reference years 2008-2020, data for the Sections P, Q, R and S were provided on a voluntary basis and K64.2 was not covered.
3.4. Statistical concepts and definitions
BD constitutes an important and integrated part of the EU Regulation 2019/2152 on European Business Statistics (EBS Regulation).
3.5. Statistical unit
Enterprise. We use the European profiling ( with IPT) and national profiling. An other approach is carried out by matching operations (bottom-up approach).
3.6. Statistical population
The target population is the private sector economy, including all active (having either turnover, employment at any time during the reference year) enterprises. In the additional datasets on employer business demography, the threshold is set to one employee at any time of the reference period. The following thresholds are used:
1 employee - population of employer enterprises,
10 employees in the beginning of the growth - population of high-growth enterprises (10%)
3.7. Reference area
Belgium
3.8. Coverage - Time
2006-2022
3.9. Base period
[Not applicable]
• The number of active, birth, death and survival enterprises, as well as high-growth enterprises is expressed in units. • The number of employees is counted as head counts and is expressed in units. • The number of persons employed is the sum of number of employees and self-employed persons. • The number of self-employed persons is the average number of persons who were at some time during the reference period the sole owners or joint owners of the statistical unit in which they work, measured in annual average headcounts, expressed in units. • Derived indicators are expressed in units or percentages
2022
6.1. Institutional Mandate - legal acts and other agreements
Before reference year 2021, EU Regulation 2008/295 on structural business statistics, Annex IX, was providing a legal basis for the BD data collection. The Commission implementing EU Regulation 2014/439 ensured data collection on employer enterprises (with at least one employee), high-growth enterprises (more than 10% annual growth over three years) and their employment.
Up to reference year 2006 data have been collected under gentlemen's agreement within the context of the development of Structural Business Statistics.
6.2. Institutional Mandate - data sharing
[Not applicable]
7.1. Confidentiality - policy
We have a law (Loi relative a la statistique publique) published on 04 July 1962 and more specifically the articles 1st (modified on 22 March 2006), 2nd and 17th.
7.2. Confidentiality - data treatment
We apply primary confidentiality based, on the one hand, on a threshold number of enterprises below which we hide employment data, and on the other hand on the dominance of one or two enterprises in terms of employment. After that, we apply secondary confidentiality in order to secure primary confidentiality.
7.2.1. Confidentiality rules (primary and secondary)
Data treatment
Remarks
Confidentiality rules applied
yes
Threshold of number of enterprises (Number)
Less than 3
Number of enterprises non confidential, if number of employments is confidential
yes
Dominance criteria applied
yes
If dominance criteria is applied, specify the threshold (in %) and the method of applying the dominance rules
90%
Secondary confidentiality applied
yes
If secondary confidentiality is applied, explain the rules and the methods used
We use the software Tau-Argus to treat the secondary confidentiality, with a usual threshold (equals to 10% or 20%) for manual safety range.
7.2.2. Measures taken to reduce the number of confidential cells
Remarks
Measures taken to reduce the number of confidential cells
yes
If measures have been taken, describe them briefly
In order to reduce the number of confidential cells, we reduce sometimes the manual safety range from 20% to 10%. Moreover, the number of enterprises are not confidential since refyear 2013. The number of employment are still confidential when a flag is set.
Impact of these measures
satisfactory
8.1. Release calendar
The policy is to release the final data during this year.
There is an electronic publication (database). Next release in September 2025.
10.3.1. Data tables - consultations
[Not requested]
10.4. Dissemination format - microdata access
No planification about dissemination of microdata.
10.5. Dissemination format - other
Data are sent to Eurostat, to be released as national data as well as to be used for EU aggregates
10.5.1. Metadata - consultations
[Not requested]
10.6. Documentation on methodology
Methodological explanations are available on the website with a short text.
10.6.1. Metadata completeness - rate
[Not requested]
10.7. Quality management - documentation
[Not requested]
11.1. Quality assurance
The Code of Practice is applied.
11.2. Quality management - assessment
BD data are complete, coherent and of good quality. Data sources are available on time. Ensurance of coherent SBS and BD created some delays in the data publication, however it the work in progress.
12.1. Relevance - User Needs
We do not have consultations with our main users.
12.2. Relevance - User Satisfaction
We do not keep track of the number of downloaded on-line publications or databases. We do not organise a regular survey related to the users' satisfaction.
12.3. Completeness
No missing data.
12.3.1. Data completeness - rate
100%
13.1. Accuracy - overall
[Not requested]
13.2. Sampling error
[Not applicable]
13.2.1. Sampling error - indicators
[Not applicable]
13.3. Non-sampling error
Information is not available
13.3.1. Coverage error
[Not requested]
13.3.1.1. Over-coverage - rate
[Not requested]
13.3.1.2. Common units - proportion
[Not requested]
13.3.2. Measurement error
[Not applicable]
13.3.3. Non response error
[Not applicable]
13.3.3.1. Unit non-response - rate
[Not applicable]
13.3.3.2. Item non-response - rate
[Not applicable]
13.3.4. Processing error
No proportions of enterprises wrongly designated active and non-active are available for the moment.
We estimate that the false rejection rate is less than 5% ; the false acceptation rate is less than 1%. Every accepted match has been checked manually ; no automatic acceptation procedure has been implemented; based on our experience, we estimate that 1% is the risk of error in the manual check procedure.
The matches that have been automatically rejected are only related to a part of the LOC-DENOM table. In this table, matches have been sorted out by decreasing matching scores values. Manual checks stop as soon as the score was such that on the last 100 proposed matches, only 5 were accepted.
13.3.5. Model assumption error
[Not requested]
14.1. Timeliness
Since the publication for ref. year T is at T+18 months, there is no risk that the data might be affected. Here below the time lags of the inflow data relevant for EBSBDS coming in our register.
Source TVA-BTW-VAT, units - ULEG, identification data, monthly, T+45 days
Source TVA-BTW-VAT, units - ULEG, identification data, monthly, T+45 days
Source TVA-BTW-VAT, units - ULEG, individual data (VAT form), weekly, T+45 days
Source TVA-BTW-VAT, units - ULEG, composition of VAT Fiscal group (VAT units), weekly T+1 day
Source ONSS-RSZ, units - ULEG, individual data (employment), quarterly, T+2months, T+4months, T+7months
Source ONSS-RSZ, units - ULEG, identification data, quarterly, T+2months
Source KBO-BCE, units - ULOC, identification data, daily, no delay
Source BNB-NBB, units - ULEG, annual accounts, weekly, T+1 week
14.1.1. Time lag - first result
[Not requested]
14.1.2. Time lag - final result
[Not requested]
14.2. Punctuality
The recent problems of punctuality are essentially due to the unification of data between BD and SBS: BD is based on administrative sources and SBS is based on a survey with sampling: the processing of SBS includes in its process numerous steps, due to non-response, changes in the activity of enterprises as well as corrections to the profiling and matching of enterprises. This processing must therefore be included in the process for BD. The SBS data must first be validated and frozen before these steps can be applied in the BD process. Since the SBS data cannot be validated and frozen before Y+18, we have to wait until Y+18 to carry out these processes. However, if the first years of reunification required time and adjustments, we are optimistic that the next years will experience fewer difficulties.
14.2.1. Punctuality - delivery and publication
[Not requested]
15.1. Comparability - geographical
[Not requested]
15.1.1. Asymmetry for mirror flow statistics - coefficient
[Not applicable]
15.2. Comparability - over time
a) First reference year available (calendar year): 2006
b) Breaks in time series and reasons for the breaks: 2008 and 2018
The main reason is due to the change in the NACE nomenclature for non comparable data before 2008.
The introduction of enterprise concept could imply disruptions between 2008-2017 and 2018-...
c) Outliers in time series: No
15.2.1. Length of comparable time series
2006-2017 (births and deaths) 2008-2017 (active enterprises)
2018-... (births and deaths) 2018-... (active enterprises)
15.3. Coherence - cross domain
The reunification of SBS and BD eliminates differences in the number of enteprises and employment between these statistics.
15.3.1. Coherence - sub annual and annual statistics
[Not applicable]
15.3.2. Coherence - National Accounts
[Not requested]
15.4. Coherence - internal
[Not requested]
[Not requested]
17.1. Data revision - policy
No data revision policy is planned.
17.2. Data revision - practice
For the differences between preliminary and final data, the main explanation is the information given by the ref. year + 2, in order to consider a unit as active and so non-death unit. The number of preliminary deaths are greater than the number of final deaths.
17.2.1. Data revision - average size
No data revision policy is planned.
18.1. Source data
a) Type of data source:
This source is based on
Identification data. a daily copy of the Crossroad bank of enterprises (BCE/KBO). The Crossroad Bank for Enterprises is a register containing comprehensive identification data related to businesses and their ‘establishment units’ (i.e. business locations). It includes data from the national register of legal entities and the trade register and VAT and NSSO information and is kept up to date by the relevant bodies which enter data there. Individual data. Quantitative data coming from Value Added Tax administration (VAT declarations: turn-over,..), ONSS (National Social Security Organization) (employment,..), ONSSAPL (local authorities administrations) (employment data), Income tax of Natural Persons (*) A regular copy to the national accounts of the National Bank of Belgium. (*)Note that the access to the Income tax of Natural Persons declarations is necessary because . until June 2009, natural persons such as lawyers or doctors (liberal professions) or accountants that do not have employment and are not registered to the VAT were not included in the KBO database. Now, the access to this source is less and less necessary because the coverage of those professions is more and more complete
b) Coverage of SBR (Statistical Business Register):
All required activities and legal forms are covered.
c) Matching, profiling or imputation:
No matching or profiling within these sources has any impact on profiling, matching or imputations as set out in the methodological guidelines.
18.1.1. Concepts and sources
Criteria for inclusion in SBR :
VAT (TVA-BTW). The conditions for inclusion in the VAT register can be found in CODE DE LA TAXE SUR LA VALEUR AJOUTEE Moniteur Belge 03 July 1969, Dernière modification, par la Loi du 29 mars 2012, M.B. du 6 avril 2012, éd. 3).
The tax is mainly due for the activities linked to the production of goods or services. Some exemptions exist (e.g. lawyers, dentists, nurses, hospitals, banks, insurances,..). A special case holds when the annual turnover is below 5580: then the enterprise is exempted to fill in the VAT declaration (régime de la franchise de la taxe). This means that the real turnover is unknown for those units in our database.
NSSO (ONSS-RSZ). The largest category of individuals reported to the NSSO as employees are those working according to the terms of a work contract. This is a contract that binds an individual to carry out services under the authority of another individual. Certain individuals are employees, but their employer need not report them to the NSSO due to the limited nature of their services. Examples can be found in Social and cultural sector, Sporting Events, Agricultural activities, for Students, Household staff – Domestics (more details on: Employers and nsso).
18.2. Frequency of data collection
Annual
18.3. Data collection
Business demography variables are compiled from the national statistical business register.
18.3.1. Data matching
a) Data matching process and tools:
The tools used to compute a similarity score between strings of characters are
The SAS function SPEDIS. For a description of this function, see for example : on website
Implementation in SAS of the TF-IDF method (Term Frequency – Inverse Document Frequency) (also called cosine similarity, see G. Salton, C. Buckley, Term weighting approaches in automatic text retrieval : Information Processing and Management, vol. 24, n° 5, 1988, pp. 513-523.).
b) Matching:
The matchings between units are based on a correspondence on
4 digits NACE code (NACE) Names of enterprises (DENOM) Addresses of enterprises (LOC). Note that our business register holds the street codes of enterprises, making the comparison on addresses more easy. These codes come from the national register.
The results of the matching program have been grouped in several tables for eventual further clerical checks:
(1) NACE_LOC_DENOM: this table holds the pairs of units with same NACE, address and names.
(2) LOC_DENOM: this table holds the pair of units with same address and names.
(3) NACE_DENOM: this table holds the pair of units with same NACE, and names.
(4) NACE_LOC: Tous les couples d’unités légales ayant même NACE et même LOC.
(5) NACE: this table holds the pairs of units with same NACE. Only mixed pairs of legal and natural persons are considered here.
All the matches found by the automatic procedure have been considered for a final decision:
0 : reject the match
1 : accept the match
2 : reject but a possible use of this match can be made for profiling complex groups of enterprises
3 : Doubt.
This decision is based either upon
a clerical check (this is the case for all the acceptations and an important part of rejects and doubts ),
on ad-hoc rules that have been progressively built upon the expertise acquired from our team.
For ref.year 2009, a set of 650000 links have been identified and analyzed, either automatically or manually. 105000 cases (16%) have been manually checked or were treated via ad-hoc rules (see g and h). Because of the introduction of automatic rules, the percentage of manual checks has drastically decreased in ref year 2010-2011.
The ad-hoc rules may depend on the activity sector : special rules hold for Hotels and restaurants, pharmacies, agriculture (even if this sector is not in the scope for SBS-Ann9, we do consider it in order to build a frame of all active units from which all our universes can be extracted), teaching, liberal professions. For other cases not falling into these special categories, we also have a set of additional rules to apply. In general, as soon as a legal person is involved in the match, the rule is more severe for accepting it. This is because we have observed that the risk of false acceptation is higher for the legal units than for the natural persons.
18.3.2. Manual checks
The number of large births and large deaths manually investigated is about 500 (the top 250 employer enterprises (with at least 1 employee) and the top 250 turnover).
This approach covers all births and deaths.
Confirmation of real births: about 64%
Confirmation of real deaths: about 77%
Percentage of real large births and deaths by legal form:
Enterprise birth: LL 44%, PA 56%, SP 0%
Enterprise death: LL 56%, PA 29%, SP 15%
Percentage of real large births and deaths by NACE sections:
Birth: B, D, K, P 0%, E, J, L, R, S 1%, Q 2%, H 3%, I 5%, M 6%, N 6%, C 9%, F 16%, G 50%
Death: B, D, S 0%, K, E, R 1%, J, P, L, C 2%, Q 3%, I, M 4%, N 7%, H 8%, F 31%, G 33%.
18.4. Data validation
Before sending to Eurostat, the following checks are performed: hierarchical, inter-variable plausibility, confidential and completeness.
18.5. Data compilation
Estimations are not done
18.5.1. Imputation - rate
[Not requested]
18.6. Adjustment
[Not applicable]
18.6.1. Seasonal adjustment
[Not applicable]
No comments.
The annual Business demography data collection covers variables which explain the characteristics and demography of the business population. The methodology allows for the production of data on enterprise births (and deaths), that is, enterprise creations (cessations) that amount to the creation (dissolution) of a combination of production factors and where no other enterprises are involved (enterprises created or closed solely as a result of e.g. restructuring, merger or break-up are not considered).
A summary of the available indicators is listed below. The data is available at EU, country and regional level, with breakdowns for type of activity, legal form and size class.
For the population of active enterprises: • Number of active enterprises • Number of enterprise births • Number of enterprise survivals up to five years • Number of enterprise deaths • Related variables on employment: 'employees' and 'persons employed' (employees and self-employed persons)
For the population of active employer enterprises: • Number of enterprises having at least one employee • Number of enterprises having the first employee • Number of enterprises having no employees anymore • Number of enterprise survivals up to five years • Related variables on employment: 'employees' and 'persons employed' (employees and self-employed persons)
For high-growth enterprises, the following indicators are available at EU and country level: • Number of high-growth enterprises (growth by 10% or more) • Number of employees of high-growth enterprises • Number of young high-growth enterprises (up to five years old high-growth enterprises) • Number of employees of young high-growth enterprise
31 March 2023
BD constitutes an important and integrated part of the EU Regulation 2019/2152 on European Business Statistics (EBS Regulation).
Enterprise. We use the European profiling ( with IPT) and national profiling. An other approach is carried out by matching operations (bottom-up approach).
The target population is the private sector economy, including all active (having either turnover, employment at any time during the reference year) enterprises. In the additional datasets on employer business demography, the threshold is set to one employee at any time of the reference period. The following thresholds are used:
1 employee - population of employer enterprises,
10 employees in the beginning of the growth - population of high-growth enterprises (10%)
Belgium
2022
[Not requested]
• The number of active, birth, death and survival enterprises, as well as high-growth enterprises is expressed in units. • The number of employees is counted as head counts and is expressed in units. • The number of persons employed is the sum of number of employees and self-employed persons. • The number of self-employed persons is the average number of persons who were at some time during the reference period the sole owners or joint owners of the statistical unit in which they work, measured in annual average headcounts, expressed in units. • Derived indicators are expressed in units or percentages
Estimations are not done
a) Type of data source:
This source is based on
Identification data. a daily copy of the Crossroad bank of enterprises (BCE/KBO). The Crossroad Bank for Enterprises is a register containing comprehensive identification data related to businesses and their ‘establishment units’ (i.e. business locations). It includes data from the national register of legal entities and the trade register and VAT and NSSO information and is kept up to date by the relevant bodies which enter data there. Individual data. Quantitative data coming from Value Added Tax administration (VAT declarations: turn-over,..), ONSS (National Social Security Organization) (employment,..), ONSSAPL (local authorities administrations) (employment data), Income tax of Natural Persons (*) A regular copy to the national accounts of the National Bank of Belgium. (*)Note that the access to the Income tax of Natural Persons declarations is necessary because . until June 2009, natural persons such as lawyers or doctors (liberal professions) or accountants that do not have employment and are not registered to the VAT were not included in the KBO database. Now, the access to this source is less and less necessary because the coverage of those professions is more and more complete
b) Coverage of SBR (Statistical Business Register):
All required activities and legal forms are covered.
c) Matching, profiling or imputation:
No matching or profiling within these sources has any impact on profiling, matching or imputations as set out in the methodological guidelines.
Annual
Since the publication for ref. year T is at T+18 months, there is no risk that the data might be affected. Here below the time lags of the inflow data relevant for EBSBDS coming in our register.
Source TVA-BTW-VAT, units - ULEG, identification data, monthly, T+45 days
Source TVA-BTW-VAT, units - ULEG, identification data, monthly, T+45 days
Source TVA-BTW-VAT, units - ULEG, individual data (VAT form), weekly, T+45 days
Source TVA-BTW-VAT, units - ULEG, composition of VAT Fiscal group (VAT units), weekly T+1 day
Source ONSS-RSZ, units - ULEG, individual data (employment), quarterly, T+2months, T+4months, T+7months
Source ONSS-RSZ, units - ULEG, identification data, quarterly, T+2months
Source KBO-BCE, units - ULOC, identification data, daily, no delay
Source BNB-NBB, units - ULEG, annual accounts, weekly, T+1 week
[Not requested]
a) First reference year available (calendar year): 2006
b) Breaks in time series and reasons for the breaks: 2008 and 2018
The main reason is due to the change in the NACE nomenclature for non comparable data before 2008.
The introduction of enterprise concept could imply disruptions between 2008-2017 and 2018-...
