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.
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.
Institutional sector classification (SEC2010) is used as selection criteria of enterprises.
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).
The reference period used is the calendar year.
Employment and/or turnover criteria were used to decide if a unit was active in the given period.
Birth and death dates from the business register have not been used to decide if units are active or not, except for provisionnal enterprise death data for reference year t.
Enterprises classified in institutional sectors S127, S13 and S15 and enterprises identified as SPE's are not covered.
Enterprises classified in NACE K642 and K643 are not covered as these enterprises are either classified in institutional sector S127 or are non-market producers.
Enterprises having some specific legal forms are not covered as these enterprises are either SPEs and/or are non-market producers (e.g. natural persons operating PV systems).
Enterprises classified in NACE M69101 (Lawyers) are only included when having at least one employee in the reference year.
Information from Social security is used to include enterprises for NACE Q8621 and Q8622 which cannot be identified by using employment and/or turnover data.
Enterprises classified in NACE Rev. 2 K652 are no more covered from reference year 2021 on (these enterprises are considered as being only financial captive entities and non-market producers).
3.5. Statistical unit
Enterprise (Enterprises in the NSBR are compiled from legal units which are available from administrative data sources; enterprises consists of one or more legal units).
3.6. Statistical population
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.
NACE Rev. 2 K642 and K643 are not covered as these enterprises are classified in institutional sector S127 (which is not covered) and/or are non-market producers.
All enterprises having at least one employee and/or turnover in the reference period are covered.
For further exclusions, please see concept 3.4.
3.7. Reference area
The whole territory of Luxembourg is covered, no regions are excluded.
Branches of foreign enterprises operating in Luxembourg are included.
3.8. Coverage - Time
Data are available from 1997 onwards for active enterprises and enterprise deaths.
Data are available from 1998 onwards for enterprise births.
Data are available from 1999 onwards for enterprise survivals.
Data on employer enterprises is available from 2004 onwards.
Data on high-growth enterprises is available from 2008 onwards.
Data on regional business demography is availble from 2021 onwards.
Break in time series:
2021: Enterprises classified in NACE Rev. 2 K652 are no more covered from reference year 2021 onwards (these enterprises are considered as being only financial captive entities and non-market producers).
2018: Enterprises classified in NACE Q861 have been classified from institutional sector S11 to S13 on request by the National accounts department and are no more covered by BD from reference year 2018 onwards.
2015: A big enterprise has been classified from NACE G479 to M7010 from reference year 2015 onwards.
2013: Enterprises classified in institutional sector S13 are no more covered in BD from reference year 2013 onwards. The following NACE divisions where impacted: 37, 38, 68, 72, 85, 90, 91 and 93. Furthermore, the institutional sector classification of some more important enterprises has been revised too. One big enterprise classified in NACE H49 has been classified in institutional sector S13 and is no more covered. Some enterprises classified in NACE Q87 and Q88 have been classified outside of institutional sector S15 and are covered by BD from that point on.
2007-2008: The small break in time series from 2007 to 2008 is due to the introduction of NACE rev. 2 in BD.
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.
2021.
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.
The overall assessment of the quality of Business Demography data is good. Data quality is in accordance with principles of accuracy and reliability, timeliness and punctuality, coherence and compatibility. See chapters 12, 13, 14, 17 and 18.
All required data has been send to Eurostat. The accuracy of the data is considered to be high.
12.1. Relevance - User Needs
There are no specific user needs as regards business demography except micro-data made available to researchers under specific conditions.
Business demography data published at national level is less detailed than the one send to Eurostat.
12.2. Relevance - User Satisfaction
No views and opinions of the users are collected. No punctual or regular survey related to the user's satisfaction regarding the availability and the quality of the data used, and whether their needs were fulfilled has been organized.
12.3. Completeness
All required variables and breakdowns have been send to Eurostat.
12.3.1. Data completeness - rate
100%.
13.1. Accuracy - overall
The accuracy of the data is considered to be high.
13.2. Sampling error
Not applicable.
13.2.1. Sampling error - indicators
Not applicable.
13.3. Non-sampling error
The proportion of wrongly designated active enterprises should be extremly low or zero as we rely only on employment and turnover data to identify active enterprises.
The proportion of wrongly designated non-active enterprises should be low and only relevant for enterprises having no employees where turnover data is not yet available or missing when compiling BD data.
All administrative data sources used for the NSBR can be matched by using national ID. All legal units which are part of an enterprise can be identified by national ID. Enterprises consist of one or more legal units and can be identified by NSBR enterprise ID. The proportion of false matches and false non-matches should be zero.
The proportion of wrongly identified real birth or death should be low as a lot of clerical checking work described later is done to validate the results and should be relevant for small enterprises only.
The proportion of non-identified real birth or death should be low and be relevant for small enterprises only.
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
All significant processing errors that have appeared in the past or now before the final data collection process have been removed.
13.3.5. Model assumption error
Not requested.
Timeliness of the source data is considered to be high. Punctuality of data transmission can be improved.
14.1. Timeliness
All administrative data sources used for NSBR and to compile BD data is updated monthly.
Turnover data for reference year t may not be completely available at t+18 months when BD data should be send to Eurostat (the missing turnover data should represent less than 1% of total turnover data available and is only relevant for enterprises having no employees).
BD data is released at national level only after data has been send to Eurostat, The national release of BD data should take place 24 months after the end of the reference period at latest.
14.1.1. Time lag - first result
Not requested.
14.1.2. Time lag - final result
Not requested.
14.2. Punctuality
Most data sets have been send to Eurostat with less than 70 days delay, EBSBDS_FHG adn EBSBDS_ALL have been send to Eurostat with more than 70 days delay.
Reasons for delays:
- Reorganisation inside the NSBR (colleagues who left the office have to be replaced which is still ongoing);
- Implementation of changes and new requirements resulting from the new EBS regulation.
Action taken or planned to improve punctuality:
- Reorganization of the NSBR should be finished for the next data collection;
- Changes and new requirements resulting from the new EBS regulation have been implemented.
14.2.1. Punctuality - delivery and publication
Not requested.
Coherence over time is considered to be very good.
Coherence between BD and BR : cannot be compared anymore.
Coherence between BD and SBS : can be improved.
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):
1998
b) Breaks in time series and reasons for the breaks:
2021: Enterprises classified in NACE Rev. 2 K652 are no more covered from reference year 2021 onwards (these enterprises are considered as being only financial captive entities and non-market producers).
2018: Enterprises classified in NACE Q861 have been classified from institutional sector S11 to S13 on request by the National accounts department and are no more covered by BD from reference year 2018 onwards.
2015: A big enterprise has been classified from NACE G479 to M7010 from reference year 2015 onwards.
2013: Enterprises classified in institutional sector S13 are no more covered in BD from reference year 2013 onwards. The following NACE divisions where impacted: 37, 38, 68, 72, 85, 90, 91 and 93. Furthermore, the institutional sector classification of some more important enterprises has been revised too. One big enterprise classified in NACE H49 has been classified in institutional sector S13 and is no more covered. Some enterprises classified in NACE Q87 and Q88 have been classified outside of institutional sector S15 and are covered by BD from that point on.
2007-2008: The small break in time series from 2007 to 2008 is due to the introduction of NACE rev. 2 in BD.
c) Outliers in time series:
There are no known outliers.
15.2.1. Length of comparable time series
1998-2007 / 2008-2012 / 2013-2021
For reasons, please see 15.2.
15.3. Coherence - cross domain
1) Coherence between BD characteristics and the same characteristics in Business Register (Number of enterprises, number of persons employed):
The reference population used for the annual inquiry on business registers is compiled of enterprises covered by Business demography, enterprises covered for EGR (EuroGroups Register) and other enterprises not covered by the former. The reference population used for Business demography can no longer be compared to the one used for the annual inquiry on Business Registers for the following reasons:
- BD uses only employment and/or turnover data to identify active enterprises;
- BD does not cover enterprises classified in institutional sectors S127, S13 and S15;
- BD does not cover enterprises identified as SPE's (others than those classified in S127);
- BD does not cover enterprises classified in NACE K642 and K643 as these enterprises are either classified in institutional sector S127 or are non-market producers;
- BD does not cover enterprises having some specific legal forms as these enterprises are either SPEs and/or are-non market producers (e.g. natural persons operating PV systems);
- BD does not cover enterprises classified in NACE M69101 (Lawyers) having no employees;
- BD uses information from Social security to include enterprises for NACE Q8621 and Q8622 which cannot be identified by using employment and/or turnover data;
- BD does not cover anymore enterprises classified in NACE Rev. 2 K652 from reference year 2021 on (these enterprises are considered as being only financial captive entities and non-market producers);
- BD does not cover inactive enterprises.
2) Coherence between BD characteristics and the same characteristics in Structural Business Statistics (Number of enterprises, number of persons employed, number of employees):
2.1) Number of enterprises
Small differences can be explained by a different date of compilation of the two populations and different selection criterias used for identifying active enterprises for the same reference year.
Big differences can be explained as follows:
Section B: small number of enterprises (<10), the difference is one enterprise;
Section H: most enterprises classified in NACE H50 are removed for SBS data (SPEs) or classified elsewhere in NACE;
Section K: BD does not cover enterprises classified in NACE K642 and K643 as they are either classified in institutional sector S127 or non-market producers;
Section M: BD covers only enterprises classified in NACE M691 having at least one employee in the reference period;
Section Q: BD covers enterprises classified in NACE Q8621 and Q8622 which are not available in NSBR (additional source used: Social Security).
2.2) Number of persons employed
Small differences can be explained by a different date of compilation, different selection criterias used for identifying active enterprises for the same reference year.
Big differences can be explained as follows:
Different source and estimation method used for estimating the number of persons employed.
Section H: most enterprises classified in NACE H50 are removed for SBS data (SPEs) or classified elsewhere in NACE.
2.3) Number of employees
Small differences can be explained by a different date of compilation, different selection criterias used for identifying active enterprises for the same reference year.
Big differences can be explained as follows:
Section H: most enterprises classified in NACE H50 are removed for SBS data (SPEs) or classified elsewhere in NACE.
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
Revision of preliminary data on high growth enterprises is done as it is planned by Eurostat to send final high growth data 6 months later.
Revision of preliminary data on enterprise deaths is done as it is planned by Eurostat to send final enterprise death data 12 months later.
Revision of data collection EBSBDS_ALL is done when data collection EBSBDS_EMP is compiled as additional flags from the latter have to be transferred to the former data collection which could not be done earlier (only number of persons employed for active enterprises).
Other revisions are done when necessary (no other revision has been done in recent years).
17.2. Data revision - practice
1) High growth enterprises:
The difference between preliminary and final data is small.
2) Enterprise death
Using the normal BD methodology to compile preliminary deaths data for reference year t would result in an overestimation of the number of death due to missing turnover data at t+18 (method 1) for reference years t+1 and t+2. By using any other information available in the BR (e.g. date of liquidation and bankruptcy in NSBR), the number of preliminary deaths would be largely underestimated at t+18 (method 2). Instead, we use the final death data from the 2 previous years to estimate a preliminary population of enterprise deaths for t. This population consists of all enterprise deaths estimated by method 2 and a sample of enterprise deaths estimated by method 1.
The differences between preliminary and final death data can be explained as follows: - The use of different methods to estimate/compile both populations; - More turnover data available for reference years t+1 and t+2 at t+30 then at t+18 months; - The small size of our economy (less than 2500 deaths per year) compared to the high number of cells to provide makes it more difficult to compile a preliminary death population.
3) EBSBDS_ALL
Additional secondary confidentiality flags have to be transferred from data collection EBSBDS_EMP to EBSBDS_ALL (only for the number of persons employed broken down by employee size class for active enterprises). As this can only be done when data for collection EBSBDS_EMP has been compiled, the revision of data collection EBSBDS_ALL is necessary for each cycle.
17.2.1. Data revision - average size
Please see 17.1.
18.1. Source data
a) Type of data source: administrative data, survey
Central population register of private and legal persons (new LEU, update of LEU characteristics)
VAT register (new LEU, update of LEU characteristics, turnover data)
Social security (new LEU, employment data)
Trade register (update of Enterprise unit (annual accounts, ownership information))
Survey on economic activities (NACE and institutional sector code, legal form changes, demographic events)
Other sources: SBS and National accopunts (update of NACE, institutional sector and enterprise unit)
b) Coverage of SBR (Statistical Business Register):
All required economic activities and legal forms for BD are covered. For some activities, coverage is not complete.
All legal entities which are not natural persons are covered in NSBR.
Natural persons having any economic activity are only covered when having any employees and/or being registered to VAT.
No threshold is applied for number of employees in NSBR.
Number of self-employed persons is not directly available in NSBR (additional data from Social Security).
VAT threshold as regards turnover: 30.000€ annual turnover. The VAT threshold means that enterprises are not obliged to register to VAT but are allowed to register on a voluntary basis. The BD population of active enterprises includes about 15% of enterprises having an annual non-zero turnover of less than 30.000€.
c) Matching, profiling or imputation:
All administrative data sources in NSBR used for BD are matched by using National ID.
Name matching is only partially done to add Trade Register ID to NSBR (used to identify legal form changes, update of enterprise unit).
Data collected by economic activity survey can be matched with other sources by using national ID.
Profiling is used to update NACE and/or institutional sector code of enterprise and compile and update enterprise units.
The legal form is part of the national ID. A change of legal form means that a new legal unit is implemted in the NSBR. The new and old unit have to be linked to the same enterprise ID in NSBR. These changes can only be identified by information collected by our economic activity survey, by using information from Trade Register ID (Trade Register ID does not necessarely change in case of legal form change) and by doing manual checking work for business demography purposes. Not identifying this changes has an impact on real birth and death populations.
Birth and death dates are included in the data sources and implented in the NSBR, but are only partially used to compile the provisional enterprise death population in t at t+18 month.
For legal entities, birth date is the date of implementation in the NSBR (which is close to the date of incorporation).
For natural persons, birth date is either the date of registration to VAT if applicable or the date of the first employee.
Death dates are one of the following:
- official date of liquidation of the unit;
- official date of bankrupty of the unit;
- end date of registration to VAT (however, end of registration to VAT is not necessarely the end of activity of the unit).
18.1.1. Concepts and sources
The NSBR covers all information needed for BD except for self-employed persons needed to calculate the number of persons employed by enterprise. Self-employed persons are compiled from additional Social Security data or estimated by using legal form. This imputation is only done for BD data.
Additional information from Social Security is used to add missing enterprises in NSBR classified in NACE Q8621 and Q8622. This imputation is only done for BD data.
18.2. Frequency of data collection
All administrative data which is needed for BD is updated monthly.
NACE and institutional sector are updated if necessary.
Enterprise unit is updated if necessary.
18.3. Data collection
Data from administrative sources is accessed as an extraction from a data base and either transmitted monthly to the NSBR or can be downloaded from the administration.
18.3.1. Data matching
a) Data matching process and tools:
Enterprise birth having either at least 10 employees in birth year or first year after birth are all manually investigated for real birth (N<150). Non-real birth are removed from birth population.
Enterprise death having either at least 10 employees in death year or last year preceding death are all manually investigated for real death (N<150). Non-real death are removed from death population.
Enterprise birth having the same NACE (4-digit) and address (city, postal code, street no) as active enterprises not part of birth population are manually investigated if necessary for real birth (N<1000). Non-real birth are removed from birth population.
Enterprise death having the same NACE (4-digit) and address (city, postal code, street no) as active enterprises not part of death population are manually investigated if necessary for real death (N<1000). Non-real death are removed from death population.
b) Matching:
Enterprise birth having the same self-employed person than any other active enterprise which is not part of birth population are maunally investigated if necessary for real birth (N<250). Non-real birth are removed from birth population.
Enterprise death having the same self-employed person than any other active enterprise which is not part of death population are maunally investigated if necessary for real death (N<250). Non-real death are removed from death population.
Enterprises having 5 or more employees in year t which were part of the population of active enterprises in year t-1 and which are not part anymore of the population of active enterprises in year t, but are part of the population of active enterprises in NSBR in year t are all manually investiged to check if they are really out of scope for BD (N<50). If necessary, changes have to be made in the NSBR.
Enterprises having 10 or more employees in year t which were not part of the population of active enterprises in year t-1 but are part of the population of active enterprises in t and part of the population of active enterprises in NSBR in year t-1 and not part of the population of birth in year t are all manually investiged to check if they are really out of scope for birth population (N<25). If necessary, changes have to be made in the NSBR. These enterprises are not considered as birth.
Enterprises having 5 or more employees which are part both of the population of active enterprises in year t and t-1 but with different NACE and/or institutional sector code are all manually investigated to check for NACE/institutional sector code to apply in year t (N<50). If necessary, changes have to be made in the NSBR or only for BD populations.
Only head offices are taken into account for matching on location.
There is only a small number of reactivations each year.
18.3.2. Manual checks
Enterprise birth having either at least 10 employees in birth year or first year after birth are all manually investigated to check for real birth (N<150). Non-real birth are found generally for enterprises having more than 20 employees.
Enterprise death having either at least 10 employees in death year or last year preceding death are all manually investigated to check for real death (N<150). Non-real death are found generally for enterprises having more than 20 employees.
Enterprises having no employees but large turnover are no more investigated.
For surving enterprises, we use the populations already compiled for surving year t-1 to compile surving populations for surving year t. For each of the five surving populations to year t, the following manual checking work is done:
- does the enterprise ID still exist in our NSBR? If not, the legal units of the enterprise have been added to an other enterprise already existing in our NSBR (e. g. in case of mergers). In that case, there is no survival of the old enterprise.
- the enterprise ID still exist in our NSBR but the enterprise was no more active in reference year t. In that case, the enterprise has simply stopped activity or has been taken over by another existing enterprise (e.g. mergers). There is no survival of the old enterprise.
- the enterprise ID still exist in our NSBR but the enterprise was no more active in reference year t and another enterprise born in reference year t has been identified to have taken over the activity of the old enterprise (e.g. legal form change, change of ownership,...). In that case, the legal unit(s) of the new enterprise are linked to the old enterprise in the NSBR. There is survival of the old enterprise and the wrongly identified new enterprise disappears from NSBR.
18.4. Data validation
1. Validation of format and file structure checks
The structure of file or format of the datasets is only checked when changes have to be implemented in the data compilation process (e.g. new EBS Regulation). Before sending the data sets to Eurostat, the structure and format of data sets are checked using the pre-validation tool available in EDAMIS.
2. Intra-dataset checks
Intra-dataset checks are only done when changes have to be implemented in the data compilation process (e.g. new EBS Regulation). Before sending the datasets to Eurostat, intra-dataset checks are done by using the pre-validation tool available in EDAMIS.
3. Inter-dataset checks
Inter-dataset checks are only done when changes have to be implemented in the data compilation process (e.g. new EBS Regulation). Consistency has to be guaranteed between the different BD populations. Before sending the data sets to Eurostat, additional inter-datasets checks are done by using the pre-validation tool available in EDAMIS.
4. Intra-domain, intra-source checks
The different BD populations are compared to the ones compiled for the previous year in order to check whether big variations from one year to another at 5 digit NACE level are real or are due to compilation errors or errors in the NSBR. Before sending the data sets to Eurostat, additional checks are done by using the pre-validation tool available in EDAMIS.
5. Plausibility or consistency checks between two domains available in the same Institution
Consistency checks are done by Eurostat by comparing BD data with the data compiled for the annual NSBR inquiry;
Consistency checks are done by Eurostat by comparing BD data with SBS data.
6. Plausibility or consistency checks between the data available in the Institution and the data / information available outside the Institution.
Only applicable for data of NACE classes Q8621 and Q8622.
18.5. Data compilation
Estimation of missing number of self-employed persons
The missing number of self-employed persons is estimated only for the BD compilation and the annual Business Register inquiry.
Estimation of preliminary deaths
Death data for reference year t-1 to send at t+30 months is final as the number of reactivations is extremly low.
Provisional death data for reference year t to send at t+18 months is a mix of real data (by using date of liquidation, date of bankruptcy of the enterprise) and estimated data by using information from the 2 previous death populations.
Computation of annual average of employees and self-employed persons in active enterprises
Please describe how you compute the annual average of employees and self-employed persons in active enterprises:
The monthly number of employees and self-employed persons for each legal unit of the enterprise is added by month to enterprise level which in turn is summed up to the year and divided by 12 month and rounded up to 2 digits after zero.
- If there are no employees, but only self-employed persons working in the enterprise, the annual average is calculated the same way as there were employees. The monthly number of self-employed persons for each legal unit of the enterprise is added by month to the enterprise level which in turn is summed up to the year and divided by 12 month and rounded up to 2 digits after zero.
- Enterprises with no any employees and self-employed persons are included (less than 10% of the population of active enterprises).
- The number of self-employed persons is estimated in case there is no information available from Social Security by using the legal form of the enterprise (e.g. for a natural person having any economic activity one self-employed person is added if no information is available from Social Security).
Annual average is calculated by taking headcount of employees or self-employed persons per month, summed up to year and divided by number of months per year (12).
The number of employees or self-employed at the end of the reference year is not used as approximation for annual average. The annual average is calculated and not the operational period average (the latter was done before the implementation of the new EBS Regulation).
18.5.1. Imputation - rate
Not requested.
18.6. Adjustment
Not applicable.
18.6.1. Seasonal adjustment
Not applicable.
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
28 February 2024
BD constitutes an important and integrated part of the EU Regulation 2019/2152 on European Business Statistics (EBS Regulation).
The reference period used is the calendar year.
Employment and/or turnover criteria were used to decide if a unit was active in the given period.
Birth and death dates from the business register have not been used to decide if units are active or not, except for provisionnal enterprise death data for reference year t.
Enterprises classified in institutional sectors S127, S13 and S15 and enterprises identified as SPE's are not covered.
Enterprises classified in NACE K642 and K643 are not covered as these enterprises are either classified in institutional sector S127 or are non-market producers.
Enterprises having some specific legal forms are not covered as these enterprises are either SPEs and/or are non-market producers (e.g. natural persons operating PV systems).
Enterprises classified in NACE M69101 (Lawyers) are only included when having at least one employee in the reference year.
Information from Social security is used to include enterprises for NACE Q8621 and Q8622 which cannot be identified by using employment and/or turnover data.
Enterprises classified in NACE Rev. 2 K652 are no more covered from reference year 2021 on (these enterprises are considered as being only financial captive entities and non-market producers).
Enterprise (Enterprises in the NSBR are compiled from legal units which are available from administrative data sources; enterprises consists of one or more legal units).
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.
NACE Rev. 2 K642 and K643 are not covered as these enterprises are classified in institutional sector S127 (which is not covered) and/or are non-market producers.
All enterprises having at least one employee and/or turnover in the reference period are covered.
For further exclusions, please see concept 3.4.
The whole territory of Luxembourg is covered, no regions are excluded.
Branches of foreign enterprises operating in Luxembourg are included.
2021.
The accuracy of the data is considered to be high.
• 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.
Estimation of missing number of self-employed persons
The missing number of self-employed persons is estimated only for the BD compilation and the annual Business Register inquiry.
Estimation of preliminary deaths
Death data for reference year t-1 to send at t+30 months is final as the number of reactivations is extremly low.
Provisional death data for reference year t to send at t+18 months is a mix of real data (by using date of liquidation, date of bankruptcy of the enterprise) and estimated data by using information from the 2 previous death populations.
Computation of annual average of employees and self-employed persons in active enterprises
Please describe how you compute the annual average of employees and self-employed persons in active enterprises:
The monthly number of employees and self-employed persons for each legal unit of the enterprise is added by month to enterprise level which in turn is summed up to the year and divided by 12 month and rounded up to 2 digits after zero.
- If there are no employees, but only self-employed persons working in the enterprise, the annual average is calculated the same way as there were employees. The monthly number of self-employed persons for each legal unit of the enterprise is added by month to the enterprise level which in turn is summed up to the year and divided by 12 month and rounded up to 2 digits after zero.
- Enterprises with no any employees and self-employed persons are included (less than 10% of the population of active enterprises).
- The number of self-employed persons is estimated in case there is no information available from Social Security by using the legal form of the enterprise (e.g. for a natural person having any economic activity one self-employed person is added if no information is available from Social Security).
Annual average is calculated by taking headcount of employees or self-employed persons per month, summed up to year and divided by number of months per year (12).
The number of employees or self-employed at the end of the reference year is not used as approximation for annual average. The annual average is calculated and not the operational period average (the latter was done before the implementation of the new EBS Regulation).
a) Type of data source: administrative data, survey
Central population register of private and legal persons (new LEU, update of LEU characteristics)
VAT register (new LEU, update of LEU characteristics, turnover data)
Social security (new LEU, employment data)
Trade register (update of Enterprise unit (annual accounts, ownership information))
Survey on economic activities (NACE and institutional sector code, legal form changes, demographic events)
Other sources: SBS and National accopunts (update of NACE, institutional sector and enterprise unit)
b) Coverage of SBR (Statistical Business Register):
All required economic activities and legal forms for BD are covered. For some activities, coverage is not complete.
All legal entities which are not natural persons are covered in NSBR.
Natural persons having any economic activity are only covered when having any employees and/or being registered to VAT.
No threshold is applied for number of employees in NSBR.
Number of self-employed persons is not directly available in NSBR (additional data from Social Security).
VAT threshold as regards turnover: 30.000€ annual turnover. The VAT threshold means that enterprises are not obliged to register to VAT but are allowed to register on a voluntary basis. The BD population of active enterprises includes about 15% of enterprises having an annual non-zero turnover of less than 30.000€.
c) Matching, profiling or imputation:
All administrative data sources in NSBR used for BD are matched by using National ID.
Name matching is only partially done to add Trade Register ID to NSBR (used to identify legal form changes, update of enterprise unit).
Data collected by economic activity survey can be matched with other sources by using national ID.
Profiling is used to update NACE and/or institutional sector code of enterprise and compile and update enterprise units.
The legal form is part of the national ID. A change of legal form means that a new legal unit is implemted in the NSBR. The new and old unit have to be linked to the same enterprise ID in NSBR. These changes can only be identified by information collected by our economic activity survey, by using information from Trade Register ID (Trade Register ID does not necessarely change in case of legal form change) and by doing manual checking work for business demography purposes. Not identifying this changes has an impact on real birth and death populations.
Birth and death dates are included in the data sources and implented in the NSBR, but are only partially used to compile the provisional enterprise death population in t at t+18 month.
For legal entities, birth date is the date of implementation in the NSBR (which is close to the date of incorporation).
For natural persons, birth date is either the date of registration to VAT if applicable or the date of the first employee.
Death dates are one of the following:
- official date of liquidation of the unit;
- official date of bankrupty of the unit;
- end date of registration to VAT (however, end of registration to VAT is not necessarely the end of activity of the unit).
Annual
All administrative data sources used for NSBR and to compile BD data is updated monthly.
Turnover data for reference year t may not be completely available at t+18 months when BD data should be send to Eurostat (the missing turnover data should represent less than 1% of total turnover data available and is only relevant for enterprises having no employees).
BD data is released at national level only after data has been send to Eurostat, The national release of BD data should take place 24 months after the end of the reference period at latest.
Not requested.
a) First reference year available (calendar year):
1998
b) Breaks in time series and reasons for the breaks:
2021: Enterprises classified in NACE Rev. 2 K652 are no more covered from reference year 2021 onwards (these enterprises are considered as being only financial captive entities and non-market producers).
2018: Enterprises classified in NACE Q861 have been classified from institutional sector S11 to S13 on request by the National accounts department and are no more covered by BD from reference year 2018 onwards.
2015: A big enterprise has been classified from NACE G479 to M7010 from reference year 2015 onwards.
2013: Enterprises classified in institutional sector S13 are no more covered in BD from reference year 2013 onwards. The following NACE divisions where impacted: 37, 38, 68, 72, 85, 90, 91 and 93. Furthermore, the institutional sector classification of some more important enterprises has been revised too. One big enterprise classified in NACE H49 has been classified in institutional sector S13 and is no more covered. Some enterprises classified in NACE Q87 and Q88 have been classified outside of institutional sector S15 and are covered by BD from that point on.
2007-2008: The small break in time series from 2007 to 2008 is due to the introduction of NACE rev. 2 in BD.
