|
|
For any question on data and metadata, please contact: Eurostat user support |
|
|||
1.1. Contact organisation | Statistics Netherlands |
||
1.2. Contact organisation unit | Demographic and socio-economic statistics – Team labour and wages |
||
1.5. Contact mail address | Statistics Netherlands | Henri Faasdreef 312 | P.O. Box 24500 | 2490 HA The Hague |
|
|||
2.1. Data description | |||
The Labour Cost Survey (LCS) provides comprehensive and detailed information on the level, structure and development of labour costs in the different sectors of economic activity. It covers businesses with at least 10 employees and all economic activities defined in sections B to N, and P to S, of the Statistical classification of economic activities in the European Communities (NACE Rev 2). The transmission of data covering small enterprises (below 10 employees) and enterprises belonging to NACE Rev. 2 section O is optional. |
|||
2.2. Classification system | |||
NACE Rev 2, NUTS, Size of the enterprise. |
|||
2.3. Coverage - sector | |||
Nace (rev.2) sectors B to S are included in the Dutch 2020 Labour Cost Survey (LCS). |
|||
2.4. Statistical concepts and definitions | |||
The framework for compiling the 2020 Labour Costs Survey (LCS) is derived from the Statistics on Employment and Earnings (SEE). The SEE is based upon a complete Register of Earnings and Social contributions (RES) from the Dutch Social Security Organization, containing information about the declaration of earnings and social contributions of all employees in The Netherlands, which originates from the Dutch Tax Authorities. As such, the SEE covers all enterprises with employees in all economic activities throughout the year 2020 in The Netherlands. As the enterprises in the sample are equal to the enterprises in the population, coefficients of variation are 0 (zero), both for annual labour costs (D), as well as for hourly labour costs (D/B1). |
|||
2.5. Statistical unit | |||
|
|||
2.6. Statistical population | |||
All enterprises with employees in all economic activities throughout the year 2020 in The Netherlands. |
|||
2.7. Reference area | |||
The Netherlands. |
|||
2.8. Coverage - Time | |||
The year 2020. |
|||
2.9. Base period | |||
The year 2020. |
|
|||
3.1. Source data | |||
The data for the LCS 2020 were derived from multiple sources. The following sources were used:
The SEE is by far the main source for the LCS 2020. In contrast with the LCS 2016 the statistics have been made more in line with LA. This is illustrated by the following:
|
|||
3.2. Frequency of data collection | |||
The frequency of data collection depends on the source. In this section, only the RES will be further elaborated upon. Both the SEE and LA statistics are based on the RES.
For the RES, with each salary payment, all employers in The Netherlands supply a declaration of earnings and social contributions of their employees to the Dutch Tax Authorities. The declaration does not only concern current employees, but also ex-employees having social benefits or pension. The declaration is redirected to the Dutch Social Security Organization, which is the controller of the RES of all (ex) employees in The Netherlands. Every week, Statistics Netherlands receives information from the RES. The information is used to make the Statistics on Employment and Earnings (SEE). The SEE is produced provisionally every month and definitely every year. LA statistics of National Accounts (NA) are produced every year and are made consistent with the national accounts’ supply-and-use tables. |
|||
3.3. Data collection | |||
See 3.2 Frequency of data collection. |
|||
3.4. Data validation | |||
The whole process of the SEE and the LA, from the RES entering SN, to the making of tables for publication, is accompanied by numerous data checks and corrections. Many control/correction steps are performed fully automatically. For example, the deletion of identical double records, or the imputation of missing information (0,14 % for 2020): for some enterprises information of one or more periods in 2020 is missing. Other control/correction steps are performed in a semi-automatic way (i.e. the possible error is detected automatically, but the solution has to be performed by hand). These checks are mostly concerned with value range (outliers, or non-existing codes) and time-relation (unlikely differences with previous or later periods). The semi-automatic control/correction is based on a top-down analysis system, which automatically brings forward the (possible) errors that have the largest effect on the output-level. In 2020, 9,7 percent of all records from the RES were corrected in some way when making the SEE and LA statistics. An important issue is the deduction of the exact population of employees from the RES. Although the RES is a complete register, a well-defined definition of employees is not evident. Therefore, the number of employees is deducted by using a set of criteria based on several tax codes, the social security or tax wage, and paid hours. Since the quality of (especially) the variable ‘paid hours’ in the RES is not yet optimal (however steadily improving), possible errors in the determination of the number of employees (and consequently in the amount of labour costs) may occur. Another possible error may occur from the fact that enterprise-units in the RES may differ from the statistical units in Statistics Netherlands’ General Business Register (GBR). To ascribe the labour costs results to NACE-groups, the enterprise-units in the RES have to be linked to the statistical units in the GBR. Although most units are linked in a 1:1 ratio, a few hundred enterprises are linked in a 1:n or m:n ratio. Consequently, the distribution of those labour costs to NACE-groups is not certain. It is estimated that at the publication level (NACE 2-digit), at most 2 percent of all labour costs are possibly ascribed to a wrong NACE-group. |
|||
3.5. Data compilation | |||
Also see 3.1 Source data, 3.3 Data collection and 3.4 Data Validation. In this section other sources used than SEE and LA statistics will be considered. As already mentioned before, 93 percent of all labour costs could directly be derived from the SEE and LA statistics. For calculating the other 7 percent several other sources were used (see 3.1 Source data). From these other sources information was collected about the relative volume of missing costs. These missing costs were expressed as a percentage of gross wages and salaries or as costs per employee. With the help of the SEE variables "gross wages and salaries" and "number of employees" the missing cost components in the SEE and LA statistics could be accurately calculated or imputed. For most of the cost components that could not be directly derived from SEE and LA statistics, assumptions had to be made (see table 1). For instance, because the data related to a higher aggregation level than NACE 2-digit × six size classes. Another reason could be because the data originated from another reference year than 2020. In the former case it was assumed that the proportions or amounts could also be applied to lower aggregation levels. In the latter case the data were extrapolated under certain assumptions. Annexes: Table 1 |
|||
3.6. Adjustment | |||
See 3.4 Data Validation and 3.5 Data compilation. |
|
|||
4.1. Quality assurance | |||
See 6. Accuracy and reliability. |
|||
4.2. Quality management - assessment | |||
See 6. Accuracy and reliability. |
|
|||
5.1. Relevance - User Needs | |||
The LCS is an obligation from Eurostat. The data are used to compare European countries. |
|||
5.2. Relevance - User Satisfaction | |||
The Dutch LCS 2020 did not have significant deviations from Eurostat’s implementing arrangements for the LCS. An extensive quality report is made to ensure high user satisfaction. |
|||
5.3. Completeness | |||
The Dutch LCS 2020 did not have significant deviations from Eurostat’s implementing arrangements for the LCS. Also, see table 1 for more information. |
|||
5.3.1. Data completeness - rate | |||
100% complete. |
|
|||
6.1. Accuracy - overall | |||
Overall accuracy is high. Accuracy was further improved compared to LCS 2016, were accuracy was also high. See the subsections below for further elaboration. |
|||
6.2. Sampling error | |||
See 6.2.1. |
|||
6.2.1. Sampling error - indicators | |||
The framework for compiling the 2020 Labour Costs Survey (LCS) is derived from the Statistics on Employment and Earnings (SEE). The SEE is based upon a complete Register of Earnings and Social contributions (RES) from the Dutch Social Security Organization, containing information about the declaration of earnings and social contributions of all employees in The Netherlands, which originates from the Dutch Tax Authorities. As such, the SEE covers all enterprises with employees in all economic activities throughout the year 2020 in The Netherlands. As the enterprises in the sample are equal to the enterprises in the population, coefficients of variation are 0 (zero), both for annual labour costs (D), as well as for hourly labour costs (D/B1). |
|||
6.3. Non-sampling error | |||
See 6.3.1 to 6.3.5. |
|||
6.3.1. Coverage error | |||
The population for the LCS 2020 is derived from the SEE and therefore from the Dutch register of earnings and social contributions (RES). Since this register covers all enterprises with employees in The Netherlands, the whole reference population is covered. It has to be noted that the reference population from the RES is slightly different from the population of Statistics Netherlands (SN) General Business Register (GBR). This difference in population is partially ascribed to administrative effects, since there is a certain time lag between the updating of information in the RES and the GBR. Besides this, not all institutions in The Netherlands are obliged to register at the Chamber of Commerce (for instance religious institutions), which is the main source for the GBR. Consequently, the RES contains enterprises, which are not (yet) known in the GBR. The information in the RES is therefore considered to be more up-to-date and more accurate. The difference is, however, small. |
|||
6.3.1.1. Over-coverage - rate | |||
See 6.3.1. Coverage error. |
|||
6.3.1.2. Common units - proportion | |||
See 6.3.1. Coverage error. |
|||
6.3.2. Measurement error | |||
The data for the LCS 2020 were derived from multiple sources. The following sources were used:
The SEE is by far the main source for the LCS 2020. In contrast with the LCS 2016 the statistics have been made more in line with LA. This is illustrated by the following:
Measurement and processing errors in SEE With each salary payment, all employers in The Netherlands supply a declaration of earnings and social contributions of their employees to the Dutch Tax Authorities. The declaration does not only concern current employees, but also ex-employees having social benefits or pension. The declaration is redirected to the Dutch Social Security Organization, which is the controller of the RES of all (ex) employees in The Netherlands. Every week, Statistics Netherlands receives information from the RES. The information is used to make the Statistics on Employment and Earnings (SEE). The whole process of the SEE, from the RES entering SN, to the making of tables for publication, is accompanied by numerous data checks and corrections. Many control/correction steps are performed fully automatically. For example, the deletion of identical double records, or the imputation of missing information: for some enterprises information of one or more periods in 2020 is missing. Other control/correction steps are performed in a semi-automatic way (i.e. the possible error is detected automatically, but the solution has to be performed by hand). These checks are mostly concerned with value range (outliers, or non-existing codes) and time-relation (unlikely differences with previous or later periods). The semi-automatic control/correction is based on a top-down analysis system, which automatically brings forward the (possible) errors that have the largest effect on the output-level. In 2020 9,7 percent of all records from the RES were corrected in some way when making the SEE. An important issue is the deduction of the exact population of employees from the RES. Although the RES is a complete register, a well-defined definition of employees is not evident. Therefore, the number of employees is deducted by using a set of criteria based on several tax codes, the social security or tax wage, and paid hours. Since the quality of (especially) the variable ‘paid hours’ in the RES is not yet optimal (however steadily improving), possible errors in the determination of the number of employees (and consequently in the amount of labour costs) may occur. Another possible error may occur from the fact that enterprise-units in the RES may differ from the statistical units in Statistics Netherlands’ General Business Register (GBR). To ascribe the labour costs results to NACE-groups, the enterprise-units in the RES have to be linked to the statistical units in the GBR. Although most units are linked in a 1:1 ratio, a few hundred enterprises are linked in a 1:n or m:n ratio. Consequently, the distribution of those labour costs to NACE-groups is not certain. It is estimated that at the publication level (NACE 2-digit), at most 2 percent of all labour costs are possibly ascribed to a wrong NACE-group.
Variables For LCS 2020 no variables could not be properly estimated due to measurement and processing errors.
|
|||
6.3.3. Non response error | |||
See 6.3.3.1 and 6.3.3.2. |
|||
6.3.3.1. Unit non-response - rate | |||
Unit response rate The 2020 SEE contains almost 118 million records of individual employees (1 record per employee for each month or 4-week period). Since occasionally some employee-records are missing for one or more periods, these records are imputed. The imputation-rate is rather small: in 2020 about 165 thousand (0.14 percent) employee-records were imputed. |
|||
6.3.3.2. Item non-response - rate | |||
Due to practical reasons, the description of non-response errors in this section will be restricted to the (full) 2020 SEE.
Item imputation rate About 9,7 per cent of all 118 million employee-records in the 2020 SEE were corrected in some way for missing or wrong data in one or more variables. Imputations of variables took place by using (a) the same variable (or a deduction from that variable) of the same employee of another period, (b) other variables of the same employee, (c) average values from other employees in the same enterprise or section. Within the individual employee-records, several variables contribute to the total annual labour costs (D). |
|||
6.3.4. Processing error | |||
See 6.3.3 Non response error. |
|||
6.3.4.1. Imputation - rate | |||
See 6.3.3 Non response error. |
|||
6.3.5. Model assumption error | |||
See 3.5 data compilation. |
|||
6.4. Seasonal adjustment | |||
No seasonal adjustment applied. |
|||
6.5. Data revision - policy | |||
The 2020 data are definitive. No revision will occur, unless major inaccuracies would be uncovered. |
|||
6.6. Data revision - practice | |||
See 6.6.1. |
|||
6.6.1. Data revision - average size | |||
See 6.5 Data revision – policy. |
|
|||
7.1. Timeliness | |||
See 7.1.1. and 7.1.2. |
|||
7.1.1. Time lag - first result | |||
The data tables for Eurostat were prepared, including confidentiality checks, and sent to Eurostat via Edamis on 28 June 2022, two days before the original deadline. Due to blocking issues a second delivery of Table A, B and C on July 12th was necessary. A third delivery of Table B and C on July 21th was necessary because another method of regionalization was preferred. The results of the Dutch LCS 2020 were published on 15 July 2022 on StatLine, the electronic databank of SN. |
|||
7.1.2. Time lag - final result | |||
Due to blocking issues a second delivery of Table A, B and C on July 12th 2022 was necessary. A third delivery of Table B and C on July 21th was necessary because another method of regionalization was preferred. The results of the Dutch LCS 2020 were published on 15 July 2022 on StatLine, the electronic databank of SN. |
|||
7.2. Punctuality | |||
See 7.2.1. |
|||
7.2.1. Punctuality - delivery and publication | |||
For compiling the LCS 2020 figures, Statistics Netherlands used a methodology in which, by using existing sources, it was not necessary to conduct a separate survey on labour costs. This methodology was developed for the 2000 LCS and is used since then. The first preparations of the LCS 2020 started in June 2020 with a preliminary investigation. This investigation was aimed at examining the use of all possible existing sources, both inside and outside Statistics Netherlands. The aim was to use the same sources as the Labour Acounts (LA), part of National Accounts (NA), were possible. As compared to the 2016 LCS, new sources of LA had to be used. Up from June 2020 until December 2020 those several sources were investigated and processed. This work was time-consuming because many basic data had to undergo further preparations. Eurostat’s deadline of 30 June 2022 was met. However, blocking issues and another preferred method resulted in a resubmission on 12 July 2022 and 15 July 2022. Also see 7.1.1 Time-lag first result. |
|
|||
8.1. Comparability - geographical | |||
See 8.1.1. |
|||
8.1.1. Asymmetry for mirror flow statistics - coefficient | |||
This section cannot fully be described. For the difference or the absolute difference of inbound and outbound flows between a pair of countries divided by the average of these two values, please refer to Eurostat, as Statistics Netherlands does not have this information.
For the Netherlands there are no differences between the national classifications and the Eurostat classifications. However the information about units for the Netherlands does not refer to local units, but to kind of activity units. By using the commuter statistics we have corrected for this. |
|||
8.2. Comparability - over time | |||
See 8.2.1. |
|||
8.2.1. Length of comparable time series | |||
The LCS data of the available can be found on the online Eurostat database. These years are largely comparable over time, however it needs to be noted that (re)sources and methods to compile the LCS have changed slightly over time. Trend breaks between two following vintages have been analyzed, but trend breaks over longer periods have not. Compared to the LCS 2016, there have been some changes in methods:
- The Dutch LCS is now, where possible, fully consistent with the Dutch Labour Accounts (LA). Only for some topics the regulations for LCS and LA deviate. This process yielded several small, neglectable changes in the output of the LCS. However, a couple of mutations were more substantial. Those were:
- In C13 (Apprentices’ paid hours) an error was corrected. For the LCS 2016, overtime hours of all part time employees were added to C13. In LCS 2020 only overtime hours of apprentices were added. Indication: C13 in LCS 2016 would go from 345 mln. hours to 242 mln. hours (-30%).
|
|||
8.3. Coherence - cross domain | |||
LCS vs. Labour Cost Index LCS2020/2016 growth rates of the hourly labour costs are compared, for all (part-time and full-time) employees including apprentices, working in an enterprise of 10 employee or more and belonging to NACE rev,2 sections B to S except O, with the corresponding 2020/2016 growth rates of the annual LCI (dataset: lc_lci_r2_a in Eurobase, variable ‘Labour cost for LCI’, NACE B to S).
Differences in growth rates between the two domains do exist, although differences between LCI and LCS are small in most sections (see table 2). The difference in growth rate in total is 1 percentage point. However, only for sector I - Accommodation and food service activities and sector R - Arts, entertainment and recreation we see a higher difference between the growth rates. These differences can be explained by the following factors: - The LCI accounts for structural effects (changes in the population over time), the LCS does not. - The LCS does not include enterprises with size class < 10 employees, the LCI includes all size classes. - The LCS used for D5 a definitive version of the register data concerning the subsidy given to companies during the covid-19 pandemic. In contrast the LCI used a preliminary version of the same data. This accounts for the majority of the differences shown in table 2 below.
LCS vs. LFS The 2020/2016 growth rate in the number of employees taken from LCS, for all (part-time and full-time) employees including apprentices, working in an enterprise of 10 employee or more and belonging to NACE rev.2 sections B to S except O is compared with the corresponding 2020/2016 growth rate from LFS (dataset: lfsa_eegaed in Eurobase, variable:‘ Total number of employees’, all ISCED levels, age 15 to 64).
Furthermore, the 2020/2016 growth rate in the number of hours worked taken from LCS for all (part-time and full-time) employees including apprentices, working in an enterprise of 10 employee or more and belonging to NACE rev.2 sections B to S except O is compared with the corresponding 2020/2016 growth rate from LFS (dataset: lfsa_ewhun2 in Eurobase, variable: ‘average number of usual weekly hours of work in main job’, employees, total worktime, all NACE sections).
Table 3 shows a difference of 1 percentage points for the growth factors of the total number of employees in the Netherlands. For the number of hours worked, this is 3 percentage points (see table 4). The highest absolute difference on sector level depicted in table 4 is 18 percentage points (sector B - Mining and quarrying). The differences in the growth factors between the LCS and the LFS can be explained by the differences in the methodologies of the two domains: - Different sources:
- Different populations:
LCS vs. Structural Business Statistics The 2020 level and the 2020/2016 growth rate in the number of employees taken from LCS, for all (part-time and full-time) employees including apprentices, working in an enterprise of 10 employee or more and belonging to NACE rev.2 sections B to N is compared with the corresponding SBS data (dataset: sbs_na_sca_r2 in Eurobase, variable: ‘Employees-number’, NACE sections B to N except K). See table 5.
Both LCS 2020 levels and LCS2020/2016 growth rates of the total labour costs should be compared, for all (part-time and full-time) employees including apprentices, working in an enterprise of 10 employee or more and belonging to NACE rev.2 sections B to N with the corresponding SBS data (dataset: sbs_na_sca_r2 in Eurobase, variable ‘Personnel costs’, NACE sections B to N except K). See table 6.
As can be seen in table 5, the differences between LCS and SBS in the growth rate of the number of employees are zero in almost every sector. Only in the sector C - Manufacturing and M – Professional and the total there is a small difference.
As can be seen in table 6, the differences between LCS and SBS in the growth rate of the total labour costs are quite small in certain sectors. However, in several other sectors (E, I , N), the difference seems quite substantial. These differences in growth rates observed can be explained by the following factors: - In the LCS, both the number of employees and the wages and salaries are derived from the same source (the SEW), which is considered to be complete (since it is based on a complete register with all enterprise and employees). The SBS also makes use of the SEW for the number of employees, but to derive the wages and salaries, the SBS makes use of a different sample survey based source (Production Statistics). - Section C has fewer employees in the SBS data (table 5) and also a lower total of labour costs (table 6), the wages and salaries per employee are higher. This can be explained by the exclusion of NACE 32991 (Sheltered workshops) in the SBS data. Employees in NACE 32991 will earn not more than minimum wage. This lowers the average wage of Section C in the LCS. - For section E specifically: This section contains a lot of employers that are non-market producers. The Production Statistics and therefore the SBS, exclude these non-market producers when counting the wages. The SEW and therefore the LCS, do not exclude these employers when counting the wages. - The SBS does not subtract subsidies from the labour costs, while the LCS does. During the covid-19 pandemic in 2020, there was a substantial subsidy present. This results in bigger differences between the LCS and SBS, especially for sections there are more effected by covid-19. - The section N includes temporary employment agencies. In the LCS all employees working for temporary employment agencies are assigned, together with their wages, to section N. In the SBS however, due to partial usage of the Production Statistics, for a part of the working employees for temporary employment agencies, the wages are assigned to other sections. - The exclusion of NACE 70101 results in a lower number of employees in the Section M of the SBS data, compared to LCS data.
Annexes: Tables 2 to 7 |
|||
8.4. Coherence - sub annual and annual statistics | |||
There are no annual nor sub annual versions of the LCS 2020. |
|||
8.5. Coherence - National Accounts | |||
Both LCS 2020 levels and LCS2020/2016 growth rates of the hourly labour costs, for all (part-time and full-time) employees including apprentices, working in an enterprise of 10 employee or more and belonging to NACE rev.2 sections B to S except O with the corresponding NA data (dataset: nama_10_lp_ulc in Eurobase, variable ‘Compensation of employees per hours worked, in euros).
Differences between NA and LCS are small in all sections (see table 7). For the whole economy (sections B to S except O) the difference is less than .09 percent. In half of the sections, hourly labour costs are a bit lower in NA. In other sections, hourly labour costs are a bit higher in NA. The results between the NA and the LCS are so similar because, where possible, the same sources and definitions are used. Differences mainly occur due to small deviations in regulations, timing issues and because the Labour Accounts are made consistent within the whole of the National Accounts.
Annexes: Tables 2 to 7 |
|||
8.6. Coherence - internal | |||
The LCS 2020 was checked extensively for internal coherence. Eurostat provides a broad set of internal consistency rules to be checked before the data is transmitted to Eurostat and which are also checked by Eurostat after transmission. Additionally, SN conducted an additional set of internal consistency checks. (Sub) totals of the LCS 2020 are consistent with other Dutch labour and earnings statistics based on the RES. For instance, see https://opendata.cbs.nl/#/CBS/en/dataset/81431ENG/table?ts=1545396406733 . |
|
|||
9.1. Dissemination format - News release | |||
No Dutch news release was published concerning the LCS 2020. |
|||
9.2. Dissemination format - Publications | |||
On August 6th 2022, a small article was published on the journal ESB. See URL: https://esb.nu/kort/20071401/groter-dempend-effect-van-coronasteun-op-arbeidskosten-in-2020. |
|||
9.3. Dissemination format - online database | |||
See 9.3.1. |
|||
9.3.1. Data tables - consultations | |||
Detailed results of the Dutch LCS 2020 are accessible via StatLine, the electronic databank of SN. The data are presented in four tables: http://opendata.cbs.nl/statline/#/CBS/nl/dataset/85292NED/table?ts=1659706671629 http://opendata.cbs.nl/statline/#/CBS/nl/dataset/85293NED/table?ts=1659706703412 http://opendata.cbs.nl/statline/#/CBS/nl/dataset/85291NED/table?ts=1659706714017 http://opendata.cbs.nl/statline/#/CBS/nl/dataset/84155NED/table?ts=1659706726295 All the information on this databank may be consulted, printed out and downloaded free of charge. |
|||
9.4. Dissemination format - microdata access | |||
There is no standardized free accessible micro data access for the LCS 2020. Contact Statistics Netherlands for the possibilities. |
|||
9.5. Dissemination format - other | |||
Contact Statistics Netherlands for the possibilities. |
|||
9.6. Documentation on methodology | |||
Concomitant with the results on Statline, a description of metadata is given (in Dutch). See the URL’s: http://opendata.cbs.nl/statline/#/CBS/nl/dataset/85292NED/table?ts=1659706671629 http://opendata.cbs.nl/statline/#/CBS/nl/dataset/85293NED/table?ts=1659706703412 http://opendata.cbs.nl/statline/#/CBS/nl/dataset/85291NED/table?ts=1659706714017 http://opendata.cbs.nl/statline/#/CBS/nl/dataset/84155NED/table?ts=1659706726295 |
|||
9.7. Quality management - documentation | |||
If there are questions left after consulting this web page, please contact Statistics Netherlands. |
|||
9.7.1. Metadata completeness - rate | |||
100% |
|||
9.7.2. Metadata - consultations | |||
This information cannot be provided. |
|
|||
The manpower needed by SN to create the LCS 2020 was 3,5 FTE (full time full year, including a post for technical maintenance). The Dutch LCS 2020 did not impose an additional burden on respondents, as the research re-used data that was already available to SN. |
|
|||
In accordance to Eurostat regulations, primary and secondary confidentiality checks and flags were applied. |
|||
11.1. Confidentiality - policy | |||
In accordance to Eurostat regulations, primary and secondary confidentiality checks and flags were applied. Confidential flagging shall allow that the confidential data should not be suppressed in the data file sent to Eurostat but should be transmitted with a flag, so that the flagged data can be used for calculations. Flagged data will not be published. Member States are required to carry out both primary and secondary confidentiality flagging of the national LCS tables (A, B and C) to be delivered to Eurostat. The individual records transmitted for Tables A, B and C consist of grossed up data. The risk of confidentiality can arise in Tables A, B and C when the number of enterprises or local units in the population is tiny for an individual record relating to a given economic activity, size class or region. Clearly, the risks are greater when the individual record relates to one or two large units. Likewise, the confidentiality risk can be higher for Tables B or C because of the additional breakdown by size class and region, respectively. Whereas primary confidential flagging seems to represent no major problem, secondary confidential flagging is a slightly more complex issue. Therefore, see the following short example for primary and secondary confidential flagging. Secondary confidentiality enters into play when a value for an aggregate is asked to be published without disclosing a value of its confidential element(s). |
|||
11.2. Confidentiality - data treatment | |||
See section 11.1. |
|
|||
No further comments. |
|
|||
|
|||
Table 1 Tables 2 to 7 |