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 Federal Statistical Office (FSO) started publishing estimates of labour costs in 2006 (reference years 2000, 2002 and 2004) and subsequently every two years until 2012 (reference years 2006, 2008 and 2010). These estimates are largely based on data collected for the Swiss Earnings Structure Survey (SESS), which is carried out every two years. Estimates of labour costs for the years 2000 to 2010 are available on the website of the Federal Statistical Office (https://www.bfs.admin.ch/bfs/en/home/statistics/work-income/wages-income-employment- labour-costs/labour-costs.html).
In accordance with the bilateral agreement on cooperation between the Swiss Confederation and the EU in the field of statistics which came into force on 1 January 2007, the LCS 2012 and subsequent was conducted in compliance with the definitions fixed in the European regulations (Council Regulation (EC) No 530/1999 of 9 March 1999 and the Commission Regulation (EC) No 1737/2005 of 21 October 2005). The adoption of the new definitions was made possible thanks to the revision of the Swiss Earnings Structure Survey (SESS), which remains the main source for data in the structural statistics on labour costs (LCS). The main changes made to the SESS which are relevant to the LCS concern the monetary variables (more detailed breakdown of wage components) and the population concerned (inclusion of apprentices).
Both the Swiss Earnings Structure Survey (SESS) and the structural statistics on labour costs (LCS) take place every two years.
2.2. Classification system
Classifications used: NACE Rev2
2.3. Coverage - sector
Private and public sectors, sections B to S of the NACE Rev.2.
LCS includes enterprises (and public administrations) with three or more employees.
2.4. Statistical concepts and definitions
Labour costs correspond to the total expenditure incurred by employers when employing labour. They include wages and salaries (D.11), social contributions paid by the employer (D.12), as well as vocational training costs (D.2) and other expenditure paid by the employer (D.3). Taxes (D.4) and subsidies received by the employer (D.5) are not pertinent for Switzerland. The definitions in Commission Regulation (EC) No 1737/2005 are taken into account.
2.5. Statistical unit
The statistical unit is the enterprise.
2.6. Statistical population
LCS applies to all activities of sections B to S of NACE Rev.2 and includes enterprises (and public administrations) of the private and public sectors with three or more employees. The survey is based on a sample extracted from the Swiss Business and Enterprise Register.
2.7. Reference area
Switzerland.
2.8. Coverage - Time
The LCS is released every second year. The transmission to Eurostat should take place every fourth year within 18 months of the reference year (2012, 2016, 2020).
A.1 "Total number of employees", C.1 "Total hours paid" and D.1 "Compensation of employees"), other FSO statistics are used in the calculations (see also chapter 2.2.4). Other sources of the LCS are:
-the Survey on continuing vocational training SBW https://www.bfs.admin.ch/bfs/fr/home/statistiques/education-science/enquetes/sbw.html and a specific study on the cost of apprentices based on a representative survey conducted in enterprises (« Die duale Lehre: eine Erfolgsgeschichte – auch für die Betriebe. Ergebnisse der dritten Kosten- Nutzen-Erhebung der Lehrlingsausbildung aus der Sicht der Betriebe », Miriam Strupler et Stefan
C. Wolter, 2012)
Sampling method for the Swiss Earnings Structure Survey (SESS)
The SESS is based on a random sample at two levels: the first level concerns the stratified enterprises, while the second level concerns employees in the enterprises.
Private sector
The SESS 2020 is based on a sample stratified into two levels. The first level concerns the enterprises, which are grouped according to three criteria: enterprise size (3 classes), economic branch (39 aggregations of NACE Rev. 2 divisions) and geographical entity (7 major regions, 8 cantons and 1 city). The stratification in the private sector comprises some 1600 cells (strata). For each strata, a sample of enterprises is chosen at random. This sample is drawn from the Business and Enterprise Register (BER). It should be noted that large businesses (50 employees or more) and enterprises working in very specialised sectors are always included in the survey for reasons of representativeness. The sample of the SESS 2020 included about 45 000 enterprises. The second level concerns the employees of the selected enterprises: the number of wage data that has to be provided depends on the size of the enterprise (enterprises with 1-19 employees: wage data must be supplied for all employees; 20-49 employees: data for one in two employees; 50 employees or more: one in three). The selection of employees must be carried out randomly by responding units (it should be noted that many enterprises and administrations, and particularly the larger ones, do not make a selection and provide data on all their employees).
Public sector
In the public sector of the Confederation and the cantons, the survey is comprehensive at the level of the enterprises and administrations surveyed. In the public sector at commune level, the survey is also based on a random sample at two levels: the commune and the employees. Communes are stratified according to size (4 sizes) and geographical unit (7 major regions). In total, the stratification of the SESS at commune level is comprised of 27 cells (strata). For each stratum, a random sample of communes is drawn. The gross commune sample is made up of roughly 300 communes although the survey is comprehensive for communes with more than 800 employees.
3.2. Frequency of data collection
The data is collected every second year
3.3. Data collection
After receiving the questionnaire of the SESS, enterprises and public administrations have submitted their data to the NSO by using one of the following means:
paper questionnaire,
electronic questionnaire (XML or Excel),
electronic questionnaire on internet (eSurvey) or
data sent automatically by means of a certified payroll accounting software (standard salaire CH swissdec - ELM).
3.4. Data validation
A range of plausibility checks were used to identify errors. Respondents were re-contacted if necessary to clarify and correct the figures provided.
3.5. Data compilation
Not applicable
3.6. Adjustment
Not applicable
4.1. Quality assurance
The concepts and classifications of the LCS are defined in the European regulations (Council Regulation (EC) No 530/1999 of 9 March 1999 and the Commission Regulation (EC) No 1737/2005 of 21 October 2005).
4.2. Quality management - assessment
The quality report is delivered to Eurostat
5.1. Relevance - User Needs
The main users of the LCS are government bodies, trade unions, employers’ associations, international organisations (Eurostat, International Labour Organization ILO), courts, researchers, students, mass media and the public.
5.2. Relevance - User Satisfaction
No specific survey on user satisfaction has been carried out
5.3. Completeness
All requested labour cost variables and breakdowns according to the LCS regulation 1735/2005 are provided.
5.3.1. Data completeness - rate
-
6.1. Accuracy - overall
-
6.2. Sampling error
-
6.2.1. Sampling error - indicators
LCS 2020: coefficient of variation (%) by NACE REV. 2 sections, size class 10+
NACE REV. 2 sections labour labour cost cost
Hourly labour cost (D/B1)
Annual labour cost (D)
BS
Total
0.38
0.44
B
Mining
1.37
1.16
C
Manufacturing
1.23
1.23
D
Production and distribution of energy
0.81
0.93
E
Water supply, sewerage; waste management
1.27
1.33
F
Construction
0.57
0.71
G
Wholesale and retail trade; repair of motor vehicles and motorcycles
1.39
1.73
H
Transport and warehousing
1.05
1.44
I
Accommodation and food service activities
0.41
0.94
J
Information and communication
2.19
2.29
K
Financial and insurance activities
0.7
0.79
L
Real estate activities
0.72
1.1
M
Professional, scientific and technical activities
0.91
1.01
N
Administrative and support service activities
0.85
1.32
O
Public administration, defence, social security
0.47
0.79
P
Education
0.41
0.66
Q
Human health and social work activities
0.52
0.72
R
Arts, entertainment and recreation
1.36
2.4
S
Other service activities
0.62
0.83
LCS 2020: coefficient of variation (%) by size class, NACE Rev. 2 B to S
Size class
Hourly labour cost (D/B1)
Annual labour cost (D)
Total (1+)
0.35
0.41
Total (10+)
0.38
0.44
1 - 9
0.72
0.81
10 - 49
0.57
0.65
50 - 249
0.23
0.29
250 - 499
0.5
0.59
500 - 999
0.56
0.68
1000+
0.35
0.43
6.3. Non-sampling error
-
6.3.1. Coverage error
Over-coverage: no over-coverage detected
Under-coverage:
Economic activity: no under-coverage
Size of enterprises: only enterprises with three or more employees were taken into consideration for the sampling. The number of employees working in enterprises belonging to the size 1 to 9 (E1_9), indicated in the LACOST_B_A4 table, has therefore been slightly underestimated (as well as, of course, the volume of hours worked and paid and the total labour costs). In comparison with the number of employees in the total economy (sections B to S of the NACE REV. 2), the under-coverage is estimated to be approximately 8%. (estimation source: Swiss Business and Enterprise Register). The enterprises surveyed who stated that they employed fewer than three employees during the month of October 2020 (reference period for the SESS) were, however, not excluded from the survey.
6.3.1.1. Over-coverage - rate
-
6.3.1.2. Common units - proportion
-
6.3.2. Measurement error
Definitions: in order to limit measuring errors, the SESS questionnaire has adopted the terms and definitions normally used in payroll accounting and/or in the wage certificate intended for tax administrations.
Electronic questionnaire (XML or Excel), electronic questionnaire on internet (eSurvey) and data sent automatically by means of certified payroll accounting software: before validation the data transmitted is submitted to automatic plausibility checks and logical controls. Paper questionnaires are checked by qualified statworkers. Missing or implausible values are added or corrected after consultation with the respondents. Before final validation, the data are subjected to a second series of checks carried out by specialised research associates.
During the course of the survey, an information service (hotline) was made available to respondents to help them with questions.
6.3.3. Non response error
-
6.3.3.1. Unit non-response - rate
The gross unit response rate for the SESS 2020 was 74%. Some 35 000 private (and public) enterprises as well as public administrations took part in the survey: data on approximately 2.1 million employees were collected and processed.
Non-response is mainly due to business closures and mergers. Furthermore, some enterprises responded too late and/or the quality of data was too poor to be exploited.
6.3.3.2. Item non-response - rate
-
6.3.4. Processing error
-
6.3.4.1. Imputation - rate
-
6.3.5. Model assumption error
Variables A, B, C and D
The values of the variables A. (number of employees), B. (hours actually worked), C. (hours paid) and D. (labour costs) were extrapolated from data provided by enterprises as part of the SESS. The reference period for this statistic is not the whole year but the month of October, considered to be representative. The result of this is that:
- the total number of employees and apprentices (A.11, A.12 and A.13) corresponds to the number of employees and apprentices with an employment contract in the month of October 2020 and who were paid for services rendered during that month,
- the number of hours paid and worked (B.11, B.12, B.13, C.11, C.12 and C.13) was calculated by annualising the hours paid and worked during the month of October 2020,
- the compensation of employees (D.1) was also extrapolated from SESS data. One part of the monetary data collected refers to the month of October (D.11111, D.121), the other (D.11112, D.1114 and D.1223) to the period during which the employee was employed by the enterprise during the reference year (information which is also collected as part of the survey). In both cases, the values were annualised in order to guarantee coherence with the variables B and C.
The consequence of proceeding in this way is that the values calculated for these variables are probably slightly over-estimated. The reasons for deciding to accept this deviation from the definitions set out in the regulation are as follows:
- the main indicator in the LCS is the hourly labour cost. Employment, volume of work and total employees' wages are covered by other statistics;
- the SESS is the only business survey which collects, at individual level, the wages and salaries (broken down into all wage components) and the corresponding working time (as well as work- time percentage). The coherence between numerator and denominator, which is decisive for an indicator's quality, is also guaranteed.
Variables B.11, B.12 and B.13 (Hours worked by full-time employees, part-time employees and apprentices)
In contrast to hours paid, the hours actually worked are not collected by the SESS: as this information is not recorded in the payroll accounting, the administrative burden on respondents would be too great. Hours worked, therefore, have been estimated on the basis of hours paid provided by the SESS and other statistical sources, using the following formula:
Hours worked = hours paid - holidays - public holidays - sick leave - absence due to other reasons (strikes, bad weather, etc.)
Sources:
- hours paid (variable C. of the LCS): Swiss Earnings Structure Survey (SESS),
- public holidays: official public holidays (by canton),
- sick leave and absences due to other reasons: Work Volume Statistic (WV). As a breakdown of days of absence by section or division of the NACE Rev. 2 (and by size of enterprise) is not available, the same value was used for all employees.
The hours worked by apprentices were calculated in the same manner, but also by deducting hours of education and training from the hours paid. According to our estimates (hours of training and education are not collected in the SESS), these correspond to an average of 30% of paid hours.
Due to a lack of additional sources, it was not possible to deduct from the hours paid, for each apprentice, the exact time spent on training and education (this varies depending on the type of apprenticeship, the year of apprenticeship and the canton where training and education take place).
Variable D.1221 (Guaranteed remuneration in the event of sickness)
This variable is not collected as part of the SESS and had to be estimated by using data (variable: annual length of sick leave) from the Work Volume Statistics (WV). As a breakdown of days of absence by section or division of the NACE Rev. 2 and by size of enterprise is not available, the same value was used for all employees.
Variable D.2 (Vocational training costs)
The vocational training costs comprise the costs for the education and training of apprentices as well as costs paid by employers for the continuing professional education and training of their employees.
As the SESS does not collect this information, the variable D.2 had to be estimated. Sources:
the Survey on continuing vocational training in companies (SBW) to estimate the costs of the continuing professional training of employees
As data broken down by division or section of the NACE Rev. 2 and by size of enterprise were not available, an average value was used for all employees and all apprentices.
6.4. Seasonal adjustment
-
6.5. Data revision - policy
-
6.6. Data revision - practice
-
6.6.1. Data revision - average size
-
7.1. Timeliness
The results of the LCS 2020 were disseminated at national level on 28 March 2023, 29 months after the end of the reference period
7.1.1. Time lag - first result
-
7.1.2. Time lag - final result
-
7.2. Punctuality
-
7.2.1. Punctuality - delivery and publication
Main stages of the LCS 2020
Date
Action
July/August 2020
Sampling procedure for SESS 2020
January 2021
Send questionnaires
End of March 2021
Deadline for answering
Mid-April 2021
1st reminder
Mid-May 2021
2nd reminder
End of November 2021
End of data collection and plausibilisation of data
October 2021 – January 2022
Weighting and analysis of data
28 March 2022
Publication of national results of SESS 2020
End of March 2023
LCS 2020 (LACOST_A_A4 and LACOST_B_A4 tables) sent to Eurostat
28 March 2023
Publication of national results from the LCS 2020
8.1. Comparability - geographical
National concepts have been defined as closely as possible in line with European concepts
8.1.1. Asymmetry for mirror flow statistics - coefficient
-
8.2. Comparability - over time
The LCS 2020 is the fifth survey (after 2012, 2014, 2016 and 2018) that has been conducted in compliance with the definitions fixed in the European regulations (Council Regulation (EC) No 530/1999 of 9 March 1999 and the Commission Regulation (EC) No 1737/2005 of 21 October 2005).
Comparison of the number of hours actually worked (in thousands) by employees (B.11 + B.12) and apprentices (B.13) with the SLFS results, reference year 2020
LCS
SLFS
Difference in %
LCS
SLFS
Difference in %
B11+12
Volume of work
(LCS-SLFS) / SLFS
B13
Volume of work for apprentices
(LCS-SLFS) / SLFS
XB_E
1'062'151'165
1'083'707'024
-2.0%
29'076'113
41'380'892
-29.7%
XF
480'281'644
507'200'396
-5.3%
29'092'406
42'378'241
-31.4%
XG
752'271'989
773'490'043
-2.7%
34'666'624
44'910'021
-22.8%
XH
348'333'153
359'667'532
-3.2%
5'944'161
9'550'706
-37.8%
XI
247'520'887
208'776'788
18.6%
5'675'276
7'218'047
-21.4%
XJ
236'520'718
246'245'537
-3.9%
3'920'925
5'359'370
-26.8%
XK
336'615'964
372'686'142
-9.7%
6'432'518
6'984'541
-7.9%
XLN
423'207'222
385'768'391
9.7%
5'440'757
9'356'173
-41.8%
XM
502'523'580
523'277'568
-4.0%
13'919'245
20'523'471
-32.2%
XO
333'689'133
304'300'349
9.7%
8'810'149
9'266'068
-4.9%
XP
360'904'305
470'402'558
-23.3%
7'423'510
11'170'206
-33.5%
XQ
790'746'759
880'141'710
-10.2%
40'473'305
51'760'087
-21.8%
XRS
188'824'743
283'607'576
-33.4%
5'631'517
7'496'471
-24.9%
XBSO
6'063'591'262
6'401'144'237
-5.3%
196'506'506
267'354'295
-26.5%
8.4. Coherence - sub annual and annual statistics
-
8.5. Coherence - National Accounts
Comparison of total labour costs (D.1, D.11 and D.12) with the figures resulting from the National accounts (NA), in millions of francs, reference year 2020
Employees' wages
Wages and pay
Social contributions paid by employers
D.1
D.11
D.12
LCS 2020
386’822.16
317’628.91
69’193.25
NA 2020
414’846.28
344'622.92
70’224.35
Difference in % (LCS-NA) / NA
-6.76%
-7.83%
1.47%
As for the LCS – SLFS comparison, there is no in-depth analysis aiming to explain these differences. The main causes are probably the following:
the different data sources: the NA is a composite economic statistic (the variable D1 is based on data from the OASI social insurance and other administrative sources), the LCS is created from a survey carried out among enterprises,
the LCS does not take into account enterprises with fewer than 3 employees,
the LCS extrapolates the job level and the volume of work for the reference year from data pertaining to the month of October (see also chapter 2.2.4) and
the NA figures also include the primary sector (divisions 1 to 3 of the 2nd NACE revision).
8.6. Coherence - internal
-
9.1. Dissemination format - News release
An update of the tables on LCS 2020 was published in March 2023
The application of the FStatA is guaranteed by the fact that no publication for an aggregated number of fewer than 60 employees or 5 enterprises is possible.
-
The Federal Statistical Office (FSO) started publishing estimates of labour costs in 2006 (reference years 2000, 2002 and 2004) and subsequently every two years until 2012 (reference years 2006, 2008 and 2010). These estimates are largely based on data collected for the Swiss Earnings Structure Survey (SESS), which is carried out every two years. Estimates of labour costs for the years 2000 to 2010 are available on the website of the Federal Statistical Office (https://www.bfs.admin.ch/bfs/en/home/statistics/work-income/wages-income-employment- labour-costs/labour-costs.html).
In accordance with the bilateral agreement on cooperation between the Swiss Confederation and the EU in the field of statistics which came into force on 1 January 2007, the LCS 2012 and subsequent was conducted in compliance with the definitions fixed in the European regulations (Council Regulation (EC) No 530/1999 of 9 March 1999 and the Commission Regulation (EC) No 1737/2005 of 21 October 2005). The adoption of the new definitions was made possible thanks to the revision of the Swiss Earnings Structure Survey (SESS), which remains the main source for data in the structural statistics on labour costs (LCS). The main changes made to the SESS which are relevant to the LCS concern the monetary variables (more detailed breakdown of wage components) and the population concerned (inclusion of apprentices).
Both the Swiss Earnings Structure Survey (SESS) and the structural statistics on labour costs (LCS) take place every two years.
Not Applicable
Labour costs correspond to the total expenditure incurred by employers when employing labour. They include wages and salaries (D.11), social contributions paid by the employer (D.12), as well as vocational training costs (D.2) and other expenditure paid by the employer (D.3). Taxes (D.4) and subsidies received by the employer (D.5) are not pertinent for Switzerland. The definitions in Commission Regulation (EC) No 1737/2005 are taken into account.
The statistical unit is the enterprise.
LCS applies to all activities of sections B to S of NACE Rev.2 and includes enterprises (and public administrations) of the private and public sectors with three or more employees. The survey is based on a sample extracted from the Swiss Business and Enterprise Register.
A.1 "Total number of employees", C.1 "Total hours paid" and D.1 "Compensation of employees"), other FSO statistics are used in the calculations (see also chapter 2.2.4). Other sources of the LCS are:
-the Survey on continuing vocational training SBW https://www.bfs.admin.ch/bfs/fr/home/statistiques/education-science/enquetes/sbw.html and a specific study on the cost of apprentices based on a representative survey conducted in enterprises (« Die duale Lehre: eine Erfolgsgeschichte – auch für die Betriebe. Ergebnisse der dritten Kosten- Nutzen-Erhebung der Lehrlingsausbildung aus der Sicht der Betriebe », Miriam Strupler et Stefan
C. Wolter, 2012)
Sampling method for the Swiss Earnings Structure Survey (SESS)
The SESS is based on a random sample at two levels: the first level concerns the stratified enterprises, while the second level concerns employees in the enterprises.
Private sector
The SESS 2020 is based on a sample stratified into two levels. The first level concerns the enterprises, which are grouped according to three criteria: enterprise size (3 classes), economic branch (39 aggregations of NACE Rev. 2 divisions) and geographical entity (7 major regions, 8 cantons and 1 city). The stratification in the private sector comprises some 1600 cells (strata). For each strata, a sample of enterprises is chosen at random. This sample is drawn from the Business and Enterprise Register (BER). It should be noted that large businesses (50 employees or more) and enterprises working in very specialised sectors are always included in the survey for reasons of representativeness. The sample of the SESS 2020 included about 45 000 enterprises. The second level concerns the employees of the selected enterprises: the number of wage data that has to be provided depends on the size of the enterprise (enterprises with 1-19 employees: wage data must be supplied for all employees; 20-49 employees: data for one in two employees; 50 employees or more: one in three). The selection of employees must be carried out randomly by responding units (it should be noted that many enterprises and administrations, and particularly the larger ones, do not make a selection and provide data on all their employees).
Public sector
In the public sector of the Confederation and the cantons, the survey is comprehensive at the level of the enterprises and administrations surveyed. In the public sector at commune level, the survey is also based on a random sample at two levels: the commune and the employees. Communes are stratified according to size (4 sizes) and geographical unit (7 major regions). In total, the stratification of the SESS at commune level is comprised of 27 cells (strata). For each stratum, a random sample of communes is drawn. The gross commune sample is made up of roughly 300 communes although the survey is comprehensive for communes with more than 800 employees.
Not Applicable
The results of the LCS 2020 were disseminated at national level on 28 March 2023, 29 months after the end of the reference period
National concepts have been defined as closely as possible in line with European concepts
The LCS 2020 is the fifth survey (after 2012, 2014, 2016 and 2018) that has been conducted in compliance with the definitions fixed in the European regulations (Council Regulation (EC) No 530/1999 of 9 March 1999 and the Commission Regulation (EC) No 1737/2005 of 21 October 2005).
