STATEC (Institut national de la statistique et des études économiques)

Unité SOC2 - Marché du travail et éducation (Unit SOC2 - Labour market and education)

STATEC

Unité SOC2 - Marché du travail et éducation

B.P. 304

L-2013 LUXEMBOURG

[Not requested]

Enterprises are grouped by economic activity (NACE2) and size group.

NACE2 sections B to S excluding O

Enterprises with 10+ employees

Concepts and definitions according to Eurostat's LCS Implementing Arrangements.

Enterprise (legal unit)

All enterprises with 10 or more persons employed in the covered sectors.

Total population: 5643 enterprises

Sample drawn: 3665 enterprises (65% of total population)

Respondents in final dataset: 2763 (75% of sample)

Luxembourg (country) (LU0)

Reference year: 2020

not applicable

[Not requested]

[Not requested]

[Not requested]

[Not requested]

[Not requested]

[Not requested]

[Not requested]

[Not requested]

The LCS has been conducted using a web questionnaire, so the respondent (in most cases the head of accounting or human resources of the enterprise) had to fill in the questionnaire on his own.
In order to guarantee the quality of the answers provided,
- clear definitions and examples had been provided in the online questionnaire and the accompanying notes
- collaborators of the national statistical office answered the questions of the respondents on the phone or by e-mail
- answers were crosschecked for inconsistencies and the enterprises were recontacted in order to correct data on a case-by-case basis

[Not requested]

[Not requested]

Summary of national core users

The most important national core user is STATEC. The other national core users include ministries, administrations, foreign embassies, researchers, media, employers’ federations, trade unions, companies from the private sector, as well as students from high-schools and universities.

Description of their main needs including an assessment of their level of satisfaction with the data offered

The users are mainly interested in breakdowns of labour cost along several variables, e.g. economic activity and size of the enterprise. Another focus of interest is the disaggregation of the labour cost in its components, such as direct vs. indirect cost, contractual wages vs. bonuses, wages and salaries vs. social contributions, etc.

There is no systematic and formal assessment of the users’ satisfaction. Nevertheless, the small size of the national user community enables STATEC to have a direct contact with the users. As a result, STATEC can be very responsive to the users’ needs. The informal feedback suggests that the users’ needs are generally satisfied. As of today, there is no information regarding redundancy or lack of completeness.

[Not requested]

[Not requested]

[Not requested]

[Not requested]

(See "6.2.1. Sampling error - Indicators")

Table 6.2.1. in annex shows the coefficient of variation (CV) and its components, as well as the confidence intervals

- for hourly labour costs (D / B1) (including apprentices)

- for annual labour cost per employee FTE (in EUR/FTE/Year) (D/A1FTE) (including apprentices)

(See details under 6.3.1. to 6.3.5.)

No problem of under-coverage is known. However, there has been some over-coverage.

No problem of under-coverage is known. However, there has been some over-coverage. The reasons for this over-coverage stem from a discrepancy between the administrative files used for the sampling and the real world, and can be put into 2 categories:

1. The legal unit has less than 10 employees in practice;

2. The legal unit does not exist. The reasons for this non-existence can be bankruptcy, merger, liquidation or discontinuance of business.

Table 6.3.1.1. in Annex gives an estimate of the over-coverage rate (in %) expressed in terms of legal units and in terms of employees. Furthermore, the table distinguishes the 2 above mentioned causes for over-coverage.

[Not requested]

A non-negligible amount of measurement errors was noticed for some working time and labour cost variables. Internal quality and plausibility tests have detected incoherencies between these variables and other linked variables. These incoherencies stemmed from a misunderstanding of the questionnaire and from typos.

Most of the errors could be corrected via direct follow-up with the legal units.

Data from education-related legal units (division 85 of NACE rev.2) should be considered with caution because variables such as “hours actually worked” and “annual holiday leave” might have been interpreted very differently from one respondent to another, which obviously affects hours worked and hourly wages.

(see 6.3.3.1. and 6.3.3.2.)

(see table 6.3.3.1. in Annex)

Unit non-response rate in terms of legal units has been 22.2% (up from 11.6% in 2016). These enterprises represented 10.9% of the employees (up from only 5.6% in 2016)

Non-response was expected to be higher due to the COVID-19 pandemic and a larger than usual gross sample had therefore been drawn.

The data collection was conducted via a web-based questionnaire. The online questionnaire was designed in a way to oblige respondents to fill in most cells and checking tools were integrated in the questionnaire, so that item non response was largely reduced. An extensive follow-up allowed reducing item non-response and missing values even further, so that item non-response is negligeable (see also point 6.3.4.1. imputation).

(see 6.3.4.1.)

As explained in section 6.3.2, many corrections were made during the data checking and follow-up process.

In some cases, a variable was missing for apprentices that had to be imputed.

Some variables were calculated on the basis of several variables in the more detailled questionnaire. In some rare cases (<= 2%), one of several informations needed to calculate the variable was missing and had to be imputed.

See table 6.3.4.1. in Annex for details.

Not applicable.

[Not requested]

[Not requested]

[Not requested]

[Not requested]

[Not requested]

The reference year for the Labour Cost Survey is 2020.

The first results have been published in December 2022.

The final data have been available in December 2022

The fieldwork started on May 3, 2021 with the launch of the online questionnaire.

A first invitation letter to the enterprises was sent on May 3, 2021, followed by three reminders sent in June, July and September 2021, the final deadline having been October 15, 2021.

The last questionnaires / corrections of questionnaires were received in February 2022.

The fieldwork officially ended with the transmission of first results to Eurostat on April 4, 2022.

[Not requested]

[Not requested]

In Luxembourg, the European concepts and definitions of statistical units, populations, reference times, classifications and definitions of variables have been used.

[Not requested]

Coverage

The Labour Cost Survey of 2020 has been the sixth survey (after 2000, 2004, 2008, 2012 and 2016) to be conducted under the harmonized European rules.

The 2000 survey covered the sections C to K of the NACE rev.1 classification.

In 2004, sections M to O of the NACE rev.1 classification were added.

The 2008 Survey was based on the NACE rev. 2 classification and covered sections B to N and P to S. However, the results were also available in a separate table based on NACE rev.1.

The 2012, 2016  and 2020 Surveys used only NACE rev.2 and covered the same sections than in 2008.

Survey Design

Stratified sampling has been applied.

The population of legal units has been stratified on the basis of the size of enterprise in terms of employees.

For each stratum, sampling has been done according to the sampling rates mentioned in table 8.2. in the Annex.

Break in series

Definitions of hours worked and hours paid by full-time employees have been changed between the 2012 and the 2016 survey.

[Not requested]

In order to evaluate the extent to which the results of the Labour Cost Survey are coherent, the suggested comparisons have been made.

A) The number of hours actually worked per employee (B1/A1), compared to the average actual hours worked in the main job according to the Labour Force Survey (LFS) (See table 8.3.A in Annex.)

The number of hours actually worked is on average 25% higher in the LFS than in the LCS. The difference has been noted in the past but has worsened in 2020.

Indications of the possible structural reasons for this difference :

- LFS is based on declarations by individuals. These declarations can be inaccurate. Respondents might e.g. not correctly remember absences at work for sickness etc. and might tend to declare contractual hours as actually worked hours.

- Difference in coverage : LCS does only cover enterprises with 10 or more employees, whereas LFS covers the whole economy. It is however known that employees in smaller enterprises tend to work more hours.

- Difference in coverage : LCS covers all employees (of which nearly half are commuters), LFS covers only residents

- NACE Sections B, D, E, L, P, R and S : number of units sampled in LCS and/or number of employees sampled in LFS are low

- Moreover, the sector of activity in the LFS is based on individual declarations. This is another source of inaccuracy, which can explain the fact that for individual activities, the differences are higher than for the total.

For 2020, hours actually worked have dropped significantly in LCS due to the lockdowns in several sectors as a consequence of the COVID-19 pandemic. However, this sharp drop is not reflected in the LFS self-declared figures.

B) The variable “Wages and salaries” expressed per employee (D11/A1) compared to the same variable, per employee of the Structure of Business Statistics (SBS) (See table 8.3.B in Annex.)

Overall (sectors B-J and L-N), LCS and SBS are consistent. Some differences subsist, mainly in the "small" sectors (NACE Sections B, D, E & L), where the number of units sampled in LCS and in SBS are both very low.

C) The average annual growth rates (between 2016 and 2020) of the variable “Hourly labour costs” (D/B1) compared to the average annual growth rate of the unadjusted Labour Cost Index (LCI) (See table 8.3.C in Annex.)

For the economy as a whole, the average annual growth rate for the period 2016 to 2020 according to LCS is higher than the rate computed from LCI.

Indications of the reasons if the values differ considerably:

In some NACE sections, especially in those with few units or some dominating units, overall results are heavily influenced by irregular payments that vary considerably from one year to another (i. e. bonuses, contributions to employees’ saving schemes, subsidies received by the employer or not, salaries in kind). This explains in part, why average growth rates are different in LCS than in LCI, as these types of payments are fully covered by the LCS but not by the social security data base used for the estimation of the yearly update of the LCI.

In NACE section B, D, E, L and R, the sample is very small.

In reference year 2020, hourly wages have been heavily impacted by the reduction in working hours due to the COVID-19 pandemic.

[Not requested]

The variable “Compensation of employees”, expressed per employee (D1/A1), compared to the same variable of the National Accounts (NA)

(See table 8.5. in Annex)

Indications of the reasons if the values differ considerably:

- NACE Sections B, D, E, L, P, R : number of units sampled in LCS is low

- NACE Section P: LCS does only consider private sector education units (which are very few), whereas NA consider all education units, including public ones (i.e. most). Compensation of employees in public education sector is higher

- LCS only covers enterprises with 10 or more employees

- there might be differences in definitions or scope (to be explored)

[Not requested]

References for core results publications, including those with commentary in the form of text, graphs, maps, etc.

Description of main actions carried out by the national statistical services to inform users about links to the data

The public will be informed through a News release dedicated to the publication of the results of the Labour Cost Survey in the fourth quarter of 2022 / 1st quarter 2023

Results will be published and analysed in short publications in a series called “Regards” (in French).

Several tables will be published on Statistiques.lu

[Not requested]

[Not requested]

[Not requested]

Description of and references for metadata provided

Metadata are published on the website Statistiques.lu

References for core methodological documents relating to the statistics provided

A detailed methodological note is published on the website Statistiques.lu

[Not requested]

[Not requested]

[Not requested]

[Not requested]

[Not requested]

[Not requested]

[Not requested]

None.