Statistics Austria

Directorate Social Statistics

Living Conditions, Social Protection

Guglgasse 13, 1110 Wien, Austria

The Labour Cost Survey (LCS) 2020, on which we present this quality report, is the sixth survey of this type in Austria conducted in accordance with the requirements of European law, following those of 1996, 2000, 2004, 2008, 2012 and 2016. The 2020 LCS is based on a Council Framework Directive[1] and three Commission Implementing Regulations[2]. A national labour cost statistics regulation[3] was also required to implement the survey in Austria.

 The LCS 2020 was conducted as a sample survey in the entire manufacturing sector (sections B to F of ÖNACE 2008[4]) and in virtually all parts of the services sector (sections G to N and P to S), irrespective of whether producers were market or non-market units. Survey units with fewer than ten employees and section O of ÖNACE 2008 were excluded in accordance with the EU framework regulation.

 The units were surveyed between April and October 2021 and the data were processed by December 2021. They were finalised in the first half of 2022 and the results were forwarded to Eurostat by the deadline of the end of June 2022.

[1]     Council Regulation (EC) No 530/1999 of 9 March 1999 as regards quality evaluation of structural statistics on labour costs and earnings (OJ L 63 of 12.3.1999, page 6 et seq.).

[2]     Commission Regulation (EC) No 1726/1999 of 27 July 1999 implementing Council Regulation (EC) No 530/1999 concerning structural statistics on earnings and on labour costs as regards the definition and transmission of information on labour costs (OJ L 203 of 3.8.1999, page 28 et seq.); Commission Regulation (EC) No 698/2006 of 5 May 2006 implementing Council Regulation (EC) No 530/1999 as regards quality evaluation of structural statistics on labour costs and earnings (OJ L 121 of 6.5.2006, page 30 et seq.).

[3]     Regulation of the Federal Minister for Economics and Labour on the labour cost statistics in manufacturing and services sectors (Arbeitskostenstatistik-Verordnung (Labour Cost Statistics Regulation); BGBl. (Federal Law Gazette) II No 126/2006), in the version of the Regulation of the Federal Minister for Economy, Family and Youth implementing the Regulation on labour cost statistics in industry and construction as well as services sectors (Labour Cost Statistics Regulation) (BGBl. II No 166/2017).

[4]     ÖNACE 2008 is the Austrian version of the EU Statistical Classification of Economic Activities NACE Rev. 2.

• NUTS (Nomenclature of territorial units for statistics)
• NACE Rev. 2 - Statistical classification of economic activities

Economic sections B - N, P - S according to NACE Rev. 2

Labour costs (D) include wages and salaries (D1) of employees excluding (D111) as well as including apprentices (D112), employer’s social contributions (D12), vocational training costs (D2), othere expenditure (D3), taxes (D4). Subsidies received by the employer (D5) are deducted from labour costs.

The number of employees (A1), hours actually worked (B1) and paid hours (C1) are respectively broken down into full-time employees, part-time employees and apprentices.

The concepts and definitions are in line with the ESS and international standards.

Survey units are enterprises, consortia, public corporations, establishments and unions of public corporations as well as associations. The data transmitted is broken down to local kind of activity units.

All enterprises in sections B - N and P - S according to NACE Rev.2 with 10 or more employees.

Austria

Business year 2020

Not applicable.

The selection framework for the sample was formed by the business register for statistical purposes of Statistics Austria. The LCS sample took the form of a stratified random sample and was drawn from the total population of the survey units (enterprises, consortia, public corporations, establishments and unions of public corporations, associations) with ten or more employees (sampling date: 30.9.2020). The divisions of ÖNACE 2008 (sections B to N and P to S) and five enterprise size bands (10-49, 50-249, 250-499, 500-999, 1 000 and over) were used as stratification characteristics. From 500 employees upwards (classes 4 and 5) all the survey units were included in the sample (full survey). The other strata were broken down in accordance with the characteristic “employees according to the Dachverband der österreichischen Sozialversicherungsträger (Federation of Social Insurances)” (ideally in accordance with Neyman-Tschuprow)^[1].

The selection rates by ÖNACE 2008 sections and enterprise size bands in percent are presented in the following table.

Source: Statistics Austria, Labour Cost Survey 2020.

The next table shows the absolute sample size by ÖNACE 2008 sections and enterprise size bands.

Source: Statistics Austria, Labour Cost Survey 2020.

Every 4 years.

Survey Data

The raw data were surveyed from the respondents using Statistics Austria's electronic questionnaire application for establishments named eQuest. The enterprises drawn into the survey sample were contacted by initial letters containing user logins for the online questionnaires and information about their response obligation. Among the submissions (see 6.3.3.1.), 99.3% of the enterprises were using the electronic questionnaire, whereas hard copy questionnaires were provided to the rest. The electronic questionnaire including explanations is attached in the annexes.

Administrative data

Wage tax data

Local kind of activity breakdown

Of the 7 323 enterprises that took part in the LCS 2020, 2 141 (29.2%) records surveyed at enterprise level had to be broken down to the level of local kind of activity units (Arbeitsstätten). This was done primarily on the basis of pay and employee data from the Short Term Statistics in the production sector (STS) or wage tax data and the Structural Business Statistics (SBS) 2020 for enterprises in the services sector. The quality of wage tax data was assessed by comparison with wages and salaries according to SBS 2020 on enterprise level. However, wage tax data was used in a lower extent than in the LCS 2012 and LCS 2016 surveys, as the assessment pointed out declining data quality in many cases, where SBS 2020 or STS 2020 data was used instead.

Wages and salaries of apprentices

From LCS 2016, wages and salaries were surveyed for all employees including apprentices (D.11) whereas compensations for the apprentices were excluded from the questionnaire and wage tax data were used instead. The figures for wages and salaries of apprentices extracted from wage tax data underwent plausibility checks. In case of missing pay slips for apprentices (despite there were apprentices reported from the enterprise) or implausible data (e.g. too high or too low figures in relation to usual tariffs for apprentices in the corresponding economic activity), these records were treated as missing data and have been estimated by statistical imputation.

Short-time working allowances

Administrative data from the Arbeitsmarktservice (Austrian Public Employment Service, AMS) on allowances paid for short-time working (Kurzarbeitsbeihilfe) were included into the computation of subcomponents of several EU labour cost variables such as D.11112, D.1211, D.1224, D.123, D.5. The data was also used for estimating the number of employees on short time due to the Covid-19 pandemic as well as their corresponding absent hours (see 6.3.4) in the trade and services sector.

Equalisation of Family Burdens Fund (FLAF)

For the second time after LCS 2016, administrative data from the Familienlastenausgleichsfonds (Equalisation of Family Burdens Fund, FLAF) on employers’ contribution to the fund were included into the computation of subcomponents of EU variables D.1211 and D.123. Before its use, the data were undergoing plausibility checks in relation with the wages and salaries indicated by the enterprise (D.11). Implausible records were replaced by estimated values.

Funding in construction sector

Administrative data from the Bauarbeiter, Urlaubs- und Abfertigungskasse (Construction Workers' Annual Leave and Severance Pay Fund, BUAK) were included into the computation of the EU labour cost variables D.1211 and D.123.

Extensive plausibility checks were used to be able to identify and correct measurement and processing errors. The data received were checked in a multi-stage procedure for completeness and plausibility. The survey data of the (few) paper questionnaires initially underwent a rudimentary check for the absolutely indispensable data even before they were input. The information in the web questionnaires was immediately checked for minimum requirements and rudimentary plausibility and flagged up incorrect information for the respondents; this prevented reports from coming in uncompleted or implausible. Subsequently, there was a detailed check in which each individual variable of a questionnaire was checked electronically for completeness and plausibility (for example with regard to totals, subtotals or logical dependencies between various survey questions). At the same time checks were made for disparities vis-à-vis external data (employee data from the Dachverband der österreichischen Sozialversicherungsträger - Federation of Social Insurances), the Short Term Statistics survey in industry and construction, the Structural Business Statististics survey 2019 and administrative data for short time work allowance). Checks were complemented or updated to cover effects through short-time work or other COVID-19-measures. Once the initial results of the Structural Business Statistics 2020 were available, individual checks of common variables were conducted again. Data sets containing errors of minor importance are not further processed. In the case of 6 147 survey units (84% of responding units) missing or implausible values were added or corrected using manual as well as automatic microediting or stochastic imputation methods. In a smaller number of cases (1 687 respondents; 23%) the respondents had to be consulted in order to obtain complete missing or to correct implausible information. On average there were 6.1 plausibility errors per survey unit.

The amount of imputation used was monitored, ensuring that the imputation rates did not exceed 20 % for each item (see imputation rates in 6.3.4.1).

Other than the surveyed data, also administrative data which were used for several items underwent formal plausibility checks. Obvious measurement errors or missing data were edited using automatic correction respectively statistical imputation programs.

After the process of manual and automatic data editing was completed, a second set of plausibility checks was performed on the complete data, listing only the outliers for further assessment.

Not provided.

Not applicable.

[Not requested]

The main national users of the 2020 LCS are the representatives of employers (Wirtschaftskammer Österreich - WKÖ (Austrian Federal Economic Chamber)) and of employees (Bundeskammer für Arbeiter und Angestellte - BAK (Austrian Federal Chamber of Labour)), the Austrian National Bank, science and research, media and other enterprises. In Statistics Austria (the national statistical institute) the LCS results are also used for other statistics (especially annual labour cost statistics, labour cost index, national accounts etc.).[Not requested]

The main national users of the 2020 LCS are the representatives of employers (Wirtschaftskammer Österreich - WKÖ (Austrian Federal Economic Chamber)) and of employees (Bundeskammer für Arbeiter und Angestellte - BAK (Austrian Federal Chamber of Labour)), the Austrian National Bank, science and research, media and other enterprises. In Statistics Austria (the national statistical institute) the LCS results are also used for other statistics (especially annual labour cost statistics, labour cost index, national accounts etc.).

Information about the level of satisfaction with the data provided is not available for Austria.

All requested labour cost variables and breakdowns according to the LCS regulation 1726/1999 are provided. Regarding the breakdowns, NACE Rev. 2 section O Public administration and defence; compulsory social security, which is optional, is not supplied.

[Not requested]

Not provided.

The survey units for the LCS 2020 were enterprises. Subaggregates for workplaces (local units) were constructed from enterprise data which represented all the existing cells – ÖNACE 2008 divisions combined with NUTS 2 regions – for each enterprise. This means that the sample can also be interpreted as a sample of these aggregates clustered at the enterprise level. For the extrapolation process, the number of employees per cell (regional breakdown x ÖNACE 2008 breakdown) was adjusted for the numbers of the total population.

Precision measures for estimating the random variation of an estimator due to sampling.

The variance was estimated using a rescaled bootstrap method (Preston’s multistage rescaled bootstrap)

The variation coefficients for annual labour costs (D) and labour costs per hour worked (D/B1) are given in the following three tables, broken down by ÖNACE 2008 sections, NUTS 1 regions and enterprise size classes.

Table: Variation coefficients by ÖNACE 2008 section

Table: Variation coefficients by NUTS 1 region 

Table: Variation coefficients by enterprise size classes 

Source: Statistics Austria, Labour Cost Survey 2020.

The variation coefficients for the annual labour costs (D) as well as for the labour costs per hours worked (D/B1) presented here (0.44) exhibit a slightly higher total spread than in the LCS 2016 (0.38). As was already the case in the past LCS surveys, there are considerably higher disparities to be found in individual ÖNACE 2008 sections (such as B Mining and Quarrying and R Arts, Entertainment and Recreation) that are caused by sampling effects. This is attributable to considerable differences in the structure of labour costs within these economic activities.

The question of under- or over-coverage is closely linked to the quality of the Statistics Austria’s business register for the purposes of statistics. Comparisons with external administrative registers and administrative sources (commercial register, trade registers of the economic chamber, tax and social security data, etc.) are made in an attempt to ensure that the business register is complete and up-to-date.

Under-coverage: As the sample was drawn at the enterprise level and ÖNACE 2008 section O Public administration and defence; compulsory social security was not included, the local units (workplaces) of enterprises in section O were not contained in the results at local unit level. This applies mainly to the sections P Education (Universities are independent survey units (with full legal standing) and are therefore surveyed in section P of ÖNACE 2008) and Q Human health and social work activities (in part public hospitals) as well as sections F Construction and R Arts, entertainment and recreation.

Units which, during the course of the survey, were found not to be required to report for the LCS 2020 (because they were inactive, non-operational or did not have any employees, etc.), were recorded as neutral non-respondents (and not replaced by other units). Survey units with fewer than ten employees remained in the sample if they filed reports and were included in the lowest enterprise size band.

Not applicable.

Local kind of activity units’ breakdown

As described in 3.3, the EU variables were attributed to local units, broken down by economic activity and NUTS 2 region (Bundesländer) for all enterprises which had several local units in various NUTS 2 regions or covered various divisions of ÖNACE 2008 on the basis of pay and employee data from the Short Term Statistics in the production sector (STS) or wage tax data and the Structural Business Statistics (SBS) 2020 for enterprises in the services sector. Depending on the availability of the data sources, in some instances one single variable was used for the breakdown of a whole set of LCS items. By this method the local unit potentially received the same labour cost respectively working hour structure as on enterprise level. However, STS 2020 data provide a satisfying number of suitable variables for the breakdown which were used for enterprises in ÖNACE 2008 sections B to F.

Moreover, the use of wage tax data at local unit level may cause biased distribution structures as some issuer of pay slips do not always indicate the local unit information correctly (e.g. all pay slips are assigned to one of the local units) or they even skip the information. In order to minimize the error, local unit data from the SBS 2020 was used instead for enterprises with insufficient local unit information.

Short-time working absent hours

While the Austrian economy was profoundly affected by the Covid-19 pandemic in 2020, short-time working played an important role and also had a noticeable effect on several LCS variables, especially absent hours and payments for days not worked. As described in 6.3.2, administrative data from the Arbeitsmarktservice (Austrian Public Employment Service, AMS) were used for the estimation of above-mentioned variables. The data were raised in the course of the application process for short-time working allowances that are funded by the AMS, containing the proposed period and the amount of the allowance. However, not all of these applications were granted by the AMS or they were later withdrawn by the applicant, whereby not all of these records could have been cleared from the data. Therefore, it can be reasonably assumed that the use of AMS data may have caused an over-estimation of absent hours as well as short-time working allowances in LCS 2020 which cannot be quantified.

See below

The LCS 2020 unit response rate was 97.6% (7 323 survey units), which was very high (see table below) and stems from the legal obligation to provide information, minimization and specification of the questionnaire programme and a series of measures to encourage response during the survey procedure (see 6.2). The unit non-response of 2.4% is made up of 17.5% non-responses which mainly occurred as a result of insolvency or the enterprise falling under the employee threshold and of 82.5% other non-responses which resulted from failure to file a report.

As regards the item non-response, the missing or implausible data were – as mentioned - estimated using statistical imputation or were corrected by means of consultations with the respondents. The imputation rates (item and total) are supplied in 6.3.4.1.

In the course of surveying and processing the data the following main measurement problems occurred:

Statistical imputations have been applied to a restricted set on items in the case of missing values:

For the extrapolation, the survey sampling was based on the number of employees as of September 2020, whilst the annual average as indicated by the enterprises was used for extrapolation. Different values between these time references can cause enterprises switching their size bands. In order to compensate the resulting bias, the basic weighting at enterprise level was formed for the stratification layers by combining the ÖNACE 2008 divisions with the enterprise size bands.

At the local unit aggregate level the combined weightings of the ÖNACE 2008 divisions and the NUTS 2 regions were applied as an adjustment factor. An iterative process was used to adjust the weights at enterprise level and at local unit level at the same time.

Not applicable.

Not provided.

Not provided.

Not provided.

The Labour Cost Survey (LCS) 2020 was prepared and implemented in good time, so that the data were forwarded to Eurostat 1 day before the deadline of the end of June 2022. Therefore the period between the end of the reference year and the first data release was 18 months. The results of the LCS 2020 were published on the Statistics Austria website in August 2022; a more extensive version was published in the November issue of the statistical journal of Statistics Austria (Statistische Nachrichten).

Not provided.

Not provided.

Not provided.

Not provided.

As “enterprises” in section O (Public administration and defence; compulsory social security) of ÖNACE 2008 – and hence their local units – were not included in the survey, the results at local unit level, especially in sections P (Education) and Q (Human health and social work activities), F (Construction) and R (Arts, entertainment and recreation) of ÖNACE 2008, are under-recorded. This may have led to over- or under-estimates of the labour costs in these economic sectors but they cannot be quantified.

As regards the reference period, scope, statistical units (Council Regulation (EC) No 530/1999) and variables (EU Regulation No 1726/1999), there are no notable discrepancies between the European terms and national implementation. Moreover, as the previous surveys the Austrian LCS 2020 results are compiled for local kind of activity units not just for local units.

Not applicable.

According to the operational definition of apprentices introduced for LCS 2016 by Eurostat, A.13 Apprentices and all related variables cover apprentices and trainee nurses only. While the definition of trainees (A.13) in accordance with the German EC Regulation (EC Regulation 1737/2005, to the definitions of variables in EC Regulation 1726/1999) was used in the LCS 2012, 2008 and in the LCS 2004, it included other trainees (such as health and trainee nurses, people acquiring work experience and other trainees). In 2000 the LCS included only apprentices as trainees. Some minor labour cost components (such as payments to employees’ saving schemes, costs for in-house and external facilities for employees) and (any) contributions from employers were not surveyed in the LCS 1996.

With LCS 2016, in the questionnaire some variables were replaced or partly replaced by administrative data (D.112 Wages and salaries of apprentices; components of D.1211 Statutory social-security contributions) or by estimation using administrative data (D.123 Employers’ social contributions for apprentices) and by a change in the questionnaire (NACE B to F: D.1113 Payments for days not worked, D.1221 Guaranteed remuneration in the event of sickness). This did very much reduce the response burden, but might have a slight effect on results.

The results of LCS 2020 are presented at local unit level as of the 2004, 2008, 2012 and 2016 surveys, whilst the presentation and survey units of the 1996 and 2000 LCS were enterprises (exceptional provision in accordance with the Annex to EC Regulation No 530/1999).

For the fourth time, economic sections were recorded on the basis of ÖNACE 2008. ÖNACE 2008 sections P, Q, R and S were first included in the LCS 2008 and these branches very largely correspond to ÖNACE 2003 sections M, N and O which were first recorded in the LCS 2004. The LCS 1996 differs from the other surveys in as much as sections F, G, H, I of ÖNACE 2003 were not surveyed

Not applicable.

1. Comparison with the micro-census Labour Force Survey (LFS)

The hours actually worked per employee recorded in the LCS 2020 were on average 0.9% higher than in the micro-census Labour Force Survey (LFS) in 2020 (see following table). While the total figures seem to correspond approximately, there are remarkable differences within certain economic activities. The difference was greatest in the section of Art, entertainment and recreation (R) where the hours actually worked in LCS 2020 were +14.9% higher than in the LFS 2020, followed by Accommodation and food service activities (I) and Real estate activities (L), where there were considerably higher estimates for the hours worked of +11.5 and +10.1% as well. In the sectors Financial and insurance activities (K) and Electricity, gas, steam and air conditioning supply (D), LCS 2020 hours worked were lower than in LFS (-6.0% and -5.2%).

As these deviances are occurring in the same NACE sections (excluding section I) and moreover in very similar magnitudes as in the comparison between the LCS 2016 and the LFS 2016, this can be seen as an indication that the differences may be caused by surveying effects of the same kind as 2016 rather than by sampling effects or effects caused by differences in monitoring short-time work in the first year with COVID-19 measures.

On the one hand it is assumed that the differences are caused by the reference period of one week in LFS which is possibly not sufficient in order to measure certain absences. On the other hand, hours worked but not paid in the LCS were not reported or not reported in full by respondents because there were no time records; there is under- or non-recording of hours paid which do not feature in the accounts (“clandestine” payments) too. It is also assumed that, in survey units with fewer than 10 employees (which are not included in the LCS), the number of hours worked is higher. Moreover, it should be noted that the data on hours actually worked in the services sector are raised differently in LCS than in LFS. As further described in 3.3, administrative data were used for deriving the hours not worked due to short-time (which were then subtracted from the hours worked) in LCS, while the data in LFS were surveyed directly. With the use of administrative data for short-time hours in LCS, there might appear a possible distorting effect as well, as pointed out in 6.3.2.

Table: Coherence with the micro-census Labour Force Survey (LFS)

2. Comparison with the Structural Business Statistics (SBS)

Enterprises with fewer than 10 employees were not surveyed in the LCS 2020; local units of these enterprises are not covered. These enterprises are, however, included in the SBS 2020 data. Furthermore, usually SBS results at enterprise level are used. In order to strip out this difference in coverage and the unit effect, a special tabulation of SBS 2020 results for local units of enterprises with 10 and more employees is presented here. It should be borne in mind that the employment figures at local unit level in the SBS include self-employed persons.

Comparing the LCS 2020 with the SBS 2020 (see following table) shows that LCS figures for wages and salaries per employee are on average 5.4% lower than SBS results. In Accommodation and food service activities (I) the differences were highest at -11.8% followed by Water supply; sewerage, waste management and remediation activities (E) at -9.9%.

Table: Coherence with the Structural Business Statistics (SBS)

One of the main reasons for these differences is the different definition of wages and salaries. The wages and salaries in the SBS also include guaranteed remuneration in the event of sickness (D.1221) and payments to employees leaving the enterprise (D.1223) which, however, are recorded under employers’ social security contributions in the LCS. In the next table these components are incorporated in LCS wages and salaries, which bring the difference between LCS and SBS down to -0.9%.

However, notable deviances remain in particular sections such as Accommodation and food service activities (I, -7.8%) and Construction (F, -6.4%) with lower wages and salaries in LCS than in SBS. Figures for the sections Professional, scientific and technical activities (M, +5.2%), Financial and insurance activities section (K, +3.4%) and Transportation and storage (H, +2.3%) are higher in LCS than in SBS 2020. The information for the LCS mainly came from personnel accounts and that for the SBS from the profit and loss accounts. The latter probably are more in line with accounting criteria than with LCS definitions. Furthermore, in SBS short-time working allowance paid by the Arbeitsmarktservice (Austrian Public Employment Service, AMS) to employers is not deducted while these allowances reduce several labour cost components. This can explain the higher differences in some branches and in Accommodation and food services (I) in particular.  

A further reason for the discrepancies is the different data compiling method: whilst the LCS is conducted as a sample survey with extrapolation, the SBS is a primary statistical survey supplemented by register and administrative data as well as other statistics for non-surveyed units.

Comparison with the Labour Cost Index (LCI)

The average annual change in labour costs per hour worked (see following table) was 4.9% in the LCS and 6.6% in the LCI (difference: -1.6 percentage points) between 2016 and 2020 for the total of the economic sectors presented. The annual rate of change differed most in the sectors Real estate activities (L, ‑4.3 percentage points) and Transportation and storage (H, ‑3.4 percentage points).

Table: Coherence with the Labour Cost Index (LCI)

The reasons for the difference mainly stem from the fact that the LCI uses other data sources. In the industries and construction sector, data from the Short Term Statistics are used, which are extrapolated using administrative data, and in the services sector it is administrative data that are used to estimate labour costs and data from the micro-census Labour Force Survey for the hours worked. The LCS is used only to a small extent for the LCI; the result levels of LCS are not used in the LCI, as the LCS does not cover enterprises with fewer than 10 employees and is available only every four years which, for the purposes of the LCI, constitutes a considerable time lag.

Differences in the definition of labour costs (D.2 vocational training costs and D.3 other expenditure paid by the employer do not feature in the LCI) make only 10% of the difference. Methodological peculiarities of the LCI, such as the inclusion of micro-enterprises and holders of a non-standard contract (freie Dienstnehmer:innen), may cause differences by comparison with the LCS too. The impact of restricting the LCI to enterprises which were also represented in the previous year's data pool and of enterprises that change ÖNACE section cannot be quantified.

If we compare the LCS 2020 with the results of the NA in accordance with ESA 2010 (see following table), the result for compensation of employees in the LCS 2020 is 6.1% higher per employee than in NA. In Real estate activities (L), the compensation of employees recorded in the LCS 2020 was 38.0% higher, in Other service activities (S) 17.0% and in Professional, scientific and technical activities (M) 15.2% higher than in the NA, whilst in Accommodation and food service activities (I, ‑17.1%), Electricity, gas, steam and air conditioning supply (D, -12.4%) and Education (P, ‑8.2%) compensations according to LCS 2020 were below that in the NA. For the comparison with the NA, employment relationships (jobs) were used, since employees in the LCS represent employment relationships by definition (employees may have been active in several enterprises).

The differences are mainly attributable to the fact that, in the LCS 2020 local kind of activity units of enterprises with fewer than 10 employees, which tend to exhibit lower remuneration per employment relationship, are not surveyed whilst the NA results cover all size bands. The results of the LCS 2020 are therefore higher than those in NA for most ÖNACE sections.

The inclusion of atypical employment relationships such as holders of non-standard contracts, domestic helpers etc., which are not included in the LCS, has an increasing effect on the different levels of the results presented.

Furthermore, the NA assesses components that belong to the clandestine economy such as tips in Accommodation and food service activities (I), Taxis (H) and Hairdressers (S), whilst tips are to be included in total wages and salaries by the statistical units in the LCS. It may be assumed that tips are under-recorded in the LCS in section I of ÖNACE 2008.

Table: Coherence with National Accounts (NA)

In Education (P), the lower compensation of employees in the LCS 2020 stems from the fact that local units of enterprises of section Public administration and defense, compulsory social security (O) of ÖNACE 2008 are not surveyed in the LCS. Hence public-school teachers in the Education (P) sector, employees of publicly run hospitals in Human health (Q), road maintenance staff in Construction (F) and employees of public facilities in Arts, entertainment and recreation (R) are missing in the LCS.

Finally, it should be mentioned that the NA are compiled at level of kind of activity units (Betriebe), whilst LCS results are presented at local kind of activity unit level (Arbeitsstätten). This difference in the statistical units, however, only has a slight influence on compensation of employees per employee.

All statistical outputs within the LCS 2020 data set are consistent,

Not provided.

The main results are presented on the Statistics Austria website under www.statistik.at > Startseite > Statistiken > Arbeitsmarkt > Arbeitskosten und Tariflohnindex > Arbeitskosten. More detailled results as well as historic data can be downloaded as OpenDocument spreadsheets (ODS).

The article “Arbeitskostenerhebung 2020” (Labour Cost Survey 2020) which appeared on pages 812-826 in issue 11/2022 of Statistische Nachrichten.

In the first half of 2023 the results will be presumably presented in an extensive print publication.

All the abovementioned publications contain metadata to differing degrees. The article in the Statistische Nachrichten presents additional graphics.

A set of detailled basis data will be provided in the public Statistical Database STATCube, which allows generating and exporting tables in different formats according to user’s individual needs.

Not provided.

LCS microdata will presumably be stored in the Austrian Microdata Center (AMDC) within the first six months of 2023 to make it accessible to researchers for scientific use.

There are no plans to automatically forward results to the reporting enterprises included in the survey and other survey units.

The metadata will be documented in a very detailed standard documentation that will be available via the internet to any members of the public on the Statistics Austria website within the first six months in 2023.

Not provided.

Not provided.

Not provided.

Not provided.

Not provided.

Not provided.

None.