Statistics Norway

Division for Income and Wage Statistics

Akersveien 26
0177 Oslo

The aim of this report is to supply information on the quality of the data and statistics from Norway that are reported to and distributed by Eurostat in connection with the Labour Cost Survey 2012.

The report only covers the aspects regulated by the regulations and do not discuss any documentation or analysis of the results from the statistics.

The purpose of the statistics is to provide an overview of the total costs of having an employee. Statistics are provided for each industry separately, broken down into cost components. The statistics are released on Statistics Norway’s website:  http://www.ssb.no/en/arbeid-og-lonn/statistikker/arbkost

Not available.
New concept added with the migration to SIMS 2.0.
Information (content) will be available after the next collection.

Not available.
New concept added with the migration to SIMS 2.0.
Information (content) will be available after the next collection.

Not available.
New concept added with the migration to SIMS 2.0.
Information (content) will be available after the next collection.

Not available.
New concept added with the migration to SIMS 2.0.
Information (content) will be available after the next collection.

Not available.
New concept added with the migration to SIMS 2.0.
Information (content) will be available after the next collection.

Not available.
New concept added with the migration to SIMS 2.0.
Information (content) will be available after the next collection.

Not available.
New concept added with the migration to SIMS 2.0.
Information (content) will be available after the next collection.

Not available.
New concept added with the migration to SIMS 2.0.
Information (content) will be available after the next collection.

-

[Not requested]

[Not requested]

[Not requested]

[Not requested]

[Not requested]

[Not requested]

Not available.
New concept added with the migration to SIMS 2.0.
Information (content) will be available after the next collection.

[Not requested]

Users are Eurostat, The Technical Reporting Committee on the Income Settlement, research institutes, employees and employer organizations, the media, business and industry.

There are no feedback from our users indicating that the statistics does not meet their needs.

There are no feedback from our users indicating that the statistics does not meet their needs.

[Not requested]

[Not requested]

-

The population is made up of all enterprises in Statistics Norway's Central Register of Establishments and Enterprises, with the exception of small enterprises with fewer than 10 employees.

A sample of enterprises were selected to participate in the survey. Sampling was done on a 3-digit NACE level. To ensure broad coverage within each stratum, both NACE code and enterprise size was taken into consideration. An absolute criteria was that at least one enterprise (min n=1) was drawn from each stratum. In addition there was random sampling in each stratum of a certain size, given by predefined percentages of both stratum and population (Appendix A). Enterprises within the sample received a questionnaire.

-

Variance of interest in this case is variance that arises because of the size and composition of the sample, more specifically the sampling model, so-called sample variance. In annex we see coefficients of variations (CV) for total labour costs and sum of hours classified by industry and size.

-

The population is based on the Norway's Central Register of Establishments and Enterprises, March 2013. The following criteria must be fulfilled to be a part of the population:

• Only enterprises with an average of 10 or more employees throughout the year constitute a part of the population.
• The enterprises must have been in operation in the period January 2012 to March 2013.
• The NACE-industry division has to be in the intervals B - S except O.        

These variables are rarely changed with retroactive effect. Since the reference and study population are approximately equal, there are no over- and under-coverage in LCS2012. 

-

[Not requested]

Measurement errors mainly occur because the respondent misunderstands what is included in, and/or consequently report wrong, each column in the questionnaire, or because information requested is difficult to obtain. To avoid this, the questionnaire uses the most common book-keeping terms and commonly known aggregates of time and hours such as normal working hours, overtime, vacation and various types of absence as far as possible. If suspected erroneous, data can be corrected by asking respondents to update the questionnaire or obtain data from other sources such as administrative registers. In cases where none of the previous mentioned methods apply, related statistics were used to establish base levels or valid boundaries/extremes, and logical controls were used for further correction and/or imputation.  

The respondents were asked to report the average number of employees throughout the year. To help generate this number, respondents were to fill inn the number of employees for each month. In some cases there were mismatch between the level of costs accumulated through the year and the number of employees. We have therefore cross-checked reported numbers of employees with the NAV State Register of Employers and Employees (EE-register) In cases where substantial discrepancy was revealed imputation/correction were made.

-

Unit non-response refers to the fact that the respondent, in this case each individual enterprise, has not completed and returned the questionnaire. In the statistics, the unit response is between 92 and 98 per cent (table in annex). The main reasons for non-response are that units have ceased to exist, been sold or transferred to a new owner, gone bankrupt or been merged. Furthermore, there is a small group whom provide data of a quality that cannot be used for statistical purposes. In cases of unit non-response, the weights of the units on which the statistics are based are adjusted to compensate for the non-response. 

The most typical for a sample survey is that the sample unit, enterprise, has not reported on all necessary items in the questionnaire.

With imputation of data we refer to the substitution of missing values in the data set. We have not separated between data errors and missing values in our data correction work. We are therefore unable to separate the two causes for imputation. The values most frequently corrected were associated with the variable payments in kind. In addition several other variables were corrected in varying degree.

The majority of the data has been reported through our web gateway (more than 99 per cent); leaving less than 1 per cent for manual recording. Both manual and automatic controls have been carried out on the material. Table in annex ("Percentage pf correction") show that the variables most often corrected are wages and salaries in kind and employers’ social contributions. Likely errors have been revealed through controls against other sources. In general many of the corrections are caused by respondents not summing up variables to a total. Several variables have solely been retrieved from register and are not included in the table. The percentage in the table in annex has been calculated with the total number of enterprises as the numerator. This implies that a low percentage of correction doesn’t necessarily imply that the data reported are of good quality. Many variables have a high occurrence of zero, and values of zero are seldom revised.

-

The sample model used is based on stratified samples. Dividing the population into groups (strata) according to certain stratification variables reduces the possibility of imbalances in the sample and assures a better coverage of certain units or group of units.

The sample consists of enterprises drawn from the population. The population includes all active enterprises in the section, with the exception of small enterprises with fewer than ten employees, which are not included in the frame population. Large enterprises (sample units), where the definition of large varies between industries, receive a sampling probability of 1. While strata that cover small and medium sized sample units are given lower sampling probabilities.

The stratification is made according to industry and size (number of employees) of the enterprises, on the assumption that labour costs and composition of these costs in large enterprises differ from those in small ones, and that there are differences according to industry. In each stratum, this sample model ensures a minimal dispersion in the main variables measured, i.e. labour costs, and especially when it comes to compensation of employees where supplementing sources exist.

• Non-response that is not randomly distributed may bias the sample. This can have influence on the statistics. Non-response in the statistics is 5 per cent for enterprises and varies between 1 and 8 per cent for the different divisions.
• Cut-off for enterprises with less than 10 employees.
• Only enterprises that were in operation throughout the entire 2012 are included in the population. This infers that enterprises that were established during 2012 are excluded.
• Since all included enterprises are operating during the entire year, there is no need for scaling enterprises that only operate through part of the year.
• For enterprises with values for payment in kind that were obviously wrong or missing, information from the register of End of the year certificates was used. 

[Not requested]

[Not requested]

[Not requested]

[Not requested]

[Not requested]

-

The reference period for the survey is the year 2012. The results of the survey were published 14^th of October, 2014.

[Not requested]

[Not requested]

The questionnaires were made available for respondents 30^th of April 2013, with deadline 30^th of May 2013. Two rounds of reminders were used (June and August). In addition, 3 rounds of compulsory fine were issued (September, January and March), until adequate response rate was reached. Statistics Norways` standard routine is to issue no more than one compulsory fine per survey when such is needed. Due to the low response rate, extraordinary measures were taken to force respondents to reply. Several enterprises were also phoned and e-mailed to ensure that their questionnaires were returned. Exact dates and response rates are given in figure in annex ("Data flow"). 

During the first quarter each year, enterprises are occupied with balancing their accounts. The questionnaires were therefore dispatched quite late in the following year, 2013. The statistics are collected in accordance to the mandate given through “The Statistics Act of 1989”, which for LCS makes response mandatory. The data processing period started in July 2013. From then it was an ongoing process of analysing and approving questionnaires. Most enterprises had reported data by September 2013, while the last data were received 17^th of June, 2014. We had then received valid data from 3 315 enterprises. 

[Not requested]

-

-

The Labour Cost Statistics for Norway is regarded as one region, at NUTS 1 level. Hence data is not broken down by geography.

[Not requested]

The Norwegian Labour Cost Survey is collected every fourth year. There have been earlier surveys for 1996, 2000, 2004 and 2008. The questionnaires were partly different for these surveys, but the grand majority of the variables are comparable throughout.

The main difference between LCS 2012 and the previous LCS surveys is that the statistics is published on the new industry classification standard only. Between 2004 and 2008 Statistics Norway implemented the new industry standard (NACE Rev.2), LCS 2008 was therefore a break year in the series. However, the survey was published according to both NACE-standards (Rev. 1 & Rev. 2) for that year. LCS 2012 is only published according to NACE Rev. 2 (SN2007), and is therefore comparable only with the 2008-publication.

As opposed to earlier years, public sector was included in the population for 2012. This was especially noticeable in the industries Human health and social work activities and Education where public sector is the (by far) largest employer.

New for 2012 was also a more thorough identification of the variables concerning time consumption/hours. By improving the questionnaire – both formulations and what was asked for - the quality of the reported data was considerably improved. The data quality for LCS 2012 is therefore considered adequate for producing the “hour variables”, both paid hours and hours actually worked, hence a fulfilment of the Council Regulation.

The applied methods and models have been subject to ongoing improvements based on increased knowledge. An important point has been the extensive use of registers to identify and correct data in the survey. 

[Not requested]

1. Coherence with the Labour Force Survey

This is a short presentation and comparison of the Norwegian LCS and LFS surveys. It is important to point out some factors that may cause the observed differences between the surveys..

The main reasons for different surveys are in most cases, to meet different needs and as a consequence the statistics are built up to satisfy the core users needs. The LFS survey monitors and documents quarterly changes in the composition and distribution of the work force. It is based on a sample survey covering individuals (the sample unit is per household), that report on their current status in the work force.

Statistics on Labour costs on the other hand are built up to answer questions concerning the level and distribution of total labour costs. The source is, as earlier described, a sample of enterprises that report for the whole unit. The populations, the sources of information and the sampling models differ. Furthermore the two surveys have different reference periods, and utilize different sources for control, verification and finally dissemination.

Both statistics are none the less used for explaining different properties of the same subject and in this capacity we can use the LFS to understand aspects within the distribution and composition of employees within the labour force. Discrepancies should, where they occur, be explained and understood as a consequence of overlapping information.

Population and sampling units

Variable definitions

Objective of the LFS and LCS statistics

2. Coherence with Structure of Business Statistics

For some main industries within SBS the owner is defined as an employee. The definition of Wage and salaries are different between the two statistics. In general the values from LCS are slightly higher than corresponding values from SBS due to the cut-off in LCS. In general, wages are higher in enterprises with many employees than in enterprises with few employees. Part-time employee count as one employee, therefore would the distribution of part-time employees have an influence on the effect from cut-off. This could either raise or lower the mean value. 

[Not requested]

We find it not relevant to compare these two statistics as LCI is based upon the same source as LCS.

The national accounts (NA) statistics are designed to provide a consistent and comprehensive statistics of the overall national economy. The annual national accounts give both a summarised description of the economy as a whole and a detailed description of transactions between different parts of the Norwegian economy.

The definitions of the variables in NA are:

Compensation of employees = Wages and salaries + Employers' social contributions

Where wages and salaries are remuneration to employees in respect of work done in production, wages and salaries are both in cash and kind. Wages and salaries in cash include, in addition to normal salary, pay for overtime, and sickness and maternity allowances. Wages and salaries in kind consist of goods and services, or other benefits, provided free or at reduced prices by employers that the employees can use at their own discretion. Wages and salaries in kind also include, inter alia, the services of vehicles, value of the interest forgone by employers when they provide loans to employees at reduced rates of interest, and free transportation for employees in some transport industries.

Employers' social contributions are social contributions incurred by employers, paid to central government and to autonomous social security and pension funds, as well as non-autonomous pension funds. They include the following sub-items: employers' contributions to National Insurance, employers' other actual social contributions (contributions to the Public Service Pension Fund, Municipal Pension Funds, other social security schemes, and other social contributions), and in addition, employers' imputed social contributions. The latter item coincides with social benefits actually paid through unfunded arrangements - from employers to present or former employees, for instance AFP-pensions.

Hours worked is defined as hours worked by employed persons (employees and self-employed) in production during one year. The hours worked refer to production within effective and normal working hours, with addition for overtime while deducting the leave of absences due to sickness, vacations and any labour conflicts.

Hours worked are also influenced by the calendar effect (movable holidays and leap years). Number of working days may vary up to three days from one year to next.

The number of employees include employed persons who, by agreement, work for another institutional unit and receive a remuneration recorded as compensation of employees. Owners of corporations (joint-stock companies etc.) if they work in these enterprises, are also counted as employees.

Figures from NA are somewhat adjusted to enable a more relevant comparison, for example are NA figures refined to include only employees (self- employees are excluded). Important factors behind the still occurring differences include the use of cut-off in LCS for enterprises with less than 10 employees, the inclusion of taxes in NA figures and differing statistical units, NA uses the local unit whilst LCS uses enterprises.  

[Not requested]

The results have been sent to Eurostat http://epp.eurostat.ec.europa.eu/portal/page/portal/labour_market/labour_costs/database.

No results are sent directly to the respondents.

[Not requested]

The statistics are published at http://www.ssb.no/en/arbeid-og-lonn/statistikker/arbkost.

-

[Not requested]

[Not requested]

-

At the same web address mentioned in the previous chapter, users can find references to a brief methodical document in the link “about the statistics”. http://www.ssb.no/en/arbeid-og-lonn/statistikker/arbkost/hvert-4-aar/2014-10-14?fane=om#content

[Not requested]

[Not requested]

[Not requested]

[Not requested]

-

[Not requested]

[Not requested]

-