Statistics Iceland

Wages, Income and Education Statistics

Borgartúni 21A

IS-150 Reykjavík

Iceland

This metadata contains the quality criteria of the Structure of Earnings Survey (SES), referred in the Annex of the Commission Regulation (EC) No 698/2006 of 5 May 2006.

Economic activities are classified by ÍSAT2008 (NACE rev.2). Data are reported with two digits.

Occupations are classified by ÍSTARF95 (ISCO-88) but data is converted into ISCO-08. Data is reported with two digits because of the small size of the labour market.

Educational attainment is based on one-digit ÍSNÁM2008 (ISCED-97) but data is converted into ISCED 2011.

The Icelandic data for the Structure of Earnings Survey 2018 includes the economic activities C, D, E, F, G, H, I, J, K, O, P, and Q (NACE rev. 2) that cover about 85% of the labour market (B-S, enterprises with 10 or more employees) according to the Pay as You Earn register (PAYE). For an overview of the sample size see attached document MD_IS_2_3 Sample size.

As the labour market in Iceland is quite small the sample is structured into 1-3 size bands (large, medium and small) as stratifying using the conventional size bands would include empty strata.

In some economic activities certain sub-groups are missing. In the economic activity J small units and units in information technology (divisions 62 and 63) are missing. Economic activities O, P, and Q include only the public sector. The public sector cover about 94% of employees in O, 87% employees in P, 76% of employees in Q. Weights are adjusted for O, P, and Q to represent full coverage of these economic activities. Weights are not adjusted for J as the sub-groups that are missing are different from the groups that are included in the sample.

Statistics Iceland is striving towards a full coverage of the Icelandic labour market. But, for now, the economic activities M, N, R, and S are not fully covered and thus, data for those activities are missing for Iceland in SES 2018. The economic activities B and L are only a remote part of the Icelandic economy.

Data and variables follows the Council Regulations (EC) No 530/1999 and No 1738/2005 and amendments No 1916/2000. The implementation is in line with Implementation Arrangements for SES2018 as described in Doc.: Eurostat/F3/LAMAS/24/17.

For SES 2018 all obligatory variables are accounted for, and only one optional variable is missing:

• B.34 Other annual days of paid absence

For an overview of explanations of variables see attached document MD_IS_2_4 Variable overview.

Local units are not available in business registers and therefore the legal unit along with the 5-digit NACE classification is used as a proxy.

All employees in local units in NACE sections B-S with 10 or more employees.

Iceland.

Reference month is October every fourth year, starting 2002.

Not applicable.

-

The data for SES 2010 is based on ISWEL survey (Icelandic Survey on Earnings, Wages and Labour cost). ISWEL is a cluster sample survey stratified by sectors, economic activity and size.

In the private sector a sample is drawn from the ‘Monthly Pay as You Earn’ data and the public sector is covered by a sample of municipalities (local government) that is drawn according to geographical location and full data from the Ministry of Finance (central government).

Probability sampling

The ISWEL sample is a stratified cluster sample, where the sample unit is the legal unit and the observation unit is the employee. Within the private sector the target population contains all legal units with 10 or more employees. The population frame is based on monthly PAYE register. PAYE register reflects the sum of the wages reported to the tax authorities.

The population is stratified in sections and subsections according to NACE rev. 2. Enterprise with more than A/m employees are selected with a probability of 1 (where A is the number of employees in the stratum and m is the number of enterprise to be selected from the stratum). For the rest of the population, a simple random sample is selected from each stratum.

In the public sector a sample is drawn from the municipalities in such a way as to achieve a certain proportion for each of the country's regions. As in the private sector, the population is defined as all wage earners working for municipalities that have 10 or more employees.

Non-probability sampling

Full coverage data were received for employees working for the central government. All employees working with in institutions with 10 or more employees are included.

Weights

Weights are assigned to legal units based on their share of yearly earnings within the stratum, rather than the number of legal units in the stratum. By using this weight the number of employees and the sum of earnings with in each stratum is better reflected than if the number of legal units were used for weighting.

The ISWEL survey is an ongoing monthly survey.

Every month, each legal unit sends a text file containing detailed information on earning structure and cost items for all employees. In addition, background data on the individuals and companies is collected. Contracts have been finalised with Iceland's leading software firms on writing computer programs that fit their own wage software and allow companies to submit standardised and detailed information electronically to Statistics Iceland on wages, labour cost, working hours and necessary background factors. Data on variables that are missing or of poor quality in ISWEL (education, length of service, type of contract) are enriched with other data sources (database on education, PAYE register and other sources).

The monthly ISWEL data is validated using a series of predefined parameters. All abnormalities are examined and corrected if needed before validation. In addition the late submission of SES allows comparison with other sources, most notably the PAYE register, tax returns, Labour Force Survey and National Accounts.

Data is compiled using a pre-programmed sequence. Imputations are done for education on a small part of the sample, were ISCO code and highest level of education are not in coherence. This can happen when education earned abroad is not in the database on education.

Not applicable.

The ISWEL survey is the main survey for all statistics on earnings and labour cost and is as such under constant observation for quality issues.

Not applicable.

-

The SES data are used by Eurostat as comparison between countries on wages and wage components. It has not been examined who the users of the data are but the most likely users, along with Eurostat, are institutions, analysts, news media and the public.

User satisfaction has not been measured.

The Icelandic data for the Structure of Earnings Survey 2018 includes the economic activities C, D, E, F, G, H, I, J, K, O, P, and Q (NACE rev. 2) that cover about 85% of the labour market (B-S, enterprises with 10 or more employees) according to the Pay as You Earn register (PAYE). All mandatory variables are submitted for every record.

Not applicable.

-

Great effort is made to ensure coherent and accurate data in the ISWEL survey. During visits to employers in the sample, technical aspects concerning the software for wage calculations are addressed along with details on entering records such as classification of occupations according to the ÍSTARF95 and ÍSAT2008 standards. The employer's wage structure is examined assigning payroll items to the wage items of the survey to ensure coordination with other employers. Efforts are continuously made to improve data quality through feedback to the employers and through updating and improvement of the validation checking system. Validation checks are made continuously and data is edited by using a set of predefined rules.

The ISWEL survey is subject to possible errors due to the sampling methods applied and the quality of the Monthly Pay as You Earn register (PAYE) which the population is based on. Sampling errors can arise if employers are incorrectly registered according to their economic activity classification in the sample frame, in cases of employers operating in more than one economic activity but only registered in one of it, if newly established companies are not a part of the survey’s pool, etc. As the SES data is based on the ISWEL survey is subject to challenges because of lack of data coverage and due to inadequate sample refresh.

Please see the attached document MD_IS_6_2_1 Coefficient of variation.

Coefficients of variation by size bands of the enterprises are not available.

Non-sampling errors such as none-response errors and measurement errors are errors that occur in the data collection process. By maintaining a constant partnership with the sampled business units, these errors are minimized. The measurement of paid hours is especially prone to measurement errors as in many business units fixed-wage contracts are common. Thus, employees are not paid for overtime work, and their working hours might be underestimated.

Please refer to chapter 2.3. for the coverage of the sample. The ISWEL sample is based on the PAYE register and therefor errors in NACE classification in the PAYE register can have an impact on the coverage of the ISWEL survey.

The sampling frame is built on the PAYE register, which in turn is the target population. Thus over-coverage is only existing if the sampled unit is not operating in the reference month (October 2018). This did not apply to any of the sampled units in ISWEL.

The sampling frame is built on the PAYE register, which in turn is the target population. Thus no common units exist.

In spite of the data collection method some measurement errors can occur. The main concerns relate to the educational level and length of service in the enterprise. These two variables are coded in the ISWEL survey but the accuracy of the coding is not satisfactory. To get a better estimation of these two variables data were sought from other sources, both registers and other surveys conducted by Statistics Iceland. However, the educational level might be underestimated as many people seek higher education in other countries and therefore do not appear in registers in Iceland. The length of service in the enterprise might be underestimated as data are only available back to 1998 in registers in a format that can be used at this time.

In addition, the inclusion process for the data for Ministry of Finance is not complete and thus some errors in ISCO classification of employees in the central government can occur.

Annual bonuses and allowances not paid in each pay period (variable 4.1.1), could be under­estimated as corrections on such payments are not included in the variable. Nevertheless, these corrections are included in, gross earnings in the reference month (variable 4.1).

Lastly, by using number of weeks which the gross annual earnings relate to (variable 3.1) and the share of a full-timer’s normal hours in the reference month (variable 2.7.1) to gross up the annual earnings to a full year earning an error can occur. This is the case as working hours in the reference month do not necessarily represent hours worked in other months of the year, due to seasonality in the labour market. This is especially true for part-time workers. In order to correct for this the variable 3.1 has been adjusted as explained in the attachment in chapter 2.4.

All business units are obligated to respond to Statistics Iceland. As information is collected directly from these sources and not from individuals, non-response errors are minimal. To further minimize the non-response errors good relations are kept with the business units and special work methods are used in the data collection and quality checking.

In spite of all efforts, non-response errors can occur. These are mainly due to two reasons. First, a technical error can occur with a certain business unit and thus it may be unable to provide data for a period of time. Second, individual transfers can be erroneous, due to human factors, and these transfers are eliminated from the dataset. Weights are adjusted, both for local units and individuals to minimize the impact of these non-response errors.

The business units response rate in SES 2018 was 89%

As every individual transfer in the ISWEL survey is regarded as one item regardless of partial existence of data or not, the item non-response rate is calculated on the dataset as whole and not the individual variable. The transfer response rate within the local units included in the dataset was 94.6%.

The data are collected directly from the software each business unit uses for calculating wages. Taking advantage of data collection by direct access minimises the bias caused by recording.

Imputation rate for educational level was 26.7%

Not applicable.

Not applicable.

Data is final but can be revised if serious errors are found in the data or processing.

If serious errors would be found Eurostat would be contacted as soon as possible and new data submitted.

Data have not been revised.

The ISWEL survey is based on a lengthy tradition of collecting data directly from the software each business unit uses for calculating wages. Every month, each business unit sends a text file containing detailed information on earning structure and cost items, together with background data on the individuals and companies. Thus, all data retrieved from the ISWEL survey are collected simultaneously and sent to Eurostat in T+18 months.

Data are submitted by the local units within 10 days after the end of the reference month and most of the variables are available at the end of T+1 month. The complete dataset can be available in T+6 months.

Not applicable.

Not applicable.

Data were delivered on time.

Not applicable.

The data, which the SES 2018 is based on, are collected monthly and complies with most mandatory points drawn up in the council regulation no 530/1999, and subsidiary commission regulations. However, due to some specific characteristics of the Icelandic labour market some incoherence can occur.

• Statistics Iceland is striving towards a full coverage of ISWEL. For now, the economic activities M, N, R and S are not fully covered and thus, data for those activities are missing in SES 2018. In addition the economic activities B and L are a remote part of the Icelandic economy.
• In Iceland each legal unit has a unique identification number and the enterprise is considered to be the same as the legal unit.
• In coherence with former submission the local unit is defined on the basis of identification number (the same as above) and NACE to a 5 digit level.
• In order to be included in the sample the employee had to have a minimum of 10 normal working hours in October 2018.
• Full-time employees are defined as those whose normal hours exceed 90% of the contractual working time.
• In coherence with former submission and definitions used in local production of statistics, variable 1.4 is coded according to sector (ESA 2010). Thus, only institutions belonging to the general government (S.13) are coded as public. Public controlled marked producers are coded as private.
• The grossing up factor for the local unit is based on the number of employees and sum of earnings within each stratum.

Iceland is considered one geographical area (NUTS).

Not applicable.

The ISWEL survey, which SES 2018 is based on, has not undertaken major changes in methods or definition of variables since 1998. Still, all processes and quality checks are under constant revision in order to improve the quality of the data and thus some changes has been made to the processing and editing of data. The economic activities C, F, G and H, in the private sector, are included since 1998, in 2002 part of J was included, in 2005 the section K was included, in 2008 sections D and E were included, and in 2018 economic activity I was added. Inclusion of the public sector is almost complete. The definition of full-time employment has changed a little since SES 2010, and is now more in line with the SES definition. Before, both normal hours and overtime hours where considered, but now only normal hours are included in the definition of full-time employment.

First data submission of SES was for the year 2002 but then only a few economic activities in the private sector ware included in the data. From 2006 and onwards data included both the private and public sector and more economic activities have been added over time as explained in chapter 8.2. Within economic activities comparison is comparable over time but total comparison should be done cautiously. 

Situation as of December 20, 2020 on Eurostat’s website. See attached document Coherence.

Domains and variables used:

Labour Cost Index (LCI)

Index on wages and salaries (LC_LCI_r2_q - D11 in I16)

When comparing data from SES to LCI it has to be kept in mind that the LCI is based on an administrative dataset that includes all employees and the divider is hours actually worked that are estimated by a statistical model.

Labour Force Survey (LFS)

Employment 15-74 (LFSQ_EGAN for Y15-74)

Average number of actual weekly hours in work of main job (LFSQ_EWHAN2 for employed persons)

When comparing data from SES to LFS it must be kept in mind that the survey designs are different. LFS the sample is drawn from the national register, stratified by gender and age. As the sample designs are different and the population of Iceland is relatively small (about 330,000) the datasets may differ.

Structural Business Statistics (SBS)

Personnel cost (1330)

Number of person employed (16130)

Data has been submitted but is not yet available on Eurostat’s website. Thus data is retrieved directly from the production unit at Statistics Iceland. When comparing data from SES to SBS it has to be kept in mind that the reference population is different as SBS refer to NACE B-N, except K, including S95.

National Accounts NA

Wages and salaries in million units of national currency (NAMA_10_GDP - D11 in CP_MNAC)

Total employment per year (NAMA_10_PE - SAL_NC)

When comparing data from SES to NA it has to be kept in mind that NA include the whole economy.

Statistics Iceland publishes every year data on average monthly earnings based on a full calendar year. This data is based on the same dataset as SES. From 2014-2018 the mean of total regular earnings (similar concept as monthly earnings in SES) rose from 519 thousand ISK to 674 thousand ISK for full time employees or an rise of 29.9%. Monthly hours paid for full time employees decreased from 184 to 183.7 a decrease of 0.2%

See chapter 8.3.

Data are final.

-

Eurostat publishes news releases based on the SES data and Statistics Iceland will inform its users with news releases based on Eurostat’s publications.

Eurostat publishes publications based on the SES data and Statistics Iceland will inform its users with news releases based on Eurostat’s publications.

Eusostat database.

Not applicable.

Via Eurostat’s microdata access or Statistics’ Iceland tailored statistics.

Not applicable.

The methodology follows the implementation arrangements published by Eurostat.

Not applicable.

Not applicable.

Not applicable.

All efforts are made to keep the cost and burden of the enterprises surveyed for SES 2018 at a minimum. As the data are based on the ISWEL survey, no additional data collection was necessary from enterprises apart from the regular monthly data collection related to ISWEL. In the ISWEL survey contracts have been finalised with Iceland's leading software firms on writing computer programs that fit the enterprises’ own wage software and allow companies to submit standardised and detailed information electronically to Statistics Iceland on wages, labour cost, paid hours and necessary background factors. Thus, the expense and nuisance companies associate with data collection and the time lost through reporting is kept to a minimum after the inclusion process is finished. During the inclusion process a specialist from Statistics Iceland visits the business units and assists with the ISCO classification and other classification matters in order to maintain coherence between enterprises. This is important, since a substantial number of business units are obliged to participate in the survey on a regular basis so as to maintain a representative sample.

-

Statistics Iceland works under the Act on Statistics Iceland and Official Statistics and the Act on Protection of Individuals with regard to the Processing of Personal Data. In addition Statistics Iceland follow the European Statistics Code of Practice and the UN Fundamental Principal of Official Statistics. For further information see Statistics Iceland’s website: https://statice.is/about-statistics-iceland/laws-and-regulations/

See Statistics Iceland’s Rules of Procedure for Treating Confidential Data. https://statice.is/about-statistics-iceland/laws-and-regulations/statistics-icelands-rules-of-procedure-for-treating-confidential-data/

-