Eurostat, the statistical office of the European Union

F3: Labour market

5, Rue Alphonse Weicker L-2721 Luxembourg

The Structure of Earnings Survey (SES) is a 4-yearly survey which provides EU-wide harmonised structural data on gross earnings, hours paid and annual days of paid holiday leave, which are collected every four years under Council Regulation (EC) No 530/1999 concerning structural statistics on earnings and on labour costs, and Commission Regulation (EC) No 1738/2005 amending Regulation (EC) No 1916/2000 as regards the definition and transmission of information on the structure of earnings. The objective of this legislation is so that National Statistical Institutes (NSIs) provide accurate and harmonised data on earnings in EU Member States and other countries for policy-making and research purposes.

The SES 2010 provides detailed and comparable information on relationships between the level of hourly, monthly and annual remuneration, personal characteristics of employees (sex, age, occupation, length of service, highest educational level attained, etc.) and their employer (economic activity, size and economic control of the enterprise).

Regional data is also available for some countries and regional metadata is identical to that provided for national data.

The Structure of Earnings Survey (SES) is a 4-yearly survey which provides EU-wide harmonised structural data on gross earnings, hours paid and annual days of paid holiday leave, which are collected every four years under Council Regulation (EC) No 530/1999 concerning structural statistics on earnings and on labour costs, and Commission Regulation (EC) No 1738/2005 amending Regulation (EC) No 1916/2000 as regards the definition and transmission of information on the structure of earnings. The objective of this legislation is so that National Statistical Institutes (NSIs) provide accurate and harmonised data on earnings in EU Member States and other countries for policy-making and research purposes.

The SES provides detailed and comparable information on relationships between the level of hourly, monthly and annual remuneration, personal characteristics of employees (sex, age, occupation, length of service, highest educational level attained, etc.) and their employer (economic activity, size and economic control of the enterprise).

Regional data is also available for some countries and regional metadata is identical to that provided for national data.

The "economic activity" is coded in NACE Rev. 2 (General industrial classification of economic activities within the European Communities) whereas the "occupation" is coded according to the Commission Recommendation of 29 October 2009 on the use of the International Standard Classification of Occupations (ISCO-08).

Information on the 'Highest successfully completed level of education and training' variable is classified using the International Standard Classification of Education, 1997 version (ISCED 97).

Regional breakdown is based on the Nomenclature of Territorial Units for Statistics (NUTS) reference year 2010. Detailed NUTS classifications are available for EU Member States as well as for EFTA and Candidate Countries.

The statistics cover all economic activities defined in NACE Rev. 2 sections B to S. NACE Section O (Public administration and defence; compulsory social security) is optional, however covered by most countries.

The enterprises included employ at least 10 employees and the size classes (corresponding to the number of employees) available are 10 to 49, 50 to 249, 250 to 499, 500 to 999 and more than 1 000. The size class of enterprises employing less than 10 employees (1 to 9) is optional and covered by some countries.

Employees are all persons who have a direct employment contract with the enterprise or local unit and receive remuneration, irrespective of the type of work performed, the number of hours worked (full or part-time) and the duration of the contract (fixed or indefinite).

Mean annual gross earnings also cover all 'non-standard payments', i.e. payments not occurring in each pay period, such as: 13th or 14th month payments, holiday bonuses, quarterly or annual company bonuses and annual payments in kind.

Mean monthly gross earnings in the reference month cover remuneration in cash paid before any tax deductions and social security contributions payable by wage earners and retained by the employer, and are restricted to gross earnings which are paid in each pay period during the reference month.

Mean hourly gross earnings are defined as gross earnings in the reference month divided by the number of hours paid during the same period.

Number of hours paid includes all normal and overtime hours worked and remunerated by the employer during the reference month. Hours not worked but nevertheless paid are counted as 'paid hours' (e.g. for annual leave, public holidays, paid sick leave, paid vocational training, paid special leave, etc.).

The compilation of structural statistics on earnings is based on local units and enterprises, as defined in Council Regulation (EEC) No 696/93, and provides information on employees in enterprises with 10 or more employees classified by size and economic activity. Information for employees in enterprises with fewer than 10 employees is optional.

The statistics cover all activities defined in NACE Rev. 2 sections B to S (excluding O) for enterprises with at least 10 employees.

The SES 2010 statistics refer to enterprises with at least 10 employees in the areas of economic activities defined by NACE Rev. 2 sections B to S excluding O.

The inclusion of NACE section O is optional, as well as the inclusion of enterprises with less than 10 employees.

The data cover EU-Member States, Turkey, the former Yugoslav Republic of Macedonia, Iceland, Norway and Switzerland.

EU aggregates are available for: EU27, EU25, EU15, EA17 and EA13.

2010.

Not applicable.

-

The data collection for LCS and SES can be obtained from 'tailor-made' questionnaires, existing surveys, administrative data or a combination of such sources, which provide the equivalent information. While accepting a degree of flexibility in the means employed for collecting the survey data, the information obtained must be of acceptable quality and be comparable between European countries.

Four-yearly.

The national surveys are generally conducted on the basis of a two-stage random sampling approach of enterprises or local units (first stage) and employees (second stage).

Data validation consists of global checks and plausibility checks. Global checks are necessary to ensure that complete data is received for microdata records. For each country, all microdata records should contain data for all mandatory variables. Missing mandatory data or codes are not accepted. Concerning optional variables each country decides which of these it is able to supply. Furthermore, plausibility checks on all variables were done to ensure that the data are reasonable and consistent with other variables. Possible deviations are reported by countries in their national Quality Report transmitted to Eurostat. For further details on the plausibility checks implemented by Eurostat are provided in the implementing arrangements document which is approved and sent to all countries prior to each round (Validation level 2).

EU aggregated are compiled, with the number of employees per country being the weighting factor for each individual country.

Not applicable.

According to Regulation (EC) 530/1999 national authorities shall ensure that the results reflect the true situation of the total population of units with a sufficient degree of representativity. National authorities are therefore obliged to provide a Quality Report containing all relevant information to enable the quality of the statistics to be evaluated.

Refer to SES2010 synthesis and national quality reports in annex.

Upon transmission to Eurostat, SES microdata are checked for completeness and consistency.

-

Among others, the most important and frequent users of SES are mainly; research centres, universities and students, the media, social partners and trade unions, private companies, national public institutions as well as international institutions. The large sample size of the SES makes it a unique source of information in which individual earnings can be linked with the characteristics of individual employees (sex, age, education level etc.) as well as to the characteristics of the enterprise they work for (economic sector, size of the enterprise, location etc.).

The SES tables published on Eurostat’s website are considered to be well followed by our users as between October 2012 and April 2013, the number of hits associated to the SES2010 (earn_ses10) datasets recorded an average of 1500 hits each month. Datasets providing information on earnings by sex, economic activity and collective pay agreement as well as earnings in quantiles are the most looked for, with information on earnings by level of education and geographic location to a lower extent.

Refer to national quality reports link in annex.

[Not requested]

-

In the majority of the countries, a two-stage stratified sample technique is adopted; first a random sample of enterprises / local units, followed by a sample of employees within the selected enterprise / local unit. 

For further details, refer to national quality reports in annex.

Refer to national quality reports in annex.

Refer to national quality reports.

Refer to national quality reports in annex.

BE:     No differences between the reference and study population can be mentioned.

BG:     Under-/Over-coverage occurred due to newly emerging companies or closing down of companies (respectively) at the time of drawing the sample.

CZ:      Coverage errors have been eliminated by the new system of data collection.

DK:     The continuous updating of the business register prevents particular problems in acquiring close to complete coverage.

DE:      Over coverage (6.3%) is due to inactive enterprises. Employees with extreme high income are under covered.NACE Section P covers only 70% as most non-public employers have less than 10 employees.

EE:      The sample frame (updated in 2009) contain an element of under-coverage due to delay between sample selection and data collection and some over-coverage due to new-born enterprises after 2009.

IE:       Due to the fact that only enterprises with three employees or more were selected, some professional sectors such as doctors, solicitors etc. may be under represented.

EL:       No information has been documented.

ES:      Employees are identified by their affiliation number within the local unit during the reference year. The problem is with apprentices due to their particular type of contract which are registered under a different type of affiliation register with different characteristics than the general file for employees.

FR:      Enterprises with less than 10 employees have not been covered by the survey.

IT:       An update of the list of population has been made in order to take into account; new units which were either left out of scope (1%), others which were added wrongly due to wrong address (4.7%) and others which had change of status (2.5%) in the sample.

CY:     Coverage errors occur due to misclassifications (incorrect classification of units that belong to the target population); under coverage (new-born enterprises or enterprises which are excluded from the sample) and over-coverage due to duplications of units or others which from sampling to data collection closed down or became inactive.

LV:      The only reason for under-coverage (2%) is due to the time lag of one year between when the sampling frame is drawn and when the actual sampling was done.

LT:      Under and over-coverage were assumed to be negligible since the sampling frame was constructed at the end of the reference period.

LU:      Whereas no problem of under-coverage is known, over-coverage stem from a discrepancy between the administrative files used for sampling and the real world.

HU:     Over-coverage may happen due to misclassification of the number of employees by bands whereas under-coverage is the result of births and mergers of new units.

MT:     Coverage errors occur due to misclassification of NACE or size class assigned to the units in the target population, and over-coverage errors of units which were included in the sample when it was drawn but which were no longer active at reference period.

NL:      The time lag in updating the register may cause minor elements of misclassification, under-coverage or over-coverage which do not influence SES outcomes.

AT:      The low rate of over coverage is due to inactive enterprises or employees not working anymore for the enterprise or not having a salary in the reference month. Under-coverage is due to exclusion of enterprises in NACE O to R from the sample.

PL:      Over coverage errors (1.9% of selected sample) relate to units which are present in the frame at time of sampling but which do not belong to the target population or do not exist in practice. Under coverage elements occur due to birth of new units between time of sampling and actual data collection.         

PT:      Due to the fact that the sampling frame is continuously updated, coverage errors for the private sector have no meaning whereas public sector units do not have considerable changes in the short run.

RO:     The main over / under coverage problems are related to the information quality concerning size class of enterprise by number of employees.

SI:       On total there was 5.5% (among business entities) and 1.9% (among employees) of over coverage. Under coverage was not detected.

SK:      The coverage errors were caused by inaccurate data which were provided by the statistical units to the Registers of organisations and establishments.

FI:       Since the survey frame refers to the middle of the reference year, some non-existing enterprises were included whereas some new / growing enterprises were not included at time of collection, reference month being October.

SE:      The reference month surveyed is September 2010, which is considered to better reflect the target population.

UK:     Areas of under coverage are explored because of some small number of low-paid jobs that do not operate PAYE scheme, and which are hence not on the PAYE register.

IS:       Despite the fact that the ISWEL does not yet include economic activities I, L, M, N and S, full economic coverage may lack precision. As the ISWEL is based on the “Pay As You Earn” (PAYE) register, errors in NACE classification of the latter can have an impact on the ISWEL survey coverage.

NO:     Errors in stratification variables, NACE activity and number of employees in the frame population could be a source of error. In order to control this potential error, local units in the sample are asked to control to pre-printed code of activity on the form and correct it if this is believed to be incorrect.

CH:     A limited influence of under-coverage exists with regard to economic activity, size of enterprise and specific categories of employees.

HR:     No information has been documented.

MK:    In order to avoid possible rate of over / under coverage, which may happen because of births, deaths, mergers and de-mergers of old units within the selected sample, the latest version of the sampling frame was used.

TR:      Over coverage occurred due to dead units (44%), units without any employees (45%), NACE section being out of scope (1.6%) and other reasons (9.3%). Information on under coverage is not possible to obtain since no external sources are used to compare the frame information.

See coverage errors (5.3.1)

[Not requested]

BE:      Several aggregated and inconsistency checks were integrated in the data collection tools, which were solved by following them up with the local units concerned.

BG:     While the electronic version of the questionnaire was facilitated with integrated data validation and plausibility checks and dialog boxes in order to facilitate respondents’ life, respondent units were directly contacted when issues of completeness, compliance and consistency arose.

CZ:      Data was checked at the data entry point when users had to install special software on their computers. In case of any mistakes, data providers were contacted directly via telephone one e-mail. Further checks are done within the NSI during data processing.

DK:     When inadequate and erroneous data is detected, enterprises are contacted.

DE:      Errors were minimised since the questionnaires were well organized and supported with explanations. The use of a social security key also avoided wrong and improper coding of ISCO and ISCED.

EE:      Logic tests were applied in order to identify all errors of magnitude. The variables which needed most corrections were the occupation code, overtime hours, holiday leave days in October, days not worked but paid, earnings in relation to overtime hours and number of hours worked and paid days to which the gross annual earnings relate. These errors were followed up with respondents and corrections were made.

IE:       Measurement errors are not applicable are information is acquired from administrative sourced which are corrected at source.

EL:       No information has been documented.

ES:      Questionnaires undergo a series of error-detecting checks with more than 400 rules.

FR:      A series of logical controls have been integrated in in-built software that monitors the incoming data. This software ensures data consistency (detecting outliers) and orders of magnitude before data is sent for further checking by Eurostat.

IT:       In addition to the structured scheme of controls established in the data collection process (web questionnaire) in order to avoid missing mandatory information, further logical as well as data processing controls were applied.

CY:     In addition to explanatory notes which supported the questionnaire, data was collected by trained interviewers to minimise errors. Nevertheless consistency checks were designed to identify any inconsistencies in the data.

LV:      Validation programs including a series of arithmetical and logical controls were established in addition to detailed instructions supplemented with the questionnaire. 

LT:      75% of respondents sent their data through electronic questionnaire which was validated by statisticians applying arithmetical and logical controls.

LU:      A minor amount of measurement errors were detected, which nonetheless were followed-up and corrected directly with the local units concerned.

HU:     Measurement errors are checked through logical checks and amended through a lengthy and thorough process.

MT:     Incoming questionnaires were checked thoroughly by trained statisticians using a number of validations and consistency checks.

NL:      For grossing up of the SES, data from the Annual Survey on Employment and Earnings (ASEE) was used.

AT:      In addition to detailed explanatory notes attached to the questionnaire, the web-questionnaire included plausibility controls. A hotline service was also established.

PL:      Detailed explanatory notes are attached to the questionnaire in order to increase clarity. Errors consist mainly of misunderstanding / misinterpretation of questions from the respondents’ side and wrong figures inputted in the computer system which are identified and corrected through arithmetical and logical controls.

PT:      No major errors exist as enterprises are used to this survey and have good knowledge of the classifications involved. The electronic survey did not allow for non-response and hence the imputation rate for demographic variables and wages is zero.

RO:      In addition to detailed explanatory notes annexed to the questionnaires, other IT applications were established for further checking to identify any kind of errors. Certain variables were also compared to data from other sources. Plausibility checks also followed and where necessary, errors were corrected.

SI:       Hard mistakes were detected and corrected by the companies themselves, soft mistakes were followed up and data was double checked with the same companies.

SK:     The data are evaluated and revised on the basis of global and plausibility checks in accordance with Eurostat’s implementing arrangements.

FI:       Validation has been mainly based on imputation of missing or conflicting variables, which share has been insignificant.

SE:      Besides the questionnaire, respondents receive guidance with explanations including FAQs and contact information for further help. All data has been validated through logical tests and respondents were contacted to validate errors and correct data.

UK:     Missing data for key variables are imputed on the basis of shared characteristics with imputation ‘donors’. A range of validation cheques are applied to identify potential errors in the data collected.

IS:       Non-response errors can occur due to technical errors, when certain business units are unable to provide data, or human errors, in which case transfers are eliminated from the dataset.

NO:     Measurement errors which may arise due to lack of information or difficulty from the respondents’ side in calculating a particular value are identified and corrected by logical automated computer controls as well as manual checks for outliers.

CH:     Following first-hand manual controls, plausibility of data was further checked by means of electronic tools.

HR:     No information has been documented.

MK:    Since the SES was conducted for the first time, most errors (in particular for monthly and annual earnings) were followed-up and corrected in the post-data collection phase.

TR:      Incoming data was validated through consistency checks and further checked against Eurostat’s validation rules.

See measurement errors (5.3.3)

[Not requested]

[Not requested]

See measurement errors (5.3.3)

See national Quality Reports.

No assumptions were made or modelling information was given for the non-listed countries.

DK:     Since earnings data are collected for a full year, some assumptions are made: An employee switching from one local unit to another, half-way through the year is double counted working as 26.07 weeks instead of 52.14 (once). In this case, the employee should not be taken into account when comparing data across countries; The number of hours paid, earnings related to overtime and special payments for shift work in the reference month are calculated from the earnings register in which yearly data is assumed to be equally divided throughout all the months of the year; Full-time (≥29.91) and part-time employees (<29.91) are distinguished on the basis of their hours worked; With regard to information on collective pay agreement, whereas most enterprises in Denmark are members of either the Confederation of Danish employers or the Danish employers’ association for the financial sector, of which all are covered by a collective agreement at an industry level (B), the remaining ones are subject to further estimations as reported in the detailed national country report.

DE:      For NACE O and P data corresponds to June 2010 except that gross monthly earnings have been estimated for October (reference month). Taxes and contributions have been imputed. Overtime and shift work have been filled as “0” because they could not be derived or calculated with adequate quality. The size class “1000+” has been assumed as there are only employees within the public service.

EE:      The main error is probably made by assuming that the distribution of non-respondents is similar to that of respondents.

IE:       The NES 2010 hours worked were unchanged to 2009, but the earnings have been adjusted to follow Revenue Commissioners income trends.

CY:     No imputations were employed but the grossing-up factors were adjusted in order to correct for unit non-response.

LV:      All local units within same stratum were assigned equal design weights and which weights were adjusted using the response level in each stratum.

SE:      Reference month in Sweden is September, which is considered to be a representative month since it does not include public holidays or lot of absences due to vacation.

UK:     Since respondents in higher earning occupations are less likely to respond, a weighting system based on 108 weighting classes is applied.

NO:     In addition to October, the month of September is also chosen as a reference month as both are considered to be stable in terms of wages and less affective by holidays.   

[Not requested]

Apart from adjustments following internal checks, the data are accepted directly as communicated by the NSI. Revisions only occur rarely.

[Not requested]

If necessary, after running a series of data validation checks, countries are asked to revise their data until it is considered fit for publishing.

[Not requested]

-

Refer to national quality reports in annex.

[Not requested]

[Not requested]

Refer to national quality reports in annex.

[Not requested]

Coherence of statistics is the extent to which they can be reliably combined in different ways and for various uses. It is, however, generally easier to identify cases of incoherence than to prove coherence.

-

Comparability of the SES data across national borders may be affected by the use of different observation units and definitions, methods or classification schemes, i.e. by deviations between national and community concepts. Geographic comparability may also be affected by new regional classification (NUTS).

For an overview on geographical comparability, refer to the synthesis of quality reports in annex.

[Not requested]

Comparability over time may be affected by new definitions and classifications used in coding the SES data (NACE, NUTS, ISCED, ISCO).

For an overview on comparability over time, refer to the synthesis of quality reports in annex.

[Not requested]

For further details refer to the synthesis of quality reports in annex.

[Not requested]

[Not requested]

For further details refer to the synthesis of quality reports in annex.

-

News releases on-line.

See Eurostat website. 

Please consult free data on-line or refer to contact details.

[Not requested]

The conditions for SES microdata access are stated in the Regulation (EC) No 1104/2006 of 18 July 2006 amending Regulation (EC) No 831/2002 implementing Council Regulation (EC) No 322/97 on Community Statistics, concerning access to confidential data for scientific purposes. For details see Access to SES microdata.

Not applicable.

SES 2010: Eurostat's arrangements for implementing the Council Regulation 530/1999, the Commission Regulations 1916/2000 and 1738/2005

For the SES 2010 synthesis and national quality reports, refer to link in annex.

[Not requested]

[Not requested]

Not available.

-

Regulation (EC) No 223/2009 on European statistics (recital 24 and Article 20(4)) of 11 March 2009 (OJ L 87, p. 164), stipulates the need to establish common principles and guidelines ensuring the confidentiality of data used for the production of European statistics and the access to those confidential data with due account for technical developments and the requirements of users in a democratic society.

In the SES, information about individual entities (employees and enterprises) is collected. The safety of these microdata has to be guaranteed to make sure that individual entities cannot be recognised through inspection of released data. The goal of disclosure control is to disseminate statistical information in such a way that individual information is sufficiently protected against recognition of the subjects to which it refers, while at the same time providing as much information as possible.

Only tabular data have therefore been published. In order to limit the disclosure risk of these tables the following measures have been applied:

• Region: restricted to the national level;
• Economic activity: restricted to NACE Rev. 2, one digit level;
• Size of the enterprise: published in size classes of employee numbers;
• Age: restricted to 5 size classes (less than 30 years, 30-39 years, 40-49 years, 50-59 years, 60+ years);
• Occupation published at ISCO-08, one digit level.

The anonymisation method for SES consists of two primary confidentiality rules:

- minimum frequency rule;

- dominance rule

Secondary confidentiality rules are applied as additional protection to protect data from recalculation.

-