Statistical Office of the Republic of Slovenia

Labour Market Department

Litostrojska 54, 1000 Ljubljana

With the Structure of Earnings Survey we wish to provide insight into the structure of earnings, i.e. to present different wage categories (earnings, bonuses, holiday bonus, paid hours, paid overtime, days of holiday leave, etc.) by different characteristic of business entities or their units (size class, activity, region, etc.) and by characteristic of employees (sex, education, occupation, etc.).

List of abbreviations: 
AJPES Agency of the Republic of Slovenia for Public Legal Records and Related Services
NR National Accounts
PRS Business Register
SES Structure of Earnings Survey
SRDAP Statistical Register of Employment
SURS Statistical Office of the Republic of Slovenia
ZAP/M Monthly Report on Paid Earnings by Legal Persons; national official data on earnings

Classifications used:

• The Standard Classification of Activities (SKD), based on the European Classification of Economic Activities (NACE Rev. 2);
• The Nomenclature of Territorial Units for Statistics (NUTS, level 1);
• The Standard Classification of Institutional Sectors (SKIS);
• The Standard Classification of Occupations, based on International Standard Classification of Occupations, (ISCO-08);
• Classification System of Education and Training (KLASIUS), based on the International Standard Classification of Education ISCED 1997 and ISCED 2011.

Data are collected for all activities A-S.

Key variables:

• Average annual / monthly / hourly earnings per person in paid employment by different wage categories, by different characteristics of business entities or their units and by characteristics of employees;

• Structure of earnings;

• Gender Pay Gap.

Local kind of activity unit (LKAU).

The Structure of Earnings Survey covered all persons in paid employment with employment contracts working for legal or natural persons or their units from the mentioned sample. The survey did not cover people working under work contracts and royalty, pupils and students on practice, workers in employment promotion schemes, posted workers and labour force hired through the student employment centre.

The whole country (Slovenia) is covered on NUTS-1 level.

In Slovenia SES data are available from 2002. For the SES2014 the reference period are data for the year 2014.

No based period. The SES statistical survey is conducted every 4 years.

Business entities (or their units) were randomly select and all employees in these business entities were observed (one stage sampling).

Sampling frame

As a basis for creating the sampling frame, data from SRDAP and PRS were used to which data from two statistical surveys – Monthly Report on Earnings Paid by Legal Persons and Monthly Report on Earnings Paid by Registered Natural Persons – and statuses from other statistical surveys were added. In this case data from the above mentioned surveys were used only as auxiliary data in determining large units that will be sampled with certainty. In the final sampling frame 76,241 units were included.

Size classes were determined on the basis of the data on the number of employees (source: SRDAP) as well as data on labour costs (source: final accounts estimated at the level of local units). Large business entities were those with 250 and more employees or with labour costs exceeding EUR 450,000. Other business entities were divided into three size classes: medium-sized (from 50 to 249 employees), small (from 10 to 49 employees) and micro (with fewer than 10 employees). As the second and third stratification variables activity (by the Standard Classification of Activities 2008 (NACE Rev. 2)) and cohesion regions were used.

Sample  

The final sample size for the SES2014 was 3,636 business entities or their units. Reporting units had to report data for all employees in the business entity or unit selected in the sample.

In the distribution of sample units by strata, optimum allocation by the number of employees was used, while in those strata where the calculation by optimum allocation yielded fewer than 8 units 8 units were sampled, i.e. if there were fewer than 8 units in the stratum, all units were sampled. Systematic sample selection was applied, sorting the units within each stratum by five-digit Standard Classification of Activities 2008 (NACE Rev. 2) codes and thus ensuring implicit stratification at the lowest level of activity.

Because in the case of the selected sampling plan it cannot be controlled for how many employees the unit will report data, it was decided to include control in the software application for reporting. If the business entities reported data for too low number of employees (according to data from SRDAP), the person reporting the data was warned.

Data are collected every 4 years.

For the Structure of Earnings Survey existing sources are used (Statistical Register of Employment - SRDAP, Business Register - PRS, data on income tax and final accounts) and data from the questionnaire sent to the business entities selected in the sample; these data are collected by the Agency of the Republic of Slovenia for Public Legal Records and Related Services (AJPES) by a web questionnaire.

The responding units were notified with a circular letter about being included in the sample, which was sent to them by mail or by post where mail was unknown. The circular letter contained the general information about the survey and obligation by Law of Statistics to fill the questionnaire and for which variables data are already collected from the existing sources. The guidelines and definitions were published on internet sites of SURS and AJPES.

The questionnaire was open for filling it in on 13 May 2015. The deadline for filling in the questionnaire was set to 5 June 2015, but the final deadline was then change to 3 July 2015. After the deadline non–responded business entities were notified by mail. In July and August just non-responded key-responders were contacted to send the data. The data collection ended in October 2015. 

From end of October 2015 to February 2016 data were analysed and some double checking with responding units were done. In March 2016 the data were weighted and basic tables were made. Through period from April until June 2016 other tables were produced and checked. First Release was published on 29 June 2016 as a provisional data. First release contained data on hourly, monthly and annual earnings per employee by activities, sex, main occupational and educational groups, age and size classes, including short methodological explanations. On the next day data were put in Eurostat's standard scheme and sent to Eurostat via Edamis.

Data collected with the questionnaire were validated trough set of logic controls. First logic control was built in application of data collection where mistakes were hard or soft (colored in red or yellow). Data could not be transferred with hard mistakes while for soft mistakes their values must be double-checked.

Data from the questionnaire and administrative data are compiled into the ORACLE database using PIN number of employee and local kind of activity unit.

Due to national specific earnings system data are adjusted to be in line with the regulation's requirements. Also separate tables with data for 1+ employee and 10+ employees are prepared.

Key quality indicators are calculated regularly.

Standard quality report is prepared for every SES survey.

The main users of the SES2014 results are the Institute of Macroeconomic Analyses and Development, Office for Equal Opportunities, the Bank of Slovenia, the Ministry of Labour, Family and Social Affairs, the Ministry of Finance, the Chamber of Commerce and Industry, the Employers' Association of Slovenia and trade unions. Important users are also units within the Statistical Office of the Republic of Slovenia (SURS). Other users of survey results are various research institutes, domestic and foreign companies, students and the media.

The most important national users are the Ministry of Labour, Family and Social Affairs (The International Cooperation and European Affairs Service), the Institute of Macroeconomic Analyses and Development and the Employers' Association of Slovenia.

Major foreign users of survey data are Eurostat, the European Central Bank, the International Monetary Fund and the International Labour Organisation. SES data also represent basis for many researchers dealing with earnings statistics.

From SES2014 anonymised data for Slovenia are also available on CD-ROM for researchers, prepared by Eurostat.

Structural data on earnings, including Gender Pay Gap, are one of the top data on earnings which are interesting not just for experts but also for general public. In general users are satisfied with data offered but two main weaknesses were mentioned; information on structure of earnings should be available more often than every four years (problem is solved by annual estimations, especially with annual Gender Pay Gap) and data availability which is almost one and half year after the reference period (for domestic users) or almost two years (for international comparison). With SES2014 final data were published T+24 months (for SES2010 T+31).

There is user service on the level of NSI. At least once per 1.5 year Statistical Advisory Committee is held, where main and other users can express their needs. Key users are treated individually on request.

In the SES surveys the articles of the EU Regulation regarding SES are fully implemented. Because holiday bonus and payments in kind (including payments for meal and for travel from/to work) are not part of the earnings system in Slovenia they are for Eurostat included in the earnings data but shown separately in national tables.

100%

Accuracy is described in 6.1 to 6.6.

See 6.2.1.

In the attached documents sampling errors are shown in percentage for variables 4.2 (Gross earnings in the reference month) and 4.3 (Average gross hourly earnings in the reference month) for different individual breakdowns.
In the tables from 1 to 5 coefficients of variations are calculated for business entities with 1 and more employees while in the tables from 6 to 10 for business entities with 10 and more employees. In all tables from 1 to 10 data are for all activities A-S.

See 6.3.1. - 6.3.5.

See 6.3.1.1. and 6.3.1.2.

There are no difference between the reference and the study population. For national purposes for business entities and employees in activities A and S data were also collected and published. Moreover also data for business entities with less than 10 employees were included.

On total there was 3.2% of over-coverage among business entities and 0.9% among employees. Higher over-coverage is the size class 'less than because of the increased number of business entities without paid wages mainly because of the crises. There were some responses from units which were not in the sample but they were in the frame. Under-coverage was not detected.

Not applicable.

Data were collected electronically by AJPES. Every question in the questionnaire contains the definition of what must be included in and excluded from the data. Methodology and definitions were published on SORS and AJPES internet sites. First logic control was built in application of data collection where mistakes were hard or soft (colored in red or yellow). Data could not be transferred with hard mistakes. Mistakes are described in the following table. In case of unit non-response reporting units were notified to send the data.

All hard mistakes detected through the system of logical controls were corrected by the companies themselves. Where a lot of soft mistakes occurred companies were contacted and data were double checked. Variables from existing sources were not controlled; they have their own checking system. Data from existing sources were put through code list to check possible miscoding.

See  6.3.3.1. and 6.3.3.2.

The overall unit response rate was 81.1%. Response rate concerning employees was 87.7%. In the sample key-responders were selected which represent units from which respond is necessary to obtain the sample quality. In case of unit non-response re-weighting was used.

For mandatory variables, including ‘Gross earnings in the reference month'; variable (4.2), data were expected while filling in the electronic questionnaire. This variable was not included in the logic control as hard mistake, because it is possible for employee not to have been paid in October. Missing data were checked at the business entities and corrected if necessary. No imputation was performed in case of item non-response.

See 6.3.4.1.

Most of the mandatory variables (those collected with the questionnaire) were included in the logic control and mostly of them were treated as hard mistakes in case of item non-response. Also the questionnaire could not be delivered if these variables were not filled in correctly. For those variables no imputation were performed, just statistical editing. For other mandatory variables missing values were found out in the latter stage of data analysing and imputation were performed. Overall imputation, editing and total rates were 1.0%, 3.3% and 2.1%, respectively.

No models were used.

Not applicable.

See link.

Provisional data were published on 29 June 2016 and final detailed data were published on 14 December 2016.

No differences between provisional and final data.

Data were published on 29 June 2016. Data were sent to Eurostat on 30 June 2016 which is 18 months after the reference period as it is stated in the EU regulation.

18 months after the reference year (29 June 2016).

23.5 months after the reference year (14 December 2016).

Data were published as planned.

0 days.

SES2014 data were collected in accordance with EU regulations with some exceptions listed below:

• Also business entities with less than 10 employees were included because of national purposes. In Slovenia there are many small business entities; by SES2014 sample frame 85.7% of all business entities, which represents 21.9% employees.
• Apprentices were excluded due to negligible phenomena and because units would face a problem filling the data.
• By the regulation payments paid by employer at a reduced rate are to be excluded. In Slovenia there are a lot of payments at a reduced rate because all sickness leave which is paid by employer (up to 30 days) is paid at a reduced rate (except in case of injuries at work). Therefore data on monthly and annual earnings were collected separately for total and payments at a reduced rate. For EU purposes payments at a reduced rate were deduct from total payments (the same procedure was applied for paid hours and paid hours at a reduced rate).
• Holiday  bonus in Slovenia is not treated as wage component. In tables for Eurostat holiday bonus was included but for national purposes holiday bonus was excluded from annual earnings data and shown separately.
• Wages in kind in Slovenia is not treated as wage component though wages in kind in Slovenia represent high share in total costs because of payments for travel between home and work. In Slovenia most of employees are entitled to receive payments for travel between home and work by different ways (e.g. as cash payments, bus or train tickets). In tables for Eurostat payments in kind were included but for national purposes payments in kind were excluded from annual earnings data and were shown separately.

Not applicable.

In comparison with previous SES almost all methods were the same. There were small changes in data collection regarding more logic controls.

In Slovenia SES is carried out from 2002.

Some variables, which were available from Tax Administration (annual wages, holiday bonus) were double checked.

Not applicable.

In some activities there are lower average annual wages and salaries per employee in SES2014 compared to NA data because in Slovenia in wages and salaries per employee in NA data also some payments are included which are not parts of wage system (e.g. retirement bonus, jubilee rewards). Data on holiday bonus and payments in kind (including payments for meals and payments for going to/from work) are also not part of wage system in Slovenia but were collected and included in SES 2014 Eurostat tables but excluded for national purposes. In the attached table holiday bonus and payments in kind are included in SES2014 data.

There reasons for the major differences between both surveys among earnings and employees by activities are from the fact that some groups of business entities or employees are included in NR but not in SES; unpaid family members in activity A and business entities without employees in activity L.

All variables containing earnings and paid hours data were checked with results from existing wage statistical surveys.

According to the SURS publishing policy, data were first published in First Release (29 June 2016) as provisional data, including explanations and short methodology.

More detailed results were published on 14 December 2016 in statistical database SISTAT.

See URL.

In case of special data inquiry special tables are made.

Individual data are also available in the safe room at SURS and from SES2014 also on Eurostat's CD-ROM for researchers.

To the reporting units no results will be sent, only on their explicit request.

Methodological Explanations are under major revision and SES 2014 Methodological Explanations will be revised by the end of the year 2017. For now Methodological Explanations for 2010 could be used (see link) because not many methodological changes occur between SES 2010 and SES 2014.

See link.

All metadata elements are in the methodological explanations.

No information on the event.

There were additional costs for SURS because of additional logic control and additional individual data protection both implemented in the electronic questionnaire. There is always burden on units selected in the sample, because data on earnings and paid hours are collected with the questionnaire. All other data are collected from the existing sources (see item 3.3.).

SURS treats all data of all our reporting units (households, persons and enterprises) confidential according to our National Statistics Act, other laws and internal acts. The dissemination of the data must not allow any identification of a reporting unit or disclosure of sensitive data, therefore the data are protected in accordance with the rules of statistical disclosure control. The access to the data is enabled only for the employees, who need the access due to the nature of their work.

Dissemination: SURS checks the data for the dissemination in terms of the risk of identification and sensitivity. Rules used for the determination of the primary sensitive data are threshold, dominance rule, p%-rule; for secondary suppression we use method Modular inside Tau-Argus, soft wear for a protection of tabular data. All links between tables are taken into account at applying SDC methods.

Microdata: Microdata are available only in the secure room or via remote access. All direct identifiers are removed from microdata sets and all outputs are checked by SURS' methodologists to prevent any identification of the unit or disclosure of sensitive data.

No comment.