ESS Standard for Quality Reports Structure (ESQRS)
National Statistical Institute
|Contact organisation unit|
"Labour Statistics" Department
|Contact person function|
head of unit
|Contact mail address|
2, P. Volov Str.; 1038 Sofia, Bulgaria
|Contact email address|
|Contact phone number|
|Contact fax number|
The Structure of Earnings Survey is aimed to give detailed and comparable at European Union level information on distribution and relationships between the level of remuneration, individual characteristics of employees and these of their employer.
Objects of the survey are: characteristics of employer (enterprise, local unit) - number of employees in the local unit, type of ownership, existence and type of collective pay agreement, size of the enterprise; individual characteristics of employees - age, sex, educational level, occupation, length of service, mode of employment (full-time/part-time), type of employment contract, annual gross earnings, annual bonuses, annual payments in kind, annual days of paid holiday leave, monthly gross earnings, earnings related to overtime, earnings related to shift work, employee's compulsory social security contributions and income tax, number of paid monthly hours, number of paid overtime hours.
· Classification of Economic Activities (CEA-2008, for international use NACE.BG 2008);
· National Classification of Occupations and Duties-2011 (NCOD-2011) - Data were collected by the NCOD-2005 as it was the acting classification in 2010 and after that recoded into NCOD-2001 to be in compliance with international classification ISCO-08, in accordance with the requirements of the Commission Regulation N1738/2005.
· Nomenclature of Educational Levels - in compliance with the International Standard Classification of Education, 1997 version (ISCED 1997);
· Classification of Territorial Units for Statistical Purposes in Bulgaria - NUTS.
Enterprises with 1 or more employees in economic activity within sections B to S of NACE.BG-2008 are covered.
|Statistical concepts and definitions|
Employees are all persons who have a direct employment contract with their employer and receive remuneration in cash or in kind for certain quality and quantity of work done, irrespective of the type of work performed, the number of working hours (full or part-time) and the duration of the employment contract (fixed or indefinite).
Gross earnings are the remuneration in cash paid to the employee directly and regularly by the employer at the time of each pay period, before deductions of any tax and social security contributions payable by employee and withheld by the employer.
The gross monthly earnings of employees include:
The annual gross earnings are the total amount of the regular payments in cash received by the employee for the work performed during the reference year, including:
· the value of annual payments in kind (goods and services) made available to employees by employer;
· all irregular payments as quarterly bonuses, 13th or 14th salaries and other gratuities not received at each pay period.
Paid hours cover the total number of normal and overtime hours to which the gross monthly earnings in the reference month relate. The number of paid hours includes: actually worked normal hours, worked and paid overtime hours, hours not worked but nevertheless paid by the employer at a full rate (annual leave, work stoppages and other hours paid such as for medical examinations).
Local units (territorial structures) with 1 or more employees belonging to enterprises with 1 or more employees.
The structure of earnings statistics relates to enterprises with at least one employee in economic activities within sections B to S of NACE.BG-2008, including section O "Public administration".
Area of Republic of Bulgaria.
2002, 2006, 2010,2014
Source of the data is a sampling statistical survey. The sampling procedure used for the SES contains two stages. In the first stage, a stratified random sample of local units without replacement is drawn. Stratification criteria used include:
· economic activity – divisions (2-digit level) of NACE.BG 2008;
· the number of employees in the local unit: 1 to 9 employees; 10 to 49 employees; 50 to 249 employees; 250 to 499 employees; 500 to 999 employees; 1000 and more employees;
· regional breakdown – level 1 of national Classification of Territorial Units for Statistical Purposes in Bulgaria, in force since 2009:
- BG3 - Severna i Yugoiztochna Bulgaria
- BG4 – Yugozapadna i Yuzhna Tsentralna Bulgaria.
At the second stage, a systematic sample of employees is taken within each of the selected local units.
At the first sampling stage 18 684 local units were selected (9.0 form total population) from which 16 429 responded units (88% from the sample) provided data for approximately 200 700 employees (9.1% from total population).
|Frequency of data collection|
Once per four years.
Data are collected through tailor-made questionnaire which consists of: part A collecting information for sampled local units and part B collecting information for each sampled employee. The questionnaire is accompanied by a dispatch note to each respondent about the purposes of the survey and explanatory notes for sampling of employees within the local units and instructions on information required. The SES questionnaire was developed on paper and electronic format, uploaded on the official web page of NSI. The Head office of NSI provides methodological assistance to the respondents and to the Regional Statistical Offices (RSOs) on completion and processing of the information.
Approximately 100 controls have been applied on micro data for validation of: completeness, plausible values, logical and arithmetic coherence between collected variables. Data are checked at three levels: (1) during the initial data entry - in the electronic questionnaire by respondents and by the RSOs from paper questionnaires. (2) at the RSOs on the joint regional micro data set; (3) at NSI level - on the joint national micro data set. Besides micro data validation, survey results are compared with other sources of similar type of information.
The data processing goes through the following stages:
· entry of the initial information into electronic format;
· data validation;
· data editing and imputation on the base of additional information from respondents and/or other statistical and administrative sources;
· weighting of the sampling data to gross-up results over the total surveyed population;
· producing of summarized table results.
The software used for data processing:
· completion and processing of the individual data by respondents - on-line based questionnaire;
· integration of individual files into a database, validation and export of outputs from the database by different dimensions, checks and formatting of the macro data to be sent to Eurostat – MS Access;
· processing of the integrated national database, table outputs and analyses – SPSS.
According to Article 2, Para 3 of the Bulgarian Law on Statistics statistical information shall be produced in compliance with the following criteria for quality: adequacy, accuracy, timeliness, punctuality, accessibility and clarity, comparability and logical consistency.
According art. 10 of Council Regulation 530/1999 the national authorities shall ensure that the results reflect the true situation of the total population of units with a sufficient degree of representativity. The national authorities submit to Eurostat at its request after each reference period a report to enable the quality of the statistics to be evaluated.
According to the Commission Regulation 698/2006, having regard to Council Regulation (EC) No 530/1999, each Member State shall prepare quality report for evaluation of the quality of structure of earnings statistics at the latest 24 months after the end of the reference period.
The user groups are defined on the base of the data requests received by NSI. The customers of the SES results can be classified as follows:
· National institutions - Ministries, Agencies, Councils, other governmental bodies and public;
· International institutions - Eurostat, ILO, OECD, UNICEF, UNECE;
· Social partners: Trade unions and employment associations
· Private institutions and businesses, incl. media
· Researchers and students
Internal to NSI: other units of NSI, e.g. dealing with LFS, Classifications, National Accounts, Household Budgets etc. for purpose of comparison or other.
NSI has not carried out a specific survey among users to know their needs of information concerning SES and whether they are satisfied with the published results. Users usually prefer more detailed data at the lowest levels of the classifications applied in the survey which is problematic due to the limited number of observations and correspondingly the lower level of precision.
The survey covers all mandatory variables according to the Commission Regulation (EC) No 1738/2005. There is full coverage as well in terms of size of the enterprises (with 1+ employees) and of economic activities (NACE.BG 2008 sections B - S, including O).
|Data completeness - rate|
The survey covers all (100%) mandatory variables according to the Commission Regulation (EC) No 1738/2005 and 5 out of 10 optional variables.
|Accuracy and reliability|
The overall accuracy of the survey results depends on:
· number of the surveyed units
To achieve a certain desired accuracy of the survey results a sampling plan of employees and local units is made. First, the total number of persons who should be observed is calculated. The calculations are made with 95% confidence level that the maximum error of the estimate shall be within a preset interval. The resulting number of persons is distributed proportionally to the population by three stratification criteria: the size of the local unit of economic activity and territory (location) of the local unit. Based on the parameters of the population and proportionally distributed number of employees that need to be observed, it is calculated how many local units must be selected from each cell.
· the survey framework
To construct the framework population from which the sample survey is to be selected data for the local units and the number of employees from the comprehensive Annual survey of employees, hours worked, wages and salaries and other labor costs for 2013 were used. For purposes of grossing-up procedures of the sampling data, the parameters of the population are updated with information for 2014, which is available approximately 12 months after the reference period.
· survey tools
The survey questionnaire is developed on paper and electronic format. The electronic questionnaire is an on-line based with incorporated logical controls allowing data to be validated while entered.
· methods of identifying and addressing possible errors
Approximately 100 checks are applied to verify data concerning: completeness of responses, data plausibility, arithmetic and logical consistency between the collected variables. Data editing is done by: a reference back to the persons filled information, use of information from administrative sources (Personal register of insured persons of the National Social Security Institute), use of other statistical surveys containing information about the surveyed units, application of statistical methods and techniques for the assessment of missing values (mean value imputation, the most frequent value imputation, etc.).
Coefficients of variation (relative standard errors) are calculated by use of the Horvitz-Tompson estimator. Coefficients of variation (CVs) are low for most of the relevant items and important classification levels.
Coefficients of variations by working time schedule and gender - %
The highest CVs appeared for small heterogeneous populations with low sampling probability and in cases of high unit non-response rate (small number of observations).
The following criteria were agreed for data publishing:
· cells with earnings (monthly/hourly/annual) showing CV between 20 and 30% were put in parenthesis;
· cells with earnings (monthly/hourly/annual) showing CV higher than 30% were hidden (deleted) and marked with sloped forward dash ‘/ ‘;
· cells with number of observations between 4 and 9 were put in parenthesis;
· cells with number of observations between 1 and 3 were hidden (deleted) and marked with sloped forward dash ‘/ ‘.
Extreme values were removed from the dataset and grossing-up factors were recalculated. From calculations were excluded 0.1% of records with the highest and 0.1% of records with the lowest hourly earnings in each division from 05 to 96 of NACE.BG 2008.
|Sampling errors - indicators|
Coefficients of variations by NACE Rev. 2 sections -%
Coefficients of variations by occupations (ISCO-08, 1-digit level) -%
Coefficients of variations by age groups -%
Non-sampling errors are described in details in the Quality Report where assessment is made of:
· over-coverage - the percentage of units covered in the survey that are out of scope of the target population;
· measurement errors - from the wrong interpretation of the survey questionnaire and respondents’ errors when not complying with explanatory notes for completion of the questionnaire - measured through the percentage of the wrong cases identified by the applied arithmetic and logic controls.
· the relative share of the non-response from sampled units is 12%. The non-responded units are grouped in two basic groups by reasons: (a) because of lack of up-to-date framework at the time of sample selection (~10.8%) - restructured, closed or units without activity during reference period; lack of contact; (b) units refused to respond (~1.2%). To neutralize the errors resulting from the lack of response from some of the units in the sample, the weights are recalculated with the number of respondents units.
· percentage of non-responses of individual variables or individual employees in the observation unit;
· percentage of corrected cases of key variables.
The sample of local units was taken from the local units’ population as of 31.12.2013. The sampling frame represented the most current situation of the Business Register available at the time of the sampling. In the sampling frame population were included all local units with 1 or more employees that belonged to enterprises with 1 or more employees within the NACE Rev. 2 sections B to S, including O.
The under-coverage refers to the situation when newly emerged or units with renewed activity with 1 or more employees within NACE sections B to S were not included in the sampling frame. The under-coverage was not quantified. To offset the errors that might arise from under-coverage and for purposes of the weighting procedure the framework population was updated where appropriate with the most recent situation of Business Register in 2014 to reflect major changes and fluctuations between NACE divisions and size classes of enterprises.
As over-coverage are referred sampled local units that during the reference period have been already closed-down, dormant units or units without employees. The overall over-coverage rate is 8.5%. When there have been cases of over-coverage, new units have not been sampled.
|Over-coverage - rate|
|Common units - proportion|
To avoid measurement errors detailed explanatory notes with illustrative examples were attached to the questionnaire. To further help the respondents a list with contact information was posted on Internet and telephone consultations on methodological and technical issues were provided. The Regional Offices were also provided by the Head office of NSI with written and telephone guidance how to process data and deal with arising problems.
Main sources of measurement and processing errors are:
•way of asking questions in the survey questionnaire. E.g. Annual days of holiday leave - in some cases respondents provided the number of days actually taken not the total number of days due to be taken.
•respondents keep data differently and do not make further efforts to comply to statistical requirements, or do not understand or read the explanatory notes. Example for such errors is var. 3.2 which is among most corrected items (10.9% of cases) because instead of number of hours paid during the representative month respondents provided: paid days during the month; paid hours during the year; working hours per day; paid hours excluding paid overtime hours (when available).
•data entry errors - these errors had very low proportion compared to the first two types.
|Non response error|
The overall unit response rate for in-scope respondents with 1 or more employees is 96.1% and for the mandatory size class of enterprises with 10 or more employees the response rate is 97.9%. The lowest is the unit response rate for small units with 1 to 9 employees - 95.3% and the rate to the total sample is even lower - 84.6%. The lower response rates for the enterprises with 1 to 9 employees could be explained with their dynamic nature featuring with frequent structural changes and instability as regards location, economic activity, financial status and employment - peculiarities for which it is difficult to maintain up to date information in the business register. The main reasons reported by the regional offices of NSI for the high non-response levels are rather “not found (out of date contact information)”, “closed down/sleeping”, “no employees in the reference period” than explicit refusals. Regional offices reported that nearly 48% of respondents were reminded for their duty to reply by phone calls, e-mail and follow-up letters. In the official period of data collection (May - June 2015) only 52% of responses were received. To improve response rate the deadline was prolonged with two months. Reminders were sent to the non-respondent units as special attention was paid to cells (NUTSxNACExSize) with low response rates.
|Unit non-response - rate|
In the table below are presented two types of unit response rates - the first one calculated to the total number of sampled units and the second one calculated to the total number of in-scope respondents (enterprises with one or more employees with earnings in October 2014). Rates are broken down by divisions of NACE Rev. 2 and by size classes - 1 or more employees, 1 to 9 employees (optional), 10 or more employees (mandatory).
|Item non-response - rate|
Normally, no item non-response (blank or zero values) has been accepted for any of the key variables. Only eight of the collected items could possibly be zero and therefore item non-response could be supposed.
The evaluation of quality at regional level was done by virtue of a questionnaire concerning number of issues. As regards measurement and processing errors Regional Offices were asked which variables have been most often corrected - wrong or missing. In the following table are listed variables that were reported by the 28 Regional offices (ROs) of NSI as being most problematic.
SES2014 variables most often corrected by the 28 Regional offices of NSI
In addition ROs reported that approximately 27% of responded units were contacted for reference on completeness, compliance and consistency of the data.
Methods applied for correction of data that were identified as wrong (inconsistent, impossible values, missing values, not corresponding to definition, wrong format) differ depending on the type, seriousness of error and willingness of respondents to cooperate:
|Imputation - rate|
In the following table are presented number of imputed cases and rate of imputed item non-response for the variables allowing blank or zero values.
Item imputation rates
In addition, 2919 employees’ records (1.48% of all cases) were imputed for some of the local units that provided data for significantly less employees than required or provided data were completely unreliable (e.g. one record duplicated number of times). As main data donor was used the Register of insured persons that contains many of key variables.
|Model assumption error|
|Data revision - policy|
|Data revision - practice|
|Data revision - average size|
|Timeliness and punctuality|
Publication of survey results at national level: august 2016
Submission of micro data to Eurostat: 13 July 2016
Issue of paper and electronic publication with detailed survey results: December 2016
|Time lag - first results|
|Time lag - final results|
Micro data were sent to Eurostat on 13 July 2016.
|Punctuality - delivery and publication|
13 days delay
|Coherence and comparability|
|Comparability - geographical|
The national and regional data (level 1 - statistical zones) are conformed to the acting Classification of territorial units for statistical purposes which is application of the European classification NUTS.
|Asymmetry for mirror flows statistics - coefficient|
|Comparability - over time|
The comparability over time is influenced mainly by changes in definitions and classifications as result of amendments of Community legislation as well as by change in coverage of enterprises.
|Length of comparable time series|
|Coherence - cross domain|
There are other statistical sources that produce information on number of employees, earnings and working time:
The results from the 2014 Structure of Earnings Survey are relatively comparable with from the above-sited surveys due to the methodological characteristics of each information source in regards to: goals, unit of observation, definitions, coverage of economic activities, statistical methods used for collection and estimation of variables surveyed.
|Coherence - sub annual and annual statistics|
|Coherence - National Accounts|
Values of annual earnings from SES for total B to S and in almost all of the NACE sections are lower than the corresponding values of wages and salaries from NA, except for section C. Although the definitions of the compared variables are similar the two sources have many methodological and conceptual differences that explain the disparities in levels of earnings.
Among the main reasons are:
• inclusion in 'Wages and salaries' of some kinds of imputed social contributions like: guaranteed remuneration in event of sickness that are paid by employer at a reduced rate (70%); compensations paid to dismissed workers (severance pay and compensation in lieu of notice);
• coverage by NA of employees working under non-labour contract (civil) and their earning in the form of fees and commissions;
• inclusions by NA of more types of wages and salaries in kind than in SES: part of daily allowances for business travelling; provision of recreation or holiday facilities for employees and their families;
• NA make adjustments for exhaustiveness on the following components: tips are estimated for activities like restaurants, bars, transport and other service activities (hairdressing and other beauty treatment); non-reported wages for employees working without any contract (informal employment).
|Coherence - internal|
Indicators within the data set are internally coherent.
|Accessibility and clarity|
Specialized detailed publication "Structure of Earnings 2014" - on paper and electronic format.
Detailed results are available to all users of the NSI website under the heading Labour Market - Structural (four yearly) statistics on earnings and labour costs - Structure of Earnings - national level, 4-year periodicity: http://www.nsi.bg/en/node/6520
Information System INFOSTAT: https://infostat.nsi.bg/infostat/pages/module.jsf?x_2=95
|Data tables - consultations|
Access to the anonymised micro data is granted according to the Rules for granting access to anonymised micro-data for scientific and research purposes set by NSI.
|Metadata - consultations|
200 to 400 in the data collection period (may-august)
|Documentation on methodology|
· Structure of Earnings survey 2014 - survey methodology:
· SES 2014: Eurostat?s arrangements for implementing the Council Regulation 530/1999, the Commission Regulations 1916/2000 and 1738/2005: https://circabc.europa.eu/sd/a/5e10d6d0-455f-4506-a338-0219ecf61551/SES2014%20Implementation%20arrangements.pdf
|Metadata completeness – rate|
A quality report is prepared according to requirements of the Commission Regulation 698/2006.
|Cost and burden|
Survey on respondents’ burden is not carried out.
|Confidentiality - policy|
· Law on Statistics (Statistics Act);
· 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.
|Confidentiality – data treatment|
Individual data are not published according to Art. 25 of Statistics Act. Dissemination of individual data is performed only according to Art. 26 of the Statistics Act.