TurkStat, Turkish Statistical Institute

Department of Annual Business Statistics 

Devlet Mah. Necatibey Cad. No:114 Room: 420 Çankaya/Ankara

The quality report prepared for the assessment of the Structure of Earnings Survey (SES) 2010 comprises information pertaining to the six components of the European Statistical System. These are relevance, accuracy, timeliness and punctuality, accessibility and clarity, comparability and coherence. SES 2010 was the second application of the series structured as a four yearly survey to provide information in accordance with the Regulation No. 530/1999 of the European Council and Parliament and Regulation No. 1916/2000 and 1738/2005 of the European Commission. The objective of this report is to assess the quality of the survey. The report will follow the contents of the Commission Regulation No. 698/2006 concerning evaluation of structural statistics on labour costs and earnings. The content of the Quality Report is as follows:

1. Relevance: Main users of the survey, main needs of the users, satisfaction of the users and completeness
2. Accuracy: Assessment of the reliability of the survey results; calculation of the estimation precision
3. Punctuality and timeliness: Key dates disaggregated in terms of phases of the survey
4. Accessibility and clarity: Information on the dissemination of the survey results
5. Comparability: Information on both spatial comparability and comparability over time
6. Coherence: Comparison of the survey with relevant surveys in terms of conceptual differences and differences arising from other reasons in case there is no/negligible conceptual differences.

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

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

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

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

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

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

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

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

In 2010 Structure of Earnings Survey, two stage stratified random sampling method was used. At the first stage of the survey, local units were selected and at the second stage, employees from the selected local units were interviewed. In the survey, local units associated with the enterprises with 10 or more employees were defined as the first stage sampling unit. The sample size of the local units was 20155. The respondent local units were 14332.

The distribution of the sample size was realized by the compromise allocation method on the level of statistical region units and economic activity of local unit and the size class of the enterprises. The size classes and the economic activity groups covered in the survey are as follows:

Size band of the enterprise which the local unit is associated:

10-49

50-249

250-499

500-999

1000 and over

Economic activity of the local unit:

2 digit Nace Rev 2 B, C, D, E, F, G, H, I, J, K, L, M, N, P, Q, R, S sectors.

The second stage sampling unit is defined as the employee within the local unit. The employees were selected by considering the employee size band given below:

Employees within each local unit were ordered alphabetically by their last names and required number of employees from these lists was selected systematically.

[Not requested]

[Not requested]

[Not requested]

[Not requested]

[Not requested]

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

[Not requested]

Main users of the survey

Internal users are; law courts, trade unions, academics, students, researchers, media, employers and investors. International users are EUROSTAT, ILO, UNECE and OECD. Internal users are generally interested with the level of the earnings in the certain occupations by sex, length of service and age group. Gender pay gap is also a demanded issue. 

Main needs and satisfaction with the data offered

No survey has been carried out about the satisfaction of user needs but when the data requests analyzed it can be seen that the survey should offer more detailed breakdown in regions and it would be more frequent then four yearly intervals.

All mandatory variables as well as some of the optional variables were obtained.

[Not requested]

-

[Not requested]

The coefficients of variation as well as the standard deviations for the variables mentioned in the Commission Regulation 698/2006 are given below. CV figures by statistical region units are not provided, as this level of calculation is not mandatory for Turkey by the Implementation Arrangements.

Please see also the attached document Coefficients of variation.

The coefficients of variation as well as the standard deviations for the variables mentioned in the Commission Regulation 698/2006 are given below. CV figures by statistical region units are not provided, as this level of calculation is not mandatory for Turkey by the Implementation Arrangements.

Please see also the attached document Coefficients of variation.

The local units associated with the enterprises with 10 or more employees in the 2009 Business Register were used as the sampling framework in the Structure of Earnings Survey 2010. The business register is updated annually on the basis of information taken from Ministry of Finance and annual structural business survey results for the level of enterprise.

The survey questionnaire asks for information on the current economic activity, the region and actual number of employees in the local unit. The results were checked with the frame information. Misclassification in terms of economic activity was observed in the survey. The overall misclassification rate was 27 %.

Regarding to over coverage, 21, 1 % of the sample was out of scope. The distribution of over coverage by reasons is as follows;

• Dead units (44, 1 %)
• Not having employee (45 %),
• NACE section not covered by the survey (1.6 %) and
• Other reasons (9.3 % of total over coverage).

Other reasons could be inactive local units, duplicated address, etc.

It is not possible to obtain under coverage, since there is no other external source to compare the frame information. 

[Not requested]

The questionnaire of the 2010 Structure of Earnings Survey was prepared by taking into account the needs of data users and producers on the labour market, views of the experts on the subject and applications of the other countries, experience from 2006 implementation together with the concerned regulations of the European Union Statistical Institute. The questionnaire has two main parts; Part A: Information on the establishment (local unit) and Part B: Information relating to the employees. In Part A, information collected on the characteristics of establishment like title and main economic activity. Part B in which information on employees was collected had been divided into three sections in terms of contents;

B1. General information on the employee                            

B2. Monthly working period and payments pertaining to the month of November 2010

B3. Annual working period and payments pertaining to the year of 2010

Data collection was performed by regional offices of TurkStat through face to face interview and e-mail. Electronic questionnaire that was suitable for filling in and sending back to regional offices was provided. Data entry was also performed by regional offices. The data entry code includes coherence checks. Item non response was not allowed. Therefore imputation techniques were not used. The checks embedded in data entry code mainly focused on;

• Incompleteness in questionnaires either in information pertaining to the establishment or those pertaining to the employees
• Lower and upper limits for hours of work
• Inconsistencies between monthly and annual figures
• Inconsistencies between payments and hours pertaining to that payment

As the data were transferred from the regional offices, data were either validated or sent back to regional offices after the application of second level checks on all variables of questionnaire. Correction or approval from the establishment was required for non validated cases. The second level checks applied in the centre cover the ones embedded in data entry code with more precise lower and upper levels and also cover checks on;

• Inconsistencies between education and age,
• Inconsistencies between occupation and education; occupation and economic activities; occupation and age and occupation and sex,
• Duplicate records for employee level,

• The contradiction between the figures in the questionnaire and legal requirements.

The final data was checked against EUROSTAT validation rules and a metafile covering following explanations was sent to EUROSTAT along with EUROSTAT data tables.

• Variable 2.2: Age; there are 12 cases remaining out of the limits set for the "Variable 2.2: Age". These 12 employees are all 80 years old or older. The data is checked and validated.
• Variable 2.6: Length of service in the enterprise; There are 16 cases remaining out of the constraint set for the Variable 2.6.  In all of the cases, employees started to work below 14. The youngest one is 10. Although this is not a common case in Turkey, it is still observed in activities in which unregistered employment is more widespread.
• Variable 3.2: Number of hours actually paid during the reference month; 3482 cases violate the plausibility check pertaining to the variable 3.2 (Number of hours actually paid during the reference month). None of the cases remaining out of the limits results from the first part of the check;

(0.9*Var.4.2/Var.4.3<=Var.3.2<=1.10*Var.4.2/Var.4.3). Because, there is no separate question for Variable 4.3 in the SES 2010 application and it is already calculated by using the two other components of this first part of the check, namely Var.3.2 and Var.4.2.

3432 of the cases pertain to second part of the check (full-time employees) while 50 of those pertain to the third part (part-time employees). The employees having the highest value for the number of hours actually paid during the reference month are mostly in certain occupations such as driver in transportation, worker in construction, waiter in hotels and restaurants, sales person in trade etc.

• Variable 3.2.1: Number of overtime hours paid in the reference month; 230 cases violate the plausibility check pertaining to the variable 3.2.1. The data is checked and validated.
• Variable 3.3: Annual days of holiday leave; there are 48 employees remaining out of the limits set for the variable 3.3. It is possible to use accumulated annual days of holiday leave from earlier years according to labour law in Turkey.
• Variable 4.1: Gross annual earnings in the reference year; there are 423 employees remaining out of the limits set for the variable 4.1. The data is checked and validated.
• Variable 4.2.1: Earnings related to overtime; there are 1791 cases remaining out of the limits set for the "Variable 4.2.1. All of cases are confirmed. Most of the cases are employees working at post offices.

The overall response rate of the survey is 90.1 %. The distribution of the total non-responded cases could be listed as follows;

• Erroneous address (65.4 %)
• Change in the location of the local unit (6.1 %)
• Refusals (2.6 %) and
• Other reasons (25.9%).

The highest rate of non-response comes from the errors in the addresses frame. Other reasons pertain to inability to contact due mainly to seasonality and inadequate address.

(a) Non-response rates by economic activity

(b) Non-response rates by size band of the enterprise

[Not requested]

[Not requested]

Cannot be measured

Imputation techniques were not used.

Not used any model.

[Not requested]

[Not requested]

[Not requested]

[Not requested]

The results of the SES 2010 were released on December 20^th, 2011. Delivery of data to EUROSTAT was on March 2012 as three months before from foreseen date in the requirements of regulations and arrangement.

[Not requested]

[Not requested]

Monthly data refers to November 2010 and yearly data refers to 2010. Preparation of SES 2010 implementation began in November 2010 with the sending of implementation paper by EUROSTAT. The timetable applied throughout the survey implementation was as follows;

Crucial survey processing dates

[Not requested]

Month of November was used as the reference month, for the other variables, EUROSTAT definitions and classifications are adopted. November was selected as a reference month in previous application since it was a month without seasonal payments and absences. To keep comparability over time, the same month was used as the reference month although this time November covered an Islamic festival which moves backward 10 days each year. The accounting or fiscal year coincides with the calendar year in Turkey.

The SES 2010 is applied in the local units associated with enterprises employing 10 and more employees as set in the relevant regulations/arrangements. While C-O (excluding L) sectors of

NACE Rev. 1.1 was used the first implementation (SES 2006), the local units of B-S (excluding O) of NACE Rev.2 are covered in the SES 2010. The compilation of the data was based on local units and enterprises as defined in the Council Regulation no. 696/93. The units were stratified by size of the enterprise, NACE and statistical region units.

[Not requested]

The first survey about the earnings structure, conducted by the TurkStat, was the 1994 Employment and Wage Structure Survey. Only industry sector was covered in this application and the survey was not repeated. TurkStat started SES series starting from 2006 application. Therefore 2010 SES was the second application of the series.

Both 2006 and 2010 applications of SES are in accordance with EUROSTAT regulations and implementation guidelines. As it is known, the results can not be comparable directly by economic activity and occupations since the updated versions of the classifications were used in 2010 application according to EUROSTAT directives.

[Not requested]

Wages and salaries per employee of the National Accounts for the year 2010 required to be compared with gross annual earnings in the reference year obtained from SES 2010 but this comparison could not be substantiated since National Accounts data by income approach is not available for the related period due to ESA 95 revising studies.

[Not requested]

[Not requested]

[Not requested]

The press release, selected tables and metadata for SES 2010 may be reached from http://www.tuik.gov.tr/PreHaberBultenleri.do?id=10718 in Turkish andhttp://www.turkstat.gov.tr/PreHaberBultenleri.do?id=10718 in English. Summary of the press release were published in social media (twitter, facebook, igoogle, frienfeed, feedburner) and our subscripted users were informed with text and e-mail messages.

The bilingual publication containing estimations in table and graphic format and methodological information was released both in printed and electronic formats.

Users can download the publication for free via our web site.

[Not requested]

The confidentially treated micro data pertaining to SES 2010 is also prepared for sale in electronic format with a protocol.

Additional information is provided on request to the users in table format. The survey results are only provided to respondents on request.

The metadata used in the production of Structure of Earnings Statistics are available in our web site. Users can reach the metadata from the page press release published (http://www.tuik.gov.tr/PreHaberBultenleri.do?id=10718 and http://www.turkstat.gov.tr/PreHaberBultenleri.do?id=10718). Users also can ask any question about survey and metadata to responsible unit Labour Cost and Earnings Statistics Team.

[Not requested]

[Not requested]

[Not requested]

[Not requested]

[Not requested]

[Not requested]

No comment