Central Statistical Bureau of Latvia

Wages Statistics Section

Business Statistics Department

Mathematical Support Division

Lāčplēša iela 1, Rīga, LV–1301,

Republic of Latvia

The Central Statistical Bureau (hereinafter – Bureau) has conducted Structure of Earnings Survey 2018. The survey was carried out in compliance with the Council Regulation (EC) No 530/1999 of 9 March 1999 concerning structural statistics on earnings and on labour costs, and Commission Regulation (EC) No 1738/2005 of 21 October 2005 amending Regulation (EC) No 1916/2000 as regards the definition and transmission of information on the structure of earnings.

In accordance with the above regulations, the survey is conducted once in four years. The previous survey took place in 2014.

The Structure of Earnings Survey aims at obtaining precise, harmonised and comparable information on correlations between the size of wages and salaries and individual characteristics of employees (occupation, gender, age, educational attainment level, type of labour contract, etc.), as well as characteristics of the employers thereof – kind of economic activity, region, size of enterprise or institution in the European Union Member States and candidate countries.

Indicators on wages and salaries are summarised on the country as a whole as well as in breakdown by economic activity and region on both reference year and October of the reference year. The data on the number or coverage of employees having collective pay agreement are compiled as well.

In compliance with the above regulations, the respondents were required to fill in the Bureau questionnaires only with the data on selected employees, since only such kind of information ensured production of the necessary summaries on the mentioned groupings. In compliance with the Statistics Law, the information on employees acquired within the survey is confidential and is used only in a form of summaries preventing identification of individual persons and enterprises.

The survey results may be used for international comparisons, research purposes, development of various earningsrelated regulatory documents, and as additional information to characterise situation in the labour market. Annual, monthly and hourly wages and salaries may be analysed by employee sex, age, educational attainment, full or part-time work, type of labour contract, economic activity of statistical unit, as well as in various combinations and groupings of the above indicators.

The survey covered full-time and part-time employees who had main or secondary job on 31st of October 2018 and received remuneration. The wages and salaries are indicated for each particular place of work. They do not cover wage of the same employee elsewhere (in a second or third job).

The survey did not cover the self-employed persons. The survey aimed at obtaining information on local units of statistical units regardless of the number of employees. Only peasant and fishermen farms employing 50 and more persons were included. Data on the each local units surveyed were

submitted by the statistical units. Data in breakdown by branches are compiled based on the kind of economic activity of the local unit.

The statistics of the 2018 SES refer to enterprises in the areas of economic activity defined by sections B to S.

The survey covered full-time and part-time employees who had main or secondary job on 31st of October 2018 and received remuneration. In the summary tables each of the employees is included as one employee. The individual characteristics of employees (age, highest educational level attained, occupation, workload, etc.) are given on 31 October 2018. The wages and salaries are indicated for each particular place of work. They do not cover wage of the same employee elsewhere (in a second or third job). The survey did not cover the self-employed persons. The survey aimed at obtaining information on local units of statistical units regardless of the number of employees. Only peasant and fishermen farms employing 50 and more persons were included. In the survey it was assumed that there were no local units in the statistical units employing fewer than 10 employees. Centralized bookkeeping offices of local governments provided data not about the local units but about following economic activities, if any: Electricity, gas, steam and air conditioning supply; Public administration and defence; compulsory social security; Education; Social care with accommodation; Social case without accommodation; Arts, entertainment and recreation; Activities of libraries, archives, museums and other cultural activities. It was assumed that each activity is one local unit. Such an approach does not produce mistakes in the SES data in breakdown by region, as local units of the local governments are always situated only within the territory of the corresponding local government.

The statistical unit is a local unit. But it was assumed that there were no local units in the statistical units employing fewer than 10 employees. Centralized bookkeeping offices of local governments provided data not about the local units but about economic activities.

Out of 904.5 thousand employees forming the population in economic activities A–S, 193.4 thousand employees were included in the sample.

Indicators on wages and salaries are summarised on the country as a whole. 

From 2002 to 2018

NA

The variables that have been corrected most often are:

-          share of full-time or part-time normal hours (variable 2.7.1.) worked by employees in PT does not correspond to the share of number of normal hours worked by employee in the local unit in FT;

-          collective pay agreement (variable 1.5) does not correspond to any of 5 categories of agreements;

-          wages and salaries in October 2018 included variables on whole 2018.

The survey is conducted once in four years. The previous survey took place in 2014.

Systematic process of gathering data for official statistics.

See 4. Quality management.

Operations perfomed accordance Eurostat’s arrangements for implementing the Council Regulation 530/1999, the Commission Regulations 1916/2000 and 1738/2005

During the arrangement of micro database the following additional procedures were done:

-          if employee’s earnings in October were calculated at reduced rates (sick leave certificate A), the employees were excluded from the sample, and the weights were re-calculated so that such employees were excluded from the sample;

-          if in October employee’s earnings were affected by unpaid absence, including the cases when employee joined enterprise in October, in order to estimate the employee’s earnings for full month, the earnings and hours worked in October were adjusted.

The data of the State Revenue Service served as a source to calculate and input values of variable B41 (Gross annual earnings), B411 (Annual bonuses and allowances not paid at each pay period)and B31 (Number of weeks to which the gross annual earnings relate)for the respondents in the size class “1–9”. A special shortened questionnaire was been developed for such respondents.

Some assumptions for these respondents are made:

-          respondent has no local units;

-          the annual income of the employee is equal to his/her gross annual earnings without annual payments in kind;

-          the pay system is very simple, and variable value B411=0.

The un-weighted imputation rate for each variable accounted for  4.8 %.

The weighted imputation rate for each variable constituted 15.2 %.

The data of the State Chancellery served as a source to calculate and input values of variable B21 (Sex), B23 (Occupation in the reference month), B25 (Highest successfully completed level of education), B27 (Full-time or part-time employee), B41 (Gross annual earnings), B411 (Annual bonuses and allowances not paid at each pay period), B31 (Number of weeks to which the gross annual earnings relate), B42 (Gross earnings in reference month), B412 ( Annual payments in kind), B421 (Earnings related to overtime), B422 (Special payments for shift work) for the respondents in State direct administration institutions and other state and local government institutions. 

The un-weighted imputation rate for each variable accounted for  9.9 %.

Data of the Business Register of the Business Statistics Department of Latvia were used for imputation of non-response data of the variable A17 (Affiliation of the local unit). The imputation rate accounted for 100 %.

Data of Latvian Population and Housing Census education database 2018 and 2019 (Internal database of the Central Statistical Bureau) and he data of the State Chancellery were used for the imputation of non-response data in the variable B25 (Level of education).

The un-weighted imputation rate constituted 0.9 %.

The weighted imputation rate accounted for 0.9 %.

The overall un-weighted imputation rate comprised 13.8 %.

In order to reduce the measurement errors the questionnaire was supplemented with detailed instructions. In the validation program for each enterprise the data of the SES survey were compared with the variables from other surveys. Web questionnaire has validation control, and it generates a short description of errors in a dialog box. For some categorical variables a drop-down menu of the dialog box was used.

Data quality, based on standard quality criteria is assed as resonable precise, excluded data on NACE Rev.2 A 

The main users of the Structure of Earnings Survey data in Latvia and abroad can be classified as follows:

-          National (internal) users:

• government institutions – Ministry of Finance, Ministry of Economy, Ministry of Welfare, Central Bank of Latvia, Agency of State Social Security, State Employment Agency of Latvia;
• associations – Latvian Chamber of Trade and Industry, Latvian Confederation of Employers, trade unions;
• research institutes, educational institutions;
• units of the Central Statistical Bureau of Latvia (CSB);
• courts;
• mass media;
• lecturers, students, independent research institutions and researchers;

International users:

• Eurostat, International Labour Organisation, International Monetary Fund.

Central government institutions for own needs are using information on the number of employees in breakdown by region, occupation, age, educational group, etc., as well as data on the average wages and salaries in various breakdowns.

Trade unions, associations and societies are interested in obtaining information on the average monthly wages and salaries received by selected occupations or occupational groups.

Mass media use several key indicators, e.g., average monthly wages and salaries in the country in selected sectors and regions.

User satisfaction surveys have not been conducted yet.

Several potential users may need data at more detailed degree of elaboration, which SES 2018 survey cannot provide.

Several potential users may need data at more detailed degree of elaboration, which SES 2018 survey cannot provide.

R software is used for calculation of standard errors with Taylor linearization method.

a)      Coefficients of variation (CV) concerning  average gross earnings in the reference month (variable B42) are broken down by:

-          full-time (separately for men and women) and part-time employees;

-          Sections of the NACE Rev.2;

-          occupation (ISCO-08 at 1-digit level);

-          age band (under 20, 20–29, 30–39, 40–49, 50–59, 60 and over);

-          level of education attained (levels 1–4 of the ISCED 2011);

-          size band of enterprise (1–9, 10–49, 50–250, 500–999, 1000+).

b)      Coefficients of variation (CV) concerning average gross hourly earnings in the reference month are broken down by:

-        full-time (separately for men and women) and part-time employees;

-        Section of the NACE Rev.2;

-        occupation (at 1-digit level of the ISCO-08);

-        age band (under 20, 20–29, 30–39, 40–49, 50–59, 60 and over);

-        level of education attained (levels 1–4 of the ISCED 2011);

-        size band of enterprise (1–9, 10–49, 50–249, 250–499, 500–999, 1000+).

The standard error is calculated according to the sampling design.

Coefficient of variation (CV) is a value for which the estimate of indicator is used as the denominator and the standard error is used as the numerator.

The survey sample was drawn at the end of 2018. Time lag between the last update of the sampling frame and the moment of the actual sampling constituted approximately one year. This is the main reason behind the coverage errors in the survey.

The estimated over-coverage rate of the reference population accounted for 11.82%.  

The response rate of local units of enterprises is 89.95%. The design weights are adjusted according non-response in each stratum.

The design weights were adjusted using the data of response level in each stratum. Weight calibrations were based on the number of employees in survey 2-Labour of 2018

see below

The estimated over-coverage rate of the reference population accounted for 11.82%.  

NA

The questionnaire for SES 2018 was designed in such a way to eliminate the survey instrument (questionnaire) errors providing explanatory notes directly in questions and detailed explanations in the supplemented instruction for filling in the questionnaire.

The SES 2018 data processing had been carried out using the Integrated Statistical Data Management System (ISDMS) where statistical metadata is the key element. The metadata module ensures maintenance of validation rules of statistical survey. Each validation rule description contains validation rule code, error message text and description of validation rule, validation rule conditions. It is possible to generate validation procedure automatically for each questionnaire directly after the entry of questionnaire. Main validation rules, both arithmetical and logical, were available during filling in e-questionnaire via website of CSB.

The validation programme consisted of 168 arithmetical and logical controls. After executing the validation procedure of respondent survey, data form with list of validation errors is available. This form contains the following information for each error – error number, error description, type (acceptable, unacceptable, not marked), error type reason (for some errors it is necessary to describe the acceptability reason). 

Every statistician had a definite number of enterprises from which the questionnaires have to be collected, entered and verified. The responsible person could not finish the data entry of the questionnaire in case if any answer on the questionnaire’s variable was missed, or (according to the rules implemented in the data entry programme) filled in incorrectly. In these cases the responsible person contacted the enterprises once more and made the necessary changes in the questionnaire.

see below

The response rate of local units of enterprises is 89.95%. The design weights are adjusted according non-response in each stratum.

The average number of non-respondents was 956 local unit. The 35% of those were enterprises not found; refused to submit questionnaire or delayed to do that comprised 65 % of non-respondents.

see below

We performed imputation from administrative data, which were not included in the report. B21 (Sex), B28 (Type of employment contract), B42 (Gross earnings in reference month), B412 (Annual payments in kind), B31 (Number of weeks to which the gross annual earnings relate) were imputed 10,5% of all persons.

B41 (Gross annual earnings in the reference year) were imputed 15,3% of all persons.

B23 (Occupation) were imputed 1,3% of all persons. B25 (Education) were imputed 13,9% of all persons. 

na

NA

No

No data revision.

No data revision.

The length of time between the collection release and the reference period in Latvia was 23 months.

The length of time between the collection release and the reference period in Latvia was 18 months.

Ffinal results send between the collection release and the reference period in Latvia was 19 months.

Design of the questionnaire was finished in October 2018.

Sample was drawn in November 2018. The last corrections of the weights were made in June 2020.

The deadline for delivering the questionnaires to the statistical office was 2 March 2019.

Part of the respondents asked to extend the submission deadline, because in March and April enterprises are balancing their accounts, therefore it was allowed to deliver the questionnaires later. Rest of enterprises received reminders. At the same time, statisticians contacted enterprises also by phone.

The validation of the micro data was finished in October 2019.

The collection of statistics “Results of the Structure of Earnings Survey 2018” was published in November 2020.

Afterwards, the data were tested also at macro levels, and the files to be transmitted to Eurostat were prepared in June 2020. The data of the SES 2018 were transmitted to Eurostat on 01 July 2020.

There are no differences between the Latvian and European concepts.

There are no differences between the Latvian and European concepts.

In Latvian SES 2002 the enterprises were used as sampling units instead of local units.

Sampling unit used in SES 2018 (as well as in 2006 and 2010 and 2014) was local unit, whereas in 2002 it was enterprise, and indicators (wages and salaries, number of employees) were calculated in breakdown by regions of Latvia.

In Latvian SES 2002 the enterprises were used as sampling units instead of local units.

Sampling unit used in SES 2018 (as well as in 2006 and 2010 and 2014) was local unit, whereas in 2002 it was enterprise, and indicators (wages and salaries, number of employees) were calculated in breakdown by regions of Latvia.

The data of the State Revenue Service served as a source to calculate and input values of variable B41 (Gross annual earnings), B411 (Annual bonuses and allowances not paid at each pay period) and B31 (Number of weeks to which the gross annual earnings relate) for the respondents in the size class “1–9”. 

The data of the State Chancellery served as a source to calculate and input values of variable B21 (Sex), B23 (Occupation in the reference month), B25 (Highest successfully completed level of education), B27 (Full-time or part-time employee), B41 (Gross annual earnings), B411 (Annual bonuses and allowances not paid at each pay period), B31 (Number of weeks to which the gross annual earnings relate), B42 (Gross earnings in reference month), B412 ( Annual payments in kind), B421 (Earnings related to overtime), B422 (Special payments for shift work) for the respondents in State direct administration institutions and other state and local government institutions.

Na

Gross annual earnings in the reference year (2018) per employee (code 4.1) constituted EUR 14336.

Wages and salaries per employee of the NA constituted EUR 14430.

The methodological disparities between the “Gross annual earnings in the references year, expressed per employee” (SES) and “Wages and salaries per employee”(National Accounts) are  following:

-          NA includes self-employed persons, but SES does not;

-          When calculating the wages in SES per employee the wages are adjusted to the full-time basis. In National Accounts this procedure is not done;

-          in National Accounts compensation of employees is adjusted based on the non-registered employment, adjustment of minimum wages in small and micro enterprises, adjustment of underreported wages and adjustment of tips.

It is found that in many activities the surveyed enterprises declare to pay average wages below the legal minimum wage. However, it is known that actual wages are not below the legal minimum. This NA adjustment recalculates average compensation per employee based on the assumption that it is not below the legal minimum wage.

For the adjustment of the calculation for underreported wages the calculated wages and salaries in cash and experts' assumption on the amount of underreported wages in each sector are used. Experts' evaluations are based on the economic indicators of each sector, also taking into account the SRS and EU-SILC data. Experts of National Accounts set the percentage for underreported wages for each NACE sector.

Adjustment of tips is estimated for output and NACE sections H, I and S.

Based on the experts' assumption, tips on average account for 0.05 % of the enterprise's turnover, both in Non-financial corporations (S.11), and Households (S.14). 

For the above NACE sectors, turnover data of enterprises are increased by 0.05 %.

NA

News releases on-line.

The SES information is available in the collection of statistics “Results of the Structure of Earnings Survey 2018”, which is also available in CSB Information Centre and National Library of Latvia and on the CSB webpage.

Please consult free data on-line or contact contact person.

Please consult free data on-line or contact contact person.

No information.

No information.

https://www.csb.gov.lv/en/statistics/statistics-by-theme/tables/metadata-structure-earnings-survey

In order to reduce the measurement errors the questionnaire was supplemented with detailed instructions. In the validation program for each enterprise the data of the SES survey were compared with the variables from other surveys. Web questionnaire has validation control, and it generates a short description of errors in a dialog box. For some categorical variables a drop-down menu of the dialog box was used. 

If it was not possible to correct incorrect values, or if it was not possible to get the data for variables not filled-in, the questionnaire was considered as not delivered and the weights were recalculated.

 One of the steps to escape input of erroneous or incomplete information into the IT applications and database covered development of data entry program that included wide range of logical controls. Each questionnaire was validated according the validation program. The validation program consists of arithmetical and logical controls. The controls ensuring all-round data verification cover:

-          reporting of arithmetical, logical mistakes;

-          comparison with the data of regular survey on labour;

-          revealing of data input errors;

-          identification of deviations from average indicators.

 Every statistician has a definite number of enterprises from which the questionnaires have to be collected and verified. If any answer on the questionnaire’s variance was missed or (according to the rules implemented in the data entry programme) filled incorrectly, the responsible person contacted the enterprise once more and made the necessary changes in the questionnaire.

The validation should be done in two ways:

-          individually for each questionnaire (after the data entry);

-          for the entered questionnaires of the regional offices and the sections of the central office involved in data collecting.

 All expressions not filed could be seen on screen or in printed form.

The variables that have been corrected most often are:

-          share of full-time or part-time normal hours (variable 2.7.1.) worked by employees in PT does not correspond to the share of number of normal hours worked by employee in the local unit in FT;

-          collective pay agreement (variable 1.5) does not correspond to any of 5 categories of agreements;

wages and salaries in October 2018 included variables on whole 2018.

NA

NA

Cost associated with the collection and production of a statistical product is not calculated. Burden per respondent constituted 5 hours.

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.

The Latvian Law of Statistics.

No comments