National Statistical Institute

Labour Statistics Unit
Business Statistics Department

2, P. Volov Str.
1038 Sofia
Bulgaria

The aim of this report is to describe the Bulgarian national practice in compiling the Quarterly Labour Cost Index, according to the Regulation (EC) No 450/2003 of the European Parliament and of the Council, the Labour Cost Index main source, its coverage and timeliness.

The data are collected from enterprises and organizations that perform economic activity in Republic of Bulgaria. Sources of information are: pay-roll sheets, presence and absence sheets, labour or non-labour contracts, management contracts, letters of appointment or discharge, applications for paid or unpaid leaves etc.

Classification of Economic Activities (CEA-2008, fully comparable with NACE.Rev.2)

Covered are enterprises from all economic activities (except NACE section A) irrespective of their type of ownership and source of financing, which have at least one employee working under labour contract.

Labour Cost Index (LCI) is defined as Laspeyres index of labour costs per hour worked with base year 2016=100. It is calculated broken-down by economic activity at NACE Rev.2 section level.

Total Labour Cost (TOT) includes wages and salaries paid by the employer in cash or in kind, employers’ actual and imputed social contribution plus taxes regarded as labour costs;

Wage and salary costs (WAG) covers wages and salaries paid in cash or in kind;

Labour costs other than wages and salaries (OTH – non-wage costs) include the employers’ actual and imputed social contribution plus taxes regarded as labour costs;

Hours Worked covers the total number of normal and overtime hours actually worked by all (full-time and part-time) employees under labour contract.

Enterprise

Labour cost index refer to the enterprises with one and more employees from economic activities from B to S (incl. O) from NACE.Rev.2.

The whole country is covered.

2000-2022

2016

The data source is the quarterly sample survey on the Number of employees, time worked, wages and salaries and other labour costs. (QLCS - quarterly labour costs survey).

Quarterly

Up to 2007 the survey is carried out exhaustively for the public sector units and by a sample for the private sector units. Since first quarter of 2008 a sample from the public owned schools and kindergartens is drawn and they are no longer exhaustively covered in the survey.

The data validation goes through three stages:

• At the data entry;
• At regional level by the NSI Regional Statistical Offices;
• At national level by the NSI Head Office.

The data processing goes through the following stages:

• On-line data entry;
• Data validation;
• Correction of errors and imputation of missing information using administrative sources;
• Calculation of weights and weighting the results;
• Calculation the Labour Cost Indices.

Labour Cost Index series are available in the following three forms:

• Unadjusted;
• Working day adjusted;
• Seasonally and working day adjusted.

According to Article 2, point 3 from the Law on Statistics the statistical information shall be produced in compliance with the following criteria for quality: adequacy, accuracy, timeliness, punctuality, accessibility and clarity, comparability and logical consistency.

The quality assessment of the Labour Cost Index is related to the quality of the source from which the index is calculated. The overall quality of the Quarterly Survey on Number of Employees, Hours worked, Wages and Salaries and Other Labour Costs (QLCS - quarterly labour costs survey) is managed through system of different types of controls.  The data collection is on-line based improving significantly the timeliness and the data quality by incorporating logical and arithmetic checks at the data entry. Additional controls for coverage and completeness are done at regional and national level. For reducing non-response administrative data are used from the National Social Security register of insured persons. Coefficients of variations as the main measure of the accuracy are calculated for the main indicators (number of employees and average wages and salaries).

The main users of the Labour Cost Index can be divided into two groups:
- International users – European Central Bank to monitor development of labour costs in European Union and Eurostat in its role of intermediate user responsible for data publication and calculation of EU aggregates;
- National users – mainly media, every press release of the LCI is covered by most of the main national newspapers and internet sites. Within NSI, Labour cost index figures are used by the National accounts in calculating the Gross Domestic Product (GDP) flash estimates and from the Construction and Investment Statistics department.

No information is available on the degree to which LCI satisfies user needs.

LCI series are available broken down by labour costs components (total labour costs; wages and salaries and other non-wage costs); NACE sections (incl. O) and seasonal adjustment (not seasonally adjusted, working day adjusted and seasonally adjusted).

The data completeness rate is 75%. Labour cost index excluding bonuses is not produced.

The overall accuracy of the QLCS - quarterly labour costs survey results respectively LCI depends of:

• Sample size. The QLCS sample is representative at 2-digit level of Nace.Rev.2. and it is stratified according to the size class of the enterprise (number of employees) into three stratas. Exhaustively are covered public sector enterprises and private owned enterprises with more than 99 employees. A random stratified sample is made out of enterprises in private sector with less than 100 employees. For 2022 there were 190 000 private sector enterprises with 1 and more employees under labour contract in the population. Approximately 15 900 of them were surveyed every quarter;
• Sampling frame. The sampling frame from which the survey sample is drown is taken from the annual exhaustive survey on Number of employees, wages and salaries and other labour costs (ALCS - annual labour costs survey) for the year t-2 (the last available at the moment of sampling). The sampling frame is updated with information from administrative sources;
• Survey tools. The statistical form for the survey is available in electronic and paper format. In 2012 the survey became on-line based allowing 24 hours data collection and incorporates logical and arithmetic controls visible immediately during the data entry;
• Methods for errors identification and correction. The data processing goes through three main stages of validation: 1) data entry - 85 checks are included in the electronic questionnaire which control the plausibility of the inserted values and item non-response; 2) at regional level, the Regional Statistical Offices of NSI perform post-validation via 25 additional controls incl. checks for coverage, extreme and missing values; 3) at national level, the Head Office of NSI performs additional validation as well for coverage and completeness. Comparability is made with administrative and other sources.

Coefficients of variation and confidence intervals are calculated for assessment of stochastic reliability of the estimation for main indicators.

For the third quarter of 2022 for the private sector the confficient of variation and confidence intervals for the number of employees, wages and salaries and labour costs per hour are as follows:

See below

Coverage of employees in comparison with National Accounts number of employees

National Accounts number of employees is based on combination of Quarterly labour costs survey (QLSC) and Labour Force Survey (LFS) data. Employees working under labour contract are taken from the QLCS, while the persons working by other type of contract (civil contract) or without any contract are taken from the LFS. Being the primary sources for the National Accounts (NA), data on employees from QLCS represent nearly 90% from NA figures. The biggest difference is due to Military personal, which is not covered by the QLCS and it is taken from LFS and added to the NA number of employees. Relatively high difference is also seen in sections A and F where the number of employees working without labour contract is higher.  

QLCS - quarterly labour costs survey employment coverage compared to National Accounts number of employees

*Preliminary data

Optional

The  administrative Register of Insured Persons from the National Social Security Institute is used mainly for reducing unit non-response. The data for 693 refusing enterprises, 3.2% from the total number of surveyed enterprises, is imputed from the administrative register each quarter.

Measurement errors occur from the wrong interpretation of the survey questionnaire and respondents’ errors when not complying with explanatory notes.

To reduce these types of errors a total of 110 controls are applied at the data entry and the post validation process. In relation to the variables used for the calculation of LCI (labour costs and hours worked) the controls check for very high or low levels of wages and salaries per hour, high levels of bonuses, high or low levels of social security contributions and high or low levels of number of hours worked per day.  The cases not satisfying the controls are verified with the respondents and if not possible compared with the data from the administrative source.

In addition contact details for methodological consultation are published on the NSI web site.

The response rate for the QLCS - quarterly labour costs survey is relatively high. The main reasons for non-response are rather due to the use of old sampling frame i.e. inclusion in the sample of dead enterprises, enterprises with frozen activity, without employees, out of date contact information etc. rather than explicit refusals.

The response rate of the survey used as a source for the LCI calculation is at about 94%. Unit non-response adjustment is done by re-weighting - sample size is replaced with the number of the responded units in the calculation of weights. Information from administrative Register of Insured Persons is used as well.

For the main variables item-non response is not allowed form the logical controls applied at the data entry. If a respondent fail to fill in information for labour costs, hours worked and number of employees its data are considered as not valid and follow up contact is made in order to obtain the full set of information required.

See below

Estimation, frequency, hours worked, administrative data

1. Estimation

1.1. Estimation of missing cost items - social security contribution related to the annual bonuses

The information collected with QLCS - quarterly labour costs survey relates to each month of the reference quarter. Every fourth quarter in the reference year a separate variable related to annual bonuses is added to the questionnaire. However separate information on social security contributions related to annual bonuses is not collected. To estimate the sum of these payments the proportion of Employers actual social contributions to Wages and salaries cost in quarter four is applied to the sum of annual bonuses. The estimated sum of security contributions related to annual bonuses is thereafter added to the other non-wage costs.

Impact of the results: The estimated sum of social security contributions related to annual bonuses accounts for 0.61% from total labour costs. Among the NACE sections it varies from 0.02% for section B to 1.80% for section D.

1.2. Estimation of missing enterprises

Each quarter NSI receive from the National Social Security Institute the administrative register of insured persons. The register contains information on number of employees in the enterprise, number of hours paid, gross wages and salaries costs and the amount of social security contributions. The register is mainly used for reducing non-response in the QLCS. Approximately for 693 enterprises (explicit refusals) the information on employees and labour costs is taken directly from the administrative data.

2. Frequency

National Statistical Institute is collecting annual, monthly and quarterly information for the following main labour cost components:

* Q – Quarterly; A – Annual; √–Available; NA – Not available

3. Hours worked

The total number of hours worked is collected with the same source as for the labour costs components (QLCS). Hours worked include hours actually worked during contractual working hours, overtime hours, short periods of rest and time spent on tasks related with the preparation of the workplace, cleaning tools, machines etc. Hours paid but not actually worked (annual leave, public holidays and sick leave) are excluded from the total number of hours worked.

4. Administrative data
Administrative data is used only for reducing non-response in the survey and for updating the population. Since second quarter of 2020 up to second quarter of 2022 administrative information from National Social Security Institute on subsidies on labour costs, paid to enterprises as a government measure for reducing the effect from COVID crises is used for D.5. The amounts paid as subsidies to enterprises are classified by NACE sections and withdrawn from total labour costs. from WAG costs and subsidies received for social security from OTH costs.

Optional

National Statistical Institute performs seasonal and working day adjustment of the labour cost index series using JDemetra+ software with TRAMO/SEATS method. The type of adjustment is an indirect, as the models are updated once a year, while the parameters are updated each time when new observation is added to the time series. A tailor national caledar is developed for working day adjustment. When publishing LCI data, seasonal and working days, adjusted series are frozen for a year. Updated adjusted series are published in first quarter for the whole time series or in case of revision of not seasonally adjusted data.

There are two surveys conducted by NSI of Bulgaria, which collect information on earnings and labour costs:

- Annual Survey on Employed Persons, Wages and Salaries and Other Labour Costs (ALCS - annual labour costs survey). It is census survey and serves as a population frame from which the quarterly survey (QLCS) sample is drown;
- Quarterly Survey on Number of Employees, Hours worked, Wages and Salaries and Other Labour Costs (QLCS - quarterly labour costs survey). It is sample survey which results are used for estimating the quarterly LCI;        

In order to improve comparability between both surveys and ensure coherence between the annual statistics (based on exhaustive survey) and quarterly (based on sample survey) the not seasonally adjusted series are revised.

The main reason for the revisions is the long period of time between the population from which the sample is made and the reference year. For example the quarterly survey sample for year (t) is selected from (t-2) population. In oreder to overcome this problem the quarterly data on labour costs is revized twice. First when the population for yer t-1 become avilable and second final revizion when the population for yer t is available.

For each quarter new weights were calculated using the general population from ALCS for the year (t), adjustments were made for fluctuations between economic activities, size classes and type of ownership.

A comparison between hourly labour cost from ALCS - annual labour costs survey and QLCS  - quarterly labour costs survey and the changes between revisions is made in the table below.

Labour costs per hour worked from ALCS - annual labour costs survey and QLCS - quarterly labour costs survey

Revision history NACE Rev.2  aggregates B_S and B_N for the period 2008-2020 (TOT_LCI) - percentage change compared to the same period of the previous year – not seasonally adjusted series

Seventeen days after the reference quarter the respondents must submit the data to the NSI. After processing and first stage validation the Regional statistical offices must sent to the NSI head office the regional databases on the 30-th day after the reference period. The main results on number of employees and their average wages and salaries are published 40 days after the reference period.

LCI is published on the 75-80 days after the reference quarter.

Optional

 LCI index series are submitted to Eurostat within the required timelines.

Delays in days in transmission the LCI series

Optional

Optional

The source used for calculating LCI is one and the same for all years to which the data refer. There are no significant changes in definitions and methods used for collecting the different labor cost items.

2000 onwards

There is another statistical source that produce information on labour costs and hours worked - Annual Survey on Employed Persons, Wages and Salaries and Other Labour Costs (ALCS - annual labour costs survey). The definitions applied, the statistical unit, the coverage of employees and size classes are the same as in the Quarterly Survey on Number of Employees, Hours worked, Wages and Salaries and Other Labour Costs (QLCS - quarterly labour costs survey) used for the LCI calculation. The difference between the both surveys is that the annual one is exhaustive while the quarterly survey is a sample survey.

The difference between annual (based on exhaustive survey) and quarterly (based on sample survey) estimates on hourly labour cost varies from 0 to 3.8%. After the revision of the quarterly data the difference is usually reduced to 0.4-0.5%.

Coherence between compensation of employees per hour worked and total labour costs per hour worked.

Although the compared variables have similar definition, the two sources have many methodological and conceptual differences that explain the disparities in levels per hour worked.  

Among the main reasons are:
- Coverage by NA of employees working under non-labour contract (civil) and their earnings in the form of fees and commissions;
- Inclusions by NA of more types of wages and salaries in kind than in QLCS - quarterly labour costs survey: 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 and hours worked for employees working without any contract (informal employment).    
- NA figures cover also military personnel, which is not included in QLCS;

-Although used for comparisons the QLCS is not the source for compilation of compensation of employees. STS personal costs are used as well data from Ministry of finance for the State budget expenses. Hours worked however are taken from QLCS and additional adjustment are made for shadow economy, persons working under civil contract and military personal;

- For NACE section K the source is the Bulgarian National Bank.

Indicators within the data set are internally coherent.

Since 2009 each data publication is accompanied by press release and LCI was included in the NSI release calendar.  

Not seasonally adjusted, working day adjusted and seasonally and working day adjusted LCI series are published regularly on the NSI web page Labour Market section. (https://www.nsi.bg/en/content/3935/quarterly-labour-cost-index-not-seasonally-adjusted-working-day-adjusted-seasonally-adjusted).

See above

Optional

Optional

Optional

Methodological paper including coverage, main definitions, data sources and calculations is published on the NSI web page.

According to Article 2, point 3 from the Law on Statistics the statistical information shall be produced in compliance with the following criteria for quality: adequacy, accuracy, timeliness, punctuality, accessibility and clarity, comparability and logical consistency.

100%

Over a year period there are 4 or 5 consultations done on the LCI matters. The clarifications are related mainly to the elements included in the total labour costs and the seasonal adjusted series.

No information is available for the respondents’ burden.

• Law on Statistics;
• Regulation (ЕC) № 223/2009 of the European Parliament and of the Council of 11 March 2009 on European statistics (provision 24 and Article 20 (4)) (ОJ L 87, p. 164), stipulates the necessity 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.

According to Article 25 from the Law on Statistics:

Individual data received and collected through statistical surveys shall constitute a statistical secret and may be used only for statistical purposes. Individual data received for the purposes of statistical surveys may not be used as evidence before the bodies of the executive and the judiciary.

The National Statistical Institute and statistical authorities and their staff may not disclose or provide:

1. individual statistical data;

2. (amended, SG No. 98/2008) statistical data which can be matched in a way that enables the identification of a specific statistical unit;

3. (amended, SG No. 98/2008) statistical information which aggregates data about less than three statistical units or about a population in which the relative share of the value of a surveyed parameter of a single unit exceeds 85 per cent of the total value of such parameter for all units in the population.

The level of aggregation for LCI does not require confidentiality treatment.

Not available