Hungarian Central Statistical Office

Earnings Statistics Section

P.O.B. 51, 1525 Budapest (Hungary)

The quarterly Labour Cost Index measures trends in average labour costs by units of labour input. The index shows the short-term development of the total hourly costs of employers. The demand for information on cost of employment from investors and other users due to the liberalised international flow of capital and labour inspired the introduction of quarterly statistics on labour costs. Nevertheless, the efficiency, the competitiveness, the competition for markets and foreign investors forces management in economic units to continuously overview their social policies, and system of social benefits. The European Union also requires measuring the cost of employment. Information on labour input, cost of labour i.e. the price of labour is very important from the point of view of international comparisons and competitiveness.

Information on the level and composition of labour cost for the whole economy is available from the results of EU Labour Cost Survey every four years. As this survey provides only multi-annual data (from 2004 in every fourth year), therefore it is not designed to measure changes in the price of labour services from collective and individual agreements. A reliable indicator - labour cost index - is identified for measuring and analyzing of trends in the price of labour every quarter. The labour cost index results are used in formulating industrial relations, wages, policies and economic analysis.

Classifications used are: NACE Rev2. and the Hungarian Classification of Units by Legal Forms. Business entities, budgetary and non-profit organisations are classified to different categories of NACE Rev2 according to their main activites.

Entreprises (legal unit) and other businesses employing at least 5 persons, all budgetary organisations, and selected NGO-s are covered by the survey providing the base data for the estimation of the LCI. 

We use SNA concept concerning wages and salaries, non-wage costs (sum of compensation of employees + taxes and - subsidies) and hours worked.

Enterprises with legal entities, other businesses, budgetary organisations, selected NGO-s. 

The Hungarian Business register's frame covers the statistical population.

The whole country.

Comparable data have been available since 2000 (total labour costs excluding bonuses since 2002).

Base period is 2016.

1. The Quarterly Labor Report (quarterly institutional survey to STS (Short-term Statistics)) covers all enterprises with more than 4 employees. The budgetary organisations are observed on a full-scope basis and certain non-profit institutions also take part in the survey. A regulation in Hungary obliges public institutions (budgetary organisations) to use a standardized system for payroll from 2001 onwards. The accession to the system is still ongoing, almost every public institutions have already launched this system. On the basis of this payroll system a central database has been set up. The database includes all types of payments including hours not worked on the level of the employees. Nevertheless the data concerning the contributions are also available on the level of institutions. This database provides all the costs paid by the employer for employing labour.

2. Labour Cost Survey, 2016 results: The survey covered enterprises employing 5 and more persons, all budgetary intituions and a set of certain non-profit institutions. The survey of  enterprises with employed persons from 5 to 19 was sampled, while the ones employing more than 19 were fully observed.

Base data for calculating labour cost index derive from the Monthly Labour Report.

3. As of 2019, we started using the National Tax and Customs Administration for more accurate calculation.

Measurement of regulations year by year for the compulsory social security contribution is as follows:

* in percentage of wages and salaries
** 27% from 1^st July
*** The employer’s contribution merged (as the labour market contribution) into the social contribution tax in 2012.

**** 17.5% from 1^st July

***** 15,5% from 1^st July

Data of wages and salaries, hours worked: derive from the Quarterly Labour Report.

Data of social costs: partly come from actual proportional measurement of employers’ compulsory social contribution regulated by the government, partly come from estimation based on the latest Labour Cost Surveys (now Community Labour Cost Survey 2012 according to NACE Rev 2.)

The deadline of the quarterly data collection is 12 days (for enterprises) and 20 days (for budgetary organisations) following the reference period. The data collection takes place electronically from 1^st of January 2013. The electronic data sending via a special interface called ELEKTRA.

28-30 days after the end of the reference, month right after the arrival of the data HCSO closes the data collection. In the following next 18-20 days:

• the loaded data are audited,
• data are prepared for publication by the experts of the Earnings Statistics Section.

The administration data collect from Tax Office 40 day following the reference period

HCSO has been improving the questionnaire and the guide of Quarlerly Labour Report. Using the results of this common work a wide range of logic tests have been introduced and used in the quarterly Labour Report 2021. Hopefully the extensive list of logic tests has revealed all errors of magnitude made by respondents. All questionnaires were approved. Re-contacts were made by the field-workers if it was necessary. The data were tested during data entry, and also at macro level. Statistics concerning the number of corrections have not been made, so there are not any data for measuring processing errors. For better checking we used data from the annual Structure of Business Survey and Common Micro Database at micro level, too.

Survey strategy

For measuring labour input the Quarterly Labour Report has been used in the Hungarian Central Statistical Office (HCSO). The survey covers sections A–S (NACE Rev. 2).

The survey is carried out partly using a statified sample. The target population for the sample includes the operating enterprises with 5–49 employees (except section D, which is fully scoped). Enterprises with at least 50 employees are observed completely. The sampling frame of the survey and the number of enterprises are provided by the Business Register (BR) of the Hungarian Central Statistical Office. All of the observed units (also including those in the sample) are enterprises in the BR belong to the appropriate NACE sections and size groups and have indication that the enterprise operates. In case of non-response, we do all our best, to have collect information, on whether the enterprise operates.  If it is so, data of non-respondents in the fully scoped part are imputed by algorithm and in the sampled part by strata average.   Otherwise data are imputed by zero values.

A stratified sample is selected for the survey.

According to NACE Rev.2 about 580 strata are formed.

The total sample size is determined by taking into account both accuracy demands and cost limitation. The sample size for the various strata is determined by a modified Neyman allocation. The sample is selected systematically, according to the decreasing order of the random numbers belonging to the various enterprises in the BR. We give a primary preference to the enterprises not belonging to the sample 3–5 years earlier and a secondary preference to the enterprises belonging to the sample in the previous year. So the probability of choosing an enterprise for the sample is lower if it was in the sample one year before.

Method of estimation

In the course of processing, a few indicators are estimated by strata for the sampled enterprises and for enterprises observed completely. In case of the non-specific characteristics (hours worked by employees, compensation of employees, wages and salaries), the population total is estimated directly from the elementary data. Within the various sampled strata the following is applied. The sampling weight

is determined, where N_j is the size of the jth stratum, n_j is the sample size in the jth stratum. Within the various strata, the population total Y_j is estimated by the Horwitz-Thompson estimator with equal selection probabilities:

 where y_j is the sample total in the jth stratum.

In case of strata enumerated completely, is the number of respondents and enterprises with imputed data, nj is the number of respondents, and Yj is the sum of the data of the Nj enterprises in the jth stratum.

Obviously, for the unions of strata

The population mean for the various strata is estimated as

The population value of the specific characteristics (annual labour costs per employee, hourly labour costs) is estimated as the quotient of the estimated population totals of the corresponding non-specific characteristics.

Non-response

For the various sampled strata and their unions, the unit response rate is . As mentioned earlier, data of non-respondents are imputed.

Calculation scheme

Wages and salaries elements (D11) of the quarterly labour cost index data are calculated from the Quarterly Labour Report. The absolute monthly data are summed up month by month, then the quarterly amount of the wages and salaries are divided by the quarterly amount of the hours worked in order to get the quarterly wages and salaries per hour. As the survey contains data only concerning wages and salaries, number of employees and hours worked, the missing data for calculating the numerator of the other elements (e.g. social costs) are estimated with a so-called ‘updating method’.

The updating method is the following:

• the compulsory social security contribution paid by the employers are computed by multiplication of the current proportions of contributions and wages and salaries (D1211); its percentage within the social costs is about 90,
• other elements’ (D1212, D122, D4, D5) derive from the latest LCS.
• additional information at 6.5 Data revision – policy, 3.1. Source data
• the social contribution allowance derive from National Tax and Customs Administration

The above mentioned two parts of the social costs and subsidies make up the amount of other elements which is divided by hours worked to get the nominator of other elements index.

In both cases (wages and salaries index and other elements index) the denominators of the index are computed as quarterly averages of hourly wages and salaries and other elements of compensation in 2012 (for enterprises employing 5-49 employees).

Quality assurance is a continous work, all HCSO's surveys have to apply the general rules of HCSO's Quality Assurance Policy.

Estimations for the labour cost index derive from the Quarterly Labour Report and the Labour Cost Surveys.

Data are checked for completeness and consistency. From 2013 data have been collected by a new electrical program called Elektra. The quality of the data is better because logic test system is built in the program, so the report can be sent by the sender only if the data are checked and coherent. The published date are revised continuosly.

The Hungarian labour concepts are harmonized with the Eurostat regulations. Besides informing the Hungarian users, the result of the surveys make it possible to meet international demands fully.

The table contains the most important characteristics of the main users of labour cost index data.

We have no special user satisfaction survey on the Labour Cost Index, however we try to evaluate user needs and their opinions. Requests concerning LCI sent to HCSO using mail, phone and other communication channels are taken into account during the questionnaire design or  preparation of publications. Our users are generally satisfied. See 5.1. Relevance - User Needs.

The implementation of Regulation (EC) No. 450/2003 of the European Parliament and of the Council concerning the labour cost index is quite good.

• Definitions are fully implemented, except the inclusion of unpaid overtime hours.
• Scope is also fully implemented taking account of NACE sections (sections B to S of NACE Rev 2.); however data of small enterprises (the ones with less than 5 employees) are missing.
• Breakdown of variables is fully implemented for the years 2008-2022.
• Sources: see 3.1. Source data.

100%

For compiling the LCI we use direct information from the Quarterly Labour Report, an estimation method using coefficients from the Annual Labour Costs Survey with reference to the previous year and the current regulations on social contributions and taxes.

We have information on sampling error of average hourly wages and salaries of full time employees based on the Quarterly Labour Report.

To have an overall view on sampling errors we show the sampling error of average hourly wages and salaries per full-time employee.

Average hourly wages and saleries standard error in % in 2022

* Most of Educational and Health institutions do not belong to the business sector, they are budgetary organs

For the various sampled strata and their aggregations, the unit response rate was:

Where:

Sampling and response rates achieved by industries (Quarterly Labour Report, September 2022)

 * We perform aggregate observations of the school district centers and local government agencies.

As mentioned earlier, data of the non-respondents in the fully scoped part are imputed by algorithm and in the sampled part by strata average. Otherwise data are imputed by zero values, i.e. data of those non-respondents, of which we supposed that they would have sent us a questionnaire with only zero values. Other non-respondents are considered to belong to the population, but not to the sample.

The estimations of the labour cost index derives from the Quarterly Labour Report. The table in QLI_LCI_annex1_2022_hu.xlsx presents the data deriving from different sources (Quarterly Labour Report, Labour Force Survey and SBS).

Comparing the LFS data the percentage of coverage in the whole economy (from sector B to S) is 87,1% in 2022. Data show some discrepancies by economic sectors. The “smallest” coverage can be found in NACE Rev.2 section O (Public administration and defence; compulsory social security: 36,8%) and section N (Administrative and support service activities: 87,1%).
As concerning the SBS survey’s estimates the overall employment coverage of the labour cost index estimates were 102,5% in 2022. The worst coverage occurred in Real estate activities (NACE L) 118,4%, in Accommodation and food service activities (NACE I) 114,8%.

Over-coverage-rate: 0,4-0,5%.

Not available.

We have no direct information on labour costs for the whole population of economic units. So see 3.4. and 6.3.5.

We have no direct information on these errors, however non sampling error and sampling errors are good measure to estimate the effect of it. See 6.2.1. and 6.3.

Unit adjusted non-response-rate: 4,7-5,1%.

Item non-response was theoratically not allowed, because of the nature of the questionnaire, imputation of different items was not possible due to lack of information. Field-workers tried to get all information, but if an enterprise refused to give answers to the whole questionnaire, it became non-response.

See below.

Wages and salaries elements (D11) of the quarterly labour cost index data are calculated from the Quarterly Labour Report. The data are summed up, then the quarterly amount of the wages and salaries are divided by the quarterly amount of the hours worked in order to get the quarterly wages and salaries per hour. As the survey contains data only concerning the wages and salaries, number of employees and hours worked, the missing data for calculating the numerator of the other elements (e.g. social cost) are estimated with a so-called ‘updating method’.

Updating method is the following:

• the compulsory social security contributions paid by the employers are computed by the multiplication of the current proportions of contributions and wages and salaries (D1211; its percentage within the social costs is about 90),
• other elements’ (D1212, D122, D4, D5) derive from the latest LCS.
• (additional information at 6.5. Data revision – policy, 3.1. Source data)

The above mentioned two parts of the social costs and subsidies make up the amount of other elements which is divided by hours worked to get the nominator of the other elements index.

In both cases (wages and salaries index and other elements index) the denominators of the index are computed as quarterly averages of hourly wages and salaries and other elements of compensation in 2012 (for enterprises employing 5-49 employees). 

Data of wages and salaries, hours worked derive from the Quarterly Labour Report.
Data of social costs: partly come from actual proportional measurement of employers’ compulsory social contributions regulated by the government, partly come from estimation based on the latest Labour Cost Surveys (now Community Labour Cost Survey 2016 reclassifies according to NACE Rev 2.)

There is quite a long time period, (theoretically) 18 months, while indirect information is available. Indirect information derive from the Annual Labour Costs Survey. Nowadays the only reliable information on social costs and taxes and subsidies is the annual labour costs survey concerning the economic units employing  at least 50 persons  in the business  sector. Our model assumption concerns the estimation of social costs and taxes and subsidies. See 3.1. However to develop our estimation model continuously we validate our estimate based on direct information derived from our Annual Labour Cost Survey.

The adjustment methods are carried out in every quarter. Method used: TRAMO SEATS. Software used: J Demetra+ version 2.2.0. Direct adjustment is used for almost all indexes (wages and salaries, labour costs other than wages and salaries, total labour costs excluding bonuses), but indirect adjustment is used for total labour cost index (total labour costs) and aggregtes (B_N, B_S etc).

The Hungarian Central Statistical Office carries out revisions every quarter. (In case of a serious error data can be corrected in the month the error becomes known. The final revision for the twelve months of the reference year takes place usually in July/August of the t+1 year when data of the annual labour survey for the reference year are available.

There is an additional checking procedure of the monthly short-term statistics, when its data are compared to SBS data in year t+1 and if it is necessary, corrections are either made on the monthly labour data or on the annual labour data. The final data are available in the inner database of the Office (dissemination database) in August of year t+1. Revision of LCI estimates follows the revisions of the STS data.

There is a special revision policy for LCI using the calculation of the latest results of the annual Labour Cost Survey. The results of the survey are available 18 months after its reference year. We calculate the LCI data with the new data deriving from LCS (which consist D1212 (collectively agreed, contractual and voluntary social security contributions), D122 (employers’ imputed social contributions), and D5 (subsidies received by the employer) for the enterprises employing at least 50 persons.

The revision is made from the given year available (however LCI data will be revised, when results of LCS for the earlier year have been available), and using the combined results of LCS for the earlier years (referring to small enterprises with 5-49 employees).

There were not any revisions in 2022.

The deadline of returning the questionnaires is 12 days after the end of the reference quarter. This deadline is suitable for the enterprises since the payroll has to be completed at the end of the reference month. One month after the end of the reference period the data are being processed and tested further at micro and macro level. First results are released 50 days after the end of the reference quarter. First releases do not include labour cost index data. LCI estimations are made every quarter after the first release of the last month of the given quarter. There is a revision policy according to the revisions and corrections are made in the labour database. LCI estimations are published in the Labour Market Trends. The results are transmitted to the Eurostat. The quarterly publications (Labour Market Trends) come out in 70 days after the reference quarter.

• First results are released 50 days after the end of the reference month. First releases do not include labour cost index data.
• LCI estimations are made every quarter after the first release of the last month of the given quarter and released after 70 days of the reference quarter.

70 days after the end of the reference period

See 7.1 Timeliness.

One of the main components of the quarterly GDP calculated by the income approach is the compensation of employees. ESA2010 data are available from 2021 onwards according to Nace Rev. 2. See QLI_LCI_annex3_2021_hu.xlsx.

Not available

Optional

There were some slight changes in methodology (changes in definitions and change of NACE classifications) in 1999 and in 2004 in the base sources (STS).
NACE: the change from NACE Rev.1 to NACE Rev.2 required the back casting of the indices. In 2008 all surveys involving STS and Labour Cost Survey were processed by both NACE classifications. Back casting of the indices for the previous years (until 2000) was based on coefficients deriving from the results of 2008 surveys.

From 2000 (total labour costs excluding bonuses from 2002).

The definition of wages and salaries are the same in the two surveys (LCI and SBS), and year by year the results of consistency between the two surveys improved at the level of enterprise, but the coverage of the datasets is different. The difference at aggregate level (NACE B-S) was 0,4% in 2022. The highest gap is 1,5% in Real estate activities(NACE L) and 1,2% in Accommodation and food service activities(NACE I).

See 8.3. Coherence – cross domain.

Problems in the coherence examinations can be explained as follows:

• Discrepancy in coverage. NA data derive from the balance sheet (accountancy records) involving all incorporated enterprises. STS does not cover the economic units employing less than 5 persons.
• The estimation of hours worked in National Accounts is based on the LFS, in which there is a possibility of misclassification between NACE section.

See 8.3. Coherence – cross domain.

Not published

Labour cost index data are available for every user on the Internet, on the homepage of Hungarian Central Statistical Office (Labour Market Trends – quarterly published).  This publication involves short analysis on recent developments of the index. Other publications (bulletins, booklets, yearbooks) are available via Internet.
http://www.ksh.hu/apps/shop.main?p_lang=EN

Because of the nature of Hungarian LCI we don't have an online database.

We publish the main data in the Stadat tables.
https://www.ksh.hu/stadat_files/mun/en/mun0160.html

Because of the nature of Hungarian LCI we don't have microdata access.

Optional

Not available.

European Statistics Code of Practice is on the website of HCSO with the Peer review on the implementation of European Statistics Code of Practice. Quality policy of HCSO is on the website of HCSO. The international legal acts (Commission Regulation (EC) No 1216/2003 of 7 July 2003 implementing Regulation (EC) No 450/2003 of the European Parliament and of the Council concerning the labour cost index) prescribe the scope, the level of detail and the deadlines, furthermore the implementation of definitions of variables. The processing of manufacture of data has been developed according to these regulations.

We don't have numerical information on the rate. Metadata completeness is quite good, and continously developed.

Optional

Estimations for labour cost index derive from the Quarterly Labour Report and Labour Cost Surveys.

Confidentiality policy of the Hungarian Central Statistical Office: https://www.ksh.hu/docs/bemutatkozas/eng/avpol_web_eng.pdf

Because of the nature of Hungarian LCI no direct rules applied for treating the data set to ensure statistical confidentiality and prevent unauthorised disclosure.

Rules have to be applied in case of the main data sources, i.e. concerning Monthly Labour Report's datasets.

Prohibition of disclosure of individual-level information. During dissemination of statistical data taking into account the relevant provisions of the law – disclosure of individual-level information is strictly forbidden.

Sensitivity rules: A cell is considered unsafe, when the cell frequency is less than 3. We use this minimum frequency rule together with a dominance rule.

Labour cost

The labour cost index covers the compensation of employees (e.g. wages and salaries) and other labour costs (employers’ social contributions, subsidies and taxes) together. Compensation of employees is defined as the total remuneration in cash or in kind, paid by the employer to the employee in return for work done by the latter during the reference period i.e. wages and salaries (in Hungarian publications called wages and salaries by SNA), and all the statutory, contracted or voluntary social contributions or costs paid by the employer (together called employers’ social contributions).

Vocational training costs, other labour cost expenditure are excluded; taxes are added while subsidies are subtracted from the compensation.

Hours worked

The definition of hours worked contains the total number of hours actually worked by persons in employment or employees engaged in second or more jobs, including the ‘idle time’ spent and accounted at the workplace. The latter involves all hours of working time spent at workplace during which no work was done due to e.g. shortage of raw material, occasional lack of power, machine stoppages, accidents etc., but which is paid in accordance with the labour contract. The time of journey from one workplace to another one and ‘idle time’ due to weather factors belong to hours worked.

Lunchtime, time of paid leave, time spent in special security service, time on duty and occasional stand-by time during which the employee does not have to to stay at the workplace are not meant in the hours worked (with the exception of the compulsory working time according to work schedule of persons working, completely or partly, in stand-by jobs).