Reference metadata describe statistical concepts and methodologies used for the collection and generation of data. They provide information on data quality and, since they are strongly content-oriented, assist users in interpreting the data. Reference metadata, unlike structural metadata, can be decoupled from the data.
Short-term Statistics Department, Short-term Statistics of Construction and Services Section
1.3. Contact name
Confidential because of GDPR
1.4. Contact person function
Confidential because of GDPR
1.5. Contact mail address
H-1525 Budapest P.O.B. 51, Hungary
1.6. Contact email address
Confidential because of GDPR
1.7. Contact phone number
Confidential because of GDPR
1.8. Contact fax number
Confidential because of GDPR
2.1. Metadata last certified
3 June 2025
2.2. Metadata last posted
3 June 2025
2.3. Metadata last update
3 June 2025
3.1. Data description
The purpose of data collection is to provide volume index of construction and give information about gross output value of building construction, civil engineering and total of construction activities. The survey covers Sections F of NACE Rev. 2. All enterprises are observed if the number of persons employed is greater than 49. For enterprises employing between 5 and 49 persons the data collection is based on representative sample survey. When classifying an enterprise only the number of employees employed by the given firm is taken into account. The total number of data providers is some 1,600. The data of enterprises having less than 5 persons employed are estimated. The sampling frame arises from a fixed file extracted from the Business Register. The selection of the sample from the sample frame is enacted through random stratified sampling. The Paasche-type volume indices are calculated on the basis of corresponding period of the previous year. Fixed-based volume indices are released too and in this case monthly average of 2021 is equal to 100.
The survey covers Sections F of NACE Rev. 2. The data of enterprises having fewer than 5 persons employed are estimated at the 3-digit level and for buildings and civil engineering works, based on the production value of the smallest observed size-group (5-9 persons employed) Enterprises with 5 and more persons employed are surveyed.
3.4. Statistical concepts and definitions
Construction production: construction of new buildings, additions, alterations (changing functional purposes, life period enhancement), improvements and renovations, as well as repairs and maintenance made on the existing buildings and on other structures. Demolition is included.
As regards the definitions of variables used in practice of the HCSO the Commission Implementing Regulation 2020/1197 is the standard source.
3.5. Statistical unit
KAU
3.6. Statistical population
The target population consists of enterprises included in the construction area in the Section F of NACE Rev. 2.0. The number of possible data providers is about 13,500.
3.7. Reference area
Territory of Hungary.
3.8. Coverage - Time
Data are available from 2000 onwards. The change to TEÁOR’08, which is the national version of NACE Rev. 2., caused the last break in the time series.
3.9. Base period
Base year: 2021
Fix based index
Month
6.1. Institutional Mandate - legal acts and other agreements
At European level
Regulation (EU) 2019/2152 of the European Parliament and of the Council on European business statistics concerning short term statistics
Commission Implementing Regulation 2020/1197 laying down technical specifications and arrangements pursuant to Regulation (EU) 2019/2152 of the European Parliament and of the Council on European business statistics concerning short term statistics (General Implementing Act)
All relevant regulations can be found in the STS section on Eurostat’s website under Statistics / Short-term business statistics / Legislation
Regulation (EU) 2019/2152 of the European Parliament and of the Council on European business statistics concerning short term statistics.
7.1. Confidentiality - policy
At European level
Regulation (EU) 2019/2152 of the European Parliament and of the Council on European business statistics concerning short term statistics
Commission Implementing Regulation 2020/1197 laying down technical specifications and arrangements pursuant to Regulation (EU) 2019/2152 of the European Parliament and of the Council on European business statistics concerning short term statistics (General Implementing Act)
All relevant regulations can be found in the STS section on Eurostat’s website under Statistics / Short-term business statistics / Legislation
HCSO ensures confidentiality for all the data reported by data providers and the exclusive use of the data for statistical purposes. We disseminate only aggregated data in full compliance with the rules of confidentiality. Individual data, as well as aggregated data consisting of fewer than 3 enterprises are regarded as confidential and therefore not published. Researchers have access to de-identified data sets and to anonymised micro data for scientific purposes with appropriate legal and methodological guaranties in place. As for the employees, they can work with datasets in their competence with registered and controlled access rights. For details see Information on confidentiality for data providers on the website of HCSO.
8.1. Release calendar
All of the features of dissemination activity are consistent with the Dissemination and Communication Policy of Hungarian Central Statistical Office. In the elaboration of this document the
(Since the last update of Dissemination Policy of HCSO a new statistical law has been adopted in Hungary and the European Statistical law has been amended, but the changes do not affect the principals of Dissemination Policy.)
HCSO has a public Dissemination calendar that contains the dissemination dates of all First releases. The public dissemination calendar (Catalogue) on the website of HCSO provided information not only the first releases but the analyses, methodological publications, promotional publications, reports, statistical reflections, yearbooks and pocketbooks. Both calendars are prepared in line with the annual dissemination programme.
8.2. Release calendar access
A public dissemination calendar of the first releases is accessible and it can be downloaded in Excel format by everyone via the website of HCSO. It provides information about publish date, the date of any modification, the title of publication, the reference period or date and the date of inclusion in the database.
8.3. Release policy - user access
The date of dissemination is based on legislation rules for STS data set. The release calendar is open for everybody via the website of HCSO. There are some selected users, e.g. ministers, management of principal governmental offices and the Governor of the Central Bank. They got the data one day before the publication after the closing the Budapest Stock Exchange.
Monthly
10.1. Dissemination format - News release
First release about construction is published regularly every month including volume indices of production of construction (original, seasonally and working day adjusted, value at current prices).
Value and volume indices of production of construction by divisions.
Value and volume indices of production of construction by main groups of constructions.
Fixed-base volume indices of production of construction by main groups of constructions.
Value and volume indices of the stock of orders at the end of the period and new orders of enterprises in construction.
states monthly gross data for Section F of NACE Rev 2, from January 2000 onwards, as well as contains seasonally adjusted data and trend data.
In Section F we do not have arrangements with other data producing agencies. HCSO itself produces construction data.
In Section F we do not send STS data to OECD or UN.
10.4. Dissemination format - microdata access
In HCSO the following four data access channels are available only for researchers for scientific purposes. The HCSO performs a researcher accreditation procedure for all data requests for these four data access channels.
The HCSO offers access to deidentified microdata sets for scientific purposes in the safe environment of the Safe Centre operated by the HCSO in Budapest.
The offers access to deidentified microdata sets for scientific purposes in the safe environment of the remote access points operated by the HCSO under the same access conditions as the Safe Centre access.
For scientific purposes, the HCSO produces the requested research outputs inside its own safe environment based on the specifications/syntax files provided by the researcher.
By using this data access channel, the HCSO provides anonymised microdata sets for the researcher for scientific purposes.
10.5. Dissemination format - other
STADAT (Data tables Stadat) (the acronym stands for statistical data), which involves mostly time series can be found on the HCSOs website. STADAT includes plenty of data in pre-made tables about numerous topics, with methodological notes. The tables can be downloaded but the user cannot transform them. (While working with the dissemination database the user can assemble a cross table, elements to the table columns and rows can be added as well as filters can be applied).
Data transmission is fulfilled to Eurostat each month to be released as national data. The transmission is carried out in SDMX format via eDAMIS 4. SDMX Converter on EU websites is used for the transformation.
Additional methodological comments can be found in STADAT system.
10.7. Quality management - documentation
User-oriented quality reports on statistical domains are prepared in the framework of methodological documentation and are published as metainformation on the HCSO website: Methodological documentation . An internal HCSO regulation is in place regarding the preparation of producer-oriented quality reports for each statistical domain on a yearly basis.
In case of some statistical domains – concerning first releases – quality check is carried out and documented each month for the management of HCSO. However, this report is not published.
11.1. Quality assurance
The HCSO Quality Policy lays out the principles and commitments related to the quality ofstatistics.The documentis consistent with the goals set out in the Mission and Vision statements andwith the principles of the European Statistics Code of Practice and is publicly available on the HCSO website.
Quality Guidelines are meant to ensure the quality of the statistical processes. The document has been in place since 2007 (1st revision in 2009, 2nd revision in 2014 and 3rd revision is currently ongoing). The latest version (2014) is available on the HCSO website.
Procedures are in place in order to ensure updated documentation on product quality. (See above about Quality Documentation in 10.7) Apart from the internal reports, quality reports are regularly provided to Eurostat as well.
11.2. Quality management - assessment
Commission Implementing Regulation 2020/1197 laying down technical specifications and arrangements pursuant to Regulation (EU) 2019/2152 of the European Parliament and of the Council on European business statistics concerning short term statistics. (General Implementing Act)
12.1. Relevance - User Needs
The concepts and methods are based on European legislation. The main international user is Eurostat. The principal domestic users are the National Central Bank of Hungary and Ministries. The relevant internal user is the National Accounts department.
12.2. Relevance - User Satisfaction
Not available.
12.3. Completeness
The scope and the level of detail meet the requirements. 12 time series are transmitted to Eurostat. From dissemination database everybody can obtain additional time series.
13.1. Accuracy - overall
Sampling error is available for total construction output (B110). It is calculated for the strata and for some aggregates every month.
Although the response rate is fairly good the non-response is the main source of non-sampling errors. There are some enterprises which send late data, but their questionnaires can be taken into consideration over the revisions. Moreover, data are revised using the results of annual statistical survey of construction activities and SBS data, so the final version of the figures of industrial production is closed the one and half years after the beginning of the data collection.
The under-coverage of the survey population is rather small because the online connection between our Business Register and Registry Court assures a reliable sampling frame. A good value about the rate of over-coverage can be calculated on the basis of Business Register. The rate of other errors (for example classification error) is noted over the data collection, and the statistician can list these cases and can determine the number of those.
13.2. Sampling error
Coefficient of variation is calculated for strata, it is between 1.82% and 2.66% in 2024. The stratification variables are the staff categories of enterprises, the two-digit levels of NACE Rev. 2. (sometimes three- or four-digit levels are used) and the region (according to the headquarter of the enterprise whether is in the capital or in countryside). The sampling frame arises from a fixed file extracted from the Business Register using random numbers for the selection. The total number of data providers is about 1,600. There are some 1,200 data suppliers in the sample and circa 400 are observed exhaustively.
Coverage errors: the BR which is the source of the sampling frame has direct online connection with the Registry Court, therefore when the sample is constructed the sampling frame is timely. The sample and the full-scope part of the survey are updated every year. Over the given year sample is permanent apart from the ceased enterprises which are deleted, but the full-scope part of the survey is updated every month. Under-coverage can be arisen if an enterprise that is element of the full-scope part starts to work after the updating. In this case this data supplier goes into the data set next month.
As for the over-coverage we can measure it. In 2024 it was around 3.9%. (Annex)
Multiple listings is not known in the industry because all of the enterprises are identified by a unique number.
Access errors: HCSO uses an IT system (called ELEKTRA) to collect data via personalized online questionnaires. Data providers can check the questionnaire before sending, so the data are consistent. Administrative source is used only for the imputation. If data are inconsistent compared to the other months the data supplier is interviewed and the error is corrected.
Non-response errors: the weighted unit response rate was between 92.9% and 97.1%, the yearly average was 94.8% in 2024, which is a good value. Editing, coding and imputation errors: there is not numerical information.
Modelling errors: there is not numerical information.
HCSO is committed to reduce the non-response errors. In order to help the enterprises, we give assistance by phone or via e-mail, send automatic warning e-mails or mails. In a final case HCSO can impose fine to enterprises which do not cooperate.
Data of index of production of construction are published not later than 45 days after the end of the reference month following STS Regulation.
14.2. Punctuality
The publications agree to the dissemination calendar. There are no time lags.
15.1. Comparability - geographical
Over the data collection and calculation domestic and non-domestic markets are interpreted according to territory of Hungary. The same statistical concepts can be applied all of the territory. There are no difficulties concerning the geographical comparability.
15.2. Comparability - over time
Dissemination database consists of time series data from 2000. Between 2000 and 2008 the time series were recalculated in accordance with NACE Rev. 2, so they can be considered as continuous from 2000.
15.3. Coherence - cross domain
The values of indicator are compared with similar data which come from SBS. The difference is not relevant.
15.4. Coherence - internal
Direct adjustment method is used for seasonally adjustment. This method treats the time series separately, so there is no guarantee the time series of total production is always between the time series of production of building construction and time series the production of civil engineering.
Cost (NSI hours per month – 2017): 1000
Burden (Respondents hours per month – 2016): 426
17.1. Data revision - policy
HCSO published its renewed revision policy in 2018. The revision policy is in line with the following conceptual frameworks:
finally, the HCSO takes into consideration the deadlines of mandatory international data transmissions and publication
HCSO considers any change in the value of already published data as data revision. Revisions take place for a number of distinct reasons, which tend to break into four groups:
incorporation of better source data (e.g. replacement of first/provisional estimates based on expert judgements, or as a result of benchmarking)
capturing routine recalculations (e.g. updating the base period)
reflection of improved methodology (e.g. changes in concepts, definitions or classifications)
correction of errors
Considering the various causes of revisions and the different frequencies of publications, the HCSO – in correspondence with international guidelines – distinguishes the following types of revisions:
Routine revisions: routine revisions are changes in published data which are related to the regular statistical business process. Routine revisions mainly occur when the incorporation of late information (new or the correction of already obtained) modifies the already published results or in the case of benchmarking. Routine revisions are conducted periodically, according to a schedule drawn up in advance. With some sets of statistics (e.g. in foreign trade), a number of revisions are needed to obtain final results, while in other cases the provisional results are replaced by final data in the course of one single revision. Routine revisions barely affect the applied methodology, and only a few periods (some months or quarters) back in time are revised and longer revisions take place at a lower frequency, e.g. annually.
Major/methodological revisions: Major revisions are changes in published data, often substantial, which are due to changes in definitions, classifications and methodologies. Updating of the weights of the base year of an index series, the availability of a new structural source that is only collected at long intervals (5 to 10 years), such as the census, and the entry into force of a new legal act may also cause major revisions.
Major revisions are planned very well in advance and users are informed beforehand on the forthcoming major revisions. They are less frequent than routine revisions and occur only every 5 to 10 years. Since major revisions affect a large part of the time series and sometimes even the complete time series, it is necessary to back cast time series, otherwise major revisions would produce breaks and inconsistencies in them.
Unplanned revisions: Unplanned or unscheduled revisions are those that are not foreseen (as opposed to planned revisions), because they are a result of unforeseeable events and therefore it is usually not possible to pre-announced them in advance. As unscheduled revisions can undermine confidence in the quality of official statistics, HCSO is committed to avoiding as much as possible unscheduled revisions and to limit them to the case of important errors (whose correction results in significant improvement regarding data quality). Unscheduled revisions are communicated to the users in a transparent manner.
HCSO makes its general and domain specific revision policies publicly available on its official website. HCSO applies the general policy’s principles to all of its statistics (including the STS data transmitted to the Eurostat). Although subject matter statistics may have specificities regarding their revision practices, they have to be fully compliant with the general principles.
HCSO notifies users about forthcoming revisions in time and indicates their date and time in the revision calendar .
17.2. Data revision - practice
Planned revision
Routine revision
Data published in first releases are preliminary; there is a routine revision in each quarter.
Why to revise: in order to publish more correct data, data suppliers may revise their data. The main sources of these revisions are the questionnaires which are arrived late or the corrections which are initiated by data suppliers.
Frequency of revision: quarterly. Revised data are published in the release of the last month of the quarter. The indicators are recalculated from the beginning of the year, e.g. revised data of January-May are published with the first estimate of June.
First estimation: t-45 days, the revision practice has not been modified for a long time, there is no need to inform users in advance. Major revision
Major revision is carried out in July August after reference year, when data are compared with SBS results and with annual questionnaire of construction production. Only data of reference year are subject to change (e.g. in August 2022 data of Jan-Dec 2021 were revised. Revision of earlier years would be an unplanned revision.
Unplanned revisions:
If necessary, according to revision policy (chapter 17.1). There were no unplanned revisions in section F in the last years.
HCSO uses uniform revision policy and data transmission is carried out after revision. See the methodological document for further information and for the most important concepts the
The main source of data is a survey. Data can be reported in an electronic way. Data providers can be divided into two parts. All enterprises are observed if the number of persons employed is greater than 49. Sample survey is used for enterprises between 5 and 49 persons employed. Systematic sampling is based on the random order of enterprises. The sample is stratified; stratification variables are divisions and classes of NACE, number of employees and location of enterprise (capital/non-capital). The number of elements of the sample is approximately 1,200. The data of enterprises having fewer than 5 persons employed are estimated at the NACE Rev 2. 3-digit level and for buildings and civil engineering works, based on the production value of the smallest observed size-group (5-9 persons employed).
18.2. Frequency of data collection
Monthly
18.3. Data collection
Data collection is based on on-line survey. The design of the questionnaire including the set of checking rules is the job of the statisticians, and the check of the operation happens in close cooperation with IT department at the end of the previous year. The IT system sends an e-mail to data providers if a new questionnaire becomes accessible, and the deadline is expired. In case of non-response:
re-contact by phone and/or
sending a warning message and/or
fine as a final tool
18.4. Data validation
Reported data are checked by an IT tool (which was developed in the HCSO). This software controls the internal coherence using rules and considerations ordered by the experts who are responsible for the given data collection. For example, software checks that the part less than the whole, or that certain values are greater than 0. A list of errors is compiled automatically. The reports with the serious mistake are sent back to the data provider asking it for correction. There can be other problems (e.g. a variable varies from the value of previous month or corresponding period of previous year) when data suppliers can give an explanation and if the responsible expert accepts the justification, data will not be corrected.
The codes (NACE, Staff categories of enterprises, Geographic area, CULF) come from the Business Register. If one of the codes of data provider is changed over the year the reason is cleared up.
Before the data transmission the credibility examination is carried out. The experts review the values and the changing of the aggregates calculated 2 digits of NACE Rev. 2.
The data are transformed into SDMX form by online SDMX converter on EU website.
Before the finalisation of the data file the values are compared to the SBS data over a major revision, therefore the validation takes into consideration format and file structure (validation level 0), intra-dataset checking (validation level 1) and inter-dataset checking (validation level 2) happens.
18.5. Data compilation
The Paasche index form is used for the index calculation. For aggregation of production indices gross output weights of 3-digit divisions are used. The source of the weights is the Monthly Integrated Survey of Economic Statistics. The weights are changed every month of the reference year. For the sample survey one part of non-respondents is imputed by zero, the other part by the average of the corresponding stratum. For the full-scope survey: data for non-respondents are imputed based on the available date of the previous month of the unit; if it is not available, by the average of the units belonging to the same size group and activity.
18.6. Adjustment
Seasonal Adjustment
At the Hungarian Central Statistical Office, the TRAMO/SEATS method with JDemetra+ is used for seasonal adjustment in accordance with the ESS guidelines on seasonal adjustment. Model revision as default takes place yearly, and the results of the new model are published with the first data of the new year. During this step a fully automated model identification is made by TRAMO/SEATS, but the final decision is based on subject-matter domain information. During the year, with each new data the coefficients of the model are re-estimated, and revision of the seasonally adjusted data is made from the beginning of the time series. No forecast is made in any step of the process.
Calendar adjustment
For the calendar adjustment country specific calendar is used, which includes all the national holidays since 1990. Working day regressor is used for quarterly time series, and for monthly time series shorter than 15 years. For monthly time series which are at least 15 years long trading day regressors are used. The leap year effect is tested and used if it is significant.
Standard Easter effect is also used in the models if it has a significant effect and/or economic reason is available.
Outlier detection
Additive (AO), level shift (LS) and temporary change (TC) outliers are used during the seasonal adjustment process. In the yearly model revision, the results of the automatic outlier detection and the economic reasons are checked. At the end of the process outliers used in the previous year are changed (especially those of which were long time ago) if it is strongly reasonable.
During the year last outliers refresh strategy of the JDemetra+ is used, it means that the outliers in the last year of the sample are re-identified.
The purpose of data collection is to provide volume index of construction and give information about gross output value of building construction, civil engineering and total of construction activities. The survey covers Sections F of NACE Rev. 2. All enterprises are observed if the number of persons employed is greater than 49. For enterprises employing between 5 and 49 persons the data collection is based on representative sample survey. When classifying an enterprise only the number of employees employed by the given firm is taken into account. The total number of data providers is some 1,600. The data of enterprises having less than 5 persons employed are estimated. The sampling frame arises from a fixed file extracted from the Business Register. The selection of the sample from the sample frame is enacted through random stratified sampling. The Paasche-type volume indices are calculated on the basis of corresponding period of the previous year. Fixed-based volume indices are released too and in this case monthly average of 2021 is equal to 100.
3 June 2025
Construction production: construction of new buildings, additions, alterations (changing functional purposes, life period enhancement), improvements and renovations, as well as repairs and maintenance made on the existing buildings and on other structures. Demolition is included.
As regards the definitions of variables used in practice of the HCSO the Commission Implementing Regulation 2020/1197 is the standard source.
KAU
The target population consists of enterprises included in the construction area in the Section F of NACE Rev. 2.0. The number of possible data providers is about 13,500.
Territory of Hungary.
Month
Sampling error is available for total construction output (B110). It is calculated for the strata and for some aggregates every month.
Although the response rate is fairly good the non-response is the main source of non-sampling errors. There are some enterprises which send late data, but their questionnaires can be taken into consideration over the revisions. Moreover, data are revised using the results of annual statistical survey of construction activities and SBS data, so the final version of the figures of industrial production is closed the one and half years after the beginning of the data collection.
The under-coverage of the survey population is rather small because the online connection between our Business Register and Registry Court assures a reliable sampling frame. A good value about the rate of over-coverage can be calculated on the basis of Business Register. The rate of other errors (for example classification error) is noted over the data collection, and the statistician can list these cases and can determine the number of those.
Fix based index
The Paasche index form is used for the index calculation. For aggregation of production indices gross output weights of 3-digit divisions are used. The source of the weights is the Monthly Integrated Survey of Economic Statistics. The weights are changed every month of the reference year. For the sample survey one part of non-respondents is imputed by zero, the other part by the average of the corresponding stratum. For the full-scope survey: data for non-respondents are imputed based on the available date of the previous month of the unit; if it is not available, by the average of the units belonging to the same size group and activity.
The main source of data is a survey. Data can be reported in an electronic way. Data providers can be divided into two parts. All enterprises are observed if the number of persons employed is greater than 49. Sample survey is used for enterprises between 5 and 49 persons employed. Systematic sampling is based on the random order of enterprises. The sample is stratified; stratification variables are divisions and classes of NACE, number of employees and location of enterprise (capital/non-capital). The number of elements of the sample is approximately 1,200. The data of enterprises having fewer than 5 persons employed are estimated at the NACE Rev 2. 3-digit level and for buildings and civil engineering works, based on the production value of the smallest observed size-group (5-9 persons employed).
Monthly
Data of index of production of construction are published not later than 45 days after the end of the reference month following STS Regulation.
Over the data collection and calculation domestic and non-domestic markets are interpreted according to territory of Hungary. The same statistical concepts can be applied all of the territory. There are no difficulties concerning the geographical comparability.
Dissemination database consists of time series data from 2000. Between 2000 and 2008 the time series were recalculated in accordance with NACE Rev. 2, so they can be considered as continuous from 2000.
