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 Business Statistics and Statistics of Transport" Department
1.3. Contact name
Confidential because of GDPR
1.4. Contact person function
Confidential because of GDPR
1.5. Contact mail address
2, P. Volov Street, Sofia 1038 Bulgaria
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
Monthly industrial production index. The main source of information is the monthly inquiry on “Sales in industry, construction, trade and services”. Data collected with that inquiry are receipts from sales (turnover), export turnover, value of goods purchased for resale in the same conditions as received and stocks of finished products (from own production) of industrial enterprises (at the beginning and at the end of the month).
3.2. Classification system
Classification of Economic Activities in the European Community (NACE Rev.2)
3.3. Coverage - sector
The survey covers NACE Rev. 2 Sections B, C, D and Division 36. The survey covers enterprises with 10 or more persons employed in the previous reference year.
3.4. Statistical concepts and definitions
Industrial Production index aims to follow the monthly change in the value of industrial production. The industrial production index is compiled from deflated value of production.
For the purpose of monthly survey the production value includes:
+ Receipts from sales of finished products, goods and services
- Value of goods purchased for resale in the same conditions as received
+/- Change in stocks of finished products (from own production)
Deflation is done with the Producer price indices, resulted from the monthly PPI survey.
3.5. Statistical unit
Reporting unit and observation unit is enterprise (Legal unit considered as enterprise).
3.6. Statistical population
The total population includes about 36 600 industrial enterprises. The sampling population includes about 7 000 enterprises with more than 10 persons employed of which sample of approximately 3 700 units are surveyed each month. The surveyed enterprises have an approximate share of 98.0% in the total turnover in industry.
3.7. Reference area
All regions of Bulgaria are covered.
3.8. Coverage - Time
The indices are calculated from January 2000 to the last month of the reference year.
3.9. Base period
2021
Index
Month
6.1. Institutional Mandate - legal acts and other agreements
National Legal basis: The Law on Statistics (Article 20) states that all persons that manage or represent legal persons, are obliged to provide the National Statistical Institute and the Bodies of Statistics with reliable data when filling-in forms and questionnaires on surveys included in the National Statistical Programme, for which mandatory participation is provided.
Obligation on units to provide data: The survey is included into the National Statistical Programme – it is compulsory. The Law of statistics includes provisions relating to fines, which can be imposed for refusal, non-response and delay. They are laid down in specific Chapter 9 “Administrative and liability provisions” in article 52.
EU Legislation:Regulation (EU) 2019/2152 of the European Parliament and of the Council on European business statistics, repealing 10 legal acts in the field of business statistics (EBS-Regulation) and the Commission implementing Regulation 2020/1197.
All EU documents following the EBS-Regulation can be found in the STS section on Eurostat’s website under Statistics => Short-term business statistics => Legislation
6.2. Institutional Mandate - data sharing
According to the EBS Regulation data is transmitted to Eurostat.
Industrial production index is sent to the International Monetary Fund and United Nations Statistics Division.
7.1. Confidentiality - policy
The Law on Statistics includes provisions relating to data confidentiality and protection of secrecy, which are in compliance with the EU legislation and standard European practice. In the Law is stated that individual data received and collected at the time of statistical surveys shall be statistical confidentiality and shall be used only for statistical purposes. Individual data, obtained for the purposes of the statistical surveys, shall not be used as evidence to the bodies of the executive powers and the judicial authorities. Requirements, obligations and measures for data protection, as well as fines that can be imposed, are treated in separate chapters, namely Chapter 6 on "Protection of secrecy” and Chapter 9 on "Administrative penal provisions”.
7.2. Confidentiality - data treatment
Article 25 of the Law on Statistics specifies the criteria, according to which certain data cannot be made public or released – individual data, data which could be indirectly identified, as well as aggregations of less than three statistical units and/or when the relative share of one parameter is more than 85% of the total volume of this parameter for all population units.
Data, which according to the Law on statistics are confidential, are not published.
8.1. Release calendar
Industrial production indices are disseminated nationally according to a predefined Release calendar presenting the results of the statistical surveys carried out by the National Statistical Institute, published at the beginning of the year on the NSI website.
The Release date for the publication of Industrial production index in the Release calendar is included in the item “Sales in industry, construction and trade”.
Data are simultaneously disseminated to the public through a news release on Industrial production index. Indices of industrial production are published on the NSI website, after which those data are simultaneously accessible to all users.
No data produced and disseminated by the National Statistical Institute are provided to government officials outside of the NSI before being disseminated to the public.
Industrial production indexes are transmitted to Eurostat at the time of national publication with the required level of detail. Indices are also transmitted to various institutes and ministries.
Monthly
10.1. Dissemination format - News release
NSI publishes regular monthly press release on Industrial production indexes 37-40 days after the end of the reference month.
10.2. Dissemination format - Publications
The Industrial Production Index is published on a monthly basis on the NSI official website.
10.3. Dissemination format - online database
Industrial Production Indexes time series (from January 2000) at 2-nd digit and higher level of aggregation of NACE Rev.2 are available on the website of NSI – in the on-line database INFOSTAT.
10.4. Dissemination format - microdata access
Micro data are neither distributed nor provided to the users.
10.5. Dissemination format - other
Industrial production indexes are transmitted to Eurostat at the same time with their national publication.
Data transmission to Eurostat is performed each month. The transmission is carried out in SDMX-ML format via eDAMIS system.
Dissemination on request: Special tables may be obtained from “Relation with users and e-services” department of NSI.
10.6. Documentation on methodology
The methodology for Industrial production index compilation is available on the NSI website.
10.7. Quality management - documentation
Information on PPI quality indicators can be found in the methadata presentaion on the NSI website.
11.1. Quality assurance
The legal framework which regulates the production of data from Short-term business statistics states that statistical data compiled within the EU system must be of satisfactory quality and produced data are comparable between Member States. Verification of data is done through the process of conducting a statistical survey - from respondents, from experts in the regional statistical offices and from national experts in NSI. Identified inaccuracies and errors are removed timely and comparable data of good quality are provided.
The quality of statistical information and its production process is ensured by the provisions of the European Statistics Code of Practice. The documentation on quality management is available on the NSI webpage.
11.2. Quality management - assessment
The quality assessment of the Index of industrial production is based on the quality requirements of the ESS (relevance, accuracy, timelines and punctuality, accessibility and clarity, coherence and comparability).
The industrial production index is published in line with EBS regulation. The major break in the time series was in 2009 due to the introduction of NACE Rev. 2. In this case the time series were recalculated (backcasted) back to 2000. Indices are revised twice - every month (current month date is considered as preliminary) and an annual revision where data are compared with SBS data.
The average response rate is about 99 per cent. The coefficient of variation used as the main accuracy of estimate measurement is about 0.2 per cent. There are no known quality problems.
12.1. Relevance - User Needs
The IPI has the following primary uses:
the IPI is used as a short-term economic indicator in its own right. The Presidency, National Assembly, Ministries, National Bank and other representatives of central and local government monitor the IPI as an important short-term indicator
the IPI contributes to the BG quarterly National Accounts. Directorate "Macroeconomic Statistics" is the main internal user of IPI
the IPI is required from Statistical Offices of the European Community (Eurostat). IPI is provided to Eurostat as well as to other multi-national organizations - EC; ECB; UN; IMF; WB.
12.2. Relevance - User Satisfaction
Not available.
12.3. Completeness
The scope and the level of detail meet the requirements. The time series starts form January 2000.
13.1. Accuracy - overall
Sampling: The sampling method used is Stratified Random Sample.
The total population includes about 36 600 industrial enterprises. The sampling population includes about 7 000 enterprises with more than 10 persons employed of which sample of approximately 3 700 units are surveyed each month. The surveyed enterprises have 98.0% share in the total turnover in industry.
Measurement errors: The common errors are measurement errors where the respondent misunderstands what information is required or incomplete performance. Most of measurement errors are detected and corrected through the logical and arithmetical controls of the Information system “Short-term statistics” (Online Information system for data collection, editing and processing of the STS information). The system carries out the statutory and warning controls and in a case there are errors, it displays a warning dialog error in the report. The system does not allow completion of the report until they have corrected these errors by statutory controls.
Data processing: Production value is deflated using producer price indices and weighted using value added from SBS survey. Errors and inconsistencies in those data sources may affect accuracy of the Industrial Production Index.
Revisions: Every month data is revised so data changes after the first release (data for the current month are published as preliminary and data for the previous month as final). For the reference year 2024 the revisions for the first month after the first release did not exceed 2%. There were no signs of bias in the first estimates based on this analysis.
13.2. Sampling error
The sample method used is Stratified random sample. Criterion for stratification of the enterprises is ‘number of persons employed’. First enterprises are stratified by groups (3-th digit level of NACE Rev.2) and then in each group they are stratifies by number of persons employed. Enterprises with 100 or more persons employed are surveyed exhaustively. The enterprises with employed from 10 to 99 are randomly sampled.
Coefficient of variation is used as an indicator on precision. It is computed as a percentage.
The Coefficient of variation for June 2024 for Total Industry is 0.3%.
13.3. Non-sampling error
Non-response: The size-weighted unit response rate for September 2024 is 0.8.
Actions to reduce the non-response include:
Sending the informative reminder letters to the respondents;
Depending on the relative importance of non-respondents in terms of their size or representativeness in the activity, the online or postal survey is supplemented by telephone calls and/or direct contact and collection of data for enterprises.
The item non-response occurred for a negligible part of the surveyed units, less than 0.1%.
Imputation: The data for non-responding units are imputed. The imputation is done only in case where there is information that the non-respondent units are active.
Technical errors: are often recognized with automated checks during the data editing process. A lot of checks are carried out automatically, but under of human interaction. Original data from enterprises for a first time are checking from experts in RSOs. Final checks of data from respondents and the aggregated data analysis are done by responsible experts in Central office.
Coverage errors: The Register of statistical units (RSU) is the source of the sampling frame. RSU has direct online connection with the Trade Registry so the main characteristics of surveyed units are on-line updated. During the year the ceased enterprises are removed from the sample and new born enterprises from the exhaustive strata are included in the sample.
Multiple listing: All statistical units in the RSU have unique identification number, so the multiple listing cannot occur.
14.1. Timeliness
The Industrial Production Indexes are firstly released at 11.00 o’clock of local time on the 37-40 days after the end of the reference month. Data for the current month are provisional when first released. Provisional data are identified in footnotes. During the current month more precise data for the previous month are collected. Each month data for the current month is released as preliminary and data for the previous month as final.
14.2. Punctuality
Industrial production indexes are published 37 - 40 days after the reference month. The data are released exactly as planned in the Release calendar.
Data are transmitted to Eurostat at the day of the national publication.
15.1. Comparability - geographical
The methodology applied for calculation of the index of industrial production is in line with international standards. The same concepts are used within the whole national territory.
15.2. Comparability - over time
The monthly indices are presented as time series from January 2000. Indexes are fully comparable over time.
15.3. Coherence - cross domain
Industrial production indexes are compared with production value calculated on the base of annual Structural business statistics (SBS) survey. Also SBS is the source of weights for calculation the STS indicators.
Data received from SBS and those receives from STS are coherent. The existing differences mainly are due to:
Coverage. SBS survey is conducted exhaustively and all enterprises are legally obliged to submit their Annual Reports to the National Statistical Institute. The STS surveys are sampling surveys. The target population for IPI survey is consists of enterprises with 10 and more persons employed.
Definition. Annually with SBS survey all elements of ‘production value’ is collected. Monthly for compilation of industrial production index only main elements of production value are collected.
Weighting. For compilation of higher-level monthly industrial production indexes are weighted by value added at factor cost taken from the structural business statistics survey for the base year. Indexes calculated from SBS data are not weighted.
15.4. Coherence - internal
All NACE levels is consistent as Industrial production indexes at higher levels of aggregation (2-nd digit level and higher) are calculated as weighted average - indexes at lower levels are weighted with the value added at factor cost structure.
In the case of WDA and SA data, direct adjustment is used.
Since the beginning of 2012 NSI has collected and calculated information online via Information System.
17.1. Data revision - policy
Industrial production indices usually are revised two times:
each month data for the current month are published as preliminary and for the previous month as final
at the beginning of each year the indexes for the previous year are reconciled on the basis of the comprehensive annual data from SBS.
There is no revision calendar. At the time of each dissemination, users are informed whether data are preliminary or revised and the reason why data are revised (e.g. availability of additional data, data are reconciled etc.).
The same revision policy is applied to STS data released nationally and transmitted to Eurostat.
17.2. Data revision - practice
Minor changes are made at the time when the revision is introduced and they are marked in footnotes in the relevant published tables.
The data for the current month are marked as ‘preliminary data’.
Only final data are available in the time series posted on the NSI website.
Mean Absolute Revision (MAR) and Mean Revision (MR) are used as revision indicators. They are calculated on year-on-year growth rates for original and calendar adjusted date and on period-on-period for the seasonally adjusted data.
MAR and MR for original, seasonally adjusted and calendar adjusted data for 2022-2024:
MR
MAR
Original data
0.26
0.29
Seasonally adjusted data
0.20
0.51
Calendar adjusted data
0.32
0.64
18.1. Source data
The indexes are based on monthly survey “Sales in industry, construction, trade and services”.
Stratified sample survey which frame is based on Register of statistical units. The total population includes about 36 600 industrial enterprises. The sampling population includes about 7 000 enterprises with more than 10 persons employed of which sample of approximately 3 700 units are surveyed each month. The surveyed enterprises have 98.0% share in the total turnover in industry.
The sampling method used is Stratified Random Sample. Criterion for stratification of the enterprises is ‘number of persons employed’. First enterprises are stratified by groups (3-th digit level of NACE.BG-2008) and then in each group they are stratifies by number of persons employed. Enterprises with 100 or more employees are surveyed exhaustively. Enterprises employing between 10 and 99 employees are randomly sampled, enterprises that have less than 9 employees are not observed monthly.
The sample is updated every year.
18.2. Frequency of data collection
Monthly
18.3. Data collection
One questionnaire is used - “Main short-term indicators”. The information collected with that questionnaire regards:
receipts from sales (turnover)
of which receipts from sales on non-domestic market
value of goods resold without further processing
stocks of finished products (value at the beginning and at the end of the month).
Data collection media:
online through Information system “Business Cycles”
paper questionnaire. The questionnaires are sent out to the respondents and collected by regional statistical offices, where the information is entered into IS BC.
To reduce non-response IS BC has functionality to send an electronic letters informing respondents about the upcoming observations, in which they are included. To enable them to attach and reporting forms in PDF format and official letters.
18.4. Data validation
Data validation starts from the filling of the reports by respondents. Control of the data can be accessed from both the regional statistical offices and from experts in the National Statistical Institute. An information system ‘Business Cycles’ significantly contributes to the validation of the data submitted by set of arithmetic and logical controls.
The system carries out the statutory and warning controls. In case there are errors, it displays a warning dialog error in the report. The system does not allow completion of the report while the errors are not corrected. Data validation system incorporates many tools for data control, estimations and detail reports on data completeness. Subsequent analysis of the experts, a phone call for further clarification of data and elimination of disparities, minimize errors in the raw data and ensure accuracy of outputs at national level.
Validation of the output data - output data are verified at group level with regard to the previous month, with the same month of the previous year and also with regard to extreme values, if necessary data are validated at the level of statistical unit.
Before transmission to Eurostat, the structure of the data file and format is checked in detail.
18.5. Data compilation
Estimates for non-response: Non-response is treated using imputation techniques, depending on specific cases. In the case of non-responding enterprises from exhaustive strata imputation is made for each non-respondent enterprise separately using different techniques (estimates based on responses from the same enterprise in previous months, estimates based on responses from similar units, estimates based on ratios within the corresponding NACE activity and size). Non-responses in sampling strata are estimated by adjusting the extrapolation coefficients to take account of response rates.
Estimates for grossing-up to population levels: The grossing up procedure follows the sampling technique. The sample is representative at the 3-digit level of NACE Rev. 2. For enterprises with 10 or more and less than 100 persons employed (sampling strata) extrapolation of the results to the population is made by using extrapolation coefficients which are based on the number of the units selected and their total number in the particular strata.
The grossed up results are used only to compile indexes, they are not disseminated in their own right. The calculation of the Industrial production index involves a two stage calculation process. In the first stage the index calculation is done on the basis of a direct comparison of the data from the current month at constant prices and data on the average monthly production value in the base year. The production value is deflated at 3-digit level of NACE Rev. 2 with the Total producer price index that is calculated as a weighted average of producer price indexes on domestic market and non-domestic markets.
The second stage is the calculation of the Industrial production index at higher levels of aggregation. Industrial production indexes at higher levels of aggregation (Division, Section, Main industrial groupings and Total Industry level) are calculated as base weighted (Laspeyres) indices, in other words they are weighted according to the value added at factor costs in the base year.
18.6. Adjustment
In accordance with EBS Regulation, the Industrial Production Indices are to be seasonally and working day adjusted. The software used for seasonally and working day adjustment is JDemetra +. The method used is TRAMO-SEATS.
The time series at different levels of aggregation (division, section, MIGs and total industry) are seasonal and working day adjusted by applying the direct method.
Revisions of seasonally and working day adjusted data:
model updated once a year including revision of historical data
monthly observations presented with revision 1 month backwards.
More information about the seasonal adjustment in Annex.
Monthly industrial production index. The main source of information is the monthly inquiry on “Sales in industry, construction, trade and services”. Data collected with that inquiry are receipts from sales (turnover), export turnover, value of goods purchased for resale in the same conditions as received and stocks of finished products (from own production) of industrial enterprises (at the beginning and at the end of the month).
3 June 2025
Industrial Production index aims to follow the monthly change in the value of industrial production. The industrial production index is compiled from deflated value of production.
For the purpose of monthly survey the production value includes:
+ Receipts from sales of finished products, goods and services
- Value of goods purchased for resale in the same conditions as received
+/- Change in stocks of finished products (from own production)
Deflation is done with the Producer price indices, resulted from the monthly PPI survey.
Reporting unit and observation unit is enterprise (Legal unit considered as enterprise).
The total population includes about 36 600 industrial enterprises. The sampling population includes about 7 000 enterprises with more than 10 persons employed of which sample of approximately 3 700 units are surveyed each month. The surveyed enterprises have an approximate share of 98.0% in the total turnover in industry.
All regions of Bulgaria are covered.
Month
Sampling: The sampling method used is Stratified Random Sample.
The total population includes about 36 600 industrial enterprises. The sampling population includes about 7 000 enterprises with more than 10 persons employed of which sample of approximately 3 700 units are surveyed each month. The surveyed enterprises have 98.0% share in the total turnover in industry.
Measurement errors: The common errors are measurement errors where the respondent misunderstands what information is required or incomplete performance. Most of measurement errors are detected and corrected through the logical and arithmetical controls of the Information system “Short-term statistics” (Online Information system for data collection, editing and processing of the STS information). The system carries out the statutory and warning controls and in a case there are errors, it displays a warning dialog error in the report. The system does not allow completion of the report until they have corrected these errors by statutory controls.
Data processing: Production value is deflated using producer price indices and weighted using value added from SBS survey. Errors and inconsistencies in those data sources may affect accuracy of the Industrial Production Index.
Revisions: Every month data is revised so data changes after the first release (data for the current month are published as preliminary and data for the previous month as final). For the reference year 2024 the revisions for the first month after the first release did not exceed 2%. There were no signs of bias in the first estimates based on this analysis.
Index
Estimates for non-response: Non-response is treated using imputation techniques, depending on specific cases. In the case of non-responding enterprises from exhaustive strata imputation is made for each non-respondent enterprise separately using different techniques (estimates based on responses from the same enterprise in previous months, estimates based on responses from similar units, estimates based on ratios within the corresponding NACE activity and size). Non-responses in sampling strata are estimated by adjusting the extrapolation coefficients to take account of response rates.
Estimates for grossing-up to population levels: The grossing up procedure follows the sampling technique. The sample is representative at the 3-digit level of NACE Rev. 2. For enterprises with 10 or more and less than 100 persons employed (sampling strata) extrapolation of the results to the population is made by using extrapolation coefficients which are based on the number of the units selected and their total number in the particular strata.
The grossed up results are used only to compile indexes, they are not disseminated in their own right. The calculation of the Industrial production index involves a two stage calculation process. In the first stage the index calculation is done on the basis of a direct comparison of the data from the current month at constant prices and data on the average monthly production value in the base year. The production value is deflated at 3-digit level of NACE Rev. 2 with the Total producer price index that is calculated as a weighted average of producer price indexes on domestic market and non-domestic markets.
The second stage is the calculation of the Industrial production index at higher levels of aggregation. Industrial production indexes at higher levels of aggregation (Division, Section, Main industrial groupings and Total Industry level) are calculated as base weighted (Laspeyres) indices, in other words they are weighted according to the value added at factor costs in the base year.
The indexes are based on monthly survey “Sales in industry, construction, trade and services”.
Stratified sample survey which frame is based on Register of statistical units. The total population includes about 36 600 industrial enterprises. The sampling population includes about 7 000 enterprises with more than 10 persons employed of which sample of approximately 3 700 units are surveyed each month. The surveyed enterprises have 98.0% share in the total turnover in industry.
The sampling method used is Stratified Random Sample. Criterion for stratification of the enterprises is ‘number of persons employed’. First enterprises are stratified by groups (3-th digit level of NACE.BG-2008) and then in each group they are stratifies by number of persons employed. Enterprises with 100 or more employees are surveyed exhaustively. Enterprises employing between 10 and 99 employees are randomly sampled, enterprises that have less than 9 employees are not observed monthly.
The sample is updated every year.
Monthly
The Industrial Production Indexes are firstly released at 11.00 o’clock of local time on the 37-40 days after the end of the reference month. Data for the current month are provisional when first released. Provisional data are identified in footnotes. During the current month more precise data for the previous month are collected. Each month data for the current month is released as preliminary and data for the previous month as final.
The methodology applied for calculation of the index of industrial production is in line with international standards. The same concepts are used within the whole national territory.
The monthly indices are presented as time series from January 2000. Indexes are fully comparable over time.
