Croatian Bureau of Statistics

Industry Statistics Departement

Croatian Bureau of Statistics, Ilica 3, 10000 ZAGREB, CROATIA

Indices of industrial production, derived from the Monthly Survey on Industrial Production Volume Index and Indices of Stocks, Persons Employed and Labour Productivity in Industry (IND-1/KPS/M).

The industrial production volume index is an important short-term indicator of business cycle that measures the monthly results of the industrial sector. Taking into account its periodicity and its rapid availability, it is a central and an up-to-date indicator of the development of the industrial sector, which is one of the most volatile components of economy.

Classification: Classification of activity for enterprises and local units by the NKD 2007 (identical on all levels with the NACE Rev. 2), classification of products and services by the KPD 2008 (identical to all 6-digit levels with CPA 2008), more detailed nomenclature for monthly industrial survey by the NIPUM 2022 (10-digit levels, identical to 6-digit levels with KPD 2008 with more detailed breakdowns to the national aggregation of PRODCOM List 2022), and Main industrial groupings by the GIG 2009 (the Commission Regulation (EU) No. 2020/1197 of 30 July 2020 laying down technical specifications and arrangements pursuant to Regulation (EU) No. 2019/2152 of the European Parliament and of the Council as regards the definitions of Main Industrial Groupings (MIGs)).

Industrial coverage: IPI covers enterprises and part thereof employing at least 20 or more persons, with the limit of 10 for small industries (mainly printing, bakery and winery) within industrial sector covering mining, manufacturing and energy supply i.e. B, C & D (except group 35.3) of the NKD 2007 (=NACE Rev.2). The index is based on volumes: elementary series representing approximately 90% of value added (3-digit section level) for coverage of 1,969 reporting units.

Product coverage: Products falling within industrial activities – defined by hierarchical structured product classification, NIPUM 2022 – Nomenclature of industrial products, as an extension of the KPD 2008 (= CPA 2008). Products are recording regardless weather consumed in the reporting units for further processing or are intended to be sold and cover both primary and secondary products and all industrial services according to the KPD 2008 (=CPA 2008). The basic product information is collected for the quantities in the current month and cumulatively (from the beginning of the current year), also stocks at the end of reporting month.

• The number of the category is identical to KPD 2008 (=CPA 2008): 798 products and services.
• The number of the subcategory is identical to KPD 2008 (=CPA 2008): 1795 products and services.
• The number of items in NIPUM 2022: 4470 products and services

Size classes covered. The monthly survey covers all enterprises, legal and natural persons, with 20 or more persons employed within NACE Rev. 2 Sections B to D (except Group 35.3), as cut-off sample. The threshold is lower for some NACE Rev. 2 Groups where the coverage would be lower than 90% in terms of the proportion of value added surveyed according to information from SBS (as cut-off sample).

The concepts and definitions used in the monthly survey are specificed in the Statistical Standards for the Monthly Survey on Industrial Production and Persons Employed (IND-1/KPS/M form) published on the website of the CBS. The concepts and definitions are harmonised with the EU concepts and definitions and especially with the Regulation (EU) of the European Parliament and of the Council No. 2019/2152 and the Commission Regulation (EU) No. 2020/1197.

IPI measure the change in the volume of production of a products (goods and services) according to the Nomenclature of industrial products for monthly survey - NIPUM 2022 as hierarchically structured nomenclature which systematically covering products within sections B, C & D (except group 35.3) of the NKD 2007. (=NACE Rev.2).  NIPUM comprises CPA 2008 with breakdown to the national aggregation of PRODCOM headings. It thus follows product approach rather than activity approach. Quality change in regard to physical quantities has being taken into account when changes occurs by updating the weights.

By the Monthly Survey on Industrial Production Volume Index and Indices of Stocks, Persons Employed and Labour Productivity in Industry are collecting each month from reporting units for each industrial product in physical quantities according to measurement units defined by the Nomenclature NIPUM 2022 in respect with their production (both sub-contracting is included, for own account and/or on contract basis), for following variables (with reference to NIPUM):

• Total (actual) production in current (reporting) month
• Total (actual) production for period from the begging of the reference year to current month (cumulative)
• State of stock of finished products at the end of reporting month (intended for sale on market)

The basis for the calculation of the IPI is physical quantities of output (sum of point 1.) which include all products (goods and services)  produced whether they are sold, otherwise used or entered into stocks for sale. IPI for various levels of the NKD 2007 (= NACE Rev.2) are calculated in two stages according to the Laspeyres formula. In the first stage, indices are calculated from quantity data on production of individual products according to the NIPUM 2022 and an adequate weighting coefficient, which represents estimated value added of individual products. Quality change in regard to physical quantities has being taken into account when change occurs by updating the weighting coefficient, respectively.

In the second stage, all indices of groups and higher levels are weighted with shares of value added of individual activities of industrial production by NKD 2007 levels (= NACE Rev.2). In this way, the relative importance of individual levels is defined and, at the same time, different levels of coverage are levelled.

The movements of industrial production volume index are presented and interpreted in relation to growth rates calculated on the basis of seasonally adjusted and working-day adjusted indices. Seasonally adjusted indices are used for the monthly comparison of movements of industrial production volume and woking-day adjusted indices for the annual comparison. 

Reporting units: Enterprises and KAUs

Observation units: Industrial eneterprises or local KAUs of non-industrial enterpirses engaged with industrial activities of NACE Rev.2 Sections B to D (except gorup 35.3)

The statistical population comprises the observation units enterprises operating in the activity classes within industrial sections B to D (except group 35.3) of the NKD 2007 (=NACE Rev.2) or KAUs as a part of enterprises which secondary activities falls also in the above sections of NACE Rev.2.

The frame for identifying the population is Statistical Business Register (SBR). Cut-off sample is drawn in March from PRODCOM survey in combination with SBR. The population is defined and continuosly updated upon the SBR and at the beginning of each year and during reporting year for new-born industrial units from all available suorces.

The total population of NACE Rev.2 Section B to D (except gorup 35.3) includes approximately about 25 000 industrial enteprises and KAUs. 

The target population consists of 1 969 statistical units with 20 or more persons employed (cut-off sample) in 2022 for sections covered by this survey. Target population in certain industrial divisions/groups includes the industrial enterprises and KAUs with less then 20 employees if they have a specific industrial production important for the final survey results. 

The geographic territory of the Republic of Croatia.

The production index has been calculated since 1952. However, due to changes of due to changes of classifications, nomenclature, coverage, methods and definitions led to breaks in the time series.

Currently, data are available from January 1998 onwards on base year 2015 as gross, seasonally adjusted, working-day adjusted data and trend-cycle.

Additionally data from January 1998 to December 2017 are also available on the previous base year 2010=100.

From January 2018 the base year is 2015. (2015=100). Previous base year: (2010 =100).

Reference period is average of previous year. Weights are updated annually in combination with chain linking.

Reference year is also previous year (currently 2021 for 2022 indices)

Indices, growth rates (%).

Month.

Legal basis on the national level the Official Statistics Act (Official Gazette, Nos. 25/20) established the responsibility for collecting, processing and dissemination of data. The Croatian Bureau of Statistics is the main producer, disseminator and coordinator of the official statistics system of the Republic of Croatia. The tasks of producing agencies in Croatia were established by the Official Statistics Act (Official Gazette, Nos. 25/20). The producers of official statistics have the right to collect data from all available sources.

Statistical activities have been established by the Programme of Statistical Activities of the Republic of Croatia 2021 - 2027 and the Development Strategy fo Official Statistics of the Republic of Croatia 2021 - 2030.

Legal basis on the european level are the Regulation (EU) of the European Parliament and of the Council No. 2019/2152 and the Commission Regulation (EU) No. 2020/1197.

Data are transmitted, in addition for Eurostat, directly to other international organisations like UNE/ECE, OECD and IMF on their request.

Nationally, data sharing between the producing agencies is regulated by the Ordinance on the Access to Confidential Statistical Data for each statistics by the individual agreements between individual producing agencies and CBS.

The statistical confidentiality of Croatian CBS is regulated by:

• Official Statistics Act, OG, Nos 25/20
• Ordinance on the Access to Confidential Data of the Croatian Bureau of Statistics within the System of Official Statistics (defines conditions of handing over of confidential statistical data of the Croatian Bureau of Statistics to other producers of official statistics in the Republic of Croatia as well as methods of applying the safekeeping measures of the confidential statistical data handed over to the Bureau by other producers of the official statistics in the Republic of Croatia)
• Instruction on the Protection Rules and Methods for the Dissemination of Official Statistics of the Croatian Bureau of Statistics (regulates the rules and methods for ensuring the dissemination of official statistical data within the scope of the Croatian Bureau of Statistics in the form and manner that protect the statistical unit from direct or indirect identification)
• Ordinance on the Conditions and Terms of Access and Use of confidetial Statistical Data of the Croatian Bureau of Statistics for Scientific Purposes (defines conditions and terms of permitting the access to confidential statistical data of the Croatian Bureau of Statistics for scientific purposes)
• Instruction on the Method of Protection of Confidental Statistical Data in the Croatian Bureau of Statistics (regulates procedures and measures for storing and protecting statistical data by applying physical, technical and organizational measures required to ensure statistical data from unauhorised access, alteration, loss or destruction, responsibility implementing measures and protection procedures and other issues relevant to the security of statistical data).

Individual data are not published and the treatment of confidential data is regulated by the Ordinance on the Rules and Methods of Protection of Confidential Statistical Data (Nos. 78/2021). The data are published and transmitted without characteristics that permit indentification of the respondents and classified into groups of at least three units. 

The Calendar of Statistical Data Issues provides a precise overview of the publications due to be released by the CBS for the entire statistical year, one year in advance. According to the Official Statistics Act, the release date of the Calendar is three months prior to the release date of the first publication of the following year. The release date is usually in October. It is bilingual and available on CBS website.

The Calendar of Statistical Data Issues for 2022 is available on the website of Croatian Bureau of Statistics, free of charge. 

All users have been granted an equal access to official statistics: this means that the dissemination dates of official statistics have to be announced in advance and released in accordance with the plan annonced in the Calendar of Statistical Data Issues and have to be available on the CBS website, precisely at 11 o'clock on the set date. No user category (incl. Eurostat, state authorities and mass media) can have access to the results of official statistical survey results before other users. Every departure from the plan disclosed in the Calendar shall be previously annonces, an explanation shall be given and new release date shall be set.

All issues of publications, first releases and including statistical database are released on the CBS website simultaneosly to all users and are available both, in Croatian and English language. Data are transmitted monthly to Eurostat on the same day as the national figures are published.

CBS informs users through wesite, official Facebook page and Twitter profile that the data are released.

Monthly (IPI)

Data are released through monthly First Releases to all users, free of charge, at 11 a.m on announced day in the Calendar on CBS web site, as well as in paper format on request.

Titled as First Release No IND-2022-3-1 - "Industrial Production Volume Index and Indices of Stocks, Persons Employed and Labour Productivity in Industry, 2022" IPI is presented in original (gross), seasonally and working-day adjusted forms and as trend cycle for Industry total (B to D, ex.35.3), MIGs, NACE Rev.2 activity divisions.

The length of the time series in the First Releases is six month for all presented data.

Statistical report: Short - term Indicators of Industry- Industrial production, Producer Prices and Turnover of Industry contains detailed short-term indicators as annual publication with the most detailed and comprehensive IPI results and time series.

Statistical Information (pocketsize) contains a summary of annual data for the Republic of Croatia and the review by Counties.

Croatia in Figures inclues the most improtant data on economic and social changes in the Republic of Croatia presented in five - year series.

Indices are also published in the online Statistical Database "PC-Axis" and in the STS Database on the CBS website.

The database is available on-line to all users. Frequency of updating the online database is monthly on a date announced in the Calendar of Statistical Data Issues (the same date when First Releases are published). The data are presented for the total industry, MIGs, industry sections and divisions of NACE Rev.2 as gross indices, seasonally and working-day adjusted indices and trend indices.

The length of time series (monthly indices) available in the Statistical Database: from January 1998 to December 2022. Additionally, in the STS Database are data available on the previous base year (2010=100) from January 1998 to December 2017.

Micro-data that makes it possible to identify individual objects is not publicly released. Access and anonymisation of micro-data are regulated by "The Ordinance on Conditions and Terms of Using Confidential Statistical Data for Scientific Purposes" (OG, Nos 103/03, 75/09 and 59/12)

In the process of drafting Ordinance on terms and methods of confidential data use for scientific purposes numerous researches were included. On the basis of previous experience regarding use of microdata for research purposes a list of researchers was defined. Draft version of Ordinance was open for their comments and suggestions on rules and protocols which can insure effective access and use of microdata for scientific purposes. All comments which were in accordance with national and European rules on protection of confidential data are accepted in the final version of the Ordinance.

Data from March 2020 are also diseminated in COVID-19 Thematic section.

Final data are transmitted to Eurostat and used in European aggregates and published as national indices.

Notes on methodology used to compile the industrial production are available at the end of each IPI First Release published on the website: 

FR IND-2022-3-1 Industrial Production Volume Index and Indices of Stocks, Persons Employed and Labour Productivity in Industry, 2022

Methodological explanations are also available in other monthly or yearly publications which includes the IPI and in Statistical Database (section Industry).

Concepts, definitions and classifications used in the survey are specified in the Statistical Standards for the Monthly Survey on Industrial production and Persons Employed (IND-1/KPS/M) available free of charge on the CBS website.

Documentation on procedures applied for quality management and quality assessment are under development within TQM framework. Quality database is developed and quality reports for IPI are available on CBS website for all users, free of charge.

To assure the quality of processes and products, CBS applies the European Statistics Code of Practise and the ESS Quality Assurance Framework (QAF). CBS is also guided by the requirements provided for in the Official Statistics Act (OG, Nos 25/20).

The Quality Database is based on the exhaustive list of quality information, which has been based on two widely accepted ESS structures, ESMS and ESQRS.

The IPI is highly relevant both as an input to quarterly national accounts and stand-alone. The type of survey and the data collection methods ensure sufficient coverage and timeliness. The level of the accuracy is high. The IPI is usually published on the 29^th day in a month for the indices from previous reference month. Publication of results are always punctual, fast and easy accessible. Notes on methodology are clear and comprehensive for demanding users’ needs.

The main advantage is comparability over time with the length of time series since January 1998. It is comparable with other statistical domains, primarily with the quarterly national accounts and PRODCOM. 

Primary users of IPI:

• International users: Eurostat, ECB, IMF, UN/ECE, OECD
• National users: Croatian National Bank, Ministry of Economy, Ministry of Finance, other government bodies, Institute of Economic Zagreb (EIZ), Croatian Chamber of Economy
• Internal users: National Accounts Department

Secondary users of IPI:

• trade unions, enterprises/business
• economic and research institutes, researchers and students,
• media and general public. 

In April 2015 Croatian Bureau of Statistics (CBS) conducted the user satisfaction survey for the second time (first time in November 2013). The aim of the research was to determine users' satisfaction with the quality of data and services, as well as their needs. The survey was conducted on a sample of the CBS data users combining web questionnaire and telephone interviewing in the period from 13 to 26 April 2015. 

The sample consisted of users that have requested statistical data in the period from the beginning of the 2013 to March 2015. The questionnaire was sent to e-mails of 2 765 users through the Survey Monkey. Links to the questionnaire were placed on the website of the CBS, as well as on CBS official Twitter and Facebook pages so that other interested users can participate in the survey. Questionnaire was completed by 952 users.

CBS's data users were mostly woman (61.7%). The majority of users (56.6%) belonged to the age group 30-49 years. 52.8% of users had university degree, 22.7% of users were researchers and the second largest group of users was from the field of economy (21.5%). 

Out of the total number of surveyed users, 88.6% collected data through CBS's website, 55.9% of users collected data for research purposes, 36.4% for general information only, 24.7% of users used data each month and the most demanded data were from the field of population (46.4%), employment (40.2%) and tourism (30.8%). 65.7% of users thought that CBS's data is of high quality and 62% of users rated the comparability of CBS' data as good. Part of users (21.2%) believed that CBS' data did not meet their needs for statistics in the Republic of Croatia and 18.8% of users were not satisfied with the way they can acquire CBS data.

The following five criteria were taken as the variables that affect the overall satisfaction of CBS' users:

• satisfaction with the employees - average grade is 4.34
• satisfaction with data - the average grade is 3.56
• satisfaction with the website - the average grade is 3.16
• satisfaction with the CBS in general- the average grade is 3.8
• trust – CBS has reliable data - the average grade is 3.96.

Total CBS average grade is 3.8 and the overall index of the users' satisfaction with the CBS is 75.28%.

Time series for IPI are available since 1998 based on NACE Rev.2 on 2-digits division and sections level as well as on MIGs.

Eurostat's requirements in terms of time series are fulfilled.

The Monthly Survey on Industrial Production Volume Index and Indices of Stocks, Persons Employed and Labour Productivity in Industry (IND-1/KPS/M) is based on a cut-off sample. Sampling error is not measured. The response rate is very high so the overall accuracy is very good.

Not available. IPI is calculating based on a cut-off survey (industrial enterprises with 20 or more persons employed).

Response rate is very high – about 98% 15 days after reporting period. Level of unit non-response is very low. Up to 2% (for 2022, average unweighted unit non-response rate was 0.8%) of small-sized enterprises (statistical units) do not fill the monthly questionnaires IND-1/KPS/M even after urging them (unit non-response)). Each non-respondent is contacted several times and then estimated, if necessary. 

Procedures for non-response (only small-size enterprises):

In the case of unit non-response, two types of estimation methods are being used prior to manual imputation (imputation rate is around 1%)

• based on previous month record/or based on production trends of this unit starting from the beginning of current year up to the end of the previous month
• based on production trends of the same product groups of other statistical units within 8-digit level  (if there is no report from the previous months by statistical unit)

In the case of item non-response the missing response are requested by telephone contacts. This case is very rare (only possible for limited number of forms send by email) since usage of on-line data collection system ensuring that all fields should be fulfilled prior to sending. Thus item non-response for IPI =0%.

Measurement errors: The data collection through the electronic form on the CBS website has highly reduce the possibility of calculation errors, incorrect codes and measurement units. Also there are incorporated various warnings when submitting a form if there is a visible oscillation between previous months. Errors by companies entering the incorrect data are also possible and these data don't pass editing process. All the errors are treated, analysed and consulted with the enterprise and corrected via direct telephone contact or by e-mail.

Over-coverage rate for 2022 is 2.0%. The data for caluculation this quality indicator are available at the end of each year when we are editing the base for the upcoming year. Then we exclude statistical units that have stopped working or no longer have industrial activity.

Timing of production observations: Frequency of data collection is monthly. The time schedule for data collection ranges each month, as a rule between 1^st and 15^th day in a month for the previous reference month data. 

Timing of publishing: IPI data are being published 29 days after reference month (t+29).

The first results are also the final results.

All the IPI First Releases for 2022 have been published at 11 o'clock on the day announced in the Calendar of Statistical Data Issues, 2022.

Since Croatian IPI is in compliance with STS requirements it is comparable among Member States.

Monthly time series for IPI are available since Januray 1998 based on NACE Rev.2 on 2-digits divisions and sections level as well as on MIGs.

The transition from NACE Rev.1.1 to NACE Rev.2 was performed by backcasting on macro level to ensure comparability over time.

Break of time series exist for IPI time series prior to 1998 due to implementation of NACE classification instead of Croatian based classification of activities which is not comparable with NACE.

The data are cordinated with the calculation of the National Accounts. The IPI is needed for the estimation of the quarterly GDP. 

Consistency with annual indicators of industrial production is achieved with:

• Annual PRODCOM survey where data are being collecting on a wider range of details for about 6 000 products, according to the Nomenclature of industrial products – NIP for annual production survey. NIP is systematically related to PRODCOM List headings with further national breakdowns since 1995. Collection annual data are based on NIP versions which are updated annually according to PRODCOM List changes; currently NIP 2022 started to use in line with PRODCOM List 2022 of the EU. NIP 2015 (= PRODCOM List 2015) products are used as weighting coefficient for products in NIPUM for the base year 2015 (NACE Rev.2/CPA 2008/PRODCOM structure);
• SBS results in regard value added on factor costs. These indicators are on annual bases, and show structure of industrial activities on the 3-digit group level of NKD 2007 under B, C & D sections (divisions from 05 to 35 excluding 35.3). It is used as value added weighs for industries on the levels of groups, divisions and sections of NKD 2007 (identical to NACE Rev.2) as well as for GIG 2009 aggregates (=MIGs 2009) weights.

Consistency of SBS survey results (derived from FINA administrative data) has been tested with PRODCOM data.

The Croatian IPI is annually chain-linked Laspeyres type index. The IPI higher level indices are annually chain-linked using an annual overlap (annual average of year t-1). The chain linking is performed to hieratical level of NACE activity classes, groups, divisions and sections as well as on aggregate sub-indices levels (MIGs) and IPI total. IPI is internally consistent within available time-series since January 1998 to December 2022 for industry total, divisions and sections of NKD 2007 (=NACE Rev.2) as well as on MIGs 2009. It is compiled as unadjusted, WDA, SA and trend-cycle indices. Up until now, the TRAMO-SEATS method was used in the Demetra (version 2.2.) package. Since January 2016, the new X13 ARIMA method in the new JDemetra+ package has been used, on a time series starting from January 1998.

IPI adjustments are made for the Industry total, MIGs, Sections and Divisions of NKD 2007 (=NACE Rev.2) using direct method i.e. IPI time series are aggregated by NACE and by MIGs and then seasonal adjustments performed independently at each aggregate level.

Coherence is achieved by compilation methods and weighting system. IPI for various levels of the NACE are calculated in two stages according to the Laspeyres formula. In the first stage, indices are calculated from quantity data on production of individual products according to the NIPUM 2022 (nomenclature of Industrial Products) and an adequate weighting coefficient, which represents an estimated value added of individual products. In the second stage, all indices of groups and higher levels are weighted with shares of value added of individual activities of industrial production by the NECE levels. Indices for MIGs 2009 aggregates are calculated in the same way according to a predominant end-use of products within the same NACE activity group.

Starting from January 2011 data collection method for Monthly Survey on Industrial Production (IPI) has been changed. WEB based "on-line" questionnaires has been introduced instead of paper questionnaires (CAWI method) to reduce both burden on statisticians and on respondents. Data collection chain has been change from decentralised (via county statistical office) to completely centralised (CBS headquarter). As a consequence, cost and burden originated from statistical data collection was significantly reduced for both, respondents and statistics.

Cost for statistics is approximated as follows: The routine costs in Croatia for updating IPI weights each year are approximately 4 persons for 2 month (December to January) intensively in normal year where no revisions occur due to classifications change. It currently involves four experts manual work in the excel sheets tables approximately 720 hours per expert. In addition, updating the software for data collection, processing and calculations (especially new database construction in SQL) approximately 1 IT persons is needed to work app. 15 days along with experts. Finally, due to updating sample of enterprises/local units) 6 statisticians are involved to contact the enterprise approximately 20 days each year January/February.

Response burden is not very high – respondents need in average 10 to 30 minutes for fulfilling “on-line” questionnaire in 2022 independence of respondent size class (large or SMEs) according to information they provided.

The General Revision Policy of the Croatian Bureau of Statistics is based on the following documents:

• European Statistics Code of Practice
• Quality Assurance Framework of the European Statistical System
• ESS Guidelines on Revision Policy

The revision policy in practice depends of size of revision.  In the case when major changes in methodology occurs first there are discussions with the Statistics Advisory Committee for Industry (users and producers) about the proposed new Statistical Standards. These are issued after agreement with description of major changes or revisions under chapter Comparability with previous data series. The Annual Implementation Plan (Annual statistical programme), 2022 also includes such information’s.

Users are informed on minor methodological changes in Note on methodology which is a part of the First release and all related publications announced in the Calendar of Statistical Data Issues, 2022 which is all available on CBS website.

Croatia has no practice to publish preliminary and final IPI results. IPI is always publishing as final results 29 days after reference period (mark as "first results" since detailed results could be found in comprehensive dedicated statistical report). If revision will be needed due to integration of new significant information for IPI unadjusted results on 3 digits NACE level (more than +/- 0.5), apart from revisions generated by seasonal adjustment, then in practice revised data will be mark and explain to users. Such case occurred in July 2017 and May 2022.

IPI revisions generated by seasonal adjustment on the past data are not explaining to the users apart from explanations within Note on Methodology (part of each IPI First release) about nature of implemented seasonally adjusted methods (size of revisions due to the seasonal adjustments was not measured so far).

Revisions due to errors are mark and explained to users almost immediately after detection in the reissued First release (available on CBS website).

The statistical survey known as Monthly Survey on Industrial Production and Persons Employed (IND-1/KPS/M) is the main source of information for calculating the indices of output for the total industry by definition of the NKD 2007. (=NACE Rev.2) sections B, C & D (except group 35.3). The survey covers approximately 90% of total value added (VA) of industry on 3-digits levels of the NKD 2007. (=NACE Rev.2).The coverage of the survey is updated yearly. The cut-off sample is drawn from PRODCOM survey in combination with SBR.

The data are obtained from a monthly industrial survey for all units with 20 or more persons employed but in particular cases less than 20 when it is not enough to cover approximately 90% of value added on NKD 2007 3-digits level. The target population for 2022 included 1 969 industrial enterprises and KAUs.

Frequency of data collection is monthly, with deadline around 15^th in the current month during 2022. Actual reporting dates are between 1^st and 15^th of the month (for previous reference month data).

The data collection method used is a web-based application (the so-called CAWI). The reporting units enter data directly into the IND-1/KPS/M e-form, which is available on the CBS website. 

Notification e-mail with the deadline for fullfiling the survey are sent to all respondents. Later, e-mail reminders are sent to increase and speed up the rate of response. 

Missing responses are requested by telephone contacts. Imputation is used to compensate the non-response.

The CBS industrial experts perform all cross-checking, necessary correction after phone call-back procedure for suspicious data to obtain data of good quality. Processing is being verified automatically through established procedures of data treatment and validation each month within production quantities and cumulatively quantities (from the beginning of the year). Validation includes: cross cutting checks, checks with previous data, evolution checks (very strict criteria).

All processing steps from validation to computing the chain-linking IPI including national dissemination formats is carrying out in automatic IT support system (SQL) and before the data are to be prepared for transmitting to Eurostat. The structure of the file and the format of the data to be sent to Eurostat are derived from national formats of final results as well as plausibility checks and consistency checks which are incorporated in national data validation system. The file format for Eurostat is formatted as Excel file as required by GESMES/TS Guide and IPI required results are manually input from national formats. Eurostat specific validation checks are not carrying out so far since national validation rules are more strictly.

Imputation is used to compensate the non-response. Two types of estimation methods are being used prior to manual imputation:

• first one, based on previous month record/or based on production trends of this unit starting from the beginning of current year up to the end of previous month,
• second one, based on production trends of the same product group of other statistical units within 8-digit level (if there is no report from the previous months by statistical unit). 

Industrial production index (IPI) in Croatia is Laspeyres type chain index, by adoption of currently 2015 as base year for weighting coefficients on products level but for activity structure uses previous year activity weights.

Compilation practice: Starting from the micro product data, total production is compiled for each NIPUM 2022 product (by aggregating. across producers). These are then multiplied by a weighting coefficient to reach a "mass" of value added for each NIPUM 2022 product, which are then summed to the 4-digit level of NACE Rev. 2. At the 4-digit level relatives are compiled. These relatives are multiplied to 3-digit Group level and higher levels of NACE and to MIGs using value added weights of the appropriate 3-digit level of NACE from SBS survey. In fact, there are two steps in calculating IPI, as follows:

• in the first step: Indices at the 4-digit level of NACE are calculated from volume data on production of individual NIPUM product and appropriate weighting coefficients using the Laspeyres formula;
• in a second step indices for all higher levels of NACE Rev. 2 and for MIGs are calculated by aggregating the indices of the appropriate 3-digit Groups of NACE weighted with the ‘Group' share of value added in the total value added.

The weighting coefficients of individual products of the NIPUM 2022. used in the calculation of industrial production index represent the value added to the unit production of a respective product. They are calculated on the basis of unit values of the PRODCOM survey on industrial production for 2015 and are revised every five years. 

The structure of industrial activities (division structure) is calculated on the basis of the gross value added of factor costs by the pure activity principle. It is calculated on the basis of the annual statistical results of the Structural Business Survey carried out by the Croatian Bureau of Statistics by processing annual results taken over from the administrative data of Fina – the Annual Accounts of Industrial Enterprises. It was further corrected by the results of the PRODCOM Survey on Industrial Production and then by the industrial production volume indices. The division structure is updated at the beginning of every year for the calculation of the industrial production volume indices in the current year (i.e. SBS 2020 is used for IPI calculation in 2022, corrected with appropriate IPI 2021).

IPI are compiled as unadjusted (gross), working-day adjusted, seasonally and working-day adjusted and trend-cycle indices. Seasonal and working-day adjustment has been done by using the X13 ARIMA method in the JDemetra+ package, on a time series from January 1998. Direct method is used for the seasonal adjustment of time series. In the seasonal adjustment the impact of the number of working days, Easter and other national holidays (National Calendar is in use) was taken into account in those time series for which the mentioned influences were statistically significant. These adjustments are made for the total industry (sections B, C and D, excluding group 35.3 of the NKD 2007.) as well as for the MIGs 2009.

Not available.