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.
Index of services production (ISP) measures monthly changes of price adjusted output. It is calculated by deflation of monthly turnover data with corresponding deflator (price index). It covers services activities defined by Nace Rev. 2: sections G, H, I, J, L, M (excl. group 701, divisions 72 and 75) and N.
Monthly turnover data are derived from Monthly Report on Services in combination with administrative data (VAT).
3.2. Classification system
NACE Rev. 2.
3.3. Coverage - sector
ISP is calculated for the tollowing Nace activities:
H Transportation and storage;
H 49 Land transport and transport via pipelines;
H 50 Water transport;
H 51 Air transport;
H 52 Warehousing and support activities for transportation;
H 53 Postal and courier activities;
I Accommodation and food service activities;
I 55 Accommodation;
I 56 Food and beverage service activities;
J Information and communication;
J 58 Publishing activities;
J 59 Motion picture, video and television programme production, sound recording and music publishing activities;
J 60 Programming and broadcasting activities;
J 61 Telecommunications;
J 62 Computer programming, consultancy and related activities;
J 63 Information service activities;
L Real estate activities;
L 68 Real estate activities;
M Professional, scientific and technical activities;
M 69 Legal and accounting services;
M 702 Management consulting services;
M 71 Architectural and engineering activities; technical testing and analysis;
M 73 Advertising and market research;
M 74 Other professional, scientific and technical activities;
N Administrative and support service activities;
N 77 Rental and leasing activities;
N 78 Employment activities;
N 79 Travel agency, tour operator reservation services and related activities;
N 80 Security and investigation activities;
N 81 Service to buildings and landscape activities;
N 82 Office administrative, office support and other business support activities.
3.4. Statistical concepts and definitions
Index of services production measures the volume of sales in services and trade activities and it is derived from turnover at current prices which is converted into constant prices by procedure of deflation using adequate deflator. Since the intention is to insure changes in the volume of gross output, data on added value from the survey on structural business statistics are used as weights in the calculation of the index of services production.
Turnover comprises the totals invoiced by the observation unit during the reference period, and this corresponds to market sales of goods or services supplied to third parties. Turnover also includes all other charges linked to the buyer, even if these charges are listed separately in the invoice. Turnover excludes VAT and other similar deductible taxes directly linked to turnover as well as all duties and taxes on the goods or services invoiced by the unit.
3.5. Statistical unit
Since January 2024, the base year has been changed to 2021 and the new observation unit is a kind-of-activity unit (KAU).
3.6. Statistical population
The target population for turnover data are business entities (legal units, tradesmen and free lances) which main activity is one of the activities that are proposed by the EBS Regulation (listed under 3.1), regardless on their size.
The frame for identifying units for the population is the Statistical Business Register.
3.7. Reference area
Geographical area covered: Republic of Croatia.
3.8. Coverage - Time
The first period for which Index of services production is available is the January 2010.
3.9. Base period
Base (reference) year: 2021.
Only index numbers are disseminated as a result of the compilation of ISP.
Reference period is month.
6.1. Institutional Mandate - legal acts and other agreements
There is no data sharing to other international institutions.
7.1. Confidentiality - policy
National legislation:
Under the Official Statistics Act (NN No. 25/20 and 155/23), the CBS can not publish, or otherwise make available to any individual or organization, statistics that would enable the identification of data for any individual person or legal entity. Individual data are strictly confidential and must not be published or reported.
European legislation:
Regulation (EC) No 223/2009 on European statistics (recital 24 and Article 20 (4)) of 11 March 2009 (OJ L 87,p.164), stipulates the need to establish common principles and guidelines ensuring the confidentiality of data used for the production of European statistics and the access to those confidential data.
7.2. Confidentiality - data treatment
Confidential data are treated according to policy rules (see point 7.1).
Data are nationally published only on the level at which they are not confidential.
8.1. Release calendar
Data are disseminated nationally with a predefined Release Calendar publicly accessible.
Release schedule for the coming year, specifying dates of release is published each December in the CBS Publishing Program. This document is available in both Croatian and English.
The exact dates of release are pre-announced in the Calendar of Statistical Data Issues. It is available in both Croatian and English.
Data are simultaneously released at the same time (11.00) to all interested parties by publishing on the CBS’s website.
There is no prior access to data before its release.
Data are transmitted to Eurostat the same day after the release.
Monthly.
10.1. Dissemination format - News release
Currently data are not published in the form of News release.
10.2. Dissemination format - Publications
Currently data are not published in any type of Publications.
10.3. Dissemination format - online database
Data are available from 2021 in Excel tables on the website of the Croatian Bureau of Statistics within the Databases under the title STS Database.
10.4. Dissemination format - microdata access
Not available.
10.5. Dissemination format - other
Data are disseminated to Eurostat in pre-defined tables through eDamis, to be used in European aggregates as well as released as national data.
10.6. Documentation on methodology
Methodological explenations on Index of service production data are available in the STS Database under Services – Turnover Indices of Service Activities – Volume Turnover Indices of Service Activities Monthly. It can be found through CBS website
International publications on ISP methodology:
Compilation manual for an Index of services production, OECD, 2007;
Guide on developing and Index of Services Production (ISP), Eurostat ISP Task Force;
10.7. Quality management - documentation
The quality report will be published in 2025 according to the Release calendar. The exact dates of release will be pre-announced in the Release calendar. It will be available in both Croatian and English.
11.1. Quality assurance
There is a written statement under the Official Statistical Act (NN, No.25/20 and 155/23) that official statistics in Croatia should be of enough quality.
CBS has started systematic approach to quality management with the adoption of the European Code of Practice and with the decision to implement TQM. This process is still in course. Data base for storing information for the purpose of assessing data quality has been developed.
Usual quality checks have been undertaken during the data editing. The outputs are improved data.
11.2. Quality management - assessment
ISP in Croatia is a new indicator and there is still room for improvement. Quality is assured by checking of all input data and comparing with value turnover in services indices.
Indices are produced in compliance with methodological requirements and standards.
12.1. Relevance - User Needs
The main users are National accounts in CBS and Eurostat.
12.2. Relevance - User Satisfaction
Not relevant.
12.3. Completeness
Indicators are complete and fully in line with the concepts and definitions in methodology based on the Regulation (EU) 2019/2152 of the European Parliament and of the Council on European business statistics.
Index series, publiced nationally and transmitted to Eurostat are complete, with no gaps or missing data parts.
13.1. Accuracy - overall
Accuracy of ISP data depends on input data and on correct processing of the data.
13.2. Sampling error
There is no sampling error because the survey is not conducted on the basis of a random sample.
13.3. Non-sampling error
Since the Index of service production uses data from Monthly turnover in services survey, non-sampling error is calculated in that survey.
14.1. Timeliness
Results are published 60 days after the end of the reference period in STS Database. On the same day, data are diseminated to Eurostat.
14.2. Punctuality
Data are delivered according to the pre-announced Release calendar and there is no time lag between the actual delivery of the data and the target date when it should have been delivered.
15.1. Comparability - geographical
Same statistical concepts are applied in the entire national territory; there is no problem of geographical comparability.
Data are also comparable with other countries because the methodology is based on the same regulations.
15.2. Comparability - over time
Data for all activities are available from January 2010. Until the end of 2023, the length of the series is 168 months.
15.3. Coherence - cross domain
Coherence between SBS data and STS data is not fully possible because of different methodology.
15.4. Coherence - internal
Data are internally coherent. Higher level aggregations are derived from detailed data according to well-defined procedures. For all NACE areas the same definition of turnover is used.
Cost of CBS: approximately 650 working hours annually.
To calculate this indicator there is burden only on CBS staff to collect input data, calculate indeks and analyze results.
17.1. Data revision - policy
The General Revision Policy of the Croatian Bureau of Statistics is based on the following documents:
European Statistics Code of Practise;
Quality Assurance Framework of the European Statistical System;
ESS Guidelines on Revision Policy for PEEIs.
The corrections of published data caused by technical errors, calculation errors, linguistic and other similar errors are not the subject of the Revision Policy of the Croatian Bureau of Statistics. When disseminating changes to already published statistical data, it is necessary to clearly indicate whether they are corrected data (due to technical or calculation errors) or revised data.
In order to ensure the transparency and availability of data, following the publication of the revised data, the Croatian Bureau of Statistics provides users with the previously published (original) data in the publishing media in which it is possible.
The "de facto revision policy" used for this survey includes:
for mayor revisions (such as change of base year) - all series are recalculated but changes are not announced in advance. Explanations are given when new data are published.
routine revisions (such as changes of SA and WDA data series) - revised data are published when new data are normally published without classified changed data as revised. There is only methodological explanation regarding usage SA and WDA tool and its implication on changes of previously published data.
if errors occur, corrected data are published as soon as possible or in the next issue of the publication.
The same revision policy is applied internally and in transmission to Eurostat.
There is no availability of a vintage database.
Revision calendar exists - information covered under item 8.1.
The benchmarking is not carried out.
17.2. Data revision - practice
The ISP data are regularly revised. Due to lack of quarterly price indices (used as deflators) at the beginning of every quarter, monthly ISP is forcasted. For example, the monthly ISP have to be produced and provided in time T+60 after the end of the reference month (e.g. for January 2020 it has to be provided at the end of March 2020). At that time the quarterly price indices are not available yet. Moreover, they are only available T+90 after the end of the quarter (end of 3rd month of the quarter - this is at the end of June 2020). This means that monthly indices for all three months of the quarter at that time point have to be forecasted without quarterly index.
Mean Absolute Revision (MAR) and Mean Revision (MR), computed values for the last 24 monthly (Jan 2023 - Dec 2024) for total services production:
Growth rates for calendar/working day adjusted data series (YoY): MAR =1.57 , MR = -0.25
Because of change of base year first published date for series from January 2023 to December 2024 is January 2024.
18.1. Source data
Main source data for ISP calculation is monthly turover from MTOS database. Monthly turnover on a micro level is used to carry out a deflation in order to get volume indices (ISP).
Source data for calculation of Index of Services Production are also different price indices. These pruce indices are:
Monthly and quarterly Services Producer Price Indices (SPPI);
Monthly Consumer Price Indices (CPI);
Monthly Producer Prie Indices (PPI);
Quarterly Agriculture Price Indices (API).
There is also inputs that are not a price index and are used for ISP calculation:
Quarterly indices from Transport statistics – phisical indicators (PHI);
Monthly indices of average gross salaries (IAS).
18.2. Frequency of data collection
Monthly.
18.3. Data collection
Concerning Monthly survey on turnover in services, survey is contadted monthly via web questionnaire and combined with VAT data.
ISP data are not collected from reporting units. Every month input data are updated in excel worksheet and than they are used for ISP calculation.
18.4. Data validation
There is no automatic data validation. Data are manualy validated as follows:
to compare with previous data;
to check minimum and maximum values of yearly indices (indices on the base of the same month of previous year);
to compare data with value turnover in services indices.
18.5. Data compilation
Statistical process for calculation of ISP includes:
1. Gathering of all the sources, needed for the ISP calculation.
2. Estimation of monthly indices for the groups for which the quarterly indices are determined as deflators. This has been done by two different procedures:
Estimation (forecast) of monthly indices on the basis of auxiliary indices or historical indices in the months when the quarterly indices are not yet on disposal;
Disaggregation of quarterly price indices when the quarterly indices are on disposal. With this procedure monthly indices for all three months of the quarter are estimated at once.
3. Deflation of the nominal turnover at micro levelin order to get the deflated turnover.
4. Calculation of the annual “real indices” (indices calculated from deflated turnover) on the level of 3-digit Nace groups (3-digit indices).
Adjustments of indices from transport statistics (physical indicators) on 3-digit level;
Disaggregation of quarterly base year index series;
Calculation of annual indices from base year indices;
Joining together annual indices on 3-digit level from two sources: real indices and adjusted physical indices.
5. Calculation of annual indices for higher level activity groups, by calculating weighted arithmetic mean of the 3-digit level indices.
6. Calculation of base year indices (for all levels) by using the annual index and base year index from the same month of previous year.
7. Calculation of all other indices from the base year time series indices.
18.6. Adjustment
Volume turnover indices in original, unadjusted form for a current month are calculated in relation to the average of a base year, then in relation to the same month of a previous year.
Indices are presented in original, non-adjusted form, as well as seasonaly adjusted, working day adjusted form and as a trend.
The term "seasonally and working-day adjusted indices" is used to indicate that gross indices are adjusted for both seasonal and working-day effects. The indices adjusted in that way and associated rates of change in two consecutive observations are used to compare data with data from the previous month.
Working-day adjusted indices imply that gross indices are adjusted only for calendar effects. The indices adjusted in that way and associated inter-annual rates of change are used to compare data with data from the same month of the previous year.
Gross, unudjusted indices are not adjusted for seasonal nor for working-day effects.
In the process of seasonal adjustment (identification and evaluation of seasonal and calendar effects), the software package JDemetra+ was used. Considering that there are two available options in the JDemetra+, TRAMO-SEATS (using Seats method for estimating seasonal effects) and X13 (using the X11 method for estimating seasonal effects), seasonal and working-day adjustment has been done by using the X13 ARIMA method. The effect of a random component in the seasonal adjustment process of the time series of indices has been excluded.
This process identifies and assesses seasonal and workingday effects in time series using statistical models and procedures. The revision of the previously published seasonally and working-day adjusted indices is performed due to several causes. The most common cause is a change in the estimated parameters of existing statistical models while conducting new estimates, which occurs due to new observations. Other causes are changes in source data or in applied statistical models. Calendar effects have a significant impact on movements in time series, the most common of which are the effect of moving holidays, the effect of the leap year and the effect of working days.
Due to the character of the implemented adjustment method, the adding of new quarterly/monthly observations for each quarter/month may cause changes in the seasonally and working-day adjusted indices for previous periods. The official list of holidays and non-working days in the Republic of Croatia is used to assess the working-day component, that is, to assess the effect of working days and holidays.
The level of detail: nominal and real series – Base indices of the various aggregates/subgroups of the National classification of economic activities (NKD) based on NACE Rev.2. on the following levels:
H_49
H_50
H_51
H_52
H_53
I_55
I_56
J_58
J_59
J_60
J_61
J_62
J_63
L_68
M_71
M_73
M_74
N_77
N_78
N_79
N_80
N_81
N_82
G
H
I
J
M_69
M_702
H_TO_N_X_K
M
N
L
The use of default vs. national calendar – National calendar of holidays for Croatia is used in X13ARIMA method.
Adjustment for moving holiday effects (e.g. Easter) – Each series is tested for presence of Easter and Corpus Christi effect (automatic procedure in ARIMA X 13) – effect is significant in part of time series.
Treatment of leap year – The effect of leap year is tested as one of the calendar regressors, it is significant in part of time series.
Lack of calendar adjustment – There is calendar adjustment, i.e. calendar effect for Croatian holidays is tested for all the series.
Type of calendar adjustment (e.g. regARIMA) – Calendar adjustment is performed by using automatic identification procedure in ARIMA x13, starting with 7 regressors and reducing their number as some may not be significant for given series.
Type of outliers detected – Level shifts (LS), additive outliers (AO) and transitory changes (TC) are tested, most common are LS's and AO's, automatic procedure is used in ARIMA X13.
Known economic outliers – At present there is not known economic explanation for persistent outliers.
Software / its version and the seasonal adjustment method: JDemetra+ 2.2.2, method is X13ARIMA.
Manual vs. automatic model or filter selection - For the majority of series, automatic model selection procedure is used; for problematic series, several models are tried in the frame of detailed analysis module until optimal model is found.
How often are models and respective parameters re-estimated? - Parameters re-estimation for ref. periods April, August and December, in other months application of forecasted factors for calculating ‘sa’ and ‘ycal’ series. Annual revision of models, filters, calendar regressors for ref. period m12.
The horizon of revisions – It depends primarily on revisions of original series. Seasonally adjusted time series are revised in April, August and December (all values backwards).
Additive vs. multiplicative decomposition of seasonal adjustment – For the majority of series, multiplicative decomposition is performed. (more details in the attached document "Seasonal adjustment_service activities").
The verification of the adequacy of the models – Adequacy of models is checked by built-in diagnostic indicators in ARIMA x13 after each new observation is added.
Critical value and automation of outlier detection – Default critical value in the automatic detection procedure of outliers is used (ARIMA x13).
The method of selecting filter length – For the length of seasonal filters, automatic procedure is used in ARIMA x13.
Seasonal adjustment over the known breaks in series – Currently there are no breaks in series.
Indirect SA and checks for residual seasonality – No, direct SA approach is used.
The consistency of different aggregation levels – Not applicable.
The critical value for outlier detection - 4.
The filter length (automatically chosen vs. user-defined)- Seasonal filters are fixed during the year (revised at the end of a year if necessary) - (more details in the attached document "Seasonal adjustment_service activities").
Whether residual seasonality is checked and from which level of detail the aggregation is started - qs test and f-test on residual seasonality are checked for every series.
Index of services production (ISP) measures monthly changes of price adjusted output. It is calculated by deflation of monthly turnover data with corresponding deflator (price index). It covers services activities defined by Nace Rev. 2: sections G, H, I, J, L, M (excl. group 701, divisions 72 and 75) and N.
Monthly turnover data are derived from Monthly Report on Services in combination with administrative data (VAT).
7 August 2025
Index of services production measures the volume of sales in services and trade activities and it is derived from turnover at current prices which is converted into constant prices by procedure of deflation using adequate deflator. Since the intention is to insure changes in the volume of gross output, data on added value from the survey on structural business statistics are used as weights in the calculation of the index of services production.
Turnover comprises the totals invoiced by the observation unit during the reference period, and this corresponds to market sales of goods or services supplied to third parties. Turnover also includes all other charges linked to the buyer, even if these charges are listed separately in the invoice. Turnover excludes VAT and other similar deductible taxes directly linked to turnover as well as all duties and taxes on the goods or services invoiced by the unit.
Since January 2024, the base year has been changed to 2021 and the new observation unit is a kind-of-activity unit (KAU).
The target population for turnover data are business entities (legal units, tradesmen and free lances) which main activity is one of the activities that are proposed by the EBS Regulation (listed under 3.1), regardless on their size.
The frame for identifying units for the population is the Statistical Business Register.
Geographical area covered: Republic of Croatia.
Reference period is month.
Accuracy of ISP data depends on input data and on correct processing of the data.
Only index numbers are disseminated as a result of the compilation of ISP.
Statistical process for calculation of ISP includes:
1. Gathering of all the sources, needed for the ISP calculation.
2. Estimation of monthly indices for the groups for which the quarterly indices are determined as deflators. This has been done by two different procedures:
Estimation (forecast) of monthly indices on the basis of auxiliary indices or historical indices in the months when the quarterly indices are not yet on disposal;
Disaggregation of quarterly price indices when the quarterly indices are on disposal. With this procedure monthly indices for all three months of the quarter are estimated at once.
3. Deflation of the nominal turnover at micro levelin order to get the deflated turnover.
4. Calculation of the annual “real indices” (indices calculated from deflated turnover) on the level of 3-digit Nace groups (3-digit indices).
Adjustments of indices from transport statistics (physical indicators) on 3-digit level;
Disaggregation of quarterly base year index series;
Calculation of annual indices from base year indices;
Joining together annual indices on 3-digit level from two sources: real indices and adjusted physical indices.
5. Calculation of annual indices for higher level activity groups, by calculating weighted arithmetic mean of the 3-digit level indices.
6. Calculation of base year indices (for all levels) by using the annual index and base year index from the same month of previous year.
7. Calculation of all other indices from the base year time series indices.
Main source data for ISP calculation is monthly turover from MTOS database. Monthly turnover on a micro level is used to carry out a deflation in order to get volume indices (ISP).
Source data for calculation of Index of Services Production are also different price indices. These pruce indices are:
Monthly and quarterly Services Producer Price Indices (SPPI);
Monthly Consumer Price Indices (CPI);
Monthly Producer Prie Indices (PPI);
Quarterly Agriculture Price Indices (API).
There is also inputs that are not a price index and are used for ISP calculation:
Quarterly indices from Transport statistics – phisical indicators (PHI);
Monthly indices of average gross salaries (IAS).
Monthly.
Results are published 60 days after the end of the reference period in STS Database. On the same day, data are diseminated to Eurostat.
Same statistical concepts are applied in the entire national territory; there is no problem of geographical comparability.
Data are also comparable with other countries because the methodology is based on the same regulations.
Data for all activities are available from January 2010. Until the end of 2023, the length of the series is 168 months.
