Statistical Office of the Republic of Slovenia

Short-term Business Statistics

Litostrojska 54, 1000 Ljubljana, Slovenia

In this statistical survey we monitor the evolution and dynamics of the value of construction put in place. The monthly calculation of the indices of the value of construction put in place is intended to monitor short-term trends and is also suitable for analyses of economic developments.

NACE Rev. 2

The survey covers construction companies (sector F according to NACE Rev. 2) as well as their units engaged in construction and some non-construction companies performing construction work. Observation units are selected based on the threshold. In this way we determine a sufficient number of units that are included in the survey in an individual year. 

Variable collected by the survey: value of construction put in place.

Value of construction put in place covers the value of the following construction works:

•          Organisation of the execution of building projects, i.e. services of collecting financial, technical and other material means for the construction of residential and non-residential buildings to be sold afterwards

•          Construction of residential and non-residential buildings

•          Civil engineering works

•          Specialised construction activities: demolition and site preparation, construction installation activities, building completion and finishing (plastering, joinery installation, floor and wall covering, painting and glazing), roofing activities and other specialised construction activities.

The reporting unit is an enterprise.

The observation unit is the KAU. 

The survey GRAD/M covers construction companies (sector F according to NACE Rev. 2) as well as their units engaged in construction and some non-construction companies performing construction work. This survey is a treshold survey; with the treshold we determine the number of reporting units that are included in the survey in individual years. In 2022, the survey covers 330 units (observation units).

The Republic of Slovenia.

Data are available from the January 2000.

2015

Index.

A month.

The survey is conducted in accordance with:

-          Annual Programme of Statistical Surveys (LPSR) (only in Slovene)

-          National Statistics Act (OJ RS, No. 45/95 and 9/01)

-          European Business Statistics (EBS) Regulation (EU) 2019/2152 of the European Parliament and Council of 27 November 2019 and the Commission Implementing Regulation 2020/1197 laying down technical specifications and arragements pursuant to the mentioned EBS REgulation (General Implementing Act). The former legal basis for the STS indicators is the Council Regulation No 1165/98 of 19 May 1998 concerning short term statistics and subsequent amending regulations.

We do not send data directly to any other data producing agency except to Eurostat.

Regulation (EC) No 223/2009 on European statistics (recital 24 and Article 20 (4)) of 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. All data collected and published by the Statistical Office are governed by the National Statistics Act (OJ) RS No. 45/95 and (No. 9/01).

All published data are aggregated to such a level that there is no confidentiality problem.

The release calendar with advanced release dates for all months of the reference year is published at the beginning of the year on a website of the Statistical Office of the RS. The release calendar is publicly accessible.

The release calendar can be found on SURS website http://www.stat.si/StatWeb/en/ReleaseCal.

The most important information channel for data publication is the website (http://www.stat.si/StatWeb/en/home). The Office publishes several types of serial publications, different series and other publications which are according to the content intended for different users (e.g. general and professional public, statistical experts). Data are available free of charge, except those data that are prepared on users' request. Statistical data and information are always published at 10.30. All releases are announced at the Office's website http://www.stat.si/StatWeb/en/ReleaseCal in the column Release dates. Data which are published in the national publications are simultaneously sent in electronic form to Eurostat via Edamis. No user has prior access to the data. Same release policy applied to national release is applied to transmissions to Eurostat.

The indices are disseminated monthly (_M).

The data on construction put in place are published monthly at the Office's website http://www.stat.si/StatWeb/en/home according to Release calendar, in the first release "Indices of the value of construction put in place" at 10:30 a.m. The First Release contains indices for total construction, construction of buildings, civil engineering and specialised construction activites. First Release is available within 45 days after reference month. With releases in the First Release series the Office disseminates first and most important findings of statistical surveys.

There is no paper publications. Indices of the value of construction put in place are published in e-Release (First Release).

SiStat Database https://pxweb.stat.si/SiStat/en → Construction → Construction works and costs → Monthly data on construction.

SiStat Database provides a modern way of preparing and exporting data for selected categories.

In Slovenia, the entities entitled to obtain the statistically protected microdata include the registered research institutions, registered researchers, and the researchers of government offices. Basic instructions concerning the access and the use of statistically protected microdata are available on the website:

http://www.stat.si/StatWeb/en/StaticPages/Index/For-Researchers

The data on construction put in place are transmitted to Eurostat according to STS Requirements under Nace Rev.2 in order to compile EU aggregates and for the dissemination of national data in Eurostat’s dissemination database.

Methodological explanations for the indices of the value of construction put in place are available on the website: http://www.stat.si/StatWeb/en/mainnavigation/methods-and-classifications/methodological-explanations

Theme: Construction, Sub-theme: Construction Works and Costs. There is an automatic link to the methodological explanations in the eletronic version of the First Release and at the SiStat Database.

Quality reports are availabe at the website: https://www.stat.si/StatWeb/en/Methods/QuestionnairesMethodologicalExplanationsQualityReports

The Statistical Office of the Republic of Slovenia operates on the basis of the National Statistics Act and Regulation (EC) No. 223/2009 on European statistics; in performing its tasks it follows the general principles of quality management, the European Statistics Code of Practice and the Fundamental Principles of Official Statistics. In line with the stated, SURS declares that it takes into account the following principles: professional independence, process orientation, quality of products and services, planning of improvements, stimulating working environment for employees, data providers-friendly official statistics, user-oriented official statistics. Quality documentation can be reached on SURS website  https://www.stat.si/StatWeb/en/Methods/QuestionnairesMethodologicalExplanationsQualityReports

Indices of the value of construction put in place are produced in compliance with methodological requirements and standards. All statistics that are required by current Regulative are available.

Data are used by different users: Slovenian Institute of Macroeconomics Analysis and Development, National Chamber of Commerce and Industry - Construction and Building Materials Association, National Bank of Slovenia, educational institutions, Eurostat (foreign institution), internal users (national accounts), government departments, researches, enterprises, media.

The principle users are asked about their needs, wishes and interests at the regular meetings, i.e. at the Statistical Advisory Commite on Construction.

All statistics that are required by current Regulative are available.

The over-coverage rate for the index of the value of total construction put in place in 2022 is on average 18.71%. The share of over-coverage indicates a relatively high number of ineligible units in the population, which is mainly the consequence of the misregistration of the activity of enterprises in the business register and demographic changes. The average unweighted unit non-response rate is 17.14%. The average weighted unit non-response is 11.44%.

The non-response rate is calculated for the index of the value of total construction put in place. No case of item non-response is involved in the survey since also in the data editing phase, units which did not fully complete the questionnaire are contacted by telephone and asked to provide the missing data. The share of imputated data is equal to the non-response rate, since all unit non-responses are imputed. We do not calculate the share of misclassified units (units with a wrong activity).

The first release of indices of the value of construction put in place is 1 month and 15 days after the end of reference period (t+45). The data are changed the most between the first release and the second release. The absolute difference (coherence) between the data published within t+45 and those that were published within t+90 for 2022 is 2.3. The bias (calculated as the sum of difference between the data on construction put in place published within t+45 days and those that were published within t+90 days divided with 12) for 2022 is 2.2. After the second release (t+90 days) the indices changed a little (±0.2 in absolute numbers) therefore the bias for data in the first release and the final release is smaller as the bias between the data of the first release and the second release.

The survey is not conducted on the basis of a random sample, therefore there is no sampling error. We use census above a treshold.

Coverage errors

Over-coverage rate is on average 18.71%. The units which represent over-coverage can be divided into ineligible units (we include them in coverage because of lack of information, but do not belong there) and units which were at the inclusion in coverage eligible, but became ineligible at the time of survey implementation as a result of demographic changes (for example, to avoid double counting subcontractors are not required to report).

Under-coverage is a result of the procedure used for the determination of frame population. The survey frame is based on Statistical Business Register (SBR). Observation units are chosen at the end of the year for the next year on the basis of treshold coverage; with the treshold we determine the number of reporting units that are included in the survey in individual years. Units in the sample are selected based on turnover for value added tax purposes. They are distributed in descending order of turnover and then as many units are selected from the beginning of the list that their turnover exceeds the selected share (around 56%) of total turnover of all units covered by the survey.

Multiple listings: We havenʾt detected yet that the enterprises are present more than once in the frame, since the survey is based on SBR. 

Measurement errors

Data are collected with the Survey on Consturction activity. Observation units fill in the GRAD/M questionnaire on the website (e-STAT application for electronic reporting of data to the Statistical Office of the Republic of Slovenia). They must complete them with data for the previous month and send them to the Statistical Office of the Republic of Slovenia not later than the 20^th day of the current month. If the respondents give the erroneous data on construction put in place they always have the opportunity to report the correct value later on. If the erroneous data are significant then we detect them through the logical controls, otherwise we detect them by coincidence. The impact of the interwiewer is present just in the case when the unit is contacted by phone because of non-response, but we do not evaluate this kind of errors.

Non-response errors

Unit non-response: The average monthly un-weigted unit non-response rate for 2022 is 17.14%. The average monthly weighted unit non-response is 11.44%. The variable for calculating weighted rate is annual value of turnover according to annual accounts.

Item non-response: No case of item non response is involved in the survey since also in the data editing phase, units which did not fully complete the questionnaire are contacted by telephone and asked to provide the missing data.

Data processing errors

For field units for which data are obtained through a questionnaire the control takes place interactively so we do not have the accurete records of the editing rate, but only on estimate. For 2022 we estimate that we checked the data for approximately 10% of field units and actually corrected data for  less than 1% of units. The share of imputed data is equal to the non response rate, since all unit non responses are imputed. The average imputation rate is 17.14%.

Model errors

Model errors are not relevant because no specific models are used in estimation.

The publication schedule is fixed and announced in advance for the First Release. The data in the First Release are published not later than 1 month and 15 days after the end of the reference month. Data are at the same time also loaded into the SiStat Database on a website of the Statistical office of the Republic of Slovenia, created for user friendly free of charge data manipulation. Data are published as provisional and they are revised monthly as the response rate increases. Data are considered as provisional 6 months after the reference period.

There is no time lag between the actual delivery of the data and the target data when it should have been delivered.

The statistics are in compliance with the STS requirements in the EBS Regulation and the EBS methodologic guidelines. This ensures a good comparability between national data and good quality European aggregates. 

Time series of indices calculated according to the current methodology are available from January 2000 onwards. Until the end of 2022, the lenght of the series is 276 months.

In 2021, a major revision of the survey was carried out. The main reason for performing the revision is the new requirement of the EBRD (European Business Statistics Regulation), which entered into force in 2021. With the introduction of the new legislation, Eurostat's requirement for reporting and publishing data has changed and data will no longer be collected according to the CC classification, but according to the Nace Rev.2 classification. With the introduction of the new classification, it will be necessary to publish data separately for specialised construction activities. In order to keep comparability with the data already disseminated, all time series for the 2000 - 2020 period were back-cast and published. The backcasting was carried out with a combination of calculation at the micro and macro levels. For the time periods between 2015 and 2021, the calculation at the micro level was used, which means that in the micro data we estimated the share of buildings and civil engineering, which is according to the new methodology  appropriately reallocated to specialised construction activities. Using these shares, we then estimated the decrease in value in the first two groups and at the same time also the nominal value in the new group. In order to estimate the respective shares, we used the data obtained for three months (February - April 2020) with a special ad-hoc survey, where we asked all observed units in 2020 about the appropriate shares of specialised construction activities, which have so far been involved in works on buildings and civil engineering. For data before 2015, we performed calculations with a macro-level model, namely we estimated the share of “shrinkage” of the first two groups for each month of the year from the previously estimated micro data. These “fixed” shares were then used on the aggregates, obtained by the old methodology, to estimate the new nominal aggregates in all three groups.

In 2022, the method of reporting data changed. It is now again comparable to 2020, as the main contractors again report data on the value of construction put in place for themselves and for their subcontractors. In 2021, the data were reported by the contractors that actually carried out the construction work, regardless of whether they were the main contractors or subcontractors. To ensure the comparability of the time series, for 2021 we adjusted the data to the new methodology. The analysis of reported data in 2021 showed the volume of reported values to be much smaller than in previous years. This was due to the fact that many subcontractors that were supposed to report values previously reported by the main contractors were not covered as they were below the coverage threshold. The value of construction put in place in 2021 was therefore recalculated by estimating the share of values lost due to changes in the reporting method using various external sources and increasing the reported data for this share in 2021. Based on these adjusted values and the reported values of construction put in place in 2021, we then recalculated all index series for 2021.

The results of this survey are comparable with the results of the monthly survey Business Tendency in Construction (PA-GRAD/M). The observation units in PA-GRAD/M are enterprises of which the main activity in the Business Register of Slovenia is classified according to the NACE Rev. 2 as construction. The observation units selected by the following two criteria: the size (the number of employees in accordance with the Companies Act), and the classification of the enterprise according to the NACE Rev. 2.

The surveys differ in methodologies, since the Value of construction put in place is a quantitative survey, aimed at monitoring value variables, while Business tendency in Construction is a qualitative survey aiming at monitoring the opinions of directors on certain economic indicators within the enterprise. The results of the Survey on Business Tendency in Construction are shown as the balance by the individual question. Due to the different nature of the data, differences between the results of one or another survey are not indicated, but the results are shown only in graphic form.

Chart 1: Confidence indicator and indices of construction put in place, Januar 2017-December 2022

Although construction was the activity least affected by Covid-19, the construction confidence indicator in 2020 was much lower than the data collected in the monthly construction survey. Therefore, the charts for 2020 differ more than they did in previous year. 

Data are internally coherent. Each level of aggregated data is compiled directly from micro data.

In 2022 the number of working hours spent for the selection of enterprises in the sample, preparing the adress book, designing an electronic questionnair and notification letters, data capture, data entry and data verification, calculation of the results and their publication, etc. was around 3,253. Numer of reporting units that submitted the data was 220. Average time spent for one questionnaire was around 23 minutes by one reporting unit.

Types of data revisions in relation to planning:

 a) Planned data revision

 Planned data revision is subject to the following reasons:

I.  Due to the needs of users for timely information, data are published that meet the criteria of the quality of official statistical data, but do not meet the quality that can be met with additional statistical procedures. Final data are based on more complete answers about the phenomenon and/or analyses and are published later on;

II. Seasonal adjustment and/or elimination of calendar effects;

III. Change in methodology and classifications.

 b) Unplanned data revision: Unplanned data revision is not part of the regular statistical process. It appears due to unpredictable changes in the methodology, unpredictable emergence of new and better data, unpredictable changes regarding reporting units that transmit their data to the Office, unpredictable obstacles in data processing and publishing, and errors in data processing and publishing (e.g. a key unit corrects its data for the past few months, an unpredictable change in the administrative data source).

 Types of data revisions in relation to time of implementation:

 a) Regular revisions: inclusion of a more complete/additional data source or a change in the data source, seasonal adjustment and/or elimination of calendar effects;

 b) Occasional revisions are a consequence of including a new/more complete/additional data source that becomes the standard in later data releases or a consequence of an unpredictable obstacle in data processing and publishing, and change in methodology.

 Types of data revisions in relation to the purpose:

a) Inclusion of a more complete/additional data source or a change in the data source;

b) Seasonal adjustment and/or elimination of calendar effects;

c) Transition to a new base period;

d) Improvement of methodology due to a change in the statistical method or a change in classifications, concepts and definitions;

e) Elimination or errors

More information on SURS web-site: http://www.stat.si/StatWeb/en/mainnavigation/methods-and-classifications/methodological-explanations (General methodological explanations - Revision of statistical data)

The release calendar for all statistical survey is accessible on the SURS web-site: http://www.stat.si/StatWeb/en/ReleaseCal

In the case of methodological changes, the data are appropriately recalculated.

The same revision policy is applied to data released nationally and transmitted to Eurostat.

Mean Absolute Revision (MAR) and Mean Revision (MR):

The values for the IPC index are:

• Growth rates for original data series (YoY):

MAR = 3.2759, MR = 0.8947

• Growthe rates for calendar/working day adjusted data are the same, because calendar effects are not significant.

Changes and revisions of methodology are announced in advance (more information in Field 8.1). Before that the Advisory Commiteee for Construction discusses the proposed methodological changes.

If a revision is necessery due to an error, a new edition with explanatory notes is published as soon as possible.

Regular revisions of the IPC are due to inclusion of better and more completed data (because of unit non-response). Regular revisions of provisional data are on a monthly basis and are quite small and on average have no impact on the highest aggregates.

Occasional revisions are due to methodological changes. If possible and if the revision has a great impact on the data series, we recalculate the whole series (for instance when NACE Rev. 2 was implemented). JDemetra + was implemented at the beginning of 2016. In 2013 there was a change in the base year, which was 2010. At the same time we changed the methodology for calculating the total index; the chaining method is now used.

Due to change of the base year data is planned every five year. Data was last revised due to change of the base year in 2018 (which is now 2015).

The same revision policy is applied to data released nationally and transmitted to our users (including Eurostat). Wheater data is final, provisional or changed because of errors or major revisions, we inform our users. At the moment we do not use benchmarking.

Data for the last 6 months are provisional. With each release the data for the last 6 moths can be corrected and supplemented with new data. Data revision is planned. Due to the needs of users for timely information, data are published that meet the criteria of the quality of official statistical data, but do not meet the quality that can be met with additional statistical procedures. Regular revisions include a more complete data source and change of data due to seasonal adjustment and elimination of calendar effects. Data are revised when newer and more quality data respectively can significantly contribute to the quality of data-based decision-making and when due to publication deadlines determined by the European legislation less accurate data are published on the basis of incomplete coverage.

Coherence between Provisional and Final Data

We show coherance for the production in construction index (IPC) for total construction, buildings and civil engineering works. 

Table 1: Provisional IPC, current month/provisional month, 2022

Table 2: Final IPC, current month/provisional month, 2022

 Table 3: Difference between the first provisional and final IPC

Statistical survey. The survey covers construction companies (sector F according to NACE Rev.2) as well as their units engaged in construction and some non-construction companies performing construction work. The survey is a threshold survey: with the treshold we determine the number of reporting units that are included in the survey in individual years. Selected enterprises represent around 56% of the total construction works done. In 2022 330 reporting units were included.

Data are collected monthly.

One standard questionnaire - Monthly Questionnaire on Construction activity (GRAD/M) is used. From January 2016 to December 2018 data were collected by postal and electronic questionnaires (before January 2016 we used only postal qustionnaires). From 2019 on data were collected only by electronic questionnaire (we stopped postal questionnaires). 

We monitor unit non-response each month and have various methods for decreasing it starting with important reporting units.

Data are checked at the micro level with various error detection (intra-dataset checks). All errors, light and heavy, need to be checked in telephone contact with a response unit.

Plausibility checks performed at makro level: comparison of the index on aggregate level for the reference month with the indices of the same months of the previous years is also made on aggregate level. A kind of detection question on demographic change of the unit is added on the questionnaire (did the unit operate in the reference month and if not why not). There is also the remark field for any explanation of the unit (especially for the case of very low or high value of construction put in place). The objective of all kind of the detection questions on the field questionnaire is to get more relevant information on the data important for the survey. For the field units we collect the contact data of the person which fulfils the questionnaire and in the case of uncertainties in the data we make an inquiry calls. The information and explanation of deviated data gather directly from the enterprises are considered as very precious. Sometimes the information from enterprises explain the unusual movement of development of data series. Sometimes the enterprises do not inform us about the demographic change automatically but we reach the information through inquiry call. Once a year we also compare data with the data received via Annual report on Construction.

We do not use special aggregates to sent data to Eurostat. Data which are published in the national publications are simultaneously sent in electronic form to Eurostat. Data are transmitted monthly in .xml format (SDMX) via eDamis to Eurostat (since 2017).

In the case of unit non-response, data are imputed automatically. The value of construction put in place is imputed when we can not acquire the data from the observation units. In the imputation phase the following methods are used (step by step):

• The historical data method with adjustments for the increase of donor. The method is used in case of units for which we have the previous month data.
• The method of logical imputs. According to this method we imput the value that is determined by a simple arithmetic rule, according to which the imputed value is derived from the other variables of the unit, for which the value is imputed.
• The structural method, by taking into account the share in the historic data. The method is used when values are imputed for the variable that is one of the values to be added in the sum of the stated variabe.
• The internal donor method for several variables. According to this method, the value of the donor is imputed; namely, the same donor is determined for all the variables that are imputed with this method at this step.

In 2006 major revision was done including introduction of automatic imputations. In 2009 the transition to the new Standard Classification of Activities 2008 was achieved. In 2013 we changed the base year for calculating the indices (base year was 2010). At the same time we changed the methodology for calculating the total index; the chaining method is now used. The mentioned method has been used for calculating the total index since 2005. The data between 2000 and 2004 are only recalculated to the base year 2010. The weights are from 2013 changed every year. We used chain weighted indices.

In 2018 we changed the base year for calculading the indices, which is now 2015.

In 2021, a major revision of the survey was carried out. The main reason for performing the revision is the new requirement of the EBRD (European Business Statistics Regulation), which entered into force in 2021. With the introduction of the new legislation, Eurostat's requirement for reporting and publishing data has changed and data will no longer be collected according to the CC classification, but according to the Nace Rev.2 classification. With the introduction of the new classification, it will be necessary to publish data separately for specialised construction activities. In order to keep comparability with the data already disseminated, all time series for the 2000 - 2020 period were back-cast and published. The backcasting was carried out with a combination of calculation at the micro and macro levels. For the time periods between 2015 and 2021, the calculation at the micro level was used, which means that in the micro data we estimated the share of buildings and civil engineering, which is according to the new methodology  appropriately reallocated to specialised construction activities. Using these shares, we then estimated the decrease in value in the first two groups and at the same time also the nominal value in the new group. In order to estimate the respective shares, we used the data obtained for three months (February - April 2020) with a special ad-hoc survey, where we asked all observed units in 2020 about the appropriate shares of specialised construction activities, which have so far been involved in works on buildings and civil engineering. For data before 2015, we performed calculations with a macro-level model, namely we estimated the share of “shrinkage” of the first two groups for each month of the year from the previously estimated micro data. These “fixed” shares were then used on the aggregates, obtained by the old methodology, to estimate the new nominal aggregates in all three groups.

In 2022, the method of reporting data changed. It is now again comparable to 2020, as the main contractors again report data on the value of construction put in place for themselves and for their subcontractors. In 2021, the data were reported by the contractors that actually carried out the construction work, regardless of whether they were the main contractors or subcontractors. To ensure the comparability of the time series, for 2021 we adjusted the data to the new methodology. The analysis of reported data in 2021 showed the volume of reported values to be much smaller than in previous years. This was due to the fact that many subcontractors that were supposed to report values previously reported by the main contractors were not covered as they were below the coverage threshold. The value of construction put in place in 2021 was therefore recalculated by estimating the share of values lost due to changes in the reporting method using various external sources and increasing the reported data for this share in 2021. Based on these adjusted values and the reported values of construction put in place in 2021, we then recalculated all index series for 2021.

Seasonal adjustment is the main part of time series analysis. With seasonal adjustment, seasonal effects and calendar effects are removed, if they are significant. In this way, the data are simplified so that they can be more appropriately interpreted, because seasonal fluctuations can blur other important movements.  

The software and version: JDemetra+ 2.2.0, TRAMO/SEATS method.

The model/filter selection: manual, but some automatic tests are used for help (test for transformation, automatic detection of outliers, automatic selection of ARIMA model).

The models and the respective parameters re-estimated: for each new release of unadjusted data, the parameters of the models are re-estimated; the models mostly remain the same, but sometimes they are changed (e.g. an outlier at the end of a time series is added). Major changes of the models usually occur every few years. 

The horizon of revisions: for each new release of unadjusted data, the whole seasonally adjusted time series are revised.

Seasonal adjustment decomposition: all 6 transmited time series have multiplicative decomposition.

The critical value for outlier detection: critical value for outlier detection depends on the time series, usually it is between 2.5 and 3.5. Most of the time series don't have the possibility of outlier detection.

The filter lenght is automatically chosen. 

The date of seasonal breaks in the series: there are no seasonal breaks in the time series.

Direct/indirect adjustment: all the time series are seasonally adjusted directly.

Residual seasonality: residual seasonality is checked when the model is selected. Afterwards, residual seasonality diagnostics are taken into account.

The consistency amongst the different levels of breakdown: all the time series are seasonally adjusted directly, so seasonally adjusted data of an aggregate are not composed of seasonally adjusted data of its components. When models are selected or changed, connection between an aggregate and its components is taken into account (similar time series have similar models …).

No comment.