Federal Statistical Office (DESTATIS)

E 21 "Business indices, seasonal adjustment"

This indicator is part of the overall STS indicator system which supports business cycle observation. The objective of the production index is to measure changes in the volume of value added.

NACE Rev. 2.

The activities of NACE Rev. 2 Group 41.2, Division 42 and 43 (Section F) are covered. In accordance with the Classification of types in construction (CC), these activities are assigned to the building sector or civil engineering.

The terms ‘production’ or ‘output’ in the sense of business cycle statistics exclude inputs from other units. The objective of the production index is therefore to measure changes in the volume of value added. Data on the change in value added are usually not available on a monthly basis. Therefore, value added is calculated for the base period only. The monthly continuation of value added in base period is done by suitable proxy values. Hours worked is an adequate indicator for continuation in case of long production processes. It is assumed that the development of outputs (products) directly correlates with the required inputs (hours worked).

Local unit.

Production activities of Section F of NACE Rev. 2, i.e. all size classes of NACE 43.2 and 43.3 as well as local units with 20 employees or more for other NACE Groups, are covered.

Country as a whole.

Since 1991.

2015 = 100

Index.

Month.

Till 2022: The laws regulating the collection, treatment and dissemination of statistical data are the Council Regulation (EC) No. 1165/98 of 19 May 1998 concerning short-term statistics with implementing acts (e.g. Commission Regulation No 1503/2006 of 28 September 2006 concerning the definitions). 

From 2022 onwards: Regulation (EU) 2019/2152 of the European Parliament and of the Council on European business statistics, repealing 10 legal acts in the field of business statistics (EBS-Regulation).

Commission Implementing Regulation 2020/1197 laying down technical specifications and arrangements pursuant to Regulation (EU) 2019/2152 of the European Parliament and of the Council on European business statistics repealing 10 legal acts in the field of business statistic (General Implementing Act) as well as the Federal Statistics Act (1987) and the law on industrial statistics (ProdGewStatG 1975).

Furthermore, ESS guidelines on revision policy for PEEIs and recommendations for Working-Day and Seasonal Adjustment in Short-Term Statistics are applied.

The FSO and the German Central Bank cooperate closely on the seasonal adjustment of the production index. Data sent to Eurostat may contain confidential data to compile the European aggregate.

Confidentiality is guaranteed by section 16 of the Federal Statistics Act. It requires that respondents must be informed about their rights and obligations in providing statistical information. This rule is strictly followed. It also applies to the notification that individual data are only used for statistical purposes and - subject to protection procedures - for scientific purposes.

Furthermore, Chapter V “Statistical Confidentiality” of the REGULATION (EC) No 223/2009 of the European Parliament and of the Council applies.

As regards the confidentiality of data, methods of primary and secondary confidentiality checking have been developed and are being used.

A release calendar containing the precise release dates for the reference periods in the following twelve months is published on the website of destatis.

https://www.destatis.de/SiteGlobals/Forms/Suche/Termine/EN/Terminsuche_Formular.html?nn=43216

Indices are simultaneously released to all parties. The data are available at the database “GENESIS-Online” for all users at the same time.

Identification of internal government access to data before release in accordance with Principle 6.7 of the European Statistics Code of Practice: The Federal Ministry for Economic Affairs and Energy and the German Central Bank are provided with an embargo copy of the data at 2 pm on the day before the official release in order to prepare their own comments.

Data are transmitted monthly to Eurostat on the same day as the national figures are published.

Monthly.

https://www.destatis.de/SiteGlobals/Forms/Suche/Presse/EN/Pressesuche_Formular.html;jsessionid=24B1C2158B976A484E29DA16C75B4D09.internet8731?nn=43216

Annually in the German Statistical Yearbook.

Indices are also published at the time series database

https://www-genesis.destatis.de/genesis/online/data?operation=sprachwechsel&language=en

The index of production in the construction sector is published for construction as a whole and for the breakdowns “Hochbau” (over ground industry) and “Tiefbau” (infrastructure industry). Regional figures are published at a more aggregated level by the statistical offices of the Länder.

Not available.

Data are transmitted to Eurostat and used in European aggregates and/or are published as national indices.

A paper dealing with the changes of the base year is published in “Wirtschaft und Statistik Heft 3/2013: Christiane Bald-Herbel: Umstellung der Konjunkturindizes im Produzierenden Gewerbe auf Basis 2010”.

A description of the seasonal adjustment method is published in “Wirtschaft und Statistik Heft 9/2002: Sandra Jung: Erfahrungen mit dem Saisonbereinigungsverfahren Census X12-Arima für den Produktionsindex“.

A detailed analysis concerning the revision policy for this indicator is published in “Wirtschaft und Statistik Heft 9/2003: Sandra Jung: Revisionsanalysis des deutschen  Produktionsindex”.

Standardized quality reports are published on the website of destatis

https://www.destatis.de/EN/Methods/Quality/QualityReports/_node.html

The data of German official statistics enjoy a reputation of high reliability in Germany and abroad. Destatis has committed itself to assure and further enhance the level of quality already attained. In order to achieve this task – also in the ESS – the activities of destatis are based on quality guidelines. To assure quality in the long term, destatis applies a variety of systematic quality assurance measures, among others the European Statistics Code of Practice.

A detailed description can be found on the website of destatis.

https://www.destatis.de/EN/Methods/Quality/_node.html

The quality of the indicator is satisfactory. The monthly survey in the construction sector is conducted in a decentralized manner. The statistical offices of the Länder are responsible for collecting, checking and processing the respective Länder data. The Federal Statistical Office is entrusted with methodological preparations and the development of the statistics, and it collects the Länder results in order to produce and publish federal results as well as to calculate indices.

Due to its periodicity, its rapid availability and detailed breakdown by branches of economic activity, this indicator is a central indicator of the development of business activity. Therefore, it is significant to various national users. These are in particular the Federal Government, more specifically the Federal Ministry for Economic Affairs and Energy and the Federal Ministry of Finance, the German Central Bank, the German Council of Economic Experts, economic research institutes, universities, trade associations, the media and also the interested public.

Destatis conducts user satisfaction surveys in various sectors of statistics. The Statistical Advisory Committee (Statistischer Beirat) represents the interests of the users of federal statistics and regularly consults with destatis.

Additionally, user satisfaction is measured by the number of monthly visits on specialized publications on the website of destatis and the respective number of printed products sold or downloads, quarterly downloads from the on-line database “GENESIS”, downloads of the Statistical Yearbook sorted by chapters and the number of external queries via the costumer management system.

STS requirements according to EU regulations and national laws are fulfilled.

Deviations between the preliminary and the final results amount to a maximum of 0.5 percentage points.

Not relevant since surveys are conducted with a cut-off threshold or data come from various administrative sources. The non-surveyed part of the population is estimated. The estimates are conducted with information of the administrative data.

The response rate 41 and 42 was about 95% in 2022 at the time of first preliminary results. The response is obligatory. Otherwise, fines can be imposed up to EUR 5,000. For 43 only about 25% of administrative data was available at the time of first preliminary publication in 2022. Therefore, the size-weighted unit-non-response is around 50% on average for 2022.

The missing values are estimated with a factor which multiplies the value of the non-respondent from the previous period by the growth rate of the current period/ previous period.

Preliminary results are published approximately 38 days after the end of the reference month (t+38). A final index calculation in regard to the annual correction is made after the completion of the 4th quarter (normally around February).

All publication deadlines are met.

Consistent time series of the recent base year are available since 1991 since Germany is covered as a whole.

Consistent time series of the recent base year are available since 1991. Time series prior to 1991 refer to the former territory of the Federal Republic of Germany.

There are no breaks in the time series.

The data are coordinated with the calculation of the national accounts. The production index is needed for the estimation of the quarterly GDP.

The index formula (Laspeyres, rebasing every five years) is used for each indicator in industry and the construction industry. Thus, the base year and the index reference period coincide. Furthermore, all indices use the same classification framework and are consistent with the definitions, scope, classifications, and timing of recording as required by the STS-Regulation.

The components CC1 and CC2 are coherent with the aggregate total construction which is valid for raw data, working day adjusted data and seasonally adjusted data.

It is not possible to determine cost and loads for the IPC separately. The cost and burden of all STS- indicators in total were transmitted to Eurostat.

The general revision policy adopted by destatis covers routine revisions, major revisions and non-scheduled revisions.

The national revision policy is the same as for Eurostat data.

https://www.destatis.de/EN/Methods/Quality/GeneralRevisionsPolicy.html

Information on tracking vintages are given on the website of the German Central Bank: https://www.bundesbank.de/en/statistics/time-series-databases

Data revision follows the revision policy mentioned above.

Routine revisions result in revised production indices which are published in the following month (i.e. after four weeks), together with the preliminary indices for this month. The revised indices take into account subsequently received production data and possible corrections. A final index calculation in regard to the annual correction is made after the completion of the 4th quarter of the reporting period. For unadjusted data MR=0.3 and MAR=1.3; for seasonal adjusted data MR=0.2 and MAR=1.4.

Note: The revision analysis was conducted based on unadjusted data since calender adjusted data on tracking vintages are no longer available on the Website of the German Central Bank

The last major revision was conducted in 2018 in the context of the new base year 2015. In the case of regular index rebases, the results and changes in the calculation methodology are presented in detail in the FSO monthly journal “Wirtschaft und Statistik”.

The data are not benchmarked. Regular sources are completions of short-term business statistics and the administrative data. There are no indicator specific characteristics of revisions.

The data set is based on statistical surveys (47.7%) and administrative databases (52.3%). The surveys are exhaustive and their coverage is updated yearly.

The frame on which the source is based is the national business register.

Data are collected monthly for main construction works in building and civil engineering ("Bauhauptgewerbe") and quarterly for finishing trades (“Ausbaugewerbe”).

Data are reported via on-line questionnaire. They are sent and collected by the statistical offices of the Länder.

Missing responses are requested by telephone contacts or reminders are sent.

Data are checked by the statistical offices of the Länder before they are transmitted to the FSO.

Plausibility checks are jointly determined by the FSO and the statistical offices of the Länder. They are carried out on the results (i.e. comparison at the level of each time series, calculation of changes in figures compared with the previous year/month) while a review of data quality has already been conducted during the process of index calculation.

Estimates for non-response are based on values of the preceding month which are still contained in the current statistical data reported. Basically, they are replaced by the rate of change of data received within the deadlines, which is then included in the index calculation.

Grossing up data to population results requires several steps. The raw figures collected monthly correspond to the number of hours worked on building sites in enterprises employing 20 persons or more. For NACE 43.2, 43.3, no information on hours worked is available on a monthly basis. Therefore, turnover gathered from administrative data is used for continuation here. Productivity factors are then applied to the development of hours worked to calculate the indices which are ideally compiled by dividing gross value added at factor costs by hours worked. Since gross value added is not available for the same periodicity, the productivity factor is approximated by using the annual accounts of the construction sector. More specific, deflated turnover is divided by hours worked.

These annual productivity factors form a step function. In order to avoid breaks at the end of each year, the productivity function needs to be smoothened. For the months at the end of the series, an expert judgment is necessary. To avoid inconsistent productivity estimation with the building investments calculation, this judgment is always coordinated with national accounts.

Turnover data for the section finishing trades come from the administrative data sources. These data are deflated with construction prices indices.

Construction indices are calculated according to a Laspeyres formula.

The weighting scheme is based on gross value added taken from the annual structure survey and from the annual investment survey.

For seasonal adjustment, version 2.2 of the JDemetra+ software is used. It is recommended for seasonal adjustment within the European Statistical System and the European System of Central Banks. The underlying mathematical-statistical method does not differ fundamentally from the previous X-12-ARIMA procedure.

The first stage of the procedure is RegARIMA modelling; unadjusted values at the ends are extended by estimates and, where required, adjusted for outliers. Extending the time series by estimates is required, for instance, for the second stage of the procedure, where centred moving averages are calculated. Seasonal adjustment as such is carried out in the second stage of the procedure. It is an iterative process for smoothing the unadjusted data by means of trend and seasonal filters which are calculated from centred weighted moving averages. The seasonal factors of a specific month are obtained by smoothing the deviations of the unadjusted values from the trend values of that month. The trend values, in turn, are smoothed unadjusted values.

There may also be calendar adjustment in the first stage. For that purpose, the number of working days in the current month is determined after deducting holidays; the result is compared with the long-term average of that period. Public holidays that are relevant only in some Länder are weighted using the number of employees in those Länder as a proportion of the total number of employees in Germany. The percentage effect of an additional working or trading day on the unadjusted value is estimated from past values of the relevant set of statistics under RegARIMA modelling. These estimates are used to eliminate the impact of variations in the number of working days in the respective month or quarter from the unadjusted data. The effects of bridging days (days between a public holiday and a weekend), movable school holidays and the weather are not taken into account.

Results based on the calendar and seasonal adjustment procedure BV4.1 and trend results according to BV4.1 are published in addition to the results adjusted for calendar and seasonal effects using the X13 JD+ procedure.

software used and version: Calendar and seasonal adjustment based on X13 in JDemetra+ 2.2

model/filter selection: manual

times the models and the respective parameters are re-estimated: monthly

horizon of revisions: revisions are completed after the yearly correction. The time series is calculated backwards up to 2015.

seasonal adjustment decomposition: The underlying model for the seasonal decomposition is multiplicative and only for a few selected time series additive. The model for the estimation of the calendar effects in the construction sector is additive.

ARIMA-Model: mostly “airline” model (0 1 1)(0 1 1)

critical value for outlier detection: 4

filter length: the Filter lengths (Henderson and seasonal filter) are user defined with the value 17 and 3X9, respectively

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

indirect adjustment via components: The production volume index for Germany for NACE section F is indirectly adjusted via its components, which are directly adjusted.

residual seasonality: The residual seasonality is checked on a yearly basis for the industrial main groups.

aggregation: the production index for construction (NACE section F) is calculated from the two-digit NACE positions. consistency amongst the different levels of breakdown: The consistency is checked regularly on a yearly basis. Deviations are moderate but can vary.

None.