Archive:European business statistics manual - dissemination

This Statistics Explained article is outdated and has been archived - for updated information please see the dynamic version of the European Business Statistics Manual at: European Business Statistics Manual

A static full version of the European Business Statistics Manual was published in February 2021: European Business Statistics Manual — 2021 edition This article discusses the channels through which business statistics are disseminated within the European Union, concluding with 2 closely related topics:

data revision
use of standard code lists.

The article is part of the online European Business Statistics Manual, which provides a comprehensive description of methodologies and background information on how business data are compiled within the European Statistical System (ESS).

Full article

Introduction

In this article, dissemination is defined as publishing data produced by data compilers (input) and making statistics available in line with user needs (output).

Publication can be divided into a range of channels, each designed to meet particular user needs. They include news releases, online publications and databases. The channels Eurostat uses to make statistics available to the public are set out in Section 2. Section 3 outlines a number of ways to improve dissemination in future in terms of user needs, as set out in the ESS Vision 2020.

Section 4 presents EU policy and practice in the field of data revision.

The classifications and breakdowns of the data are harmonised and conventionalised using standard code lists so as to ensure that the same coding and definitions are used across data sets. Section 5 provides further background information.

Channels for publishing statistics

Increasingly, EU business statistics are published online, with paper publications becoming the exception. In fact, there are no longer any paper publications devoted exclusively to business statistics. However, the print publication ‘Key Figures on Europe’ does include some selected business statistics.

Business statistics are made available in a range of ways at EU level. The main channels are:

News releases
Free online databases
‘Statistics Explained’

Eurostat’s business statistics are linked with metadata, which describe the quality, background and structure of the data using a common standard. These metadata can be found directly as hyperlinks connected to the respective data set.

Safeguarding rules on data confidentiality are explained in Statistical Disclosure Control. These rules guarantee that data sent to Eurostat as confidential information are not published.

’Accessing European statistics’ gives a more detailed overview of the various channels through which statistics in general are published.

Business statistics are also covered by a number of 'dedicated sections' to guide users interested in data or publications in a particular area:

‘Structural business statistics’ is divided into sub-domains, including business services, globalised businesses, small and medium enterprises, business demography, foreign affiliates, and international sourcing.

Regional data in the fields of structural business statistics, the information society and science and technology are also available; see the more comprehensive 'dedicated section' on regional data.

The above domains include hyperlinks to the associated data sets. These data sets are shown by means of the Data Navigation Tree and are explained by the accompanying methodological notes, which describe all relevant metadata at a glance. These notes also explain release policy, such as time of release. There is a detailed release calendar for short-term statistics.

Improving the dissemination of statistics in future

In 2015 the ESS launched DIGICOM as an ESS Vision 2020 development project. DIGICOM’s purpose is to improve the ESS’s agility and its responsiveness to users’ needs, promoting European Statistics as a reliable framework for evidence-based decision-making and an unbiased picture of society.

DIGICOM is scheduled to run until 2019. It comprises the following building blocks:

a user profiling exercise and a new social network for users of EU statistics, to improve identification of user needs and cooperation with stakeholders
a new biennial ESS flagship product, based on an EU data pool, that will be dynamic and interactive, and new reusable tools for customised data consumption, to enrich ESS’s innovative and shareable products and tools
better automated access to EU aggregate data for heavy data users and redistributors, to improve open data dissemination; promotion of redistribution and improved access to anonymised microdata
a renewed communication and marketing strategy to further promote EU statistics; setting up of a policy framework for disseminating experimental statistics.

For further details, see DIGICOM’s business case (pdf).

The improvements to be developed by DIGICOM are general in nature, not specifically targeted on business statistics or businesses themselves.

ESS Vision 2020 (page 9) provides further guidance on how to disseminate business statistics more effectively in the future:

Enterprises are both information providers and users. Enterprises will be consulted to identify their data needs and, if asked to do so, develop and provide tailor-made analysis and services to enable a comparison to be made with competitors in the same line of industry (benchmarking)

In the course of 2018-2019 complementary measures for disseminating business statistics are being explored and implemented that are designed to meet the ESS’s enterprise-oriented objective as well as addressing user needs for cross-domain information on businesses. One of the tools being developed for this purpose is the Business Sector Profile (BSP). The new tool provides an overview for each business sector, covering a wide spectrum of short and long-term indicators on revenues and expenses, labour market, investment and technology, prices and productivity, entrepreneurship, international business, product markets and other business related aspects. A user test of a first partial beta version has been conducted in May/June 2018, and the official launch of the tool on the Eurostat website is foreseen for end of 2018 covering as a first step the sectors of manufacturing industry. During 2019 its content will be enriched with sectors from energy, construction, trade and services. Besides the Business Sector Profile, focussed on sectors at the national level, another new tool will be developed targeting at international business information country-to-country.

Data revision policies and practices

According to the European Statistics Code of Practice, which is part of the main framework for European statistics, the revision of data should be based on the following principles:

There must be advance notice of any major revisions or methodological changes.
Revisions must follow standard, well-established and transparent procedures.
Revisions must be regularly analysed to improve statistical processes.

As a next step, in 2009 the ESSC approved the following set of data revision principles, which are to be applied by the ESS and Eurostat:

Each statistical institution within the ESS defines, communicates and publicly releases well documented general and domain-specific revision policies, applicable to European statistics under its responsibility.
Domain-specific revision policies should be kept consistent across statistical domains and countries, as far as possible, as well as stable over a sufficiently long time period.
Statistical institutions within the ESS should define a common strategy, for each statistical domain, for communicating qualitative and quantitative information on data revisions of European statistics.
Routine and annual revisions should be published in the context of well defined, synchronised and regularly updated release/revision calendars at national and European level. Releases of European and national data aggregates should be synchronised as far as possible.
Major revisions should take place only at fairly long intervals. They should be announced in advance, backwards implemented and coordinated across statistical domains and institutions.
Unscheduled revisions should be reduced over time to revisions of errors and unforeseeable accidents occurring in the production process. Corrections should be released without waiting for scheduled revisions. They should be accompanied by appropriate explanations.
Domain-specific revision policies should rely on sound and homogeneous methodological choices covering scheduling of revisions, possible use of thresholds, length, depth, and seasonal adjustment whenever applicable (the list is not exhaustive).
As far as appropriate, each statistical institution within the ESS carries out and disseminates regular revision analysis at statistical domain level. For this purpose, the appropriate vintage databases consistent with release/revision calendars should be maintained and disseminated.

Further technical guidelines needed to implement ESS revision policy, which are in line with the above principles, have been provided in the context of:

Quality reporting applicable to all business statistics. The 2014 edition of the ESS handbook on quality reports explicitly includes revision as a quality indicator. The quality reports associated with the data sets and drawn up by national statistical institutes and Eurostat should cover revision policy, the number of revisions (both scheduled and non-scheduled), the average scale of revisions, the main reasons for revisions, and the extent to which revision improves accuracy.

Where unscheduled revisions (i.e. corrections of errors) are concerned, there should be an analysis of the nature of errors over the past few years and of what has been done to avoid future errors. The general reporting requirements state how revisions should be reported in the annual quality reports. The domains have the discretion to decide how scheduled and unscheduled revisions are to be treated, provided that general ESS-level principles are respected. In the area of short-term business statistics, the treatment of revisions is subject to detailed guidelines (see below).

Principle European Economic Indicators (PEEIs): these are a comprehensive set of infra-annual macroeconomic indicators that describe the economic and labour market situation, as well as price developments in the euro area and the European Union, which are of particular importance to economic and monetary policy. Many such indicators are drawn from the STS domain of business statistics. The ESS guidelines for revision policy for PEEIs provide detailed guidance on how to treat revisions of short-term business statistics covered by the PEEIs.
Seasonal adjusted data: see ESS guidelines on seasonal adjustment, in particular part 3 (page 24) on revisions. The preferred option is that ‘the revision period for the seasonally adjusted data must at least cover the extent of the raw data revision period. Due to the property of filters, it is normally acceptable to revise the seasonally adjusted data from a point of 3-4 years before the beginning of the revision period of the unadjusted data; earlier data should be frozen’. This guideline can be seen as a further refinement of the PEEI guidelines on revisions.

These guidelines/handbooks have been endorsed by the ESSC and thus represent standard practice. As regards individual statistics, information on revision is published as part of the metadata files linked to the data sets. As revisions are particularly important for short-term business statistics, they are clarified in detail in a separate article, Short-term business statistics – Revisions.

Where Eurostat data publications are revised to remove errors (unscheduled revisions), Eurostat's internal dissemination error management policy applies.

Standard code lists

Statistical code lists are structural metadata on statistical concepts which are in general used to specify the dimensions of datasets (i.e. tables). Multi-dimensional tables on Eurostat’s webpage generally use several such code lists. Where appropriate, the lists are based on official statistical classifications such as NACE and NUTS, or on other code lists relating to businesses and their economic performance and products.

The codes used in dissemination may differ from those used in data collection (particularly those based on the SDMX standard). Some such differences are technical in nature, arising from requirements with respect to software or tools. They may also arise from differences in readability requirements; in automated data collection, machines require absolute singularity of concepts and their associated codes, whereas in dissemination human-readable codes are promoted. Eurostat plans to try to reduce these differences where possible; where this is not possible, mapping with due explanations will continue. The results of this work will be incorporated in this article in due course.

Contacts

For questions or comments on this article, please contact ESTAT-EBS-MANUAL@ec.europa.eu.