/ EUROPEAN COMMISSION
EUROSTAT
Deputy Director-General
Unit DDG-02: Statistical governance, quality and evaluation
Directorate B: Statistical methods and tools; Dissemination
Unit B-3: Statistical information technologies /

XBRL Pilot Project Task Force 9 April 2008

Overview of the XBRL Pilot Project

Authors:Giuseppe Sindoni
John Allen
Claudia Junker

Document date:25.3.2008

1

1.Purpose

The purpose of this document is

  • to provide the Task Force with an overview of the XBRL Pilot Project;
  • to summarise the results of the Pilot Project;
  • to set out some tentative conclusions for future lines of work

2.Background

XBRL is the eXtensible Business Reporting Language, a technical standard for electronic communication of business and financial data. It is developed and maintained by an international non-profit consortium of more than 300 companies, organisations and government agencies. It is based on the XML and XLink technical standards.

Eurostat (units B3 and DDG-02) coordinated an XBRL Pilot Project in the European Statistical System, reporting to the Accounting and Statistics Task Force, which is a joint Eurostat-ECB group comprising representatives of NSIs and Central Banks.

The general aim of the XBRL Pilot Project Member States was to test the feasibility of exploiting for statistical purposes the use of XBRL in company reporting (for example, the company accounts which are delivered to national regulators and central banks), in order to lower the reporting burden to enterprises.

The XBRL Pilot Project was one of the work areas of X-DIS (XML for Data Interoperability in Statistics), the statistical project funded under the EU IDABC[1] programme. The budget for the XBRL Pilot Project was 375.000 euros over 2 years (budget years 2005 and 2006).

No grants were made to Member States; the IDABC Decision does not provide for grants.

3.History

The Accounting and Statistics Task Force decided in March 2004 to organise a stock taking survey on the use of XBRL in the National Central Banks and the National Statistical Institutes in the EU, candidate and EEA countries.

The results of the survey showed that a limited number of institutions had operational taxonomies in place, albeit that these taxonomies are used on a limited scale; several institutions had plans for the near future; generally speaking the comments made by the institutions that already worked with XBRL were quite positive; whereas the start-up cost can be considerable, the running cost of the use of XBRL is low as long as the reporting environment remains stable; this could pose a problem as International Financial Reporting Standards (IFRS[2]) are subject to continuous change; Germany, the Netherlands, Belgium, the UK and Spain were most advanced, but in the vast majority of the Member States XBRL was not used at all.

In 2004-2005, PriceWaterhouseCoopers conducted for Eurostat a feasibility study into a proposed project on the extraction of economic statistical data from XBRL reports produced by accounting software packages. The study showed that, for statistics, the successful exploitation of XBRL depends on the availability of local GAAP[3] taxonomies[4] and/or widespread use of the IFRS taxonomy.

4.XBRL Pilot Project

4.1.Objectives

The objective of the pilot project was to come to clear conclusions whether or not XBRL can be used efficiently in the statistical reporting chain and whether or not reporting with XBRL can gain widespread acceptance among European enterprises and Statistical Institutes. The pilot focuses on the reporting by enterprises to NSIs and not on the reporting from NSIs to Eurostat.

These objectives were to be reached by accomplishing the following tasks: mapping of existing taxonomies and the requirements in the relevant statistical legislation; monitoring of XBRL developments in existing XBRL projects in the statistical and accounting domains; creation of a statistical taxonomy; testing taxonomies in a live environment with the cooperation of four NSIs; assessing conditions for XBRL acceptance by NSIs; investigating whether it is feasible that enterprises would no longer send their information to the NSI, but that enterprises post data on their websites where they would be ‘picked up’ by the NSI.

The XBRL Pilot Project started in March 2006 and ended in December 2007. The taxonomy mapping task has been accomplished and a statistical XBRL taxonomy was developed. The initial focus was on the Structural Business Statistics framework and in 2007 the taxonomy was extended to the Short-Term Statistics framework.

4.2.The XBRL Pilot Project Task Force

Eurostat formed a small Task Force to contribute actively to the above tasks and to monitor project progress.

A Task Force preparatory meeting was held in Luxembourg in March 2006, with the participation of the NSIs of Belgium, Ireland, Spain, the Netherlands and the United Kingdom, and the Central Banks of Germany and Spain.

After this meeting, the NSIs of Belgium, the Netherlands and Spain, and the Central Bank of Spain were invited to join the Task Force as active members, whose tasks would be to provide Eurostat with all the needed information and support to accomplish the project tasks.

The XBRL Pilot Project Task Force met for the second time on 27.9.2006, with Luxembourg and Sweden as observers, together with XBRL Europe and the ECB. Member States confirmed the planning of activities up to February 2007, and agreed that the IFRS are the best basis for developing a European Statistical Taxonomy. There was optimism in the Task Force about the potential of XBRL for reducing the response burden of enterprises, but it was agreed that this was a medium or long-term benefit; no dramatic short-term benefits should be assumed.

The Task Force met again on 19.4.2007, to discuss the results of the first phase of the project and the plans for the second phase. It was agreed that although the results were not as good as expected, all the findings have been helpful and the hopes were pinned on the active members' commitment in Phase 2. During the meeting Ireland (CSO) decided to join the group of active members for Phase 2. Poland (GUS) was also present at this meeting.

As of November 2007, the Pilot Project Task Force members were:

Active members / NSIs: BE, ES, IE, NL, SE
Observers / NSIs: LU, PL
CBs: DE, ES

4.3.Results

Tests performed in the Phase 1 of the project (Q1 2006 - Q1 2007) showed:

  • heterogeneity between countries in the degree of reusability of financial and balance sheet concepts for statistical purpose and on the architectural approach to the introduction of XBRL in the statistical process
  • low degree of XBRL acceptance by data providers (companies)
  • unclear plans for XBRL compliance of the most popular accounting software systems
  • not very high availability of companies to participate in pilot exercises.

Nevertheless, a successful pilot exercise on using a reference XBRL software architecture for primary data collection, developed by Eurostat, was conducted with Belgium (Statbel), who are now considering to use and extend this software for their XBRL-based data collections.

The results of Phase 2 (Q2 - Q4 2007) showed:

  • low degree of coverage of the statistical variables by the accounting and financial reporting standards; as a consequence, there is a low degree of correspondence between the XBRL taxonomies to be used by companies for reporting to the regulators and the European statistical taxonomy developed by Eurostat
  • as a consequence of the above point, the tests demonstrated that, currently, it is not realistic to propose the reuse, for statistical purposes, of XBRL reports sent by companies to regulators.

The activities and results of Phase 2 are set out in the contractor’s final report[5], and are further discussed below.

5.Conclusions from the Pilot Project

  • It was feasible to create a European XBRL taxonomy incorporating the SBS and STS variables, as such a taxonomy was developed within the Pilot Project.
  • Although there are a number of statistical surveys of enterprises (e.g. labour cost, structure of earnings, vocational training, ICT, tourism statistics) the Pilot Project was not able to identify any areas of statistics other than SBS or STS where the use of XBRL could be investigated in the short term, as these other areas involve data which are quite separate from financial reporting.
  • The level of successful mapping between the SBS and STS variables and the test instances provided countries was quite low:

Case/Taxonomy / Framework / Successful Matching Percentage
NL case
NTP version 1.1 / SBS / 32.9%
STS / 52.2%
BE case
CBSO_NBB version 1.1 / SBS / 5.26%
STS / 8.69%
IE case
Irish GAAP version 2.1 / SBS / 9.21%
STS / 13.04%
ES case
ES-BE-CB version 1.1 / SBS / 9.21%
STS / 17.39%
  • As seen here, only in the NL case could the level of mapping be said to be encouraging: the NTP taxonomy version 1.1 is a reasonably good candidate for obtaining IFRS-based statistical information.
  • The GAAP-based taxonomies used for the mapping exercise in BE, ES and IE seemed not to offer a satisfactory basis for generating SBS or STS statistics.
  • Even the low level of mapping needs to be further checked in what concerns the exact definitions and concepts behind accounting data and statistical variables to identify whether the definitions behind the variables are identical or whether an identical content can be derived by calculations.
  • It was noted that it is not possible to extrapolate the findings of a mapping exercise to different versions of the taxonomies involved (as in the case of NTP 1.0 and 1.1). Even where taxonomies are fundamentally similar, the mapping process is affected by the differences between versions; in any future exploitation of XBRL for statistical purposes, it will be necessary to avoid versioning issues resulting from linguistic or terminological changes.
  • There was little evidence of existing or spontaneous use of XBRL for reporting by companies. This was true for all the countries participating in the Pilot Project, even though the group included countries where XBRL had been promoted as a tool for reporting.
  • The reasons for the lack of real instances of XBRL reports from companies is not clear. Several factors may be relevant, for example:

–accounting software and management information systems at companies are not capable of producing XBRL output;

–accounting software and management information systems at companies are capable of producing XBRL output but the rather complex state of national XBRL taxonomies has made it difficult for companies to exploit this capability;

–companies see no added value in producing reports in XBRL; there is as yet not European initiative corresponding to the US SEC programme for XBRL reporting.

The opinion of the Pilot Project Task Force on this point would be useful.

  • The diversity of national XBRL taxonomies reflects the complexity of existing financial reporting arrangements, where different companies are required to follow GAAP and/or IFRS and where local GAAPs may substantially differ from IFRS and among each other. The latter would impose difficulties to create European wide taxonomies for given statistical areas with a high matching of concepts.
  • The results of the Pilot Project demonstrate what could be considered self-evident, which is that a technical tool such as XBRL cannot impose uniformity and efficiency on a rather heterogeneous set of reporting requirements but it can potentially trigger a development in such a direction.

6.Next steps

It is hoped that XBRL Pilot Project Task will further elaborate the conclusions tentatively reported above. These will be reported to the Accounting and Statistics Task Force, which will be asked to place these conclusions in a wider context.

It will be necessary to formulate proposals about further work with XBRL, probably within the MEETS programme[6], where XBRL is mentioned as one of the possible instruments for Objective 3.

It is important that any future actions are designed in such a way as to overcome the obstacles which have been identified during the Pilot Project, and to build on the achievements of the Pilot Project.

It can be noted here that Unit DDG-02 plans to start, in 2009, a project to develop a taxonomy linking accounting and statistical concepts for business reporting, with a specific focus on small and medium-sized enterprises (SME), which will make use of the work done in this area by the XBRL Pilot Project. This project will thus address the challenge to first develop common taxonomies for accounting data and statistical variables and check in detail the definitions and concepts behind these data and variables to identify whether they are similar or not.

Some general principles can be tentatively advanced here:

  • while XBRL continues to offer prospects for lowering costs and response burden for business statistics, the results so far obtained in the Pilot Project suggest that Member States have not yet made much progress in the exploitation of XBRL. It seems likely that these benefits would not be realised without strategic decisions and significant investments by Member States, aimed at more homogeneous reporting requirements within and between Member States, independent of the use of XBRL or any other technology;
  • while working on the XBRL Pilot Project, Eurostat has identified some opportunities for a common ESS approach to develop and maintain elements of the technical infrastructure which could support future reporting requirements. These elements include:

–the European XBRL taxonomy

–an “XBRL Reference Environment” which could be used as a basis for NSI’s IT infrastructure for XBRL.

  • however, the technical infrastructure should not be developed in advance of the improved reporting frameworks.

1

[1] IDABC: Decision 2004/387/EC of the European Parliament and of the Council of 21 April 2004 on interoperable delivery of pan-European eGovernment services to public administrations, businesses and citizens (IDABC). For more information, see the IDA web-site at

[2] IFRS: International Financial Reporting Standards. A set of international accounting standards stating how particular types of transactions and other events should be reported in financial statements.

[3] GAAP: Generally Accepted Accounting Principles. The common set of accounting principles, standards and procedures that companies use to compile their financial statements. GAAP are a combination of authoritative standards (set by policy boards) and the commonly accepted ways of recording and reporting accounting information.

[4] XBRL Taxonomies are the dictionaries which the language uses. These are the categorisation schemes which define the specific tags for individual items of data (such as "net profit").

[5] X-DIS XBRL Pilot Project Phase 2 Final Report, Version 1.0, 31.1.2008. See especially section 6.2.4.

[6] Proposal for a Decision of the European Parliament and of the Council on a programme for the Modernisation of European Enterprise and Trade Statistics (MEETS), COM(2007) 433 final