XBRL and Statistics

Uwe Kunzler and Alice Zoppè

August 2003

 

What is XBRL?

XBRL (eXtensible Business Reporting Language) is a standard data format for electronic business reporting, based on XML (eXtensible Mark-up language: see http://www.aicpa.org/pubs/jofa/may1999/hoffman.htm and XML standardisation for data collection and exchange for more information on XML and Electronic Data Reporting).

Companies are required to report more and more data to more and more internal and external authorities, including statistical institutions. One way to improve both the production and collection processes is to reduce the number of reporting media and formats: ideally, every report would be generated and sent electronically, in an automated way, and in one reporting format.

It is in the interest of both reporting businesses and collecting administrations to digitalise and automate data reporting, and to use standard data formats.

XBRL is a candidate for such a scope. Behind its development there is XBRL International (www.xbrl.org), a non-profit consortium of companies and organisations, representing various elements of the business reporting chain.

 

XBRL: Foundations and taxonomies

The foundations of XBRL are the XML recommendation of W3C (currently version 1.0, second edition, of 6 October 2000 http://www.w3.org/TR/REC-xml) and the XBRL specification of XBRL International (currently version 2.0a of 15 November 2002: http://www.xbrl.org/resourcecenter/specifications.asp?sid=22 ).

On this basis, XBRL taxonomies can be constructed to define standard document types for specific reporting purposes. There is, for example, an XBRL taxonomy for IAS reporting published by the International Accounting Standards Committee Foundation (http://www.xbrl.org/taxonomy/int/fr/ias/default.htm).

XBRL documents with actual data, called instance documents, are created or interpreted by software using XBRL taxonomies. XBRL software solutions are being offered by a growing number of companies, including Fujitsu, Hitachi, Microsoft, SAP, Software AG, and many others.

XBRL taxonomies are set up by XBRL jurisdictions and approved by XBRL International. National XBRL jurisdictions specify national XBRL taxonomies, like Generally Accepted Accounting Principles (GAAPs). There can be only one national XBRL jurisdiction in a country. There may also be trans-national XBRL jurisdictions for internationally recognised business reporting regimes, like the International Accounting Standards (IAS).

XML-enabled off-the-shelf software – ERP systems, databases, browsers, and so on – should be able to cope with XBRL documents.

Who supports XBRL?

The key to successful harmonisation is acceptance. XBRL will be widely used if it is accepted and implemented by the major players.

The XBRL consortium has currently (August 2003) more than 200 active member organisations around the world, including – just to name a few – Bank of America, Canada Customs and Revenue Agency, CBS Netherlands, Statistics Canada, Ernst & Young, IBM, KPMG, Microsoft, Oracle, Peoplesoft, Reuters, SAP, and many more.

Formally established XBRL jurisdictions include Australia, Canada, Germany, International Accounting Standards Board, Japan, Netherlands, New Zealand, United Kingdom and United States.

Provisional jurisdictions have been formed in Ireland, Korea, Singapore and Spain. XBRL interest and activities are increasing in Argentina, Austria, Brazil, China, Colombia, Denmark, Hong Kong, India, Italy, Mexico, South Africa, Sweden, Taiwan, and Venezuela.

 

How can XBRL be used for statistics?

Statistical agencies collect more and more data from more and more data sources, including businesses.

XBRL may play an important role in improving electronic data collection from enterprises (see, e.g., the proceedings of the joint UNECE/Eurostat work session on Electronic Data Reporting (EDR) which took place on February 2002: http://www.unece.org/stats/documents/2002.02.edr.htm ).

 

Statistics Canada and Statistics Australia are experiencing the use of XBRL for the collection of data from enterprises: see http://xpository.com/statscan/demo/index.html and http://www.unece.org/stats/documents/2002/02/edr/20.e.pdf.

 

Discussion topics

Official statisticians are now requested to answer the basic question:

Should XBRL be used for statistical purposes?

 

If the answer is positive, the following question is:

Should there be specific statistical taxonomies?

 

For statistical purposes, there are two options related to taxonomies:

  1. Existing XBRL taxonomies, like IAS, are exploited for statistical purposes, or
  2. the statistical community creates their own XBRL taxonomies for statistical reporting. In this case, a trans-national statistical jurisdiction could prove useful. Trans-national statistical taxonomies could improve the international comparability of statistical collections.

Both approaches have their pros and cons:

  1. Using existing taxonomies could result in incomplete data, as not all statistical needs may be covered.
  2. Creating specific statistical XBRL taxonomies bears the risk of non-acceptance.

 

A further question is then the following:

Is there a need for a trans-national statistical jurisdiction?

 

These issues require some in-depth reflection before decisions be taken. However, if the statistical community is hesitating too long, the train may be missed.

 

 

 

Links and references