Who is using XBRL?
by Mike Willis, PricewaterhouseCoopers, LLP
'Who is using XBRL?' is a very common question. To answer the question concisely; allot of people and many of them may not even know it as they are just using data and the source and/or format is largely transparent ant thereby irrelevant to them. However, the question of 'Who is Using XBRL' may not really be the proper question (What is using XBRL?); but before we go there; here are some examples of current users:
>EDGAR-Online's IMetrix platform
http://www.edgar-online.com/OnlineProducts/IMetrixProfessional.aspx enables collaboration by retail analysts on modeling concepts. XBRL not only standardizes the data but it also standardizes formulas or collections of formulas (e.g. models). Last time I asked, over 50k retail analysts were using IMetrix and the XBRL it consumes and the XBRL formulas it enables.
>Morgan Stanley's Modelware platform
http://www.morganstanleyiq.ch/DE/binaer_view.asp?BinaerNr=141 does the same thing as IMetrix and is used by Morgan Stanley analysts around the world since it was launched over 4 years ago. The problems analysts have around analysis is: 1. timely, complete, accurate, relevant data and 2. the usefulness of the model or formulas to analyze the data. XBRL helps with both the data and the formulas.
>PwC's iDP platform on KCurve
http://idp.pwcinternal.com does the same thing as IMetrix and Modelware and we have 65k monthly internal users.
>There are also a very broad range of data aggregators who are consuming (directly and indirectly) the freely available SEC EDGAR RSS XBRL Data feeds
http://www.sec.gov/spotlight/xbrl/filings-and-feeds.shtml
>The SEC Staff are using IMetrix and other XBRL enabled tools to analyze companies and enable collaboration across analyst groups and individuals within the Commission.
>Some companies with built-in implementations who are consuming XBRL from the EDGAR RSS Feeds directly into their reporting process for peer group risk assessments and benchmarking of both numeric and narrative concepts. An example is available via this Webinar: "Automating the Last Mile: The Role of XBRL in Streamlining Financial Reporting"
https://www302.livemeeting.com/cc/pwclivemeetingroom/view?id=4FQZP6 Just enter your name.
>There are banks (ABN AMRO and Rabobank) using XBRL for credit risk assessments.
http://xbrlplanet.org/wp/?page_id=395
>There are institutional analysts not only using XBRL but blogging about the quality of the company reports; here is one example: a blog by an institutional analyst who suggests just that companies should pay close attention to their use of company specific extensions:
http://institutionalrisk.blogspot.com/2011/02/xbrl-usability-part-2-checking.html
>Here are more generic examples of XBRL use:
http://xbrl.us/research/pages/data.htm
There is more to this than 'Who is using XBRL'........ as the more relevant question may be 'What is using XBRL?'. As part of the background, it may be useful to read: "Algorithms Take Control of Wall Street"
http://www.wired.com/magazine/2010/12/ff_ai_flashtrading/all/1 on the current level of automated processes found in today's trading environment. It is very interesting and thought provoking article and a few comments may bet your my attention include:
* "But many of the professional investors subscribing to Lexicon aren’t human - they’re algorithms"
* "They just want data - the hard, actionable information"
In forming the question: 'Who is using XBRL?' is there an inherent conclusion often reached by executives asking:
* if anyone has accessed the company XBRL instance on the company website; and/or
* analysts if they are using XBRL; and/or
* corporate webmaster about the traffic hits on the company XBRL instance on the company website.
Should financial executives have an expectation that:
* they can detect an application automatically consuming the XBRL instances from the SEC RSS feeds; and/or
* analysts actually know where the data comes from that populates their models; and/or
* human beings will surf company websites looking to download XBRL instances (a machine readable format) rather than automatically accessing the XBRL instances via RSS or WebService from EDGAR.
A more appropriate question may possibly be something along the lines of: 'WHAT is using XBRL?'
* What types of applications are accessing and automatically reusing the XBRL instances? (e.g. Modeling applications)
* How should companies expose their XBRL reports to maximize the capabilities of consuming applications? (e.g. Website posting and/or RSS Feed?)
Given the level of automation outlined in the article above; should companies be paying more attention to the quality of their XBRL reports? The institutional analyst blog provided above suggests just that:
http://institutionalrisk.blogspot.com/2011/02/xbrl-usability-part-2-checking.html
In considering this topic, here are a couple of questions for companies to ask their analysts on this topic:
- Do you know how the data gets into your analytical models? If the answer is no; then they may not know if they are using XBRL or not. If the answer is yes; ask them how the data gets from company report into their analytical models?
- Have you ever compared the disclosures in your analytical models with those that in our company reports? [[This question will reveal if analysts are using data aggregators to populate their models. Most analysts models are populated by data purchased from data aggregators. Their parsing processes typically include an approximately 25% error rate for financial statement table disclosures and a much higher omission rate for note disclosures. You can prove this out by going to any of the major search sites (Google Finance, MSN, Yahoo Finance) who also use data aggregator data and compare their data to what is actually on company reports. Best to do this with an adult beverage nearby as many of the company specific disclosures are either normalized or omitted.]]
- How quickly are your models updated for our company disclosures? If the answer is anything other than immediately, then they are using data aggregator data which typically takes hours or days to reach their models. If may be useful to suggest that they subscribe to the EDGAR RSS feeds for your company reported disclosures.
- How do you incorporate company press release information within your analytical models? It may be useful to consider using the XBRL Corporate Actions Taxonomy to structure these disclosures so that they can be accurately, completely and automatically consumed by analyst models.
Hope that these are useful.
