Measuring Accounting Disclosure Complexity with XBRL

by FEI Daily Staff

A recent study finds that the more eXtensible Business Reporting Language (XBRL) tags and the more extensions a filings has, the more complex the accounting and the lower its quality.


The study, by Rani Hoitash of Bentley University and Udi Hoitash of Northeastern University,  uses the number of XBRL tags and extensions to examine the complexity and quality of accounting information.

Highlighted by the U.S. Securities and Exchange Commission's Advisory Committee on Improvements to Financial Reporting (ACIFR, SEC [2008]) , understanding and measuring accounting disclosure complexity is important because increased accounting disclosure complexity can lead to poor financial reporting.

Using XBRL filings, the authors propose two measures of accounting disclosure complexity:

  • The first, based on the volume of accounting concepts, is measured using the number of US GAAP XBRL Taxonomy tags reported in company annual filings.
  • The second is based on the number of customized (extensions) XBRL tags that are created by companies when the Taxonomy does not include suitable tags to represent company-specific accounting concepts. The study relies on detailed XBRL data provided by Calcbench.
The authors concentrate on fiscal year 2012 because it is the first year most public companies were required to tag the “face” financial statements and the notes in Item 8 of the 10-K filings. Their sample includes 3,678 firms covering a wide range of US firms and is not subject to specific sample selection. On average, companies in the sample report 338 distinct tags , of which, on average 268 (79%) are Taxonomy tags and 69 (21%) tags are unique company-specific (extended) tags.

Looking at the industry level, they find that certain industries, such as Finance (Utilities) have more complex accounting disclosures with 327 (330) Taxonomy tags and 99 (112) extended tags. This is compared to less complex industries such as Healthcare (Wholesale-Retail) that use on average 204 (228) Taxonomy tags and 51 (48) extended tags. These descriptive results are interesting in their own right because measuring accounting disclosure complexity at the industry level is not well understood.

The study validates the two accounting disclosure complexity measures by associating them with a number of proxies for financial reporting quality and with audit effort/risk. They find that these two measures are associated with higher likelihood of issuing financial statement restatements, disclosing material weaknesses in the internal controls over financial reporting, and are positively associated with abnormal accruals and higher audit fees.

The proposed accounting disclosure complexity measures are distinct from proxies for operating complexity and financial reports readability, and exhibit more consistent association with measures of poor financial reporting quality.

Ariel Markelevich, Ph.D., CMA is an Associate Professor at the Sawyer School of Business in Suffolk University, located in Boston, MA. Dr. Markelevich is also the Chief Product Development Officer at Calcbench.