RECENT POSTS
Monday, January 21, 2019
Differences in Earnings Releases and 10-Ks

Wednesday, January 16, 2019
The Importance of Textual Analysis

Tuesday, January 8, 2019
A Look at Climate Change Disclosures

Wednesday, January 2, 2019
Quants: Point-in-Time Data for Backtesting

Friday, December 28, 2018
Now Showing: Controls & Procedures

Thursday, December 27, 2018
A Reminder on Non-GAAP Reporting Rules

Monday, December 17, 2018
Researching PG&E’s Wildfire Risk

Wednesday, December 12, 2018
Tracking Brexit Disclosures

Thursday, December 6, 2018
Campbell Soup: Looking Behind the Label

Sunday, December 2, 2018
SEC Comment Letters: The Amazon Example

Wednesday, November 28, 2018
Measuring Big Pharma’s Chemical Dependency

Monday, November 26, 2018
Analysts, Can You Relate? A True Story

Monday, November 19, 2018
Digging Up Historical Trend Data: Quest Example

Sunday, November 11, 2018
Cost of Revenue, SG&A: Q3 Update

Monday, November 5, 2018
Lease Accounting: FedEx vs. UPS

Saturday, November 3, 2018
New Email Alerting Powers

Wednesday, October 31, 2018
PTC and Two Tales of Revenue

Tuesday, October 30, 2018
10-K/Q Section Text Change Detection

Sunday, October 28, 2018
Finding Purchase Price Allocation

Sunday, October 21, 2018
Charting Netflix Growth in Three Ways

Archive  |  Search:

We all know that bad data in means bad data out.  For that reason, since our founding, Calcbench has had an unwavering commitment to enhancing the quality of XBRL data, so much so that we’ve built it into both our platform and our processes. This way we ensure that we don’t pass any XBRL data quality problems on to our users.

As part of this dedication to data quality, we recently set out to quantify all of the errors we’ve identified and corrected in XBRL filings filed with the US Securities and Exchange Commission (SEC) over the past several years. Through this study, we examined several types of mistakes most likely to throw off an end users’ analysis, including errors in the Document and Entity Information (DEI), scale errors, and sign switches. (Please note, these are NOT the only types of errors a filer can make!)  What we found was somewhat encouraging: over time, companies seem to be getting better at catching common errors before they file. Yet, clearly a lot more can be done.  And we remain committed to providing as much feedback as possible back to corporate reporting professionals about their own filings, so they can learn from past filing errors (something we do via our complimentary Filer Portal service).

We know that XBRL data holds significant benefits, including easy access to significant amounts of valuable data.  However, our findings suggest that most XBRL filings contain errors, most of which are relatively easy to correct and may be the result of a learning curve by both filers and third parties creating the XBRL filings. Nevertheless, the prevalence of errors suggests that filers should pay more attention to their filings as these errors can expose filers to further actions from the SEC and to potential litigation from financial statement users.

You can read a more comprehensive overview of our findings in our recent report. Some of the key insights include:

  • The most frequent errors are sign switches (about 50% of filings), followed by scale errors (about 8% to 12% of filings) and DEI errors (about 3% of filings).
  • Approximately 3% of all filings have DEI errors. These errors seem to be consistent over time, and we only observe a slight decrease in recent filings.
  • The rate of scale errors is significant (up to 12% of filings in Q4 of 2012) but it is decreasing. It seems that scale errors are made by both small and large filers and appear in both the face financials and the notes to the financials.
  • Most scale errors occur in tags associated with shares. However, a significant number of scale errors are made in tags like Revenues, Net Income, and Total Assets, which are frequently consumed by users of financial data. Scale errors are relatively easy to detect and correct.
  • Sign switches do not seem to have decreased over time. About half of all filings have at least one sign switch. Sign switches seem to have been more prominent with larger filers earlier, but the smaller filers are the ones more likely to have them in later years.

To learn more about the current state of XBRL and read more of our findings, we encourage you to read our report, “What Filers Should Know about the Quality of XBRL Filings”.


FREE Calcbench Premium
Two Week Trial

Research Financial & Accounting Data Like Never Before. More features and try our Excel add-in. Sign up now to try the Premium Suite.