Skip to content
All library documents

Detecting and Handling Errors in SEC Company Filings

Article Quant Q&A · Author: mbmast

Summary

The document raises a data-quality problem for researchers using company share-count fields from SEC filings. It describes Apple filings around a stock split where reported common shares outstanding appear far below the scale of nearby reports for a period, with later filings reflecting the split. The example highlights how an erroneous reported value can persist in a dataset when subsequent filings do not promptly correct it.

The author asks how such errors should be handled and whether a more accurate source exists, but the document contains no answers or proposed validation method. Its practical lesson is that regulatory filings are primary sources, yet extracted fields still require checks for scale, corporate actions, filing dates, and consistency across periods. The example is specific to one issuer and reporting episode, so it does not establish how common these errors are or which alternative source is preferable. Researchers would need to verify the underlying filings and document any adjustments before using affected observations.

Key ideas

  • SEC filings can contain reported values that appear inconsistent with surrounding periods.
  • Stock splits complicate comparisons of shares outstanding across filing dates.
  • Researchers should validate extracted fields against filing context and corporate actions.
  • The example poses a data-cleaning question but does not supply a definitive correction procedure or alternative source.

Tags

Full text
# How to Handle Error in SEC Filings


# How to Handle Error in SEC Filings












Sometimes I see obvious errors in the 10-K or 10-Q filings that appear to go uncorrected in subsequent filings for the same period.

For example, on June 9, 2014 Apple, Inc. did a 7-to-1 stock split. If you look at all of the `CommonStockSharesOutstanding` tag field values in all of the 10-K or 10-Q filings leading up to (but not including) the split, you'll see numbers around 940,000,000. But then things get strange.

On April 24, 2014 (before the split) there are two 10-Q filings (with different DDATEs and the same ADSH, 0001193125-14-157311) that report `CommonStockSharesOutstanding` of 889,213 (wrong by a factor of about 1,000). There is no subsequent filing to correct this for 3 months until two July 23, 2014 filings (with different DDATEs and the same ADSH, 0001193125-14-277160) which reports `CommonStockSharesOutstanding` of 6,294,494,000 and 5,989,171,000, respectively, reflecting the June 9 7-to-1 split.

Thus, for three months (April 24, 2014 to July 23, 2014) the value of `CommonStockSharesOutstanding` was around 900,000 instead of the approximately correct value of 6,300,000,000 up to the split date.

How should these errors, and specifically this type of error, be handled? This data comes straight from the SEC so I don't think there's a more accurate source. Is there? What to do...?

Shown in full with attribution under the source's licence. Licence: CC BY-SA 4.0 (Stack Exchange)

This summary was written by Stratmill's research agent from the original; it is not a copy of the source.