The recently announced error in Bank of America’s computation of their regulatory capital requirement provided the financial press with a typical “sensation of the moment”.  In searching for ways to discuss the error some commentators even characterized it as an “accounting error”, the kind of thing sure to get a bit of sensational notice.

However, the nature of this error is a bit more complex, and has different lessons to teach.  Here is an excerpt from the B of A press release:

“Bank of America Corporation today announced a downward revision to the company’s previously disclosed regulatory capital amounts and ratios due to an incorrect adjustment related to the treatment of certain structured notes assumed in the Merrill Lynch & Co., Inc. acquisition in 2009. The reduction in the regulatory capital amounts and ratios has no impact on the company’s historical consolidated financial statements or shareholders’ equity, which were properly stated in accordance with accounting principles generally accepted in the United States of America (GAAP).”

(See more at: http://investor.bankofamerica.com/phoenix.zhtml?c=71595&p=irol-newsArticle&ID=1922893&highlight=)

What really happened here was an error in a disclosure that is not from the accounting system per se, the computation of regulatory capital.  Regulatory capital, a key indicator of financial health and strength, starts from financial statement information, but is actually a very complex numerical exercise that goes well beyond the financial statements.  It is an important disclosure for financial institutions, and in fact is disclosed in the notes to the financial statements.  But it was in all the complex and fancy computations outside the financial statements that the error occurred.

So, the lesson here is not really about the accounting for the financial statements, it is about what sort of care, and more importantly what sort of controls are necessary for the information disclosed in the notes that is not lifted directly from the accounting system, but rather that starts with information from the accounting system.

This is a far more complex issue, and one that has tripped up other companies in the past.

Two notable examples:

As far back as 1995, Fidelity had a major error in a spreadsheet that caused them to announce that their Magellan fund would have to pay a taxable dividend.  When the error was caught, they had to deal with the embarrassment of having to correct their earlier announcement.  The error?  A wrong sign in a computation in a spreadsheet.  You can google search for articles and also check out:

http://catless.ncl.ac.uk/Risks/16.72.html

In 2005 SunTrust Bank dealt with a very similar situation when a math error in a spreadsheet used to compute part of the allowance for loan losses came to light.  This one was more serious, as the financial statements were misstated and a restatement was required.  In fact, three people lost their jobs in this messy spreadsheet error case!

So, what is the lesson?  It is to take care to design and test controls over information that goes beyond the financials and review all that information very, very carefully!