Please Note: Blog posts are not selected, edited or screened by Seeking Alpha editors.

IBM Accounting: Why I Think It's OK

|Includes: International Business Machines Corporation (IBM)

So IBM disclosed the SEC is investigating their revenue recognition - again. The first episode was closed with no action. The stock tanked and is now recovering some of the lost ground. I'm pretty sure IBM's accounting is fundamentally sound, for the following reason.

Back in 2013 I decided to test their accounting for revenue recognition, and settled on an analysis of where else they could park a billing/receipt, if not in revenue. The main place would be deferred revenue. So I compared IBM to ORCL, SAP and HPQ on the ratio of deferred revenue to total assets.

At the time, IBM was highest on this metric, and a reversion to the mean would have increased reported revenue and earnings. The other companies were lower and a reversion would have decreased reported revenue and earnings for ORCL. For HPQ it would have been a small increase. SAP didn't have enough deferred revenue to be relevant.

I also looked at deferred revenues (non-current), which had been steadily increasing as a percentage of total assets.

My conclusion at the time was they might be "sandbagging;" that is, holding something in reserve so they could tap it to make the next quarter's earnings if it became necessary.

So if they have since stopped sandbagging, big deal.

I worked for an outfit once where the head guy in accounting was into sandbagging. I called him (not to his face) the sandman. Anyway, he instructed me orally to do a Journal Entry that was egregious sandbagging. I didn't argue: I made notes of the meeting, and did the entry.

In due course the auditor questioned the entry and I produced the notes. I was doodling at the time, and for whatever reason had a picture of a boomerang on the notes. The auditor discussed it with my boss, in a way that stung and hurt his ego.

In due course I found it expedient to seek other employment.

Disclosure: I am/we are long IBM.