I get an email from Seeking Alpha pretty frequently with links to the text of earnings announcement calls. Sometimes I like to click on a few, and do a control+f for the word (of course) “tax”. I like to see what companies are saying about this most important of subjects. On my most recent iteration of this exercise, I noticed something interesting. Dell notes in their recent earnings announcement conference call that “We expect our non-GAAP tax rate to be 16% plus or minus 100 basis points.” Abercrombie & Fitch notes “Third quarter adjusted operating expenses of $493 million and an effective tax rate in the mid- to upper 20s.”
In just casually listening to this, which company seems more sure about their tax future? I may just be easily fooled, but to me, the computer-like precision of Dell’s “plus or minus 100 basis points” really feels like they have a better handle on the tax future than Abercrombie’s casual, laid back, “mid- to upper 20s”. It feels like Dell has a detailed calculation with a well-thought out confidence interval underlying their statement, whereas Abercrombie is just throwing out a wild guess after consulting a Magic 8-Ball.
Thank goodness for math. My instantaneous reaction was incorrect. For Dell, the tax rate could be 15, 16, or 17%. For Abercrombie, I would interpret what they say as a possible tax rate could be anywhere from 24% to 29%. Standardizing the two estimates by taking the span of the range divided by the midpoint of the range, Dell’s range is 18.75% of the base (3%/16%), while Abercrombie’s is 18.86% (5%/26.5%). Pretty much exactly the same.
I have no idea if I am alone in my perceptions of the precision of the language used above by the two companies, or if these perceptions have any consequences. I don’t know if companies try to actively manage these perceptions in the words they use. It seems like if so, empirical testing would be able to detect that. But I do think it is interesting how my instantaneous reaction about the two estimate’s relative precision was so different than what the math actually indicates.
