GuruFocus Premium Membership

Serving Intelligent Investors since 2004. Only 96 cents a day.

Free Trial

Free 7-day Trial
All Articles and Columns »

Hewlett-Packard and Regression to the Mean

March 10, 2013 | About:

Chandan Dubey

99 followers
When Hewlett-Packard was getting pummelled in November 2012, and deservedly so, I discussed the situation in the article Simon & Garfunkel and the Situation of Hewlett-Packard.

HP guided much below the analysts' estimates for Q1 2013 and then on November 20, it announced that:
"HP today announced a non-cash impairment charge of $8.8 billion related to Autonomy in the fourth quarter of its 2012 fiscal year. The majority of this impairment charge, more than $5 billion, is linked to serious accounting improprieties, misrepresentation and disclosure failures discovered by an internal investigation by HP and forensic review into Autonomy’s accounting practices prior to its acquisition by HP."
In response the stock hit its multi-year low of $11.35, valuing the company at $22 billion. Meanwhile, I made following trades:
20 Nov 2012: HPQ 18Jan14 $18 Call +200@0.4120 Nov 2012: HPQ 17Jan15 $10 Call +100@3.05
I was quite certain that the market was over-reacting and given time the stock will go up. Since then the stock has recovered quite nicely to $20.5, a nearly 80% increase in price.

This does not mean that I was astute or the turnaround of HP is on its way. This is what the statisticians call “regression to the mean.”

Regression to the mean is a fancy way to say that things get back to normal. Extreme circumstances do not last forever. This has several implications.

Suppose that in a given year an actor has done very well. All his movies have done good business and his acting was appreciated by critics and the viewers alike. He also got the Oscar for the best actor, and magazines ran cover stories on him. If one has to guess his situation for the next year, what would be the best way?

The success of the actor in a given year can be thought of as having two components: Ability and luck. It is not enough to have great acting prowess -- one also needs to be lucky. Luck is fickle by nature. If you are very lucky today, it is unlikely that you will be lucky tomorrow, too. So, the actor's standing in the next year is going to be (more likely) worse than this year.

Daniel Kahneman has a very good way to explain it. To arrive at a prediction of the actor's next year we start with:

Success = ability + luck


For the next year, we will start with the baseline success of the actor and then adjust the prediction according to the ability of the actor in question. We will give no weight to the luck because the actor cannot count on the luck being with him two years in a row.

This situation gives rise to something called “Sports Illustrated Jinx.” The jinx says that a player/team featured on the cover of "Sports Illustrated" is likely to have a disappointing year the following season.

But we know that a player/team feature on SI is likely having a great year in terms of luck and is on “top of its game.” And the most likely direction from “the top” is down.

I sold these options at a handsome profit.
Jan 3, 3013 HPQ $18Jan14 Call SLD 200@1.49 = profit $214

Mar 5, 2013 HPQ $10Jan15 Call SLD 100@10.5 = profit $745

About the author:

Chandan Dubey
I invest because I want to be free by the time I reach 40 years of age i.e., 2025. My investment style is to find a small number of bets with large margins of safety. I pay a lot of attention to management and their incentive. Ideally, I like to buy owner operator businesses. I am fortunate to have a strong inclination towards studying. I aid my financial understanding by extensive reading in psychology, economic, social sciences etc.

Rating: 4.0/5 (19 votes)

Comments

vgm
Vgm - 1 year ago
"Regression to the mean is a fancy way to say that things get back to normal."

Questions which arise from this are: What is "normal"? How long does it take to get there? And how sure can we be that things will "get back"?

The stock of companies like HPQ, DELL, RIMM (now BBRY), and NOK have been on a downward slope for the past 5 years. A couple have had a recent bounce off the bottom. I'm not sure we can talk meaningfully about their prospects to revert back to a "mean" or a "normal". How would we define it? Your postulate that HPQ mean-reverted is retrospective; and it's also speculative, since it could rise or fall substantially further.

I wonder if what we're really talking about here is valuation and Ben Graham's observation that 'In the short term the market is a voting machine, but in the long term it's a weighing machine.'

I read the HPQ article you wrote when the stock was getting hammered. You were confident that your analysis was valid and that HPQ was being dramatically undervalued by Mr Market (even as Seth Klarman ran for the exits!). You were courageous and contrarian - and arguably correct.

Your statement at the top seems to suggest that mean-reversion is a given. I'm not so sure. But I do feel sure that stocks eventually reflect their inherent value.

Thanks as ever for the stimulation.
cdubey
Cdubey premium member - 1 year ago
>> What is "normal"? How long does it take to get there? And how sure can we be that things will "get back"?

Very valid points. Hard to say what "normal" is. You can only say that at the point of maximum pessimism, things only get better. Do I know where the point of maximum pessimism is ? No. HP could have taken a nose dive from $11.42 too ... and this is why one should avoid unless one is willing to put more money in or take the loss. I was willing to buy more if it dropped.

I agree with rest of the points you make in the comment.

HP, DELL, RIMM, NOK all good examples. If one has done his research and knows the liquidation value ... then one can opportunistically buy when the stock drops 20% in a day (which happened to RIMM, NOK) and 10% for HP and DELL. Buying DELL at $8.9 was quite safe for me (and I did).

I don't know much about NOK and RIMM but I can understand if someone does a similar thing there. As long as you are comfortable in the positions you take.

Reversion to the mean does not apply always as you point out. It could be that the company is going to zero and in this case there is no reversion. It is always hard to say what the "norma" or the "fair value" is.

If the business model is not dying and the stock has fallen too far - it will get better. A good example was MCD a few months ago when it dropped to $85 and AAPL now. Although in case of AAPL, I am still not comfortable enough to take a position.

Thank you for the comment. It made me think of things I had not thought of while writing the article. You always come up with something penetrating to say. Congratulations for the keen eye !
superguru
Superguru - 1 year ago
Buying broken businesses or industry undergoing major change for cheap or good business for fair prices? eternal value investing debate.

May be a mix. Chou style. 70% good business + 30% basket of CRAP.

Please leave your comment:


Get WordPress Plugins for easy affiliate links on Stock Tickers and Guru Names | Earn affiliate commissions by embedding GuruFocus Charts
GuruFocus Affiliate Program: Earn up to $400 per referral. ( Learn More)
Free 7-day Trial
FEEDBACK