If women ran the world we wouldn’t have wars, just intense negotiations every 28 days. –Robin Williams

Return Measures by Damordaran 2007 (More on ROIC)

**REGRESSION TO THE MEAN**

This is a key concept to learn along with EV/EBITDA, MCX, and cheapness wins.

REGRESSION TO THE MEAN A good read by an Australian Graham & Dodd-like Investor.

When an investor turns to the research on *regression to the mean* and investors overreacting to poor company performance/bad news in Richard Thaler research, he or she sees that prices of the winner and loser portfolios take three-to-seven years to revert. See also *The New Finance: The Case Against Efficient Markets* by Robert A. Haugen and *Inefficient Markets* by Andrei Schleifer.

Illustration by S of Reversion to the Mean

We next progress to *Chapter 5: A Clockwork Market, Mean Reversion and the Wheel of Fortune* in * Deep Value*.

From there we will read chapters 3 and 4 in **Quantitative Value.**

The subjects of ROIC and reversion to the mean remind me of one of the quibbles I have with Toby’s chapter 4.

To recap, Toby presents his statistical analysis of the Greenblatt “magic formula” approach and Toby then sets out evidence that the ROIC component does seem to add much to the results. As to the latter, Toby’s statistics show that using a naive selection model based on EV/EBIT alone produces results which are marginally better than his version of the “magic formula”. The chapter then goes on to explain the success of the EV/EBIT criterion by reference to the published research relating to reversion of the mean, etc.

There are a few things that bother about Toby’s thesis. To summarise:

1. It doesn’t make sense to me that the success of the EV/EBIT criterion alone is referable to the reversion to the mean phenomenon. You see, I have observed many situations in which companies suffer setbacks which require them to go through various degrees of restructuring before they can get back to their prior levels of profitability. Those sorts of changes typically involve a fairly lengthy process, often spanning several years. The problem here is that Toby’s back-testing on the EV/EBIT criterion assumes ONLY a 12 months holding period. I haven’t expressed this very well. Essentially, what I’m try to say is that Toby’s back-testing is based on a such a very short holding period that I don’t think there is enough time for the typical corporate turnaround to make much progress, so I have some doubts that his results can be explained by reversion to the mean.

2. Toby’s figures suggest that EV/EBIT works only slightly better than his version of the “magic formula”. The point that he seems to gloss over is that the “magic formula” would usually select companies which have a fairly good ROIC to begin with. Those companies would normally be scoring an ROIC better than the average company. Putting aside the back-testing for a minute and just focusing on the logic, I think it is hard to accept that the success of the “magic formula” is attributable to a company’s recent poor results improving to former levels, when the “magic formula” tends to select high performing companies to begin with. By the same token, since the EV/EBIT criterion produces results that are roughly similar to those of the “magic formula”, I just don’t see how Toby can say that it is reversion to the mean which explains the success of the EV/EBIT selection criterion.

I can’t prove this but I suspect that the main driver behind the success of the “magic formula” or EV/EBIT alone is that they identify stocks which have been subject to “overshooting” as Haugen would call it (or “reflexivity” as Soros grandly titles the effect).

I may be completely wrong on the above. Happy for anyone to set me straight !!!

Cheers,