"CAPM is ‘C.R.A.P.’"

So says this post in AllAboutAlpha which cites a report by James Montier, a Kleinwort Benson economist, which in turn cites a good deal of impressive academic work from folks like Eugene Fama (a University of Chicago professor and one of the luminaries of modern finance).

Now why is this important? Well, much of how asset pricing is done depends on CAPM, which is short for “capital asset pricing model.” CAPM posits that investments with greater risk must give a greater return to induce investors to fund them. CAPM requires analysts to determine the beta of each investment, which is the correlation of its returns with market returns.

CAPM-based models are used throughout Corporate America. Just about every capital budgeting proposal and M&A analysis is based on a discounted cash flow model. The discount rate used is chosen to reflect the project’s required expected market rate of return. Companies often do that by finding out what implicit or explicit discount rate securities analysts are using for companies in lines of business similar to the proposed investment, and when appropriate adding a premium to reflect the fact that the investment is not a liquid, tradeable stock.

Now if CAPM isn’t valid, then this methodology isn’t either. Of course, one can argue that if discounted cash flow models can still provide a good relative ranking of projects, it’s still a good tool. But without going into why it isn’t as simple as all that, the DCF approach says that all projects with a positive present value should be pursued. If we have serious theoretical problems with CAPM, this decision rule fails as well.

As anyone who has worked with discount models can tell you, they have some pretty serious practical problems. The financial forecasts are prepared to sell the deal or project, rather than present one’s best guess as to how things will play out. All models use a terminal value, which is when the project goes into a steady state (the theory says when it stops growing faster than the economy, but in practice, the terminal value is determined in year 5-7, unless you are looking at an infrastructure project, which has a longer forecast horizon. The use of a terminal value multiple assumes that the cash flows will continue in perpetuity, but if they use a higher- than-GDP-growth multiple, that ain’t valid, because in the long run, nothing grows faster than the GDP). The net present value calculation depends heavily if not entirely, on the terminal year cash flow and multiple assumptions. In other words, it’s arbitrary.

So it’s amusing to learn that not only does DCF modeling not work all that well in practice, but it may not work in theory either:

John Mauldin is a big fan of James Montier, the 34 year old Dresdner Kleinwort economist who literally wrote the book on behavioral finance….

In this comprehensive yet very readable article Montier rails against the CAPM, calling it “(C)ompletely (R)edundant (A)sset (P)ricing”. While his facts are irrefutable, their interpretation leaves room for disagreement. In other words, the exact fecal count of CAPM remains unknown. But whether you agree with his conclusion or not, you will probably find this essay an interesting read.

Montier points to Fama & French’s 2004 research showing that high beta stocks performed no better than low beta stocks since 1928. Indeed, the abstract to that study says:

Unfortunately, the empirical record of the model is poor – poor enough to invalidate the way it is used in applications…the failure of the CAPM in empirical tests implies that most applications of the model are invalid.

In that study, Fama & French goes on to describe two anti-CAPM camps: the “behavioral irrational pricing” camp (which believes prices aren’t rational at all) and the “rational risk story” camp (which aims to find a better CAPM mousetrap):

Among those who conclude that the empirical failures of the CAPM are fatal, two stories emerge. On one side are the behavioralists. Their view is based on evidence that stocks with high ratios of book value to price are typically firms that have fallen on bad times, while low B/M is associated with growth firms… The behavioralists argue that sorting firms on book-to-market ratios exposes investor overreaction to good and bad times… When the overreaction is eventually corrected, the result is high returns for value stocks and low returns for growth stocks.”

The second story for the empirical contradictions of the CAPM is that they point to the need for a more complicated asset pricing model. The CAPM is based on many unrealistic assumptions. For example…it is reasonable that investors also care about how their portfolio return covaries with labor income and future investment opportunities, so a portfolio’s return variance misses important dimensions of risk. If so, market beta is not a complete description of an asset’s risk, and we should not be surprised to find that differences in expected return are not completely explained by differences in beta. In this view, the search should turn to asset pricing models that do a better job explaining average returns.

Naturally, Fama & French fall into the second camp, the “better mousetrap” camp. But they also concede that CAPM may never be proven or disproven since the “market portfolio” is impossible to define (see related posting on Rob Arnott’s Fundamental Indexation)….