Ken Fisher's "Debunkery"

50 "myths" (or "bits of bunk") exposed and debunked in 4-5 pages each -- I suspect that's a good format for many readers with short attention spans or busy lifes full of interruption, though I personally like good old-fashioned reading and in-depth analysis and explanation, so I found myself somewhat annoyed by the glib, facile, over-simplified tone of each short "bunk". However, while I do have many objections to the format, exposition, and style, somewhat surprisingly to me (while reading critically and in a somewhat annoyed/peeved mindstate from the above issues;-), I found myself more or less in substantial agreement with 49 of the 50 points (far more than I would have expected a priori) though not necessarily with the slant the author gives to his "debunking" explanation.

Let me give an example of me disagreeing with the slant...: Fisher's debunking of beta's mythical status (bubk #19, "Beta measures risk") entirely focuses on the fact that it's backwards-looking, because, he spouts, "past performance is never indicative of future results" -- now that is unadulterated bunk. (Or would Mr. Fisher like to get me to perform surgery next time he needs some? I have no "past performance" while he could get a surgeon with extremely good past performance, but if he truly believed what he so stupidly spouts, that should count for nothing at all with him... harder to think of a more stupid attitude to life: past performance is no guarantee of future results, but to jump from "no guarantee" to "never indicative of" is, truly, totally absurd).

No, the #1 reason to debunk beta is what he himself erroneously states (100% wrong, but similar to what most people just as wrongly believe -- I'm still angry with Fisher because I'm sure he knows better, and just lets "glib and facile" overwhelm "correct and precise"!): that an equity with a beta "Lower [[than 1]] ... was less volatile than the market" (!). False! Bunk!

What β<1 means is that the covariance of that equity with the market is less than the market's variance: in other words, that the equity's volatility (which per se may be high, low, or middling) tends to be less correlated with the market's volatility, than the market itself. You can perfectly have an equity with extremely high volatility and a very low beta (even negative!) -- all it takes is for that equity's peaks and troughs to tend to coincide more with the overall market's throughs and peaks rather than viceversa.

Adding some of that equity if the rest of your portfolio is "the whole market" (e.g. via a S&P 500 index, as that's usually what beta's computed with respect to) will lower your total portfolio's volatility... in as much as the correlation of that equity with the market doesn't change drastically in the future. But unless you bother to compute beta with respect to your specific portfolio (and hardly anybody ever does), even that is not much use.

Other perfectly valid (and important) objections to beta as a measure of risk include the fact that it's based on variance -- which treats the "risk" of outperforming the market exactly as severely as the risk of underperforming it, at total variance with the psychological impact of "risk" and the way everybody uses the word in question in real life (including investing); semivariance, alas much harder to treat mathematically!, while still imperfect, would be much closer to people's understanding of, and psychological reaction to, "risk". And many others... but focusing exclusively on the "backwards looking" aspect, and capping that with that absurd proclamation claiming the past has nothing at all to teach us (as if half of the book itself wasn't about looking at historical information...!!!) is really objectionable.

Of course, the one point of the 50 I think is total bunk is the one ferociously attacking covered call writing, where not only is the argument objectionable, but the implied conclusion is totally absurd and unwarranted. But, I guess that will have to be explained in some future post!