Reversion To The Mean
The question had to do with “Reversion to the Mean” in investment settings, but, in fact, it extends to management of firms, in general. Simply put, it sought a defense of reversion to the mean as a reliable and actionable model. The answer, in my view, requires an understanding of physics, psychology, philosophy, economics, and management. Here is the original question and my response.
Stuart posed the question:
Would you like to give me a rationale for supporting reversion to the mean in the newspaper business?
I could find several decades of beautiful numbers for Knight-Ridder, Gannett and Washington Post.
I would more quickly advocate looking for disruptive events or trends that make reverting to the mean more difficult.
While not prepared to apply reversion to the mean to Knight-Ridder, Gannett, and the Washington Post, I feel comfortable defending it as an investment influence as follows:
Mark Twain asserted that, in life, there are but two certainties — death and taxes.
In all else, we must deal in probabilities and likelihoods. In fact, the reliance on the “gold standard” for medical science of a p-value of 0.05 or less is a recognition that some degree of uncertainty is rarely avoidable — even in laboratories and with the practical application of physics. A p-value of 0.05, for example, allows for 5 percent uncertainty and, conversely, 95 percent reliability.
When it comes to economics, that level of expected certainty declines, just as it does with the social sciences. In both cases, we are dealing the human animal and its psychology. P-values of 0.6 are often allowed and accepted with the social sciences (with multivariate regressions, for example), while the minimum standard in business decision-making is some figure besting 0.3 — with a preference for 0.1 or better. In other words, we would prefer to make decisions when confronting a 10 percent level of uncertainty or less.
With each, however, it is always the case that we are compelled to make a decision, regardless of the uncertainties. This was the advance authored by Blaise Pascal in “The Wager,” which introduced the concept of the null hypothesis. Even when electing to make no decision (to not take action), we have, in fact, made a decision and taken an action.
With reversion to the mean, it is not possible to assert that it applies in every case, with absolute reliability and certainty. The second principle of thermodynamics has it that everything tends to degrade over time. I know this anecdotally by standing in front of the mirror with my high school year book in hand, comparing my picture as a graduating senior with the real-time image of the 50-year-old fellow before me. Entropy is at play and is tangibly evident.
If this is the case, however, why have we emerged from the caves and advanced — developing art, music, parenting, medicine, etc. — and done so successfully? The answer comes us from complexity theory, Kuhn’s “Structure of Scientific Revolutions,” Jay Forrester’s Systems Dynamics, Peter Senge’s “Fifth Discipline,” and Robert Fritz’s application of physics to psychology. Each relies on feedback loops and our capacity to identify undesirable outcomes and, through trial and error, to seek their reform. Entropy is a linear degradation (we are older and incrementally less young with each passing day), while step-wise advances follow an outcomes delay — where the aggregated slope of the steps exceeds the force / slope of entropy.
This is why we have turn-arounds in business and why businesses fail. Most business failures occur in the first five years, when, due to uncertainty and chaos, entropy has the upper hand. Most public companies, however, are older than five years (more experienced) and better capitalized (with some combination of equity and debt). If their principle product is dieing, they can morph toward other products, services, etc. Turnarounds occur more reliably — aided by the debt forgiveness of first-round bankruptcy and the option of M&A. In short, all of the strategic management options become available to fully established and better-capitalized firms, while many of those options are not available to start-ups and those less-well capitalized.
This, of course, does not apply in every case, and some companies go out of business … never to return.
While this description provides the story behind the concept, it doesn’t support it quantitatively. For that you’ll need to watch Lew Sanders’ presentation before Bruce Greenwald’s Value Investing class at Columbia University — http://merlin.gsb.columbia.edu:8080/ramg…
