what my intuition tells me now: Statistics Are Not the Truth, Just a Mathematical MethodPosted by Jason Apollo Voss on Apr 5, 2011 in Best of the Blog, Blog | Comments Off
One of my continual beefs with our Western World and my criticisms of capitalism is our overemphasis on numerical and statistical methods in analysis. In fact, these two forms of analysis are so dominate that the word analysis has become a synonym for numerical and statistical methods.
In conducting analysis as an investor for almost twenty years I can tell you that analysis requires equal parts left brain and right brain. This is the very reason that I wrote my book The Intuitive Investor: A Radical Guide for Manifesting Wealth. Namely, I recognized that there is no such thing as a future fact, yet investing unfolds in the future. Statistics, by definition, is using the past to make some probabilistic statement about the future. But probabilities are far from reality. Oops.
To be clear, the problem is not with numerical and statistical methods, it is with their blind application in the absence of reason. Such a bias treats statistics as truth, as opposed to just a mathematical method for describing reality. Mathematics is just another language used to describe a deeper, more complex reality.
Do any of us confuse the words “you are a valued employee” (the form) for the actions, such as increased pay, benefits and responsibility, that make such a statement real (the substance)? Yet, that is precisely what happens with numerical and statistical method: form is mistaken as substance.
Now thankfully the United States Supreme Court agrees with me, as do a growing number of statisticians themselves.
Investors of Matrixx Initiatives, a pharmaceutical company, sued the business because it did not disclose to shareholders that its over-the-counter drug, Zicam Cold Remedy, caused a loss of sense of smell.
Matrixx argued that the reason it did not disclose the reports was that the side effects being reported were not “statistically significant.” Eventually the Food and Drug Administration warned consumers to not use Zicam.
Financial losses from that were great enough to warrant a law suit from shareholders. In a ruling last week the U.S. Supreme Court rejected the argument of Matrixx Initiatives and is now allowing the lawsuit to proceed in a lower court.
Critical to this ruling was the testimony of many statisticians who themselves have been warning for years about the limitations of numerical and statistical analysis. Sadly it has taken a case where people’s quality of life – Matrixx’s customers’ sense of smell and Matrixx’s investors’ wallets – is on the line in order to highlight the limitations of numerical and statistical methods.
Numerical and statistical methods are used to make many decisions for decision makers. But numerical and statistical methods are just one of many decision making tools that form a complete decision making tool kit. For example, application of experience, wisdom and intuition are other modes of analysis.
I once heard a speech given by former Treasury Secretary and Goldman Sachs risk assessor, Robert Rubin. At the time of the speech he was serving as co-Chairman of Citigroup. And at that very moment the (ultimately disastrous) AOL-Time Warner deal had been announced that very day. It was the height of the dot.com era’s madness, and Citigroup had been the investment banker behind the deal.
In the midst of worldwide media and investor attention what did Robert Rubin choose to speak about to a group of the world’s money management community hungry for “deal of the century” news? The limitations of statistical methods to fully model, and therefore properly describe, risk.
Robert Rubin pointed out that whatever statistical model/probability distribution curve you used never had fat enough tails. This is a fancy way of saying that risks are always higher than you think they are. This is a fancy way of saying that statistical significance is not the only way to measure risk.
Mr. Rubin said that in his capacity as Goldman Sachs’ executive in charge of risk assessment that he came up with a simple heuristic that was just as powerful as any model. That heuristic? “By definition the greatest risk is the one you didn’t account for.”
As I pointed out in November, solely analytical methods are doomed to failure. At the heart of my argument is that human beings invented mathematics. Therefore, the human mind is greater than mathematics and therefore, the entire mind needs to be used when making important decisions.
I continue to be a supporter and user of numerical and analytical methods applied with consciousness. That is, with cognizance of their strengths and their weaknesses.
So the next time you receive a call from your broker and there is discussion of “momentum,” “support at this level,” “reversion to the mean,” or any other statistical hokum, I grant you license to use your full mind and reason, and not just the narrow, limited context of numerical/statistical methods that so many investors seem to prefer. In short, use your head!