Hello Ellen and Hannes,

You are thinking of risk only within a MPT framework, which is what this article is about (breaking free of that context), and this article fits within an entire series. Variation around a mean is not risk, it just isn't. It was borrowed mathematically by Markowitz so he could author his thesis, and fast forward 70 years later and we are still using it as a flawed measure. Someone in this thread said that Markowitz realized the error of his ways and now uses a measure that is commensurate with loss only, and not gain, too.

If you would like to intimately understand the issue then calculate standard deviation by hand, which I am guessing you have either never done, or have not done since you studied statistics. What you will discover is that variance, and its digits-copasetic twin, standard deviation, are weighted averages, not just averages. The result is that significant numbers above the mean (i.e. in our example, an active manager that kicks ass) look riskier because of the weight they receive. This makes no sense. How is outperformance, or even performance above a risk-free rate of return, or a required rate of return counted as risk? The insurance industry knows a thing or two about risk since they have been underwriting them for 500 years, and they define risk as the chance of loss, not as variation around a mean.

My own preference for risk is to evaluate risk the way business people do - and I am not referring to the Mandarins of finance or banking. Here I am talking about the seemingly magical ability (if you are a lover of MPT) of business people to anticipate risks to their business models and build in to the business plan appropriate responses to anticipated risks, as well as their ability to respond to risks if they occur. Most militaries around the world have scenario plans for different events that may never unfold, but if they do they have an appropriate response. Not only that, but they alter force structure so that they can execute if one of the scenarios occurs. Why is it that when lives are the line, the risk of death or being conquered, militaries do not use standard deviation when assessing risk? In each of the example I have provided it is not any different than in finance. Human actors who have preferred outcomes recognize that the future may not unfold as they might like and they build a response to those possible non-preferred outcomes. No one except Markowitz and those who believe in MPT use standard deviation as their sole description of risk.

Yours, in service,

Jason