Friday, September 4, 2009
Why is risk SO hard?
You might begin by asking "who said risk is hard?" Well, for starters, I did! And why would I say such a horrid thing?
Let's begin with a simple question: what is risk? That is, how would YOU define risk? The answers tend to vary, but include "volatility," "uncertainty," and "potential loss." Volatility, as a definition, dates back to the late 1950s, if not earlier. Uncertainty was taken up as long ago as the early part of the 20th century. Most dictionaries define risk as the possibility of losing something. The most commonly cited definition today, however, relates to the inability to meet an objective. Space doesn't permit me to go into much more detail on these alternative definitions right now, but we may take this topic up at greater length in our newsletter. But you should be able to right away see that we have at least four very different views on what risk is! How's that for a way to make this topic complicated before we've even gotten into the measurements?
Okay, so let's discuss the measurements for a bit. Most measure volatility, which appears to be the most criticized definition. Let's say that you agree that the inability to meet your objective is the best definition -- okay, so how does standard deviation, tracking error, or beta relate to this definition? To paraphrase the CIO of Yale University's endowment, David Swensen, quantitative measures of risk have a lot to be desired.
Let's consider one of the assumptions that many risk measures have: that returns are normally distributed. There's been countless research efforts done that have confirmed that this isn't the case; what does that therefore say about these measures?
And next, consider the issue of ex ante risk ... that is, forward looking risk. Most such measures base their predictions on what has occurred in the past. But depending on how you view the past and how far back you go, you can get very different results. And, who's to say that the past is a good predictor of the future? What in our past would have predicted the most recent shock we had to the market and economy?
Therefore, I stand by my original statement that risk is hard. Quite a challenge. But, a critically important part of our business. In spite of its difficulties, we have to do the best we can to get our arms around the risks that have been taken as well as the risks we're taking now. A single risk measure won't do: we need as many as we can get to try to understand the risks that are being taken.