Hurricanes, Earthquakes and Financial Markets ... Part Two
Over the past few days, we have witnessed increased levels of market volatility and a historically overdue market correction. So in short, this event isn't all that surprising. Nonetheless, you'll see dramatic headlines running across your screens and so-called financial pundits trying to make sense of it. Along these lines, this is yet another reason why we discourage anyone from using the media as a source for investment guidance. But this week of volatility also is an opportunity to revisit why attempting to predict market performance is often an exercise in futility.
Consider the difference between hurricanes and earthquakes. Hurricanes are incredibly destructive, but because we have the ability to predict them well in advance, there is ample time for most people to evacuate. Earthquakes, on the other hand, are almost completely unpredictable and strike without warning.
Weather events can be forecast very accurately several days or, to some degree, even weeks in advance. This is because the state of the atmosphere is directly observable, and we have a wide range of sensors and imaging systems designed to provide near real-time data for the computer models that have been developed to incorporate it into a prediction algorithm. These models are constantly being tested and refined.
Earthquakes, however, occur when forces deep within the Earth’s crust build up over long periods of time, eventually resulting in the sudden slip of a fault plane and massive release of energy. Because it is virtually impossible to directly observe the stress that exists miles beneath the Earth’s surface, predicting when or where an earthquake is likely to strike is also almost impossible.
Financial markets behave more like earthquakes than hurricanes. While it’s possible to measure key ratios that indicate a relative market valuation, this has not proven to have much predictive power – high valuations can persist for long time periods, much like tectonic stress. What cannot be directly observed are the thought processes of market participants. We know that selloffs can be triggered by news events, but they are also a result of herding behavior and exactly what type or magnitude of the event is required as a catalyst is poorly understood. We can all point to events that seemed calamitous that were “shrugged off” by an ebullient market.
So, an interesting observation, perhaps – but, what’s the point? Simply that making predictions about financial markets is extremely difficult and all that this implies when considering strategies based on them. I’m certainly not the first to notice this and Nate Silver discusses some of the same points in The Signal and the Noise – Why so many predictions fail, but some don’t, an excellent book that we recommend reading. (Click here to see our recommended reading list.) Silver points out that markets are extremely complex systems, driven in the short-term by momentum, feedback loops, skewed incentives, herding behavior and interactions between all of the above. Complex systems can seem predictable in the long-term and yet highly unpredictable in the short term. “We have a very good idea of the long-run frequency of a magnitude 6.5 earthquake in Los Angeles. And, yet they are essentially unpredictable from day-to-day." Complex systems move quickly from orderly to chaotic and back again. As noted by many observers, from Eugene Fama to Nassim Nicholas Taleb in The Black Swan, stock price movements are characterized by very occasional and very large swings both up and down. Mathematically, the frequency of stock market crashes follows a power-law distribution. Not coincidentally, so does the frequency of earthquakes.