This is an early draft of a 250 essay. I’m posting it here so I don’t forget some of the ideas that were cut from the final draft.
Main concern: Are there two types of gray swans? 1) Conceivable, Unpredictable but forecast-able and 2) Conceivable, Unpredictable in the long-run but predictable in the (very) short term. Earthquakes are in the first category. Terrorist attacks are in the second category. 9.0 Earthquakes and 9/11 are both conceivable (and NOT unknown unknowns) but intelligence analysts had enough info to predict 9/11. Earthquakes are inherently unpredictable.
Terrorist attacks and earthquakes have a lot in common—at least to Aaron Clauset. In The Signal and the Noise, Nate Silver explains that according to Clauset, earthquakes and terrorist attacks both follow power-law distributions. If you don’t have a mathematics degree, that means that small earthquakes (0-4.9) are common, medium-sized earthquakes are less common (5.0-7.9), and large earthquakes (8.0 or greater) strike a few times a century. Terrorist attacks, in terms of total fatalities, behave similarly.
Of course, we can control terrorist attacks, and that’s one reason why they are different than earthquakes. However, I’d rather be a seismologist than an intelligence analyst. While there are only a few different types of earthquakes, there are thousands of ways a terrorist could strike. The challenge for the intelligence analyst is to scrutinize a large number of hypotheses—hijackings, chemical attacks, bombings, shooting sprees.
National security disasters in the United States are often the result not of a dearth of information but imagination. That is, no one considered the correct hypotheses. For instance, at Pearl Harbor analysts never reviewed the possibility of a full blow aerial attack even though evidence suggested that it was possible—even likely. According to the 9/11 Commission, we had enough information to anticipate September 11th, but the C.I.A and F.B.I never seriously examined the idea that terrorists would hijack four planes and crash them into high-profile buildings. As the Nobel-Prize winning economist Thomas Schelling put it, we tend “to confuse the unfamiliar with the improbable.”
What can earthquakes and terrorist attacks teach us about business? More than you think. Technology also follows power-law distributions. In history, a few revolutionary inventions (the wheel, the printing press, the steam engine) are surrounded by thousands of less groundbreaking inventions (thermometers, zippers, CD-Rom), which are surrounded by millions of small inventions (snuggies and post-it notes). Revolutionary inventions disproportionally affect the economy in the same way that large earthquakes and terrorist attacks inflict a disproportionate amount of damage. The wheel was a 9.0 magnitude invention.
Earthquakes are not predictable. Are inventions like the wheel predictable like 9/11 was? Yes and no. We cannot predict inventions (or terrorist attacks) that will occur in the distinct future. But we might be able to predict an invention that will emerge within the next 2-5 years (if there is one). For instance, all of the components of the iPod existed for a few years before Steve Jobs fit them into your pocket. According to Scott E. Page, all of the components of the steam engine existed for two decades before the steam engine. There was a small window in which someone had enough information to predict the iPod and the Steam Engine.
But remember the lesson from Pearl Harbor and 9/11. They were failures of the imagination and not failures of information. “The signals just weren’t consistent with our familiar hypotheses about how terrorists behaved, and they went in one ear and out the other without our really registering them,” says Silver.