We have posted an article by Nassim Taleb about “Black Swans” in the past. David Ignatius of the Washington Post connects the dots to Long Term Capital Management:
Taleb’s black swan concept and the LTCM collapse are two expressions of the same underlying truth: the events that matter most in financial markets are precisely the events that standard models assign the lowest probability to. LTCM’s models were built on historical data and assigned small probabilities to scenarios outside the historical range. The 1998 Russian debt default and subsequent global flight to quality was outside that range. The model said it was nearly impossible. It happened. And because LTCM was leveraged to an extreme degree on the assumption that the impossible was indeed impossible, the fund was destroyed.
Taleb’s framework explains why this pattern is not just the LTCM story. It is the story of every financial model that assumes the distribution of future events will resemble the distribution of historical events. Black swans, by definition, are outside the historical sample. They cannot be assigned correct probabilities from historical data because they have not occurred yet. Any model that relies on historical frequency to estimate future probability systematically underestimates the probability of events it has not seen before. The model says the black swan is impossible. The black swan arrives.
The practical question for traders is what to do with this knowledge. Taleb’s answer, to be long optionality and short fragility, is one response. Trend following’s answer is structurally different but compatible: do not model the probability of black swans at all. Instead, build a system that responds to whatever price does when the black swan arrives. When a black swan event produces a large, sustained price movement in any direction, the trend following system enters in the direction of that movement and follows it for as long as it continues. The system does not need to have anticipated the event. It needs only to be positioned to respond to the price movement the event produces.
LTCM was destroyed by a black swan. The traders on the other side of LTCM’s forced liquidation captured one of the largest trend following opportunities in years. The same event that destroyed a model-dependent convergent strategy created the price movements that a reactive trend following strategy captured. Both outcomes from the same event. The difference is entirely in the approach: one tried to predict and was wrong in a catastrophic way, the other responded to what happened and was right in a profitable way.
For more on this connection, see the LTCM page, the Gladwell on Taleb page, and the broader discussion of fat tails and outlier events in trend following.
Frequently Asked Questions
What is a black swan in Nassim Taleb’s framework?
A black swan is a high-impact event that was outside the range of historical experience and therefore could not be assigned a meaningful probability by models built on historical data. Black swans are not merely rare events. They are events that the models in use at the time treated as essentially impossible. Their occurrence invalidates the models that dismissed them and produces outcomes far larger in magnitude than those models could have predicted.
How does the LTCM collapse illustrate the black swan problem?
LTCM’s models assigned very low probability to the simultaneous adverse movement of all its positions that occurred in 1998. The models were based on historical correlations that did not include a scenario like the Russian debt default followed by a global flight to quality. When that scenario occurred, LTCM’s models failed and its leveraged positions produced catastrophic losses. The black swan was precisely the event the model could not see.
How does trend following respond to black swan events?
By responding to price rather than predicting events. When a black swan produces a large, sustained price movement, the trend following system enters in the direction of that movement based on its entry rules, which are triggered by current price behavior rather than by any prediction of the event. The system does not need to have anticipated the black swan. It needs only to respond correctly when the black swan’s price effects materialize. This reactive stance is the structural opposite of model-dependent convergent strategies that require the future to resemble the past.
Trend Following Systems
Want to learn more and start trading trend following systems? Start here.
