Efficiency-Fragility Fallacy
Taleb argues that visible efficiency gains can conceal larger losses when optimization makes a system fragile. A cost-benefit comparison should consider exposure to errors and disruptions as well as routine operating costs.
THE IDEA
WHERE THE “EFFICIENT” IS NOT EFFICIENT
We can easily see the costs of fragility swelling in front of us, visible to the naked eye. Global disaster costs are today more than three times what they were in the 1980s, adjusting for inflation. The effect, noted a while ago by the visionary researcher on extreme events Daniel Zajdenweber, seems to be accelerating. The economy can get more and more “efficient,” but fragility is causing the costs of errors to be higher.
The stock exchanges have converted from “open outcry” where wild traders face each other, yelling and screaming as in a souk, then go drink together. Traders were replaced by computers, for very small visible benefits and massively large risks. Taleb contrasts dispersed trader errors with failures that propagate through computerized systems. He cites the 2010 market disruption (the “flash crash”); for August 2012, he reports Knight Capital Group losses of $10 million a minute and a total of $480 million.
And naive cost-benefit analyses can be a bit harmful, an effect that of course swells with size. For instance, the French have in the past focused on nuclear energy as it seemed “clean” and cheap. And “optimal” on a computer screen. Then, after the wake-up call of the Fukushima disaster of 2011, they realized that they needed additional safety features and scrambled to add them, at any cost. In a way this is similar to the squeeze I mentioned earlier: they are forced to invest, regardless of price. Such additional expense was not part of the cost-benefit analysis that went into the initial decision and looked good on a computer screen. So when deciding on one source of fuel against another, or similar comparisons, we do not realize that model error may hit one side more than the other.
When to use it
Evaluating investment strategies that emphasize high returns but ignore tail risks.
Designing software systems that prioritize speed over fault tolerance.
Policy decisions on energy sources that consider only immediate cost without accounting for potential catastrophic failures.
Personal productivity systems that optimize for every minute but leave no buffer for emergencies.
FROM THEORY TO EVERYDAY LIFE
In practice
Flash Crash of 2010
Taleb uses the 2010 flash crash to illustrate the possibility that failures can propagate through computerized markets rather than remain confined to individual traders.
Knight Capital Group Loss (2012)
Taleb’s account reports that Knight Capital Group’s system malfunction in August 2012 caused losses of $10 million a minute, totaling $480 million.
French Nuclear Energy Decision
Taleb uses additional French nuclear-safety spending after Fukushima to argue that an initial cost comparison can omit later costs. The example concerns unequal exposure to model error across the alternatives being compared.
A MOMENT TO REFLECT
Put it to work.
How might a system you rely on be efficient but fragile?
When evaluating a cost-benefit analysis, what hidden fragilities might be ignored?
Think of a past decision that prioritized efficiency over robustness. What were the consequences?
Often confused with
References
- Research on extreme events — Daniel Zajdenweber
CONTINUE THE CONVERSATION
Source texts
- Antifragile (Nassim Nicholas Taleb)