Black Swans (Nassim Taleb) are extreme events that have system wide consequences (9/11, the Internet, financial collapse, etc.). These events typically occur infrequently enough to escape near/mid term analytical methods (as in, most of our analysis is over time horizons that don't measure the occurrence or probability of these events). As a result, we think Black Swans are exceedingly rare (as in, this can only happen once in a thousand years). In contrast, Taleb's primary claim re: Black Swans is that they occur much more frequently than typical analysis indicates. Why? The distributions of systemic outcomes aren't normal/Gaussian, but rather fat tailed (see graph inset). This is a truism, Mandelbrot's work backs this up. This led to the aha! moment that sold 1.5 million copies of his book.
His and my second claim, and one that is harder to prove, is that there are reasons to believe that changes to the underlying global system have made these events much more likely (most worrisome, more prone to negative black swans).
A systems approach, often discussed on this blog and by Taleb, argues that a combination of hyper-efficiency (JIT supply, financial flows, and tight coupling of all sorts), extreme leverage (high debt), and excessive complexity (derivatives, interlocked supply chains) have introduced dynamic instability into the global economic system. That instability, if it isn't constantly corrected by an intelligent control system (if it is even possible to build one on a global scale), can lead to catastrophic failure. For example,
high performance aircraft are dynamically unstable by design. They require constant sub-second control system inputs to bring them back into line or they become smoking craters or lawn darts.
Another way to is to analyze the stability of our global system is to look at it as a network. "
Cascades of Failure and Extinction in Evolving Complex Systems" 2006, by Paul Ormerod and Rich Colbaugh is a good attempt at doing this. They used autonomous agents with heterogenous measures of fitness (diversity of goals/needs) to model an evolving network (similar to our economy or global trading/financial system). Unsurprisingly, they found as they scaled the connectivity of the network it became more fit. Simply, it allowed each of the agents in the system to better meet their goals for fitness (health, wealth, etc.). They also found that the network was also able to shrug off more small shocks than ever before (robustness increased).
However, this general improvement in fitness and resistance to minor shocks came at a cost. The more highly connected the network became, the more prone it was to catastrophic failures/extinction events (fragility increased). Worse, this relationship appears to be non-linear, in that with each increase in connectivity there is a accelerating increase in chance of catastrophic failure.
This is interesting from the perspective of societal decision making. Improvements in robustness and increasing returns over the short to medium terms, drives society (from common wisdom to academic doctrines) to assume that our situation will/can only get better and better. This experience has also led to the assumption that radical increases in connectivity (Internet, financial integration, etc.) will only accelerate this process of betterment. However, network analysis indicates that these massive increases in connectivity have made the possibility of an extinction event (cataclysmic Black Swan) much, much higher than in the past. Therefore, assumptions that we will eventually progress (something we have to assume), should be moderated by the fact that our future may be increasingly (more frequently) punctuated by periods of and extreme and prolonged system failure. In that regard, the connectivity we have in this Century makes it truly unique in all of history, and therefore any comparisons to historical experience should be used with extreme caution.
NOTE: Both Taleb and I conclude that the best way to deal with this risk is to simplify our global economic system. Taleb recommends that we
turn back the clock to a simpler time. In contrast, I recommend that we organically evolve into a more efficient decentralized system that concentrates on
community resilience.
Resilient communities involve intentional decentralization (lower connectivity or an ability to suffer a loss in connectivity without damage), which would radically reduce the chances that large network failures would result in cataclysmic black swans. We would be able to reboot quickly and with little long term damage.
Sem comentários:
Enviar um comentário