Sunday, November 16, 2008

Black Swans Flying Over the Financial Crisis

That the shadow of complexity darkens the prognosis of the current financial crisis is a point to which I have repeatedly alluded. A subset of this general difficulty is the likelihood of a black swan (see Nassim Nicholas Taleb, The Black Swan) dramatically altering all bets.

A totally unrelated black swan (landing of Martians or the "Big One" sending Los Angeles to the bottom of the Pacific) could occur, but I wish to address black swans about which some productive predictive assessment might be made, i.e., Black Swans provoked by current human behavior. The general theoretical assumption is: to the degree that we, by our actions, choose to push society closer to the edge of chaos, we set ourselves up for a surprise with which we will not be able to cope. More precisely, we do things that 1) raise the likelihood of a surprise and 2) decrease our ability to cope with it.

What actions might increase the likelihood of a black swan?


  • Antagonizing opponents might provoke them to do something highly risky out of anger or desperation.
  • Greed carried to the extreme of, say, leveraging investments way past a reasonable level of risk, or corruption that, say, hides the fact that required collateral is not being maintained might transform a minor glitch into a real black swan event, e.g., transform the normal diurnal variations in the market into a cascade of failure.


What actions might decrease our ability to cope with a surprise?


  • Exhausting all readily available resources supporting a failing system (e.g., bailouts to corrupt or irresponsible financial institutions or to outmoded corporations or other narrowly-defined special interests) might both fail to resolve the financial crisis by evading fundamentals and leave us with a lowered ability to combat a new challenge.
  • Committing all military forces to current operations on the assumption that no new enemy will emerge could find us dangerously overcommitted.
  • Encouraging people to continue living the lifestyle that got us in trouble in the first place, on the assumption that traditional crisis-management techniques will enable us to get out of the problem we are in, might prove to be dangerously shortsighted.

These are examples of living at the edge of chaos (i.e., maximizing effectiveness of current performance by choosing to perform to the max, spending nothing on redundancy, keeping nothing in reserve). We should not expect our current predicament to be immune to the impact of our previous behavior. To phrase it differently, chickens coming home to roost may attract black swans.

It may be worth noting that the above discussion implicitly treated black swans as negative. That of course need not be the case. Something wonderful might fall in our laps: say, the sudden discovery of a cheap, clean way of generating solar power or the arrival of Martians bearing gifts. I would just make two points: first, no need to try to forecast such a miracle. If it occurs, fine; problem solved. Second, don't bet your mortgage on it, because we are queering the system by the mistakes in crisis management that we are making, lessening the already small probability of a free lunch.

The general problem is that choosing to live at the edge of chaos (running full speed along the edge of a cliff where the view is great, on the assumption that you will not stumble) is dangerous because if you stumble, the consequences are dire. The specific black swan problem is that even if you do not stumble, something might take you by surprise, and on the edge of the cliff running at top speed, responding effectively will be tricky. The whole problem of how crisis management might be impacted by a surprise that hits us "out of the blue" is a big and neglected issue. These are just a few thoughts to provoke contemplation.

Might we perform crisis management in a way that would reduce the likelihood of a black swan?

Might we perform crisis management in a way that would enhance our ability to deal with one?

No comments: