The financial crisis: complexity, chaos and the Singularity
November 5, 2008
Not long ago, I was looking for a new job. A job in London, in an investment bank. Never a job hunt was more ill timed than mine.
The financial crisis hit. The very bank I wanted to work for got nationalized. My job opportunity evaporated. That’s a sad outcome for me, but not a catastrophe. I try to never put all my eggs in a single basket. I was building alternatives right while I was preparing for the interview, to have a better negotiating position and skew the ZOPA toward my side.
The financial crisis is however too interesting not to comment. It is the first one partially caused by machines operating autonomously, whose predictions and decisions were beyond the understanding of those very humans who made them and were meant to control them.
It will not be the last one. The financial one is only a hint of the Singularity meta-crisis that awaits us. As progress and innovation speeds up and our predictability horizon shortens, we are bound to fall more and more to judgement errors caused by Taleb distributions: situations in which there is a high probability of small gains and a low probability of huge losses. The final outcome is often catastrophic, but it is masked by a stream of low risk and steady returns. It is quite difficult to stop betting on the wrong side, even if you know it’s wrong, when you’ll keep getting rewarded to do it. You always hope to be able to detect the change in winds before the rest of the herd and reap the ultimate reward.
Financial crisis and chaos are connected. Trade in derivates and exotics boomed in the last years, making the financial world much more interconnected that it was before. This linkedness, pitched as a tool to hedge risks, made the system more stiff, able to transmit gains deeper down the chain of economic players. It did wonders for looses too. It turned the financial market into a chaotic battleground.
Chaos is the sensitivity to initial conditions. “Does the flap of a butterfly’s wings in Brazil set off a tornado in Texas?” once wrote a mathematician. We’ll have to replace it with “does a defaulted homeowner in Texas unable to pay his subprime mortgage set off a financial avalanche that burns 30 billions USD in Brazil (and several trillions worldwide)?”. It just happened.
[...] days ago, I wrote about our shrinking prediction horizon — what some people call The Singularity — and its role in the current financial [...]
wow R. I had fun reading this.
Although I don’t do economy lectures (can’t! tried), I will say something about this grim future you’re foreseeing.
Chaotic systems can only fluctuate. Eventually they reach a stable state. Without a stable state there is no pattern and no chaos.
And I don’t believe that singularity or any ‘final end of something’, can even exist.
If a “defaulted homeowner in Texas” buys or sells stocks depending on her mood, then our only issue here is keeping her happy.
relax KL., remeber the stock market is all in our heads, same as everything else.
I let some months pass, to see if I was right or wrong to expect a grim future. I don’t know yet. Doesn’t it look gloomier now than in November?
Let me answer your other points.
Chaos and markets. I don’t think that the market or the economic are — as a whole — chaotic. I do think that the global financial markets become too connected, too “stiff” in the last five years. A local crisis or a boom reverberates quickly on the whole system. The financial instruments that contribute to the stiffness are also some of the most leveraged ones, they act as amplifiers or — in the case of derivatives through netting — chains of amplifiers. Instead of reducing risks, like they were meant to be when they were created, they multiplied them and spread them around.
If the system doesn’t have parts to dampen these crisis and boom waves, it becomes fragile and small events can bring down the whole thing. That’s what happened the with commercial paper market, Lehman-Brothers and AIG. That’s what is still going on as the crisis waves ripple through the global economy.
That markets do act somewhat like chaotic systems when they are in a stable condition, that is when it’s business as usual. The problem is when a Taleb’s black swan shows up, when an highly improbable event with game-changing effects happen. The more efficient the system is in spreading gains and dividends, the more efficient it will be to spread havoc — over a certain threeshold, it will kick the system out of its chaotic equilibrium toward something else… it might be another attractor, if there is one that is reachable; or it will be a systemic crash, if there is none.
(Side note: I recently discovered Hector Sabelli’s work on biotic systems. They may be better tools than chaos to model markets. I don’t know how markets behave well enough and I haven’t yet fully understood Sabelli’s theories.)
I never implied that a defaulted homeowner in Texas buying or selling stock created the crisis. My point is that the system amplified crises, every kind of them. We couldn’t know beforehand where the problems would start, but we knew that some local crisis would escalate sooner or later. That’s what I meant by chaos.
The Singularity. I think you got me completely wrong on it. The singularity is not an event, it’s not a final end. It’s a period of accelerating changes, that will push our human abilities to predict change to their cognitive limits. We will rely more and more on computers to do the data analysis for us, but they can only provide information not wisdom.
We will be more and more fallible to black swans, to game-changing events, as the innovation game speeds up and more people are able to play.