May 20, 2009

Financial and Aviation Disasters: A Comparison

While reading Richard Bookstaber's memoir, the compulsively readable A Demon of Our Own Design, I started thinking about his point that engineering has improved safety in every field except in finance.

Before I go any further, I think there is a problem with this statement in that calling the invention of new trading vehicles and trading methodologies "engineering" is more a marketing ploy than anything else. Financial engineers are not really trying to design a machine, they're just trying to make money. If those products and strategies happen to bring a social good, great, but that's not really the goal.

However, for the sake of argument, let's pretend that a trading operation is comparable to an airplane, say. Then a quant trader would be the equivalent of a test pilot who would also be the engineer in charge of designing and improving airplanes. There no denying the fact that flying has become safer and safer whereas running a trading operation has become more and more accident prone with worse and worse consequences when accidents do happen. The aeronautic equivalent of what has been going on in the financial world would be a situation where ever bigger, faster and more complex planes crash more and more often and when they crash, they do so over more and more populated areas and therefore where fatalities per crash are ballooning. That's clearly not what is going on in the aviation industry. Why is that? What accounts for the drastic difference in the safety profile of flying and designing an aircraft on the one hand and managing a trading operation on the other?

I guess the first difference between the fields is that there is no such thing as a single individual responsible for designing an aircraft, testing it in flight, then improving on its design and finally flying it on a regular basis, with passengers . There might have been such individuals in the aviation industry's infancy (think the Wright brothers) but nowadays these are totally separate functions taking place in totally separate organizations. By contrast, in trading, the birth of quant traders in the 1980s has meant that the same person (or team) usually designs the product or strategy, tests it theoretically and in the markets, and then manages the trading book on a continuous basis. Not only is the trader in charge of the whole process involving a given financial product but he or she is usually one of very few individuals in the firm who really understand the product and its market. 

One could also say a big difference between the two fields is that the primary goal of aircraft engineering is to improve safety whereas the primary objective of trading is to improve returns. But that's not exactly true. In theory, aircraft engineering's primary goal is to produce better planes that fly faster, carry more people on the non-negotiable condition that it also improves (or at the very minimum does not decrease) safety.  Any plane that decreases safety even an iota, no matter how incredible its other attributes, will not make the cut (let's not mention corruption and forged tests, that's a whole other subject). In finance, theoretically, the main players try to maximize returns for a given risk. But in practice, what we often see are traders and organizations that repeatedly ignore, misprice and/or neglect the risk aspect. Why is that?

I guess one reason is that the built-in incentives for traders on the one hand and for pilots and aircraft engineers on the other differ wildly. Should the plane crash, its pilot usually dies, the engineer's reputation is shot forever and the aircraft maker most probably goes bankrupt. Should the trading book implode, the trader might have to leave his or her firm but will most likely be very much in demand at other firms (see Meriwether, John). His or her reputation will not really suffer. Even if the firm goes belly up, the trader won't. In fact, the trader doesn't even have to return his or her previous bonuses. Whatever he or she made during the fat years is his or hers for the keeping, even if it turns out the fat years were an illusion because the models were wrong and the risks severely underpriced. 

Bookstaber actually brings up the airline industry and more specifically the 1996 ValuJet crash in the Everglades as a case study to look, not for differences, but for similarities to financial disasters. He makes a crucial point when he observes that "flying an airplane is by nature fraught with tightly coupled processes. You can't pause the flight in midair to do some reengineering if something goes wrong." "Tight coupling means that components of a process are critically interdependent; they are linked with little room for error or time for recalibration or adjustment".

As in finance, when an accident occurs in the aviation industry (or, in Bookstaber's other example, in the nuclear power industry), it always seems, ex post facto, that the precise sequence of events that led to disaster had an infinitesimal probability of happening (100-year floods anyone?) and, if not for the worst luck, should not have happened. But the reality is that in such highly complex, tightly coupled environments, there is almost an infinite number of potential sequential failures. In other words, things can go wrong in a million different ways, each very unikely. As Bookstaber says, "the probability of any one event is small enough to be dismissed, but together, with so many possible permutations and with combinations beyond comprehension, the odds of one or the other happening are high". Moreover, perversely, these types of accidents are "borne of complexity, so adding safety checks to try to overcome these accidents can be counterproductive, because they add to this complexity"

Obviously I haven't even begun to cover the subject and I'm sure I overlooked a hundred pertinent points of comparison between financial and aviation disasters. Come to think of it, this kind of ties in with a point Nassim Taleb has been making incessantly, that the financial system, as it has evolved, inevitably leads to recurrent catastrophes because it is inherently vulnerable to black swans. 

1 comment:

Dennis The Menace said...

I would like to comment about the current condition of the american economy. This is one of the very hardest periods in the countries history. Their does not seem to be any consensus about the trends for the economy one week the economic news is good the next week its bad. Their seems to be no consistency what so ever when it comes to economic matters. Mcdonald’s recently hired fifty or sixty thousand people out of one million that applied maybe mcdonald’s should change their saying you deserve a break today at mcdonald’s to you deserve a job today at mcdonald’s. As far as those banksters go I say lets exchange those three piece suits and briefcases for a pick' a shovel' a bucket' and some pinstripes. Inflation Is the primary reason for much of the growing income inequality between the rich and the poor. It is also I believe the cause of the decline of the middle class. When ever the employing class and by Employing class I Mean anyone or any company That hires personal And gives them a regular paycheck. Their is always a tendency to undercompensate your personal less and less over time simply because when prices rise wages generally lag increases in prices at least for a substantial portion of the working population. Workers do not have much ability to control their wages and benifits. But companies that employ personal have much to say about the wages and benifits that their employees receive. Companies have been undercompensating their personal for decades in an attempt to increase their bottom lines. They have been systematically undercompensating their personal less than the increase in prices on purpose. The result is many workers have little income left over for any purpose other than basic needs food' rent' necessary clothing' utilities' medical bills' Its no wonder that the economy is in such serious trouble.