The Demon we did NOT design

A picture of the internet 

I have re-read A Demon of our Own Design by Rick Bookstaber, which is doing the econo- and market blog rounds at the moment, and while a second reading is an honour given only to books which I find particularly interesting, the reason I found it so interesting is because I realised I disagreed with it profoundly. SoI had to blog it before I retire to spend the rest of the month in my gite in the ex-cathar sarf of France and later, a crayfish fuelled sojourn in the Swedish Archipelago. Strictly no blogging during this period. Well, maybe a bit in Sweden. As an aside, I have to point out that here I have one over the Dealmaker, who for all his familiarity with Saint-Tropez seems to summer in er, Winnipeg. Perhaps the deals are thin on the ground at the moment.

What’s wrong with what Bookstaber has to say?

First off, he makes the point that markets are doomed to periodic crises as financial innovation outpaces our understanding and our ability to adjust. Crises are getting worse and more common. Well, what if they are not more common? Bookstaber merely asserts this; it is not backed up with evidence of the period before Bookstaber’s career. We have just lived through the Great Moderation where volatility has come down to unprecedented levels (wiping out Taleb among others). Now it seems to be going up again. To be sure, in just the past 18 months, we have had 3 mini “crises”, May 06, Feb-March 2007, and of course the present difficulties. But I cannot even remember what caused the first two; on a long term chart you would hardly notice them, if at all as mere bumps. I hope the “subprime crisis” will end the same way (I took off the last of my hedges today and am betting on a bounce). In Bookstaber’s much longer career we have had 1987, 1998 and 2000-2002. But is a crisis every 5-6 years so very bad? Were the 1980s, 1970s and 1960s so devoid of incident? I suspect not. Periodic “Sterling crises” would have certainly had UK investors very excited in that period.

Later on in the book, after an extended semi-biographical semi historical discussion, Bookstaber gets to his key thesis: markets are both “tightly coupled” and “interactively complex”, and as such susceptible to “normal accidents”. He uses the examples of a nuclear accidents, Space Shuttle disasters, and an airplane crash. Tightly coupled, complex systems can be sent out of kilter by one or two small breaks in the chain. The plumbing in Chernobyl stops working during a test, making it so that processes further down the chain break down, pressure builds up, and a devastating explosion is the result. Small changes, imperceptible at the time, can have major ramifications later on, he says, citing the archetypal (and rather tired) chaos theory example of the fluttering butterfly in Peru causing the hurricane in the Caribbean. We build in safety procedures, but they can make things worse by adding to the complexity. Bookstaber writes:

Risk controls, putting on layers of regulation and organisational oversight cannot always fix the problems that arise from the complexity and  tight coupling we have designed into the markets. Indeed, it might just make matters worse. . . a better approach for regulation is to reduce the complexity in the first place, rather than try to control it after the fact.

This is my problem: Bookstaber’s examples are most unsuitable; they are LINEAR processes and they are DESIGNED. We design, build and switch on Chernobyl because we want to have cheap power; we design, build and fly space shuttles because we like to fly people into space with an exciting frisson of danger; we do the same with airliners to travel distances quickly. We have not designed financial markets; they are classic spontaneous orders. We have ineffectually tried to regulate them, every now and then try and corner them, but generally we do not rule them; rather they rule us. They are not linear systems, the means of achieving single goals. I can think of at least 3, almost certainly more, reasons to participate in markets: investors and speculators buy, sell and hold paper to build capital; brokers provide liquidity for a fee, and companies buy and sell paper to raise capital. While a nuclear power station has a single “end ” then, the market simply does not. The ends or aims of the players may be in conflict; the owners of the shares I buy in an IPO want to sell at a high price, I want to buy at a lower one. I have likened the operation of markets to the action of a vast computer, computing nothing but itself. This puts the gains and losses individuals or funds make in the market in context. We, and Bookstaber, think of the market as “functioning” when it makes us money; when it runs in an orderly way, when it “goes up”. In reality it will be “functioning” just as well as it drops like a stone as it did this morning, wipes us out and blows us up. The market in this sense is like Spinoza’s god; it is pitiless, it does not care about us. Calling the subprime dislocation or a 10% drop in the equity market a “normal accident” makes no sense to me; it is normal functioning, it is prices clearing, it is the wiping out of excess, it is the passing on of information, it is the reallocation of resource in favour of the most efficient.

My final bone to pick is that Bookstaber says price movements, at least in the short term, are primarily driven by liquidity rather than information flow. Hence he makes a convincing case that most of the crises we have seen in the past 20 years (1987, 1990, 1998) have been liquidity driven. Now, Ciena reports on the 30th of August, while I am sunning myself on a Baltic island. The report will in all likelihood be excellent, far above the artificially low consensus, and the stock will probably go up 10%. Does Bookstaber really want me to think this movement will be liquidity-, rather than information-driven? It strikes me the two work in tandem to drive prices, the one being the context for the other; new information drives the demand for stock up, supply is dependent on the liquidity, or willingness to sell/provide stock at the current price. Picking on liquidity drivers and ignoring the informational strikes me as unbalanced.

To be fair to Bookstaber, he is a clearly intelligent observer, writing from the perspective of a risk-manager, almost a regulator. This colours both his subject matter and his understanding of the market. I see things from the perspective of a fund manager; unlike him, I do not devise or supervise new products or new markets. I don’t know how to fix things when they go wrong, I try and understand what has so I can profit from it or preserve my capital. His conclusions mostly concern regulation. But almost all the major crises he refers to in markets in the past 20 years were not fixed by regulation, they were “fixed” by printing vast amounts of money and shovelling it into the financial system, by simply feeding the beast. I do not understand exactly how Bookstaber can hope to “reduce the complexity” in financial systems in an effective way through regulation, and I don’t know that what would happen if he did would be good.

We have the impression that we have created markets as they currently function. But we did not. We cannot properly fix what we don’t understand, and what we did not create ourselves. Sophisticated markets have been around for hundreds of years. I do not see how a more spontaneous, complex system, so long as it is not a linear, single function system, need be less stable than another. I still think it might be more stable, containing as it might a series of built-in redundancies and “loose linkages”. The periodic adjustments needed to keep the system computing might seem traumatic, “perfect storms” to the participants, but, really does the system itself break? I think not.