31st August 1997

Money: a Telling Lore

Money isn't real. It may be a very important means whereby we maintain our sense of identity, but it isn't real. It's simply a set of ideas some of which we share with other people and some which are our own. We might both agree that this piece of metal is a pound coin, but is it a lot or a little to pay for a pint of milk? Value, like beauty, is in the eye of the beholder.

So too are risk and confidence, those ideas on which the market turns. I don't risk a pound on a lottery ticket because I think 1 in 14 million to be poor odds, but many people don't. Confidence is the opposite of fear, and the greatest fear we can know is that of a threat to our sense of identity. If we want to understand why the market does what it does we need to understand ourselves, that is, our own private logic.

It we want to understand why the market does what it does we need to understand ourselves, that is, our own private logic.

Nothing pleases me more than when I find some scientific research which supports my theories, especially my theory that to understand why people behave as they do we need to discover each person's private logic which is, in effect, their sense of identity. Research reported by the Guardian recently did just that. Here researchers used artificial life methods to create computer software working as robotraders to examine classic market theory.

Classic market theory or the competitive equilibrium model assumes that buyers and sellers behave rationally and thus, in a free market, produce an equilibrium between supply and demand. In real life markets might produce something like an equilibrium, but they also produce much volatility, bubbles and crashes, events which classic theory finds difficult to explain.

Economists, like psychologists, have always been strong on theory. Again, like psychologists, they have steadfastly ignored what real people actually do, because real people ruin theories. Some people might do what the theory predicts, but many won't. Instead of asking why this happens, economists and psychologists, right from the beginning of their disciplines in the nineteenth century, have seen their task as being that of scientists, of uncovering the orthodox truths which are immanent in nature. In the search for such universal laws, what individuals think is irrelevant. For psychologists no word implied such complete rejection and scorn as subjective. The ideal of science towards which they strove was physics, but in such striving they managed to overlook Heisenberg who, in describing his uncertainty principle in 1927, said that the experimenter was always part of the experiment.

What this means in terms of theory is that the particular theory which a person creates or espouses may or may not reflect the real world, but it will always reflect the individual personal meanings created by those who hold it. In psychology and economics theories are never proved or disproved. They simply come into and fall out of fashion. Freud's theories used to be all the rage while Jung's were seen as the product of an unworldly nutter. Now Freud is deplored or ignored while Jung is extremely popular. In economics Maynard Keynes went out of fashion and Milton Friedman came in, but now Milton is on the wane and Maynard is becoming increasingly fashionable.

By ignoring the essence of how we live our lives - the personal meanings we each create - economists and psychologists have developed theories about the world and the people in it from which all sense of personal experience has been removed. Psychologists have developed models which represent us as being nothing but puppets manipulated by the strings of behaviourist stimulus-response, or of genes, or of unconscious drives. Economists, wrote Paul Ormerod, an economist who is extremely critical of classical economics, see the world as a machine. A very complicated one, but nevertheless a machine, whose workings can be understood by putting together carefully and meticulously its component parts. The behaviour of the system as a whole can be deduced from a simple aggregation of these components. A lever pulled in a certain part of a machine with a certain strength will have regular and predictable outcomes elsewhere in the machine. The Science Museum now houses a model of such a machine made by Bill Phillips, an engineer turned economist. This wonderful contraption has levers to pull, buttons to press, sluice gates and liquids of different colours which, representing economic forces, would rush around the system's tubes. Now, alas, the liquids have been removed and only the skeleton remains.

Such economic machines were not assumed to be thinking. Computers might not think for themselves, but they can contain software which simulates thought. This is what the robotrader computers did, or, at least, these were models of how their makers, the computer programmers, thought people thought. They gave their robotraders the ability to adapt to market conditions by using learning algorithms which allow them to try out different buying and selling strategies to optimise their portfolios. They each reacted individually to events using the trading rules with which they'd been programmed. These rules corresponded to what we can call shared, public logic. However, these robotraders also had a kind of private logic.

We all share a public logic, but each of us has our own private logic which is the set of conclusions which we have drawn from our own individual experience. No two people ever have the same experience so why no two people ever see anything in exactly the same way. The robotraders' programmers created a kind of private logic for their robotraders by running a genetic algorithm, a technique which creates a kind of Darwinian evolution in the software, cross-breeding sections of the code to create new strings, some of which then made changes in the external logic rules. We human beings do this all the time. Your mother might have laid down the rule, On your way to school cross the road only at pedestrian crossings. You then change that rule to, Use a pedestrian crossing only when Mother can see you. Your changed rule is part of your private logic.

We use our private logic all the time. These robotraders could do so only when the researchers used the genetic algorithm. The researchers ran their experiment 25 times, involving each time some 250,000 trades. The results were clear. Under certain conditions the efficient market of classic theory emerged. Under other conditions the market which emerged was just like real life, lots of activity, lots of changes in prices, prices bubbling up and then crashing down, the kind of market which pension fund managers and market watchers find so hard to predict. And what were these different conditions? The classic market emerged when the robotraders were using only their external logic, but when the genetic algorithms were used frequently this private logic produced a market very close to real life.

This is a simple scientific replication of what goes on all the time in real life. In dealing with money, as in dealing with every aspect of our lives, we use both the shared, external logic, which is contained Aristotelian logic, scientific method and laws, mathematics and agreed rules and laws, and our own private, individual logic of the structure of meaning which gives us our sense of identity.