PsyFi Search

Thursday 28 July 2011

Perpetual Novelty, Santa Fe Style

Learnings from Science

By the late 1980's there was a growing recognition that the existing understanding of financial systems was flawed. Not only did markets not behave as the economic theories predicted but they often exhibited behaviour that didn't seem to have any pattern or cause at all.

In response to this a number of economists began looking at some of the research emerging from physics, biology and computer science in the area of complex adaptive systems and this led, in 1987, to a group of economists and scientists getting together at the Santa Fe Institute. The program of work that came out of this seminal event is still unfolding today, but suggests why academics and traders have had such different views on markets: one set lives in the real world, and the other doesn't.  Wanna hazard a guess as to which is which?

Suspect Quants

To the physicists involved in the original Santa Fe seminars it was obvious that the basis of economic quantification was highly suspect: indeed it was perfectly clear that the assumptions behind the models – stuff like people being perfectly rational – had been introduced to make the maths work, rather than being realistic assumptions. Of course, this wasn’t entirely unfamiliar to physicists, as Max Plank introduced the idea of the quantum for exactly the same reason. Only thing was, Plank turned out to be right, even though he had a hard time accepting this, as we saw in Quantum Consciousness is Market Uncertainty.

At the heart of the problem with economics was the focus on equilibrium, the idea that there’s some perfect steady state of the economy in which everything is in balance (see Exit the Walras, Followed by Equilibrium). Essentially equilibrium economics defines the desirable outcome as a static position, then asks what conditions can give rise to this, then assumes that these conditions are present. This is a bit like deciding that riding a bicycle is a good thing, identifying the balance of forces needed to achieve this successfully and then declaring that this balance is the natural order of things.

And then ignoring what happens if you try to go cycling in a typhoon.

Perpetual Novelty

This approach was endemic to the subject. As Brian Arthur explains:
“In other words, [economics] asked what patterns in the economy would look like if they were at equilibrium—were consistent with the micro-behavior (actions, strategies, expectations) that creates them. Thus, for example, general equilibrium theory asks: What prices and quantities of goods produced and consumed are consistent with—would pose no incentives for change to—the overall pattern of prices and quantities in the economy’s markets?”
Yet, as we all know from everyday experience, life is full of what Arthur calls “perpetual novelty”: we live in non-equilibrium conditions, a constant sea of change. The Santa Fe meetings led to an interest in these non-equilibrium situations, the type of economic state where things are forever changing, rather than the type where they’re static.

Reflexivity, Stupid

Underlying this change in approach is a fundamental problem. Equilibrium economics is designed to ignore psychology, because it assumes that a person never changes their mind: they’ve already considered all possible outcomes and rejected all but the most economically rational. In the new economics we can’t assume this: we must work on the basis that if the world changes then a person’s expectations also change. This is a reflexive world, not a static one.

Suddenly, human behaviour matters in economic models in a way it never did before. Psychology, having been banished out of the front door, has crept in the back and has started cooking itself a three course dinner with all the trimmings.

The outcome of the Santa Fe meeting was a focus on non-equilibrium approaches to economics. In fact non-equilibrium economics doesn’t claim that equilibrium conditions don’t occur, only that they’re not guaranteed. The reason for this is human behaviour: we’re not external to the economic machinery, we’re part of it, and our expectations about the future have a critical impact on how it develops. All of which means that the future is not predictable and models that assume it is so are just wrong.

The El Farol Bar

Brian Arthur created a thought experiment around his favourite bar to illustrate the issue. This, the famous El Farol Bar Problem, requires that individuals decide whether or not to go to their favourite bar. If they believe that there will be more than 60 attendees they don’t go, if less then they do. Of course, if everyone is doing this then the result is that everyone knows that everyone’s decision is dependent on everyone else’s predictions: you cannot rationally deduce the correctly answer.

When this situation is modelled by making the assumption that people chose to attend based on their most recent experience of the bar, the average attendance rapidly converges on 60. However, this is an emergent property of the system, not something that can be precomputed: equilibrium is not preordained.

The Santa Fe Stockmarket

In fact, it turns out that the reason this stability is achieved is because everyone has a stable hypothesis about the conditions under which they will attend the festivities. When Arthur and colleagues modelled a simple stock market by allowing investors to modify their expectations in response to market changes - which were themselves formed by expectations in an endless recursive fashion - they found that if people didn’t change their future expectations very often the simulation once again settled down to an equilibrium in line with the classical models, under the rational expectations approach. However, if their expectations changed more frequently the market never reached equilibrium and weaved about randomly, oscillating between periods of stability and periods of great volatility.

As the researchers state, this helps to explain:
"One of the more striking puzzles in finance: that market traders often believe in such concepts as technical trading, "market psychology", and bandwagon effects, while academic theorists believe in market efficiency and a lack of speculative opportunities. Both views, we show, are correct, but within different regimes."
Adaptive Markets

This latter condition looks very similar to what we see in real markets, where short-term trading opportunities are exploited by groups of investors who take positions betting that other investors will attempt to exploit certain conditions: generating positive feedback and sudden surges in markets. The same can happen in reverse, of course. This theoretical market is sensitive to changes in people’s beliefs, which ripple through markets, affecting everyone.

The problem is that, just as with the El Farol Bar, there is no single solution to the problem in the second state because we can’t analyse the problem ahead of time, because the solution isn’t preordained. There is no equilibrium so any model that starts by assuming that there is one is simply addressing the wrong question.

The reality is that people are reflexive and markets are adaptive. One begets the other and in a world where traders are forever trying to anticipate what other traders are going to do the last thing markets will ever be, in the short-term, is efficient and in equilibrium. But, of course, the trick to riding a bike is not falling off, no matter what the weather throws at you.

Related articles:
Related books:
Origin of Wealth: Evolution, Complexity, and the Radical Remaking of Economics Simply Complexity: A Clear Guide to Complexity TheoryCritical Mass: How One Thing Leads to Another

No comments:

Post a Comment