Toward Agent-Based Models for Investment

s far as I know, no one currently uses agentbased models for investment. In the context of this presentation, an agent-based model involves a model for price formation and a model for agent behavior, including the models’ information inputs and how the inputs behave, which could involve learning. From an investor’s point of view, agentbased models represent a new direction that may or may not result in more practical investment models. From an academic’s point of view, agent-based models are fascinating new tools that make it possible to investigate problems previously out of reach. The models presented here do not provide practical tools, but they do offer a taste of several new directions that may or may not turn out to have practical value in the future. One of the things that attracted me to financial prediction is the possibility of forecasting the behavior of people. People have free will—you never know what they will do. Of course, we make predictions about people in our day-to-day lives. We all know people whose behavior is pretty consistent and whose responses are predictable. But making quantitative predictions about people in general is a real challenge, and little success has been apparent so far. Because prices involve people, predicting prices is different from predicting the weather. When the weatherman says it is likely to rain, this prediction has no effect on whether it will rain. In contrast, a statement by George Soros that gold prices are going to rise can have a significant effect on the price of gold. Predicting prices is more like telling people’s fortunes. If a fortune-teller says you will take a long trip and have a terrible accident and you have any faith in her forecast, you will stay home. If you do stay home and no terrible accident occurs, was she right? The point is that forecasts regarding humans, who have free will and can alter their behavior, can be self-invalidating. This point is the idea underlying the theory of efficient markets: If someone discovers a pattern, it diminishes as it is exploited. If enough people discover the pattern, it will disappear. Informational efficiency (the lack of arbitrage) is a central idea of modern finance. Anybody who has spent time building market models knows that markets are arbitrage efficient at some level. Building models that make good market predictions is not easy. But markets cannot be perfectly efficient; otherwise, why would so many smart people waste their time investing in them? Milton Friedman originally pointed out the paradox of efficient markets: If rational speculation makes markets efficient, then because the market is efficient, no profits can be made and all the rational speculators should leave, thus causing the market to revert to an inefficient state. The theory of efficient markets is inherently self-contradictory. Finding a self-consistent theory to replace it is an important problem that demands a solution. As a physicist, I would state the situation as follows: At first order, markets are arbitrage efficient, but at second order, they are not. For markets to function, there Although agent-based models are not yet ready for practical investment application, they can yield powerful insights about market behavior, particularly in regard to the “second order” inefficiencies that create profit-making opportunities. When practical use of agent-based models becomes possible (perhaps within the next five years), their effectiveness will cause securities prices to change.