Handbook of Exchange Rates

4.4 Conclusion

In this chapter, we have presented a modeling strategy of the foreign exchange market that departs from the prevailing REs paradigm. Mainstream models have not been able to give a fully convincing explanation of several empirical regularities observed in the foreign exchange market. In this chapter, we have presented a prototype model as an example of an agent-based model. The model assumes that agents have cognitive limitations, that is, they are unable to comprehend the complexities of the underlying model. As a result, they use simple forecasting rules (heuristics). They are rational, however, in that they wish to learn from their mistakes. Thus, they switch to alternative rules if it turns out that these alternative rules perform better than the one in use. This switching mechanism provides the basis for the dynamics of the foreign exchange market. In particular, it offers a simple explanation for several phenomena typically observed in the foreign exchange market.

The simulated exchange rate in our model is systematically disconnected from its fundamentals, and it is characterized by excess volatility. At the same time, the model produces fat-tailed distributed exchange rate returns, and moreover the null of a unit root in the exchange rate level cannot be rejected. These results are driven by the switches in the trading strategies of traders. Traders try a trading strategy and when they find out that there is a more profitable alternative, they switch to it, ...