3.2 Endogenous speculative bubbles

A rather new research area in the theory of exchange rates are theories of endogenous speculative bubbles (Frankel and Rose, 1995).

A number of researchers have deviated from the rational expectations paradigm within this research area. Theories of endogenous speculative bubbles differ from theories of rational speculative bubbles because the latter models are based on rational expectations. Dominguez (1986) and Ito (1990, 1994), for example, use survey data on exchange rate expectations at various horizons to question the hypothesis of rational expectations in the foreign exchange market.

Specifically, the models within theories of endogenous speculative bubbles start from the proposition that the forecasts of the market participants are drawn from competing views, i. e. different kinds of actors are introduced into the models. This proposition is justified by Cutler et al. (1991), Ito (1990) and Taylor and Allen (1992) who provide empirical evidence for the importance of considering heterogeneous agents in exchange rate theory. In a seminal paper by Frankel and Froot (1986), three kinds of actors were introduced; fundamentalists, portfolio managers and the aforementioned chartists.

These three actors behave differently when forming their expectations about the future development of the exchange rate because they have different information sets. Fundamentalists have macroeconomic fundamentals in their information set, e.g. the exchange rate in long-run equilibrium, and thus base their expectations according to a model that consists of macroeconomic fundamentals only, e.g. the Dornbusch (1976) overshooting model. Chartists, on the other hand, have information sets containing only the time series of the exchange rate itself.

Portfolio managers, the only actors who actually buy and sell foreign currencies, form their expectations about the future development of the exchange rate as a weighted average of the expectations of chartists and fundamentalists:

where denotes the rate of depreciation of the domestic currency,

and

denote expectations of chartists and fundamentalists, respectively, and which is a weighted average of these expectations where w determines the weights, denotes expectations of portfolio managers. Thus, each of these three actors behave in a "rational manner", but they behave differently because of their different information sets.

For example, the portfolio managers in Frankel and Froot (1986, 1990) update the weights over time according to whether the chartists or the fundamentalists have recently been doing the better forecasting.

De Grauwe et al. (1993) provide another example of a model with heterogeneous agents. They showed that the interaction of chartists and fundamentalists in the foreign exchange market can give rise to a chaotic behavior of the exchange rate.

Prof. Mikael Bask

Next: Summary of Paper [iv]

Summary: Index