Information and exchange rate dynamics
Theoretical models of the exchange rate are developed where information on the model is not fully available to agents. It is an application of Benjamin Friedman's (1979) theme that full rational expectations may be a possibility only in the long-run, even for completely rational individuals. The thesis attempts to develop the theory of exchange rate behaviour by considering some neglected informational issues. The three substantive chapters each consider specific aspects of relevance to the determination of the exchange rate from an asset market view of perfect capital mobility. These are the possible current account inter-relationship, the persistence of interest rate differentials between the two currencies and the subjectivity of and the regress in beliefs across a decentralised market. Generally, limited information on the model will give rise to erroneous beliefs, on the one hand, and encourage the acquistion of information and the revision of beliefs, on the other. Erroneous beliefs will cause correlations between variables, which may not normally occur inside full rational expectations. The revision of beliefs will bring a particular source of non-stationarity to the data. And the stability of certain learning forms may require limitations on the degree of capital mobility. These conclusions would suggest that any empirical work on modelling the exchange rate may gain from relaxing certain a priori restrictions, which properly belong to models with stronger assumptions on the availability of information.