Semi-rational expectations and exchange-rate dynamics

The Rational Expectations Hypothesis (REH) asserts that, on average, the economic agents are accurate in predicting future economic developments. The paper demonstrates, however, that in a world of costly information, individual rationality may result in consistent and persistent forecasting biases. A distinction is drawn between perfect foresight or efficient forecasting-which is consistent with the REH-and myopic perfect foresight-which is the profit maximizing, and thus the rational one from an individualistic point of view, even though the latter may result in persistently biased forecasting. These concepts are illustrated in a model of exchange rate dynamics which introduces myopic or 'semi' rationality into Dornbusch's familiar model.