Date of Award

Spring 2007

Project Type


Program or Major


Degree Name

Doctor of Philosophy

First Advisor

Michael D Goldberg


The monetary authorities of many open economies regularly intervene in foreign exchange markets with the aim of limiting swings in exchange rates. A survey of exchange rate models reveals, however, that the literature lacks a model that provides an explanation for exchange rate swings and at the same time offers a role for official intervention and parity announcement to affect exchange rate movements. Furthermore, empirical studies suggest that official intervention can have a significant effect on exchange rate movements, but only in the short run (up to 3 months). In this dissertation, I build on the premium model by Frydman and Goldberg (2007), which is based on endogenous prospect theory, and construct a model that provides for swings and a role for intervention. The model implies that intervention aimed at pushing the exchange rate back to some announced or even unannounced parity level at unpredictable moments in time can lead to more limited swings in exchange rates compared to regimes without such policy through an uncertainty premium channel. This new channel follows from one of the key implications of endogenous prospect theory that market participants require a premium as compensation for their greater sensitivity to losses before they are willing to speculate in currency markets. This so-called uncertainty premium depends positively on the gap between the forecast of the future exchange rate and its perceived historical benchmark level. The main contribution of this research is empirical. I first test for the positive relationship between the premium and the gap using a contingency table analysis and regression analysis based on an autoregressive distributive lag model. Then I investigate whether central banks can take advantage of this positive relationship between the gap and the premium in limiting exchange rate swings. My results provide evidence that the gap and the premium are positively related, and that central bank intervention supporting some parity level can lead to swings of smaller magnitude. My empirical analysis is based on survey data on exchange rate forecasts, exchange rate regime classifications and data on changes in reserves, and includes 24 currencies from developed and developing countries.