Exchange rate modelling and forecasting
The objective of this thesis is to assess the current state of exchange rate modelling and forecasting. The thesis has four distinct essays, each one analysing a current interest topic in this wide and vibrant area of economic research. But a common thread runs through all four: to determine whether it is possible to use the results of this research to develop trading strategies that can add persistent value to international investment portfolios with significant exposure to the foreign exchange market. This market has a daily turnover of $1.9 trillion (BIS, 2004) and is the most liquid financial exchange in the world, by some distance. Nonetheless, we argue that the market is also inefficient, in the sense that profitable trading opportunities persist and that prices do not reflect all available public information on a continuous basis. If we are correct-and we present simulation results that suggest we are-then the opportunity to derive and test plausible trading rules for the management of international investment portfolios though rigorous academic research is enormous. Yet all too often academic exchange rate research appears to be conducted in a cocoon, with the result that conclusions are sometimes difficult to apply in a practical context by portfolio managers. These difficulties reflect the computational requirements of implementing highly intensive trading strategies, associated trading costs and size limitations, and the practical limitations on implementation raised by publication lags and general data limitations. We aim to address these difficulties throughout this thesis. By assessing the merits of various theoretical models that collectively encompass all of the main themes on the current research agenda, we will be in a position to appreciate both the statistical and economic value of existing academic research, isolating areas of real merit for the investment community, and suggesting areas for further attention.