Exchange Rate Risk and Forecasting

Sammanfattning

Since the collapse of the Bretton Woods system, the system of fixed exchange rates amongst principal industrial countries, in the early 1970s, a new era began, introducing the floating exchange rate regime. Since the inception of the floating rate regime, the general interest in forecasting exchange rate movements has grown considerably. Yet, forecasting the exchange rate is extremely difficult, and it is often referred to as an impossible task to execute successfully over a long period of time - predictions generated by forecast models are conventionally claimed to be inferior to the classic random walk in its predictability power. However, some researchers have proven to reject these claims by introducing combinations of models to forecast the Forex market, essentially utilizing the synergy between the models rather than applying them as standalone techniques. The core of this project is designed by introducing a combination of two forecasting models, of which one focuses on traditional economic theory to forecast future values, and the other one on extrapolating historical patterns from the exchange rate data itself into the future. Three different currency pairs are analyzed on a one, two, and three year horizon, and I find promising results in favor of the predictability of the Forex market. More specifically, I find that each model contains information complementary to each other and that they therefore are fit to be applied jointly. Additionally, I present a strategy with the aim of lowering the volatility of returns, essentially decreasing the exchange rate risk involved as the investor is exposed to the currency market.