A Comparison of the Monetary Model and Artificial Neural Networks in Exchange Rate Forecasting

Exchange value is one of the significant tools for investors in decision making. Since exchange values are volatile and they change within short periods, investors need an effective method to minimize the risk. This study compares the prediction performances of artificial neural networks, which is recently being used as an effective tool of prediction, and the monetary model, which is one of the methods to predict structural exchange rates. In this study exchange rates of Turkish Lira against US Dollar and Euro are predicted. In models, inflation levels for domestic and foreign countries, money supply, interest rates and economic indicators are used. The time period between 1986 and 2010 is covered for the USD and the time period between 1999 and 2010 is covered for the EU. Results of this study show that ANN, which is recently being used for the prediction problems, reached a high level prediction performance.