An application of exchange rate forecasting in Turkey
In this study, exchange rate forecasting is studied which plays a key role in free market systems. Official daily data of Central Bank of The Republic of Turkey (CBRT) are used for USD/TL ($/TL), EURO/TL (€/TL) and POUND/TL (£/TL) pars. Moving averages (MA) method, single exponential smoothing method, Holt’s method,Winter’s method and ARIMA models are applied to the each pars Performance of the models are assessed with the performance criteria of mean absolute percentage error (MAPE), root mean square errors (RMSE) and mean square error (MAE). As a result of study, successfully application of the methods based on trend analysis is exhibited for exchange rates in Turkey. These methods are evaluated according to MAPE, RMSE and MAE criteria and the best results are obtained by Winter’s method which means that Winter’s method is an useful method to forecast exchange rates for the given time interval in Turkey.
Keywords: Time series analysis, Box-Jenkins approach, ARIMA, exchange rate forecasting, techniques based on time series analysis