Open AccessForecasting Foreign Exchange Rate And Consumer Price Index With Arima Model: The Case Of Turkey
Author(s) -
Abraham Deka,
Nil Günsel Reşatoğlu
Publication year - 2019
Publication title -
international journal of scientific research and management
Language(s) - English
Resource type - Journals
ISSN - 2321-3418
DOI - 10.18535/ijsrm/v7i8.em01
Subject(s) - autoregressive integrated moving average , exchange rate , econometrics , inflation (cosmology) , economics , lira , order (exchange) , foreign exchange market , effective exchange rate , box–jenkins , foreign exchange , statistics , time series , mathematics , monetary economics , finance , physics , theoretical physics
The high and increasing rate of uncertainty in the world’s Foreign Exchange Market (FEM) is one that poses a great concern to the market players, traders and policy makers. There is need to come up with reliable and sophisticated models foreign exchange rate and its determinants in order to predict their future values hence reduce the risks. This paper makes use of the Autoregressive Integrated Moving Average (ARIMA) model to forecast the foreign exchange rate of Turkey and inflation a major determinant of foreign exchange rate. This paper provides that ARIMA(3,1,3) is the best ARIMA model for forecasting foreign exchange rate of Turkey and that ARIMA(1,1,4) is the best ARIMA model for forecasting Turkey’s inflation (CPI). The paper also postulates that ceteris Paribas the foreign exchange rate of the lira against the dollar will be stable in the short run future. However, with the passage of time the suggested model of forecasting foreign exchange rate and inflation of Turkey should always be updated with current data. ACF, PACF, AIC and BIC together with forecasting performance measures like MAE, MAPE, Bias proportion, RMSE and Theil U statistics are very useful in the process of best model selection.