A Comparative Study of Forecasting Models for Daily Exchange Rates of Seven Major Currencies against the Thai Baht
Keywords:
อัตราแลกเปลี่ยน, ตัวแบบพยากรณ์อนุกรมเวลา, ARIMA, วิธีโฮลต์-วินเทอร์ส, การทดสอบ Diebold–MarianoAbstract
This research aimed 1) to examine the characteristics and trends of the daily exchange rates of seven foreign currencies against the Thai baht, namely the US dollar (USD), pound sterling (GBP), euro (EUR), Japanese yen (JPY), Hong Kong dollar (HKD), Malaysian ringgit (MYR) and Singapore dollar (SGD), and 2) to compare the performance of six statistical forecasting models: simple moving average (SMA), simple exponential smoothing (SES), double exponential smoothing (DES), additive and multiplicative Holt-Winters (HW) and ARIMA. Secondary data were obtained from the Bank of Thailand from 2 June to 30 October 2026, giving 102 trading days per currency, which were split into a training set of 82 days and a test set of 20 days. Forecast accuracy was measured by MAPE, MAE, MSE and RMSE, and differences in accuracy were tested with the Diebold–Mariano test. The results showed that the baht depreciated against six currencies, most strongly against the yen (4.94%), and appreciated slightly against the euro (0.56%). All series were non-stationary in levels but stationary in first differences, and the selected ARIMA model for every currency was ARIMA (0,1,0), consistent with a random walk. All models produced MAPE values below 0.30%, indicating highly accurate forecasts. The lowest MAPE was obtained by SMA for USD, HKD, MYR and SGD, by additive HW for GBP and JPY, and by multiplicative HW for EUR. However, the Diebold–Mariano test indicated no statistically significant difference at the .05 level between the best model and ARIMA for any currency. The findings suggest that simple models are sufficient for one-day-ahead forecasting and can support international trade decisions and exchange rate risk management for Thai businesses.
Downloads
Published
How to Cite
Issue
Section
License
Copyright (c) 2026 Chalermkarnchana Academic Journal

This work is licensed under a Creative Commons Attribution-NonCommercial-NoDerivatives 4.0 International License.