Forecasting gold prices in Malaysia: a comparative analysis of simple exponential smoothing, Holt’s method and Arima model

Gold plays a vital role in the global economy as a safe-haven asset, widely used for investment, jewellery and as a hedge against inflation and currency fluctuations. In Malaysia, gold price trends are closely watched by investors and analysts due to their sensitivity to market and global events. This study forecasts monthly gold prices in Malaysia using three univariate time series models: Simple Exponential Smoothing (SES), Holt’s Method, and ARIMA. Using 122 monthly observations from April 2015 to May 2025, the models’ accuracy was compared based on error measures such as RMSE, MAE, MAPE and MASE. Holt’s Method outperformed SES and ARIMA (2,1,1) by achieving 91.14% accuracy with the lowest errors especially using a 60:40 training-testing split. The 12-month forecast shows a rising gold price trend from June 2025 to May 2026. The study highlights Holt’s effectiveness in capturing linear trends and suggests future research on advanced or hybrid models incorporating external factors like inflation and global events for better accuracy. It also recommends future research to explore more advanced or hybrid forecasting models and incorporate external factors like inflation and global economic events to further improve prediction accuracy.