A Statistical Investigation and Forecasting of Gold Prices through Box-Jenkins ARIMA in Pakistan

Abstract

Gold plays a significant role in the financial market, monetary policy, and various industries, including jewelry, pharmaceuticals, and electronics; however, its price is often volatile. This study focuses on forecasting daily gold prices of Pakistan using various ARIMA models to identify the most suitable approach for accurate gold price prediction. Gold forecasting is a critical and intriguing area of study, and the Box-Jenkins methodology, which employs ARIMA models, is highly effective for analyzing and forecasting time series with diverse patterns of variation. The ARIMA modeling process involves defining, estimating, diagnosing, and forecasting stages, making it well-suited for this application. Based on the analysis, the ARIMA (4, 1, 4) model was the best fit for the dataset, evaluated using SIGMASQ, AIC, and SC criteria, which yielded the smallest values and were confirmed through the Box-Jenkins. Using the ARIMA (4, 1, 4) model, the study successfully forecasted daily gold prices of Pakistan from October 2028 to June 2022.