The United States (US) presidential election in November 2024 created a significant external shock to the global economy, particularly the world gold market. The appreciation of the US dollar that followed led to a sharp fall in gold prices, reaching 2,582.1 USD, after previously recording highs in early October 2024. Gold, as a safe haven asset, is inversely related to USD fluctuations, which highlights the urgency of applying accurate modeling to capture the impact of sudden political or economic events. This study applies pulse function intervention analysis to model and forecast global gold prices, using weekly data from January 2023 to March 2025, with the intervention point in the second week of November 2024.
The pre-intervention data was best modeled using ARIMA(0,2,1), which passed diagnostic checks of parameter significance, residual white noise, and normality. Intervention analysis through the Cross Correlation Function (CCF) identified the best intervention order of b = 1, r = 0, s = 0, indicating that the post-election effect was immediate but temporary. The final intervention model produced accurate results with MAPE = 1.51%, MSE = 0.0002037, AIC = -530.394, and SBC = -525.030, demonstrating strong predictive performance. Model validation using out-of-sample test data confirmed its reliability, while forecasts suggest that world gold prices are likely to rise steadily until the end of July 2025, potentially surpassing 3,300 USD.
The findings highlight the effectiveness of pulse intervention analysis in capturing short-term shocks from political events, while also providing valuable insights for investors and policymakers. By showing that sudden dollar appreciation significantly affects gold prices, this study emphasizes the importance of hedging strategies and market monitoring. Overall, the research deepens our understanding of gold price dynamics and supports better decision-making in times of financial volatility.
Author: Sediono, Drs., M.Si.
