Forecasting Gold Futures Prices: An Empirical Analysis Using the Vector Autoregression Model

This study presents an empirical examination of the Vector Autoregression (VAR) model's efficacy in predicting gold futures prices, a critical analysis for investors and financial analysts amidst global economic fluctuations. The research leverages a dataset spanning from January 4, 2010, to June 7, 2024, derived from Investing.com, encompassing a multitude of financial indicators such as gold and silver futures, the US Dollar Index, and Brent Crude Oil futures.The methodology includes a comprehensive data preparation phase, ensuring the stationarity of time series data through Augmented Dickey-Fuller (ADF) testing and cointegration analysis. The VAR model is meticulously estimated, identifying the lag order that optimizes model performance while mitigating overfitting risks.The empirical findings underscore the VAR model's high accuracy in forecasting gold futures prices, with significant influences attributed to silver futures prices, the US Dollar Index, and Brent Crude Oil futures. Despite minor discrepancies in capturing short-term price volatility, the model adeptly reflects overall market trends.The study acknowledges limitations, such as the potential exclusion of influential variables and the scope of data timeframe, suggesting future research should expand variable inclusion and integrate advanced prediction techniques. The paper concludes that the VAR model is a robust analytical instrument for predicting gold futures prices, offering substantial support for strategic financial decision-making.