© 2016 Elsevier Inc.Gold is special as it is influenced by a wide range of factors such as commodity prices, interest rates, inflation expectations, exchange rate changes and stock market volatility. Hence, forecasting the price of gold is a difficult task and the main problem a researcher faces is to select the relevant regressors at each point in time. This model uncertainty in combination with parameter uncertainty is explicitly accounted for by Dynamic Model Averaging (DMA) which allows both the forecasting model and the coefficients to change over time. Based on this framework, we systematically evaluate a large set of possible gold price determinants and find that DMA (1) improves forecasts compared to other frameworks, (2) yields strong time-variation of gold price predictors and (3) favors parsimonious models. The results also show that typical in-sample features of gold such as its hedge property are weaker in an out-of-sample context.