This paper proposes a full Bayesian nonparametric procedure to investigate the predictive power of exchange rates on commodity prices for 3 commodity-exporting countries: Canada, Australia and New Zealand. We examine the predictive effect of exchange rates on the entire distribution of commodity prices and how this effect changes over time. For this purpose, a time-dependent infinite mixture of normal linear regression model is proposed for the conditional distribution of the commodity price index. The mixing weights of the mixture follow a Probit stick-breaking prior and are hence time-varying. As a result, we allow the conditional distribution of the commodity price index given exchange rates to change over time nonparametrically. It is shown that exchange rates do not have consistent predictive power for commodity prices in all countries considered in this paper. We find that exchange rates do have predictive power in some cases, but their impact tends to be constant over time. On the other hand, the intercept in the regression and the lagged dependent variable show signs of parameter change over time in most cases, which is important in predicting both the mean and the density of the commodity prices one period ahead. The results also suggest that for all countries considered, a significant source of time variation in the conditional distribution of commodity prices comes from the variance.