eThesis - School of Business electronic theses

In this thesis I study the impact of physically-backed precious metal ETPs on the related precious metal prices. I focus on gold, silver, palladium and platinum as they attract majority of the physically-backed ETP assets. Previous research has focused mainly on the on the impact of speculators' derivatives positions, so this thesis is among the first ones to study the purchases and sales of ETPs in the commodity spot market. The thesis contributes to the heated debate on the increased amount of speculators in the commodity markets described as the financialization of the commodity markets.

The data set consists of daily net changes in the precious metal ETP holdings and the related precious metal spot prices from 25 April 2007 to 24 January 2014. I study the relationship between ETP holdings and the related spot prices through hypothesis testing and employ Augmented Dickey-Fuller and Phillips-Perron unit root tests, Johansen test for cointegration, Granger-causality test, and OLS-regression. For robustness, I use two different lag selection criteria (Akaike information criterion and Schwarz criterion) and two time periods. The regression model is tested for heteroskedasticity and autocorrelation.

The empirical results show that the precious metal ETP holdings and the related spot prices are not cointegrated supporting the previous research. This implies that there is not a long-term equilibrium that would avoid these two from drifting apart from each other. I also find only weak evidence supporting the causal relationship of the ETP holdings and the related spot prices. The changes in gold ETP holdings Granger-cause gold spot price in the post-crisis period. For the full-sample period and for the rest of the metals I find no evidence on causality. On the contemporaneous relationship I find new evidence. The OLS-regression results show that the daily net changes in gold ETP holdings have an immediate impact on gold spot price. An increase/decrease in gold ETP holdings increases/decreases the gold spot price. The impact is strongest on the concurrent day but proves significant for the subsequent day also. The regression results for the rest of the metals are similar but statistically insignificant.

My findings for gold suggest that the commodity prices are no longer simply determined by their true supply and demand. Instead, financial innovations such as physically-backed ETPs may play an important role in the price determination if they grow large enough in relative to the size of the market.