Abstract:

In recent years, interest on the link between prices of Oil and Gold has grown. Theoretically, these two essential commodities have in common the inflation rate as a similar long-run driver. The aim of the paper is to estimate the long-run relation between cyclically-adjusted time series for both gold and oil prices over the 20th century.

A common trend in the variation of these two assets can be estimated by using cointegration method (Engle and Granger (1987)). However, this common trend will be more effective if we concentrate the analysis on relations of structural nature. Thereby, we extract the cyclical element from the two times series to disentangle short run dynamics related to business cycle fluctuations (Using H-P filter). This permits to contain the issue of reverse causation in interpreting results. Therefore, dating break points in the relation between oil and gold prices will allow us to investigate the nature of events that influence their co-movement. So, we can estimate adjustment term after a shock that causes a deviation from this long-term relationship.

For this purpose, we use data of gold prices and crude oil prices with annual periodicity from 1900 to 2010. Prices are in real terms and the US CPI is taken as a deflator. After the unit root test of the tow variables, we proceed to the extraction of the cyclical elements from the time series and we proceed to a cointegration test with unknown break points (Gregory and Hansen (1996)). In order to estimate short-run dynamics after chocks, we use a vector error correction model (Vector Error Correction Model).