The economic determinants of gold’s price :a time-series approach -

Abstract

A model for the price of gold is presented based on separation of short-run and long-run economic determinants. Economic theory, as well as the properties of gold as a commodity, is used to derive the determinants of the price. A linkage between the price of gold and inflation in the long-run is examined for the U.S. economy. Short-run variables, which are derived from supply and demand mechanics, account for deviations from long-run equilibrium. They are the U.S. Dollar exchange rate, the beta coefficient, the credit risk default premium, the leasing rate for gold (proxied by the real interest rate), and the level of income in the economy. Variable selection is carried out so as to best represent the underlying economic relationships between each variable and the price. In each case, the variable of interest is preceded by a relevant analysis which warrants its inclusion in the model. The techniques and methods employed in this project are time series econometrics. Unit-Root tests, Vector Autoregressions, Cointegration techniques were used to put theory to the test by using appropriate testing procedures. The Engle-Granger and Johansen test procedures are carried out to detect the presence of a long-run relationship between the price of gold and inflation. A Zivot-Andrews test is also applied to the series to detect structural breaks. The findings provide ample support for a one to one long-run relationship between the price of gold and the general price level. A model for the price of gold is specified in first differences along with an appropriate Error Correction Model. Strong evidence is also found for the significance of all short-run variables as well as for the ECM term in the model estimation phase.