opular opinion asserts that South African politics and oil prices are major exoge-P nous forces that move the gold market. Recent political tension in South Africa and its possible fallout in the gold market is of particular concern because South Africa produces a significant share of the world's supply of gold. The problem is that political tension may have an adverse effect upon South Africa's gold production which could lead to significant price fluctuations in the gold market. This article indexes political unrest in South Africa and demonstrates that it is possible to incorporate empirical measures of political unrest that provide insight to movements in the spot and futures prices in the gold market. This is an important step in bringing the econometric work in asset pricing closer to the reality of market forces that determine prices.
The link between oil and gold is not as obvious. Demand for gold is thought to fluctuate with the price of oil. There are at least two avenues through which this happens. First, rising oil prices are viewed as inflationary. The demand for gold as an inflation hedge increases with oil prices. Second, fluctuations in the demand for gold by oil exporting countries are identified as a source linking oil and gold prices. As oil revenues fluctuate, the demand for gold by oil exporters fluctuates in order to smooth consumption and/or government expenditures. The purpose of this article is to empirically estimate a time varying risk premium in the gold futures market and identify factors that may contribute to such a risk premium. Fama and French (1985) discuss, there are two popular theories of commodity futures markets. The theory of storage has been popularized by Working (1948), Brennan Richard Baillie, Graciela Kaminsky, and two referees of this journal provided valuable suggestions on an earlier draft. We remain responsible for all errors. Melvin acknowledges the support of the Center for Financial System Research at Arizona State University.
THEORETICAL FRAMEWORK As