A short-run pricing model for a speculative asset, tested with data from the gold bullion market. 1983.
Koutsoyiannis (1983) take these ideas forward by developing a very short run model of the gold price, over one day. Data used is daily, as opposed to the previously monthly data used, between Jan 1980 and March 1981 based on the similar data to previous papers – US dollar, US interest rate, inflation expectations (simplified to observed inflation), author calculated index of political tension as well as including other assets which could be considered as substitutes for gold for the first time: Silver, Dow Jones Index, Oil. A variety of models are tests with all variables in levels raising more unit root questions.
At a daily frequency the levels regression shows a high R2 at 95% but this is probably be due to a unit root problem. Inflation, the US Dollar and US interest rates are all significant variables with the expected signs. Silver is shown not to be a close substitute, a finding confirmed in latter studies of precious metals markets. Oil has a positive relationship and stocks markets a negative one indicating portfolio allocation choices. The adhoc nature of the political tension index adopted adds little to the discussion. As the analysis took place at a time of rapidly rising prices (going from $460 to $800 an ounce) the authors also split their sample and run a chow tests for parameter stability and find that the sensitivity of the god price to economic variables is not constant.