Tests of the random walk hypothesis for London gold prices. 2002.
This 2002 paper looks at whether gold prices follow a random walk, and while this condition does not necessarily imply that the gold market is efficient in the EMH sense. Instead it can be said to be informationally efficient – today’s prices cannot be used to forecast tomorrows.
The authors reject the random walk hypothesis for the Am and PM fixings but accept it for the London closing price.
Data: Daily London Am and PM fixings as well as the London closing price, 1990-2001
Methodology: Lo and Mackinaly (188) variance ratio test
Citation: Smith, Graham. “Tests of the random walk hypothesis for London gold prices.” Applied Economics Letters 9.10 (2002): 671-674.
Abstract: The hypothesis that London gold prices follow a random walk is tested for three prices, those determined at the morning and afternoon fixings and the closing price, using the multiple variance ratio test. For the prices set in the twice-daily fixings, the random walk hypothesis is rejected because of autocorrelation in returns. However, the closing price, which is determined by additional information and involves many more participants in the market, follows a random walk.