摘要:The statistical properties of daily closing futures prices for nine commodities are studied. Two hypotheses are examined: Price changes are normally distributed, and prices follow a random walk process. Normality is tested by estimating kurtosis, the R/S statistic, and characteristic exponents. The Gaussian hypothesis is rejected in a large proportion of cases. Randomness is tested by using the turning point test and the phase length test. Both tests reject the random walk hypothesis.
关键词:futures prices;Gaussian distribution;stable Paretian distribution;random walk