Investigating the Long time Memory in the Future Market of Gold
本文用ARFIMA模型和三种估计方法分析美国黄金期货日价格数据，发现价格存在短期记忆而非长期记忆，表明市场非有效且风险是价格行为的重要决定因素。
This paper concentrates on investigating the long memory property of the gold future markets in the United States. The data set consists of daily future prices, and long memory tests based on ARFIMA with use of three quantitative methods: Non Linear least Square (NLS), Modified Profile Likelihood (MPL) and Exact Maximum Likelihood (ML). The results of the ARFIMA model don show strong evidence of long memory, but it implies short memory. So prices follow a predictable behavior, which is inconsistent with the efficient market hypothesis. The evidence of short memory in prices, however, shows that uncertainty or risk is an important determinant of behavior of daily future prices in the gold future market of the United States.