Three petroleum futures contracts are examined over a ten-year period from 1986 to 1996. Intertemporal changes in futures prices and the net open interest positions of three trader types are compared to determine what, if any, market timing ability the traders have. Seasonal variation is considered and a simple trading rule is adopted to determine the dollar-return potential for market participation and shed light on issues of market efficiency.;This study utilizes a methodology proposed by Cumby and Modest (1985) and applied to futures markets by Hartz-mark (1991) and Leuthold, Garcia and Lu (1994). Forecast consistency is measured by univariate and multivariate logistical regressions and Henriksson-Merton's conditional probability estimation procedures. Forecast conviction is measured by univariate and multivariate OLS regressions modified by White's heteroskedasticity correction procedure.;The ten-year analysis allows examination of long cyclical patterns in crude oil futures and shorter seasonal patterns in heating oil and unleaded gasoline futures. Tests are conducted on each model's sensitivity to measurement interval, trader position and the coincidental observation of reported positions.;For all three petroleum futures contracts, univariate tests exhibit systematic evidence of a priori forecast power information in the net open interest position of large speculators. Of additional significance is the consistently perverse signal sent by the large hedger's position. Notably, nonreporting traders do not appear to play a significant role in the petroleum futures market.;Multivariate tests, on the other hand, indicate caution is warranted in concluding that returns to large speculators are due to superior information. The notion that a trader's position is determined independently is not supported. Seasonality does not appear to be a significant factor, though transaction costs do. Additional research is needed to advance the findings of this investigation.
展开▼
机译：在从1986年到1996年的十年中，研究了三份石油期货合约。比较了三种价格类型的期货价格的时空变化和净未平仓头寸，以确定交易者具有什么（如果有的话）市场时机能力。考虑季节变化，并采用简单的交易规则确定市场参与的美元收益潜力，并阐明市场效率问题。本研究采用了Cumby和Modest（1985）提出的方法，并通过以下方法应用于期货市场： Hartz-mark（1991）和Leuthold，Garcia and Lu（1994）。预测一致性是通过单变量和多元逻辑回归以及Henriksson-Merton的条件概率估计程序来衡量的。预测信念是通过怀特异方差校正程序修正的单变量和多元OLS回归来衡量的。十年的分析允许检查原油期货中的长周期模式以及取暖油和无铅汽油期货中的短季节模式。对每种模型对测量间隔，交易者持仓和报告仓位的巧合观测的敏感性进行了测试。对于所有三个石油期货合约，单变量测试显示了大型投机者净未平仓头寸中先验预测力量信息的系统证据。另一个重要意义是大型套期交易者的头寸发出的持续不正常的信号。值得注意的是，不报告的交易者似乎在石油期货市场中没有扮演重要的角色。另一方面，多变量测试表明，得出结论认为大型投机者的收益归因于优越的信息，因此值得谨慎。不支持独立确定交易者头寸的观点。尽管交易成本确实如此，但季节性似乎并不是一个重要因素。需要进一步的研究以推进这项调查的结果。
展开▼
