中文摘要：

大量研究表明，当时间序列属于非平稳过程时，判断变量间的因果关系将十分困难。而作为计量经济学的新流派，协整理论及误差修正模型很好地解决了时间序列的非平稳问题。通过实证分析的方法，证实了上海期货交易所期铝价格与现货铝价格以及氧化铝价格3个时间序列满足同阶平稳，通过向量自回归（VAR）模型证实了它们之间存在着多重协整关系，进而得出3者之间的因果关系。

英文摘要：

Causal analysis would be difficult if the time series of non-stationarity process. Nonstationarity theory and vector error correction model can be used for this type of the problem. Multicointegration and causal relationship were introduced. The multi-cointegration among Aluminum Futures, Aluminum spots and oxide-Al at Shanghai Metal Exchange were studied by the above models. The results showed that the three price series followed the multi cointegration relationship by the VAR model. The casual relationship among the three prices series were proposed.