中国大宗商品现货价格指数的构建及预测能力研究

本文首先选取了我国国民经济中八个主要行业，首次提出并采用成分商品的“三重筛选”标准，选取了72种成分商品，先以成分商品的表观消费量为权重构建了行业商品现货价格指数，然后以行业规模以上工业企业的工业总产值为权重综合行业指数得到中国大宗商品现货价格指数。其次，基于已构建指数，采用谱分析方法研究发现：中国大宗商品现货价格指数领先CPI、PPI、CI三个国内宏观经济指数的期数依次为2.48个月、0.66个月、0.89个月，领先BPI、MyBCIC、CCPI三个国内大宗商品现货价格指数的期数依次为1.54天、1天、1.21周，领先上证综合指数的期数为13天，说明中国大宗商品现货价格指数具有很好的预警和预测能力。最后，本文用时差相关系数法对CSCPI的预测能力进行了定量检验，检验结果与谱分析结果一致，说明谱分析结果是稳健的。

Construction and Predictive Ability Research on China’s Staple Commodity Price Index

t selects eight major industries from the national economy in our country, and puts forward a “triple selection” criteria of composition goods for the first time to select 72 kinds of components. It constructs the industry commodity spot price index with apparent consumption of composition goods as the weight, and then establishes the China’s Staple Commodity Price Index (CSCPI) with gross output value of industrial enterprises above designated size as the weight. Based on the constructed index, it introduces the spectral analysis method to do the empirical analysis, and finds that CSCPI is respectively 2.48, 0.66, and 0.89 months ahead of CPI, PPI, CI of domestic macroeconomic indexes, respectively 1.54 days, 1 day and 1.21 weeks ahead of BPI, MyBCIC, CCPI of domestic commodity spot price indexes, and 13 days ahead of the Shanghai composite index. It indicates that CSCPI has a good predictive ability. Finally, it employs the cross correlation method to carry on quantitative tests about the predictive ability of CSCPI, and the results consistent with the spectral analysis. So the spectral analysis results are robust.