多方法改进视角下的中国核心通货膨胀测算和效果评价

Measuring and Evaluating of Core Inflation in China from the Perspective of Multi-Method Improvement

本文在探析核心通货膨胀测算方法和评价方法原理的基础上,结合中国价格变动特点对部分方法进行适当改进,并将其应用于中国2001-2013年核心通货膨胀测算,然后采用无偏性检验、相关性检验、趋势追踪能力检验、未来预测能力检验四种准则对所有测算结果的货币政策应用效果做出评价。结果发现,没有任何一种方法是绝对最优的,相对而言,SVAR法、指数平滑法、不对称截尾均值法的效果较为优良。这一结论对中国的货币政策执行和核心通货膨胀测算方法研究具有重要的启示意义。

Based on a review of measures and evaluation methods of core inflation, this paper first combines the characteristics of price changes in China to improve existing core inflation measures and applies them to monthly data from 2001 to 2013 in China. Then, it carries out a systematic evaluation to assess the monetary policy application performance of all measures including unbiasedness test, correlation test, tracking trends ability test and forecasting ability test. It turns out that no single method is absolutely perfect while SVAR method, exponential smoothing method and asymmetric trimmed mean method are more satisfactory than others. The conclusion has important implications for implementing monetary policy and exploring core inflation measures in China.