我国货币政策松紧度综合指数的构建——基于网络爬虫和文本挖掘

The Construction of China’s Comprehensive Index of Monetary Policy Tightness ——Based on Web Crawler and Text Mining

本文通过网络爬虫采集数据,基于文本挖掘,从货币政策客观因子与货币政策主观因子角度出发,探索构建货币政策松紧度预期指数。主观因子由两个子因子构成,即从新浪微博文本挖掘得到的货币政策专家指数因子与根据百度指数整理计算得到的货币政策网络指数因子。客观指数也由两个子因子构成,即从不同期限的同业拆借利率得到的Shibor主成分因子与从不同期限的国债收益率得到的国债利率均值因子。研究发现,将专家指数因子拆解为横向判断货币政策取向现状与纵向预期货币政策方向选择,并且赋予主观因子较大权重,赋予客观因子较小权重,以此拟合得到的两个货币政策松紧度综合指数有较好的现实解释力和实践参考价值。

The paper explores the construction of the expected index of monetary policy tightness from the perspective of objective factors and subjective factors of monetary policy by collecting data from web crawlers and text mining.Subjective factors are composed of two sub-factors:the index factor of monetary policy experts obtained from Sina Weibo text mining and the index factor of monetary policy network calculated from Baidu index collation.The objective index is also composed of two sub-factors:the Shibor principal component factor from the interbank lending rate of different periods and the average factor from the yields of Treasury bonds of different periods.It is found that the expert index factor is decomposed into the horizontal judgment of the current situation of monetary policy orientation and the vertical expectation of the direction of monetary policy choice,and the subjective factor is given a greater weight,while the objective factor is given a smaller weight.The two monetary policy tightness composite indices fitted by this method have better practical explanatory power and practical reference value.