经济时间序列的频率转换是计量经济分析领域的一个重要研究问题。本文首先对不同经济指标类型(流量、存量和指数)及传统频率转换方法进行了系统梳理;在此基础上,重点介绍了3种低频向高频转换的前沿方法:Denton方法、Chow-Lin方法和Litterman方法,并给出了流量、存量和指数3种类型变量由低频(季度)向高频(月度)转换的实例;最后,对3种频率转换方法的数据转换质量进行了比较分析。研究显示,频率转换后的月度数据都较好地反映季度数据的变化趋势和波动特征,从而通过频率转换方法可以很好地解决由于收集到的数据类型不一致而无法建模的问题。
The frequency-converting of economic time series is of most importance in the field of econometrics analysis. In this paper, based on different indicator types, such as the flow, stock, index and the traditional frequency-converting method, we introduce three different frontier methods respectively, namely Denton, Chow-Lin and Litterman methods. Furthermore, examples of low-frequency (quarterly) convening to high-frequency (monthly) are presented. By comparing and analyzing the data conversion quality of three methods above, the trends and volatilities of quarterly data are well reflected in monthly series. In conclusion, frequency-conversion can effectively solve the model specifying problems due to insufficient data of inconsistent frequencies.