基于Benford法则的社会消费品零售总额统计数据质量检验

Quality Test of China’s Total Retail Sales Statistics Based on Benford Law

社会消费品零售总额是表现国内消费需求最直接的数据,其准确与否涉及到能否正确反映我国消费发展状况。基于国家统计局官方公布的月度数据,文章采用Benford法则的第一位数检验法和差分检验法分别检验了社会消费品零售总额统计数据质量,得到截然相反的结果。检验结论表明:Benford法则检验的应用研究需要进一步推进。从总体看,社会消费品零售总额首位数字的实际频率和Benford期望频率偏差有缩小趋势,差异值较小,可以认为调查统计的数据质量较为可靠。

Total retail sales of social consumer goods are the most direct data showing domestic consumption. Whether it is accurate relates to whether it can correctly reflect the consumption development in China. Based on the monthly retail sales data officially released by the National Bureau of Statistics, this paper uses Benford first digit test method and differencing test method to respectively test the quality of statistic data of total retail sales of consumer goods, only to get the opposite result. The test results show that the application research of Benford law for test needs to be further promoted. On the whole, the deviation between the ac tual frequency and Benford's expected frequency of the first digit of total retail sales of consumer goods has a narrowing trend, and the difference value is small. Thus it can be considered that the data quality of the survey statistics is relatively reliable.