Benford法則:中國宏觀經濟統(tǒng)計數(shù)據(jù)質量評價的一種新范式

摘要： 中國宏觀經濟統(tǒng)計數(shù)據(jù)的真實性和數(shù)據(jù)質量一直以來飽受質疑,特別是2008年美國金融危機以來,更受到世界各國和許多國內研究者的批評.引進、鑒證并擴展了Benford法則的分布擬合和宏觀統(tǒng)計數(shù)據(jù)的數(shù)據(jù)隨機性質量評價能力,對我國統(tǒng)計系統(tǒng)公布的宏觀統(tǒng)計數(shù)據(jù)進行了客觀、真實和有效的評價.研究表明,在0.05的置信水平下,我國的國民經濟核算、政府財政統(tǒng)計、金融業(yè)、國際收支平衡這四個宏觀經濟部門的主要經濟指標不存在人為操縱和修正,統(tǒng)計數(shù)據(jù)質量有了顯著提升.
Data quality is the lifeline of the statistical work,however,the authenticity of China's macroeconomic statistics and the data quality has been questioned from many aspects.especially since the US financial crisis in 2008,which receive more criticism from the researchers from the world and the inland.This thesis introduces,verifies and broadens distribution fitting and the competence of evaluating the randomness quality of macroeconomic statistical data.What is more,it evaluates the macroeconomic statistical data of China's statistics system objectively,authentically and effectively.The research has improves that,in 95% confidence level,the four macroeconomic economic departments' main economic indicators do not exist to human manipulations and corrections including China's National Economy Accounting,Government Financial Statistics,Financial Industries and Financial Balance of International Payments.Further more,the data quality of statistics department has been improved continuously and significantly.