恩格尔定律统计测度研究:理论反思与方法比较  

The Adaptability of Engel’s Law in China:Theoretical Reflection and Retest

作　　者：陈梦根[1] 战楠 周元任 CHEN Menggen;ZHAN Nan;ZHOU Yuanren(School of Statistics,Beijing Normal University,Beijing 100875,China)

机构地区：[1]北京师范大学统计学院,北京100875

出　　处：《统计与信息论坛》2023年第11期3-16,共14页Journal of Statistics and Information

基　　金：国家社会科学基金重大项目“数字经济对我国经济社会发展的影响效应测度与统计评价”(19ZDA118);北京市自然科学基金面上项目“北京市数字经济高质量发展水平测度、路径选择与治理体系研究”(9232012)。

摘　　要：以往研究大多基于宏观数据从线性相关的角度检验恩格尔系数与收入之间的关系,但从统计上看这种方法在表征恩格尔定律的内涵时并不完全准确。因此,以中国家庭收入项目(CHIP)调查中城镇、农村以及外来务工住户的收支数据为样本,采用四种更为恰当的统计方法进行实证检验,考察恩格尔定律在中国的适用性。研究表明,恩格尔定律实际应表述为收入与食品支出份额之间的反向随机关系,总体上恩格尔定律在中国是适用的,随着居民收入的增加,家庭食品支出占比呈显著下降趋势。部分群体中收入与食品支出占比的关系呈现先递减后上升的U型趋势,相关关系存在一定的收入阈值,且不同群体的阈值存在差异,这种U型变动在发达地区的高收入群体中表现更加明显。同时,城镇地区中收入与食品支出占比的负相关程度大于农村地区,东部地区的收入与食品支出占比的负相关程度大于中、西部,家庭中工作人数占比越大、家庭人口越少、女性占比越少、金融资产越多,食品支出占比随收入增加而下降的幅度越大。进一步的研究发现,恩格尔定律的适用性主要受群体间消费习惯、文化习俗等因素影响。Previous studies on Engel’s law mostly tested the relationship between Engel’s coefficient and income from the perspective of linear correlation,but from a statistical point of view,this method was not completely accurate in characterizing the connotation of Engel’s Law.In fact,Engel’s Law should be expressed as a reverse random relationship between income and food share.Using the income and expenditure data of urban,rural and migrant households collected from 2018 China Household Income Project(CHIP)survey,four appropriate statistical methods is employed to conduct empirical tests,including descending regression,Kendall’sτcoefficient,Lehmann quadrant dependence and Tukey stochastic decreasing conditional distribution function,and intends to empirically explore the relationship between food share and income,as well as the applicability of Engel’s Law in China.The results show that Engel’s law should be expressed as the negative stochastic association or dependence between income and food share.In general,Engel’s law is applicable in China.The specific conclusions are as follows:In some groups,the relationship between income and food share presents a U-shaped trend,which firstly decreases and then increases,indicating that there is a certain“income threshold value”for stochastic association,and the threshold values of different groups are different.The U-shaped trend is more obvious among high-income groups in developed regions.At the same time,the negative stochastic association between income and food share in urban area is greater than that in rural area,and the negative stochastic association in the eastern region is greater than that in the middle and western region.The larger the proportion of working people in a family,the smaller the number of people and the proportion of women in a family,and the more the financial assets of a family holds,the greater the food share in a family decreases with the increases of income.Also,for households in urban areas and eastern regions with a higher

关 键 词：恩格尔定律 收入 食品支出占比 反向随机关系 统计检验 

分 类 号：C812[社会学—统计学]