摘要
在国内国际双循环相互促进的新发展格局下,加快消费升级对于我国经济高质量发展具有重要战略意义。本文以京津冀地区家庭作为研究对象,采用机器学习模型和可解释机器学习算法,深入分析京津冀地区家庭消费升级的动态演化过程及其作用因素的非线性影响。研究结果表明,机器学习模型可以很好地拟合京津冀地区家庭消费升级,且非日常消费的具体细分支出信息对于家庭消费升级的拟合具有决定性作用。在模型拟合过程中,非日常消费细分支出与家庭经济信息、人口信息以及家庭所在城市信息形成了较好的信息互补关系。可解释性分析结果发现,非日常消费细分支出变量的影响存在非线性变化过程,并且不同细分支出对于家庭消费升级的影响方向存在差别。本文认为,为实现京津冀地区消费驱动经济高质量发展的战略目标,有关部门应在提高家庭收入和财富水平的同时,进一步在对家庭消费升级有正向作用的非日常消费细分领域发力,也应当根据所在地家庭的支出情况及其所处影响曲线的不同位置,制定差异化的激励政策,以便更加有效地推动京津冀地区家庭消费升级。
Under the new development pattern of domestic and international dual circulation policy,accelerating the upgrading of household consumption is of great strategic significance for China's high-quality economic development.This paper attempts to take the CFPS household consumption data of Beijing-Tianjin-Hebei region as a typical case,deeply analyzes the dynamic evolution process of household consumption upgrading in Beijing-Tianjin-Hebei region and the nonlinear impact of its influencing factors,based on machine learning models and interpretable machine learning algorithms.The empirical results show that the machine learning models can effectively fit the household consumption upgrading data of Beijing-Tianjin-Hebei region,and the sub-expenditures of non-routine consumption sub-expenditure play decisive roles through the fitting process.The sub-expenditure of non-routine consumption forms a good information complementarity relationship with household economic information,population information and household city information.The interpretable analysis results show that the impact of sub-expenditures of non-routine consumption has a nonlinear change process with various influencing directions on household consumption upgrading.This paper suggests that,in order to achieve high-quality economic development driven by consumption in Beijing-Tianjin-Hebei region,local governments should increase household income and wealth levels while further encouraging the sub-expenditures of non-routine consumption with positive effects on household consumption upgrading.In addition,local governments should formulate differentiated policies to encourage non-routine consumption according to the expenditure situation of local households and their different positions on the impact curve,so as to efficiently promote the household consumption upgrading process."
作者
俞剑
蔡铉烨
黄林
王言泽
JIAN YU;XUANYE CAI;LIN HUANG;YANZE WANG(School of Economics,Central University of Finance and Economics;School of Statistics and Mathematics,Central University of Finance and Economics;Rural Revitalization Department,China Development Bank)
出处
《工信财经科技》
2023年第4期14-30,共17页
Review of Financial & Technological Economics
基金
北京市社会科学基金一般项目“居民财务杠杆升高对北京市家庭消费增长及结构升级的影响”(项目编号:19LJB001)的阶段性成果。
关键词
家庭消费
消费升级
机器学习
京津冀地区
动态演化
household consumption
consumption upgrade
machine learning
Beijing-Tianjin-Hebei region
dynamic evolution
