消费需求分析中的模型选择--以农村居民商品与服务消费需求为例  

Model Selection in Consumption Demand Analysis--A Case Study of Rural Residents’Consumption Demand for Goods and Services

作　　者：问锦尚 朱文博 郑志浩 Wen Jinshang;Zhu Wenbo;Zheng Zhihao(Institute of Agricultural Economics and Development,Chinese Academy of Agricultural Sciences,Beijing 100081,China;Institute of Rural Development,Chinese Academy of Social Sciences,Beijing 100732,China;Beijing Food Safety Policy&Strategy Research Base,China Agricultural University,Beijing 100086,China)

机构地区：[1]中国农业科学院农业经济与发展研究所,北京100081 [2]中国社会科学院农村发展研究所,北京100732 [3]中国农业大学北京食品安全政策与战略研究基地,北京100086

出　　处：《统计与决策》2024年第5期56-61,共6页Statistics & Decision

基　　金：中国农业科学院科技创新工程项目(10-IAED-08-2023);中国社会科学院创新工程项目(2024NFSB09);中国社会科学院“青启计划”项目(2024QQJH109);北京市社会科学基金研究基地项目(18JDGLB046)。

摘　　要：文章基于2000—2021年我国31个省份农村居民商品与服务消费支出的面板数据,对施加了理论约束进而能够反映消费者选择公理的双对数(Double-Log)模型、线性支出系统(LES)模型、鹿特丹(Rotterdam)模型、近乎完美需求系统(AIDS)模型和二次近乎完美需求系统(QUAIDS)五个完整需求系统模型进行了样本内和样本外的商品与服务消费需求预测评估研究。结果显示,QUAIDS模型的预测表现最好,AIDS模型和Dou-ble-Log模型次之,LES模型和Rotterdam模型相对较差;模型函数式方法的预测误差明显大于弹性均值方法的预测误差,弹性均值预测方法优于模型函数式预测方法。根据经济变量和非经济变量弹性绝对值大小、显著性以及正负符号是否符合经济学理论和现实观察,有助于判断模型选择的恰当性,但不能据此确定模型的优劣,需要通过样本内和样本外的预测评估,特别是样本外的预测评估选取适用的需求分析模型。Based on the panel data of rural residents’consumption expenditure on goods and services in 31 provinces of Chi-na from 2000 to 2021,five complete demand system models,including Double-Log model,LES model,Rotterdam model,AIDS model and QUAIDS,which can reflect the axiom of consumer choice by applying theoretical constraints,are analyzed in-and out-of-sample goods and services consumption demand forecast evaluation research.The results go as below:The prediction per-formance of QUAIDS model is the best,followed by AIDS model and Double-Log model,and LES model and Rotterdam model are relatively inferior.The prediction error of the model function method is obviously greater than that of the elastic mean method,and the elastic mean method is better than the model function method.Whether the absolute values,significance and positive and neg-ative signs of the elasticity of economic variables and non-economic variables conform to economic theory and reality helps to judge the appropriateness of model selection,but it cannot be used as a basis to determine the merits of the model.Therefore,it is necessary to select an appropriate demand analysis model through in-sample and out-of-sample prediction evaluation,especially out-of-sample prediction evaluation.

关 键 词：模型选择 弹性均值方法 消费需求 农村居民 

分 类 号：F063.2[经济管理—政治经济学]