投入产出模型的合并偏差问题研究  

Researches of Aggregation Bias in Input-Output Model

作　　者：郑正喜 

机构地区：[1]上海财经大学统计与管理学院

出　　处：《数量经济技术经济研究》2016年第1期110-126,共17页Journal of Quantitative & Technological Economics

基　　金：国家社科基金重点项目(11ATZ002);教育部哲学社科研究重大课题攻关项目(11JZD019);福建省社会科学规划项目(2014C065)资助

摘　　要：本文探讨合并偏差的理论性质,指出合并程度不能作为衡量合并偏差幅度的合理标尺,在辨析不同偏差度量方式及其合理性的基础上,分别从产业关联效应和综合度量角度实证研究合并偏差的量化特征。结果发现,产业关联效应标准下,合并偏差的总体幅度取决于被合并产业的异质度和要素重要度,其局部分布则取决于与被合并产业的相关程度;综合度量标准下,综合异质度能够很好地解释综合偏差度,但也不排除存在其他次要影响因素。This paper discusses the theoretical characteristics of aggregation bias, and points out that the degree of aggregation behavior can＇t be served as a objective gauge for ag- gregation bias. On the basis of analyzing different measurement methods of aggregation bias and their reasonability, the paper launches the empirical studies from the angle of industrial linkage effect and comprehensive measurement respectively. The results show that as following. Firstly, under the standard of industrial linkage, the overall aggregation bias depends on the hetero- geneity degree and the element importance of corresponding aggregated sectors, and the lo- cal distribution of aggregation bias rests with the relevance between the object and corre- sponding aggregated sectors. Secondly, under the standard of comprehensive measurement, CHI （Comprehensive Heterogeneity Index） is the main determinant of CBI （Comprehensive Bias Index）, but it doesn＇t rule out other minor factors.

关 键 词：投入产出模型 合并偏差 产业关联效应 综合度量 

分 类 号：F223[经济管理—国民经济]