摘要
金融资金流量矩阵表具有重要的应用价值,然而我国资金流量表只采用账户表式,而且数据滞后严重,使其应用价值难以实现。本文在对资金流量矩阵表表式结构、账户分类和平衡关系进行深入理解的基础上,将计量经济模型中的多元回归模型、时间序列模型、状态空间模型及RAS模型有效地组合在一起,设计了一套延长金融资金流量矩阵表的组合模型,然后应用该模型将我国金融资金流量矩阵表从2009年延长至2012年。经后验法检验,数据质量达到了一个较为理想的状态。这套模型可以作为官方统计机构或民间研究机构延长资金流量表数据的通用方法。
Flow of Funds Matrix has important application value, but in China only account tables are adopted and the data lag so seriously that it is difficult to make its practical significance come true. Based on the deep understanding of table structure, accounts classification and balance relations about flow of funds matrix, this paper designed a combined model which is used to forecast flow of funds matrix by an effectively combining of econometric models such as multiple regression, time series model, state space model and RAS method, and then we used it to extend China' s flow of funds matrix from 2009 to 2012. The posteriori testing result suggested that the forecasting data have a comparatively good performance. This combined model can be generally used by official statistics agencies or private research institutions to extend flow of funds data.
出处
《统计研究》
CSSCI
北大核心
2014年第6期3-11,共9页
Statistical Research
基金
国家社会科学基金重点项目"中国社会核算矩阵研究"(10ATJ001)
教育部人文社会科学青年项目"中国‘部门×部门’社会核算矩阵的编制与应用--部门间经济流量关系及变化趋势研究"(11YJC910001)
全国统计科学研究(计划)重点项目"我国金融资金流量矩阵表的编制与分析"(2009LZ014)的资助