政府支出乘数  

Government Spending Multiplier

作　　者：王国静[1,2] 田国强[2,3] 

机构地区：[1]华安证券研究所 [2]上海财经大学经济学院,200433 [3]美国德州A&M大学经济系

出　　处：《经济研究》2014年第9期4-19,共16页Economic Research Journal

基　　金：中国自然科学基金(NSFC-71371117);上海财经大学教育部重点实验室资助

摘　　要：本文考察了动态随机一般均衡(DSGE)框架下估计政府支出乘数时所不能忽视的三个重要特征。这三个特征分别为:政府消费和私人消费之间的埃奇沃斯互补性、政府投资的外部性和财政政策规则的内生性。本文首先利用一个简单的理论模型说明政府消费乘数和政府投资乘数的不同,并指出忽略财政政策规则中的内生性会对支出乘数的估计造成严重的偏误。然后在简单模型的基础上构建一个引入以上三个特征的大型DSGE模型,并利用贝叶斯方法对模型的结构参数进行估计。对不同设定的模型进行比较发现,包含以上三种特征的模型能最好地拟合中国经济。估计得到的长期政府消费乘数和政府投资乘数分别为0.7904和6.1130。脉冲反应分析表明政府消费冲击和政府投资冲击对经济的影响有着很大的不同。This paper investigates three important features on the estimated government spending multiplier within a dynamic stochastic general equilibrium framework. These three important features are Edgeworth complementarity between private consumption and government consumption, externalities produced by government investment, and endogenous fiscal policy rule. First, we work out a simple theoretic model to illustrate the differences between the government consumption multiplier and the government investment multiplier, and point out that the severe bias would result from omitting endogenous fiscal policy rule. Then, based on this simple model, we use Bayesian method to estimate the structure parameters of a large-scale model which augmented with the three important features. By comparing models with different specifications, we find that the model which features the three important effects provides a better description of China＇s economy. The estimated long-run government consumption multiplier and government investment multiplier are 0. 7904 and 6. 1130. The impulse response analysis shows a big difference between a government consumption shock＇s impact and a government investment shock＇s impact.

关 键 词：政府支出乘数 政府消费 政府支出 DSGE建模 

分 类 号：F812.45[经济管理—财政学] F224