本文首先提出了面板数据动态门限回归模型的二阶段合并最小二乘(2SPOLS)估计方法;其次,基于中国29省市自治区1978-2008年的面板数据,对中国通货膨胀和经济增长之间关系的实证分析发现,在一定程度上,我国通货膨胀率对经济增长率的作用存在两个门限值的"双门限效应",其门限值分别为3.2%和15.7%。所以,通货膨胀率位于(0%,3.2%]时,温和通货膨胀对经济增长率存在"托宾效应"。通货膨胀率超过3.2%时,通货膨胀率对经济增长率存在阻碍经济增长的"反托宾效应",尤其当通货膨胀率高于15.7%后,恶性通货膨胀严重阻碍经济"软扩张"。因此,我国通货膨胀率的最优目标区间是(0%,3.2%]。
This paper firstly proposed the two-stage pooled ordinary least square estimations(2SPOLS) of the dynamic threshold panel data regression model and then made the empirical analysis based on the panel data of China's 29 provinces during the period from 1978 to 2008.The results showed that,to a certain extent,there is "double threshold effect" in the relationship between inflation and economic growth and the threshold values are 3.2% and 15.7%.Therefore,mild inflation rate that is less than 3.2% is beneficial to economic growth.However,inflation rate that is more than 3.2% will produce "Anti-Tobin effect".Especially,the inflation will hamper economic growth if it exceeds 15.7%.Therefore,the optimal target interval of the inflation rate is(0 %,3.2%].