浅谈经济增长和税收在Grnger因果关系浅谈经济增长和税收在Granger因果关系.doc

浅谈经济增长和税收存在Granger因果关系 Introduction to economic growth and tax revenue is Granger causality 本文试图经过选取变量构建经济增长与税收之间的回归模型，经过剖析两者之间的Granger因果关系来考证经济增长对税收增长的决议作用。 This article attempts through selecting variables to build regression model between tax and economic growth, through analyzing the Granger causality between the two to research economic growth effect on revenue growth of resolution. 选择变量。本文所选择的变量分别是福建省地域消费总值GDP和税收收入TAX。变量的原始数据GDP和TAX均来源于福建省统计局网站统计年鉴，时间区间为1995—2009。 Choice variables. In this paper, the choice of variables are in fujian province regional GDP and total cost of the consumption TAX of income TAX. Variables GDP and TAX are derived from the primitive data of fujian province bureau of statistics website statistics yearbook, the time interval of 1995-1995. 单位根检验。不存在协整关系的非平稳变量不能停止Granger因果关系检验。因而，首先判别所选取的时间序列数据能否平稳性，对GDP和TAX停止单位根检验。由表一可知GDP和TAX均为非平稳时间序列，因而在对这一组变量停止Granger因果关系检验之前要进一步判别这两者能否存在协整关系。 Unit root test. There is no co-integration relationship between non-stationary variables cant stop the Granger causality test. Therefore, first discriminant can select the time series data of stationarity, on GDP and TAX stop unit root test. Table 1 illustrates the GDP and TAX are non-stationary time series, and so on a set of variables before stop the Granger causality test to further distinguish the two whether there is a cointegration relationship. 检验方式（c,t,m）中c、t、m分别代表ADF检验中包含的常数项、时间趋向项和滞后阶数，0代表不包含,D2代表停止二次差分；时间序列平稳均是在临界值为5%时的结论。 Test methods (c, t, m) in the c, t, m represent the ADF test, paragraphs contained in constant and time trend and lag order number, 0 (does not contain, D2 represent stop secondary difference; Stationary time series are on the threshold of 5% at the time of the conclusion. 协整检验。由于对双变量模型停止协整检验请求两个变量具有相同的单整阶数。从上述单位根检验的结果中，我们曾经得出GDP和TAX都是二阶单整序列的结论。因而，笔者将运用E—G两步法对其停止协整关系检验。第一步，构建估量方程。首先运用OLS法估量协整向量GDP和TAX，可得如下方程： Cointegration test. Due to stop the bivariate model cointegration test request two variables with the same order the wh