外文翻译---外商直接投资能长期促进经济增长吗(编辑修改稿)

外文翻译---外商直接投资能长期促进经济增长吗(编辑修改稿)内容摘要：

o hypothesis that FDI does not influence economic growth can be rejected which implies adoption of the opposite hypothesis, ., that FDI affects economic growth. Similarly, a nonrejection of the zero hypothesis that economic growth does not influence FDI implies that we cannot adopt the opposite hypothesis that economic growth affects FDI. In short, the assumption that FDI affects economic growth is confirmed, reverse causality, . that growth attracts FDI, is rejected. Table3 Granger Causality Tests Sample 1:64 Lags: 1 Null Hypothesis Observations FStatistics Probability GDP does not Granger cause FDI/GDP 56 FDI/GDP does not Granger cause GDP Sample 1:64 Lags: 2 Null Hypothesis Observations FStatistics Probability GDP does not Granger cause FDI/GDP 48 FDI/GDP does not Granger cause GDP Despite significant statistics of the standard Granger causality tests there are two major drawbacks. First, pooling implies that the underlying causal structure is the same for each crosssectional unit or each candidate country。 if this were not the case, we would get inconsistent estimates. Second, the standard Granger causality test does not provide for the sign of the relationship. Sims version of the causality test which allows for estimating ‘individual effects’ in the form of a countryspecific intercept or slope is therefore applied by estimating the following two equations: rGDP = a0 + a1FDI–1 + a2FDI + a3FDI+1 + a4GDP0 + ajDUMj (1) FDI = b0 + b1rGDP–1 + b2rGDP + b3rGDP+1 + b4GDP0 + bjDUMj (2) where rGDP is the growth rate of gross domestic product, FDI is the share of foreign direct investment in GDP, GDP0 is the initial GDP per capita in percentage of EU average, and 6 DUMj are country dummies. The results are in Table 4. Country specifics were largely absorbed by GDP0 uttering the level of development。 the country dummies were therefore insignificant except for Estonia and Poland in the first equation, and Czech Republic and Poland in the second equation. In short, the results confirm the causality link found by the standard Granger test。 the tvalue of the coefficient of the lagged FDI ratio is significant at the 1 percent level and higher than the tvalue of the coefficients for lead and current FDI, and the opposite is true for the coefficients for lag and lead of rGDP. Table4 Causality Tests RGDP=f (FDI, X j ) ai ta FDI=g ( rGDP, Xj ) bi tb Constant Constant FDI1 RGDP1 FDI rGDP FDI+1 rGDP+1 GDP0 GDP0 CzechRepublic Estonia Poland Poland R2 and DW R2 and DW To estimate the robustness of the causality tests and of the coefficients in the simple regression model with growth being the dependent and FDI the independent variable, eight sub samples (in each one country was omitted) were formed. The results are in Table 5. Table5 Granger Causality Tests and Simple Regression on Narrower Sets 7 Null hypothesis rGDP does not cause FDI RGDP = a + bFDI RGDP = a + bFDI–1 FDI does not cause rGDP Lag 1 Lag 2 Const. Slope Const Slope eight countries without Czech R. without Estonia without Hungary without Latvia without Lithuania without Poland without Slovakia without Slovenia Despite the robustness of the negative sign for FDI and the solid rejection of the reversed causality the results could be disputed by the arguments of FDI endogeneity a。