1
APPENDIX
It is the result after running the random effect model in Eviews. But it is not the final result because the Hausman test shows that we should choose the fixed effect model instead of the random effect model.
Appendix 1 Random effect
Dependent Variable: FDI Sample: 1997 2005 Included observations: 9 Cross-sections included: 28
Total pool (balanced) observations: 252
C 57.34623
(218.0496)
GDP 0.117577**
(0.045996)
WORKERS -1.292034
(1.041550)
WAGE -0.036582
(0.015614)
EDUCATION -53.75008* (29.16003)
EXPORT 0.912433***
(0.055945)
TRAFFIC 0.438605
(0.054582) LOCAL
AGGLOMERATION
0.002995 (0.004457) INTERNATIONAL
AGGLOMERATION
0.124049***
(0.032980) Weighted Statistics
R-squared 0.809693 Mean dependent var 894.8827 Adjusted R-squared 0.803427 S.D. dependent var 2003.750 S.E. of regression 888.3931 Sum squared resid 1.92E+08 F-statistic 129.2352 Durbin-Watson stat 0.244751 Prob(F-statistic) 0.000000
Unweighted Statistics
R-squared 0.854696 Mean dependent var 2600.612 Sum squared resid 7.19E+08 Durbin-Watson stat 0.065247
* Coefficient is significant at the 0.1 level.
** Coefficient is significant at the 0.05 level.
*** Coefficient is significant at the 0.01 level.
Appendix 2
In this table, we should focus on the p-value which is 0.0000 here. It means that the
omission of fixed effect has significant effect on the consistency of the estimates. So,
2
the null hypothesis, here, is the random effect model is rejected and it’s better to use fixed effect model.
Hausman test
Correlated Random Effects - Hausman Test Test cross-section random effects
Test Summary Chi-Sq. Statistic Chi-Sq. d.f. Prob.
Cross-section random 211.855391 8 0.0000
Appendix 3
It’s the result after running the fixed effect model in Eviews. As we choose the fixed effect model, this is the final result.
If the coefficient is positive, it indicates that there is a positive relationship between FDI and the independent variable. And if the coefficient is significant at the 0.05 level, we will accept the hypothesis which means the relationship is supported in my test.
Fixed effect
Dependent Variable: FDI Sample: 1997 2005 Included observations: 9 Cross-sections included: 28
Total pool (balanced) observations: 252 C 712.9904***
(237.8416)
GDP 0.087374***
(0.021352)
WORKERS -0.880644
(1.218218)
WAGE -0.002854
(0.004999)
EDUCATION -2.132996 (9.701810)
EXPORT 0.548159***
(0.064772)
TRAFFIC 0.093311**
(0.038871) LOCAL
AGGLOMERATION
0.017097***
(0.003502) INTERNATIONAL
AGGLOMERATION
0.219245***
(0.022968) Weighted Statistics
R-squared 0.987847 Mean dependent var 3388.271 Adjusted R-squared 0.985878 S.D. dependent var 4119.853 S.E. of regression 489.5932 Sum squared resid 51775521
3 Durbin-Watson stat 0.696532
Unweighted Statistics
R-squared 0.977133 Mean dependent var 2600.612 Sum squared resid 1.13E+08 Durbin-Watson stat 0.320379
* Coefficient is significant at the 0.1 level.
** Coefficient is significant at the 0.05 level.
*** Coefficient is significant at the 0.01 level.
