**Relationship BETWEEN FDI AND EXPORTS**

**Mean Differences Using the Independent sample T-test**

Mean H | Mean L | Mean difference | T ratio | Sig. | ||

1990 | ||||||

FDI per capita | 51090.40759 | 2622.03559 | 48468.37200 | 2.460 | 0.000 | |

Export per capita | 59056.64042 | 4997.21796 | 54056.64042 | 2.960 | 0.000 | |

GDP per capita | 43264.0650 | 333389.84420 | 399252.2208 | 2.221 | 0.001 | |

2007 | ||||||

FDI per capita | 333296.0003 | 49712.56448 | 305105.5897 | 2.799 | 0.000 | |

Export per capita | 205232.8941 | 28190.41055 | 155520.3296 | 2.519 | 0.004 | |

GDP per capita | 1067668.551 | 166582.3215 | 901086.2295 | 2.044 | 0.005 | |

2014 | ||||||

FDI per capita | 389665.3619 | 66275.77171 | 389665.3619 | 2.357 | 0.01 | |

Export per capita | 261173.7262 | 90509.4320 | 170664.2942 | 1.835 | 0.07 | |

GDP per capita | 1242393.99 | 385942.9448 | 856451.0452 | 1.455 | 0.07 |

The independent t-test shows the mean difference between the groups L and H, the T ratio the mean and the level of significance is used determine when to reject a null hypothesis while testing the significant difference.

H_{0}: There is no significant difference in means H and L

H_{1}: There is a significant difference in means H and L(Libguides.library.kent.edu, 2017).

In all the cases there is a difference in the means of H and L, but is the difference statistically significant?

In 1990 and 2007, the sig value is less than 0.05. Therefore, we reject H_{0} and conclude that there is the significant difference in the means H and L of the categories FDI, GDP and Exports per capita.

In 2014, Export per capita and GDP per capita had no significant difference since 0.07>0.05, we, therefore, do not reject the null hypothesis while the FDI per capita has a significant difference between the means since 0.01<0.05(Hinton, 2004).

**Correlation between the FDI, GDP, and Exports per capita Through the Years**

FDI per capita | Exports per capita | GDP per capita | ||

1990 | ||||

FDI per capita | 1 | 0.884 | 0.928 | |

Exports per capita | 0.884 | 1 | 0.842 | |

GDP per capita | 0.928 | 0.842 | 1 | |

2007 | ||||

FDI per capita | 1 | 0.697 | 0.938 | |

Exports per capita | 0.697 | 1 | 0.766 | |

GDP per capita | 0.938 | 0.766 | 1 | |

2014 | ||||

FDI per capita | 1 | 0.681 | 0.287 | |

Exports per capita | 0.681 | 1 | 0.851 | |

GDP per capita | 0.287 | 0.851 | 1 |

Conclusion

The stated correlation coefficients indicate the change in variable 1 as a result of a unit change in variable 2 (Anon, 2017).

In 1990 and 2007, all the variables have a strong positive correlation with each other. This is because the correlation coefficient is greater than 0.05. FDI per capita has a strong positive correlation of 0.884 with the Exports per capita(Hinton, 2004).

In 2014: FDI per capita has a strong positive correlation with the Exports per capita.

FDI per capita has a weak positive correlation of 0.287 with the GDP per capita.

GDP per capita has a strong positive correlation of 0.851 with the GDP per capita(SPSS 15.0 brief guide, 2006).

**Regression on the ln of the variables**- Regression on FDI only

Coefficient values | Significance of coefficients | R^{2} |
|||

B_{0} |
B_{1} |
||||

1990 | |||||

lnFDI per capita | 3.618 | 0.628 | 0.000 | 0.577 | |

2007 | |||||

lnFDI per capita | 0.197 | 0.945 | 0.000 | 0.691 | |

2014 | |||||

lnFDI per capita | -1.332 | 1.064 | 0.000 | 0.707 |

Comments

H_{0}: The coefficient is not significant in predicting the Exports per capita

H_{1}: The coefficient is significant in predicting the Exports per capita

Using the linear regression model, I was able to check the significance of coefficients in predicting the Exports per capita variable and the same time finds the coefficients of the various variables.

The sig value of the years 1990, 2007 and 2014, 0.000<0.05, we thereby reject the null hypothesis and state that all the coefficients are significant in predicting Exports per capita.

TheB_{0 }indicates the values the change in Export per capita whenever the FDI per capita is zero. B_{1} shows the change in Exports per capita for every unit increase in the FDI per capita.

R^{2} depicts the percentage of the model that predicts the Y variable. In 1990, R^{2}=57.7% indicating that the model explains 57.7% of the Exports per capita. In 2007, R^{2}=69.1% indicating that the model explains 69.1% of the Exports per capita. In 2014, R^{2}=70.7% indicating that the model explains 70.7% of the Exports per capita.

- Regression on FDI and GDP

Coefficient value | Significance of coefficient | R^{2} |
||

1990 | ||||

Constant | -1.725 | 0.000 | 0.875 | |

lnFDI per capita | 0.103 | 0.000 | 0.875 | |

lnGDP per capita | 0.915 | 0.000 | 0.875 | |

2007 | ||||

Constant | -2.614 | 0.000 | 0.897 | |

lnFDI per capita | 0.016 | 0.000 | 0.897 | |

lnGDP per capita | 1.091 | 0.000 | 0.897 | |

2014 | ||||

Constant | -2.537 | 0.000 | 0.821 | |

lnFDI per capita | 0.737 | 0.000 | 0.821 | |

lnGDP per capita | 0.520 | 0.000 | 0.821 |

H_{0}: The coefficient is not significant in predicting the Exports per capita

H_{1}: The coefficient is significant in predicting the Exports per capita

The sig value of the years 1990, 2007 and 2014, 0.000<0.05, we thereby reject the null hypothesis and draw the conclusion that all the coefficients are significant in predicting Exports per capita.

The Constant indicates the values the change in Export per capita whenever the FDI per capita is zero. lnFDI per capita indicates the change in Exports per capita for every unit increase in the FDI per capita all factors held constant. lnGDP per capita indicates the change in Exports per capita for every unit increase in the FDI per capita all factors held constant(Huizingh, 2007).

In 1990, R^{2}=87.5% indicating that the model explains 87.5% of the Exports per capita. In 2007, R^{2}=89.5% indicating that the model explains 89.5% of the Exports per capita. In 2014, R^{2}=82.1% indicating that the model explains 82.1% of the Exports per capita.

**Regression per Sample**

L countries

Coefficient value | Significance of coefficient | R^{2} |
||

1990 | ||||

Constant | -1.371 | 0.000 | 0.859 | |

lnFDI per capita | 0.154 | 0.000 | 0.859 | |

lnGDP per capita | 0.844 | 0.000 | 0.859 | |

2007 | ||||

Constant | -2.925 | 0.000 | 0.903 | |

lnFDI per capita | 0.210 | 0.000 | 0.903 | |

lnGDP per capita | 0.958 | 0.000 | 0.903 | |

2014 | ||||

Constant | -2.734 | 0.000 | 0.877 | |

lnFDI per capita | 0.848 | 0.000 | 0.877 | |

lnGDP per capita | 0.393 | 0.000 | 0.877 |

H countries

Coefficient value | Significance of coefficient | R^{2} |
||

1990 | ||||

Constant | -2.717 | 0.000 | 0.858 | |

lnFDI per capita | 0.030 | 0.000 | 0.858 | |

lnGDP per capita | 1.098 | 0.000 | 0.858 | |

2007 | ||||

Constant | -2.537 | 0.000 | 0.859 | |

lnFDI per capita | -0.111 | 0.000 | 0.859 | |

lnGDP per capita | 1.197 | 0.000 | 0.859 | |

2014 | ||||

Constant | 1.010 | 0.000 | 0.855 | |

lnFDI per capita | 0.083 | 0.000 | 0.855 | |

lnGDP per capita | 1.025 | 0.000 | 0.855 |

From the tables above, it is clear that the L models predict the Exports per capita better than the H values. This is due to the increasing R^{2} values from L to H. The coefficients remain significant in predicting the Y variable.

**E. Regression as per the Dummy Variables.**

Comments on L

H_{0}: The coefficient is not significant in predicting the Exports per capita

H_{1}: The coefficient is significant in predicting the Exports per capita(Anon, 2017)

The sig value of the years 1990, 2007 and 2014, 0.000<0.05, we thereby reject the null hypothesis and conclude that all the coefficients are significant in the prediction of the variable Exports per capita.

The Constant indicates the values the change in Export per capita whenever the FDI and GDP per capita are zero. lnFDI per capita indicates the change in Exports per capita for every unit increase in the FDI per capita all factors held constant. lnGDP per capita indicates the change in Exports per capita for every unit increase in the FDI per capita all factors held constant(Anon, 2017).

In 1990, R^{2}=85.9% indicating that the model explains 85.9% of the Exports per capita. In 2007, R^{2}=90.3% indicating that the model explains 90.3% of the Exports per capita. In 2014, R^{2}=87.7% indicating that the model explains 87.7% of the Exports per capita(Liu et al., 2003).

Comments on H

H_{0}: The coefficient is not significant in predicting the Exports per capita

H_{1}: The coefficient is significant in predicting the Exports per capita

The sig value of the years 1990, 2007 and 2014, 0.000<0.05, we thereby reject the null hypothesis and conclude that all the coefficients have significance in the prediction of Exports per capita.

The Constant indicates the values the change in Export per capita whenever the FDI per capita is zero. lnFDI per capita indicates the change in Exports per capita for every unit increase in the FDI per capita all factors held constant. lnGDP per capita indicates the change in Exports per capita for every unit increase in the FDI per capita all factors held constant.

In 1990, R^{2}=85.8% indicating that the model explains 85.8% of the Exports per capita. In 2007, R^{2}=85.9% indicating that the model explains 85.9% of the Exports per capita. In 2014, R^{2}=85.5% indicating that the model explains 85.5% of the Exports per capita.

**References**

Anon, (2017). [online] Available at: https://statistics.laerd.com/spss_tutorials/linear-regression-using-spss-statistics.php [Accessed 4 Apr. 2017].

Hinton, P. (2004). SPSS explained. 1st ed. London: Routledge.

Huizingh, E. (2007). Applied statistics with SPSS. 1st ed. London: SAGE.

Libguides.library.kent.edu. (2017). LibGuides: SPSS Tutorials: Independent Samples t Test. [online] Available at: http://libguides.library.kent.edu/SPSS/IndependentTTest [Accessed 4 Apr. 2017].

Liu, R., Kuang, J., Gong, Q. and Hou, X. (2003). Principal component regression analysis with spss. Computer Methods and Programs in Biomedicine, 71(2), pp.141-147.

SPSS 15.0 brief guide. (2006). 1st ed. Chicago, Ill.: SPSS Inc.

Anon, (2017). Using SPSS for Correlation. [online] Available at: http://http://academic.udayton.edu/GregElvers/psy216/spss/cor.htm [Accessed 4 Apr. 2017].