英语论文网

tribute to the FDI enterprises which may have contributed more than 40% to its growth. Figure 2 shows a positive relationship between FDI and GDP.

Additionally, between these periods 1980-1988, output of the economy grew however, it is evident that the trend for China has been growing at a continuous rate and is currently counted as one of the fastest developing nations. Figure 2, 3, 4 & 5 also show a positive relationship with GDP.

5.2. Ordinary Least Square (OLS) Regression

The next step is the regression analysis; it explains the movements of the dependent variable in this case real GDP in terms of a set of other variables (independent or explanatory variables) through the quantification of a single equation.

The correlation matrix (as seen in the appendix) shows a strong linear positive statistical association or relationship between the real GDP explanatory variables. The L in front the variables signifies that they are logged, as in LGDP, LFDI, LLF, LH, and LGCF. The reason variables are logged is simply because log variables are invariant to scale of the variables (see equation 3.1 and 3.2) and they give a direct estimate of elasticity.

The main focus of this study is to find the relationship between FDI and economic growth (GDP) hence, it is essential to focus on FDI and GDP first before considering others.

Initial estimation of the relationship between FDI and GDP reported in Table 1 revealed evidence of serial correlation (as shown in Table 2) given by the Breusch-Godfrey Serial Correlation LM Test. Similarly there was evidence of heteroskedasticity given by the White test (as shown in Table 3) however, Figure 6 shows that the errors are normally distributed.

Heteroskedasticity is the violation of the classical assumption model which has severe consequences for the OLS. It occurs when random variable have different variances therefore causing OLS to underestimate the variances and standard errors of the coefficients. The t-Statistic and F-statistic scores become higher than they should be. The white test is the most used test for heteroskedasticity. It is tested using the 'F-statistic' to test the overall significance of the equation, if any of the coefficients is significantly different from zero. There is evidence of heteroskedastic patterns in the residuals. Accepting the null hypothesis (H0) implies that there is no evidence of heteroskedasticity.

To correct the heteroskedasticity and serial correlation we carry out a White Heteroskedasticity-Consistent standard Errors & Covariance white test as shown in Tables 4 & 5. (A.H. Studenmund, Using Econometrics, A Practical Guide, 5th Edition, Pearson International Edition)

The results in Tables 4 and 5 show that is the regression estimate of the relationship between FDI and GDP. The coefficient on FDI is significant at a 1% level and it shows that a 1% increase in FDI leads to a 0.35% increase in real GDP. The adjusted R2 is 0.848 which means that 84% of variation in real GDP around its mean is explained by FDI.

Finally, other variables i.e. human capital, labour force and gross fixed capital are included in the regression model and the results are reported in Table 6. Analysis of the residuals show that there is evidence of serial correlation (see Table 7) but no heteroskedasticity (see Table 8) also the residuals are