CHAPTER 4: ESTIMATION RESULT ANALYSIS AND DISCUSSION
4.2 Regression results and discussion
4.2.2 Impact of financial development on economic growth
In order to investigate the impact of the financial development (FD) on economic growth (EG), we treat economic growth (EG) as the dependent variable. The independent variables are the set of determinant variables of the economic growth (EG), such as Foreign direct investment (FDI); General government consumption (GOV); Trade openness (TRAO). Determinants of the financial development are also included in the right hand side of the estimated equation (5).
In table 5, the result suggests that the variables of the Ratio of Broad money to GDP (M2/GDP), Ratio of credit offered by the bank to the private sector to GDP (BANKCREDIT/GDP), Ratio of domestic credit to the private sector to GDP (DOMCREDIT/GDP), and Gross domestic savings as percentage of GDP (GROSSDSA) are highly correlated amongst themselves for most developing countries in the sample. Therefore, we intend to estimate four separate regressions to explore the impact of financial development on economic growth as suggested by Hansan et al., (2011).
Moreover, we implement test of multicolinearlity problem by using correlate command and VIF command to view the cross correlation between the variables of Broad money to GDP (M2/GDP), Ratio of credit offered by the bank to the private
Page 60
sector to GDP (BANKCREDIT/GDP), Ratio of domestic credit to the private sector to GDP (DOMCREDIT/GDP), Gross domestic savings as percentage of GDP (GROSSDSA), and Trade openness (TRAO). The output of VIF test is 17.30 and the ratio of the correlation of IMPORT and EXPORT is high (i.e. 97.15%) (see Appendix C.3.4). these results indicate that IMPORT and EXPORT variables are highly correlated. This high correlation leads to the multicolinearlity problem in models.
To diminish the multicolinearlity problem, we intend to exclude IMPORT and EXPORT variables in the regression models. We first obtain the estimation with IMPORT and EXPORT variables in the regression models to compare them with the estimation without IMPORT and EXPORT variables. The regression results, shown in Appendix C.3.5, indicate that the coefficient of the regressors (DOMCREDITLAG1, GROSSDSA, M2GDP BANKCREDITGDP, GDPGRTHLAG1, IMPORT, EXPORT, GOV, FDI ) are not all that different from the estimate excluding IMPORT and EXPORT variables. We conclude that the multicolinearlity problem in this case is acceptable.
Therefore, we modify the equation (5) as follows:
Model 1:
EGit = αo + α1DOMCREDIT/GDP it + α2FDIit + α3GOVit + α4IMPORTit + α5EXPORTit-1 + εit Model 2:
EGit = αo + α1BANKCREDIT/GDP it + α2FDIit + α3GOVit + α4IMPORTit + α5EXPORTit-1 + εit
Model 3:
EGit = αo + α1 M2/GDPit + α2FDIit + α3GOVit + α4IMPORTit + α5EXPORTit-1 + εit
Model 4:
EGit = αo + α1 GROSSDSA it + α2FDIit + α3GOVit + α4IMPORTit + α5EXPORTit-1 + εit
Page 61
Table 7: Regression Results of Economic Growth and Financial Development Dependent variable:
economic growth (EG)
Model 1 Model 2 Model 3 Model 4
C 5.128***
(0.563) p-value=0.000
5.280 ***
(0.576) p-value=0.000
5.292***
(0.573) p-value=0.000
-0.2455 (0.7867) p-value=0.755 DOMCREDITLAG1 0.0216 **
(0.008) p-value =0.007
BANKCREDITGDP 0.0127**
(0.006) p-value=0.035
M2GDP 0.0272***
(0.0054) p-value=0.000
GROSSDSA 0.2443***
(0.0198) p-value=0.000
FDI 0.01988***
(0.0528) P-value =0.000
0.226***
(0.0232) p-value=0.000
0.2100***
(0.0535) p-value=0.000
0.1884***
(0.0531) p-value=0.000
GOV -0.1669***
(0.0416) p-value=0.000
-0.1857***
(0.0419) P-value=0.000
-0.2230***
(0.0427) p-value=0.000
-0.0903**
(0.0419) p-value=0.032
IMPORT 0.0557**
(0.0229) p-value=0.015
0.0588**
(0.0232) p-value=0.012
0.0556 **
(0.0235) p-value=0.019
0.2429***
(0.0291) p-value=0.000
Page 62
EXPORT -0.0627***
(0.0194) p-value =0.001
-0.0599**
(0.0197) p-value=0.002
-0.0656***
(0.0203) p-value=0.001
-0.2552**
(0.0281) p-value=0.000
R-squared 0.0983 0.0952 0.1288 0.2792
Robustified DWH (testing endogeneity)
0.113956 (p- value = 0.7358)
Hausman test Chi2(9) = 39.00
(p- value = 0.0000)
Observations 457 479 479 483
Note: * Significant at 10% ; ** Significant at 5% ; *** Significant at 1%
The results from the four equation regression in Table 7 suggest that the determinants of financial development (M2/GDP, BANKCREDIT/GDP, DOMCREDIT/GDP and GROSSDSA) are significant at the 1% level and 5% level respectively, and positively impact on the growth rate of GDP per capita. This result is consistent with the previous empirical findings that if the system of finance is healthy, it will have a positive impact on economic growth due to the efficient allocation of savings and investment, which enhance higher economic growth (Zagorchev et. al.,2011; Masten et al., 2008; Calderón and Liu, 2003; Christopoulos and Tsionas,2004;Hassan et al.,2011).
The ratio of general government final consumption expenditure (GOV) is significant at the 1% level in four models. This indicator is negative for growth rate of GDP per capita as expected. The negative coefficient of GOV implies that the degree of government spending declines due to crowding out effect on private investment and consumption. In other words, to enhance the private investments, developing countries are likely to reduce the amount of government expenditure over the years and privatize the state owned enterprises, which are not an efficient operation (Zagorchev et.al., 2011).
Page 63
The ratio of foreign direct investment as percentage of GDP (FDI) is significant at the 1% level in four models and positive for growth rate of GDP per capita as expected.
This indicator implies that it has a positive effect on economic growth if the financial system of the recipient countries have reached a relatively high level of financial development (Lee, C.C., & Chang, P. C. 2009). According to Zagorchev et. al., 2011, FDI enhances the financial integration process and then further the development of the financial sector. Therefore, the flow of FDI support GDP growth. FDI positively impacts on financial development (Zagorchev et. al., 2011).
The ratio of trade openness to GDP (TRAO). This ratio consists of the imports of goods and services as a percentage of GDP (IMPORT) and the exports of goods and services as a percentage of GDP (EXPORT). In four models, IMPORT is significant at the 1% level and 5% level respectively, and positively impact on the growth rate of GDP per capita. But it is surprising that EXPORT has a significant and negative effect on economic growth at the 1% level and the 5% level. This result contradicts the finding in Hassan et al.,(2011) that trade openness is expected to have a positive impact on economic growth through free international trade. This result may lead to a misunderstanding that EXPORT does not enhance growth.
The p-value of the Durbin-Wu-Hausman test is 0.7358 (> 0.05). We do not reject the null hypothesis (H0) that the variables are exogenous. This indicates that there is no endogeneity occurred. Therefore, it is suggested that Generalized method of moment with instrumental variables approach is less efficient than OLS estimators. Hence, the OLS estimation technique in this case is more appropriate than Generalized method of moment with instrumental variables approach (Baum et.al., 2003; 2007; Cameron, A.
C., & Trivedi, P. K., 2009).
Page 64