The financial liberalization thesis has experienced what Arestis and Demetriades (1999) have dubbed post hoc theoretical revisions. The devastating results of liberalizations in the 1970s and 1980s induced a first round of revisions. It was linked to the problem of macroeconomic instability and inadequate bank.
For the empirical analysis, I used a time series data and a panel data, both annual data. The time periods covers from 1975 to 2008. As I mentioned earlier, the variables that were used in carrying out the analysis includes, degree of openness, gross domestic product, foreign direct investment, financial deepening.
Before testing carrying out the analysis, I checked for stationarity in the variables, using their major test, and they include the Augmented Dickey Fuller test, Phillip- Perron test and the KPSS Test. I took the first difference of the variables and found out they were stationary. In order to know if the variables tend to move together in the short run, I used the Engel Granger 2 step methods to check for cointegration, and the result showed there were co integrated which means that I can go ahead and run a regression on the variables for the data analysis.
The tables below presents the unit root test for all the variables for Nigeria and Ghana.
Unit root using the augmented Dickey -Fuller test (Nigeria)
VARIABLES
ADF TEST STATISTICS
PROBABILITY
DOP
-1.744
0.4006
FD
-2.209
0.2068
FDI
-0.063
0.9452
GDP
-1.764
0.3906
GNS
1.209
0.9981
Using the p values for all the variables at the level, it showed that they were not stationary, at 5%, 1% and 10%. Thus we do not reject the hypothesis that the variables have unit roots.
Unit root test using the Phillips Perron test (Nigeria)
VARIABLES
PHILLIPS- PERRON TEST STAT
PROBABILITY
DOP
-1.526
0.5078
FD
-2.415
0.1453
FDI
-0.035
0.9483
GDP
-2.005
0.2833
GNS
1.827
0.9996
For testing the unit root using the Phillips Perron test, the criteria for judging was the p values, at 1%, 5% and 10% significance, it showed that the variables are not stationary .Thus we do not reject the hypothesis that the variables have unit roots.
Unit root using the KPSS test (Nigeria)
VARIABLES
KPSS TEST STATISTICS
DOP
0.5852
FD
0.1170
FDI
0.6553
GDP
0.6713
GNS
0.1560
The reason for using the KPSS test was to check if the results from the Phillips and Perron test was actually true or it makes sense, according to Brooks(2008), to make the conclusion of the Augmented Dickey Fuller and Phillip- Perron test more robust, the conclusion you get from these test must be the same as the conclusion from the KPSS test.
The result I got from the KPSS test goes with the result I got from the Augmented Dickey Fuller and Phillip Perron test.
Unit root using the Augmented dickey-fuller test (Ghana)
VARIABLES
ADF TEST STATICSTICS
PROBABILITY
DOP
-2.109
0.2445
FD
-2.193
0.2122
FDI
-0.382
0.9009
GDP
0.140
0.9365
GNS
1.394
0.5725
Using the p values for all the variables at the level, it showed that they were not stationary, at 5%, 1% and 10%. Thus we do not reject the hypothesis that the variables have unit roots.
Unit root using the Phillips Perron test (Ghana)
VARIABLES
PP TEST
PROBABILTY
DOP
2.974
0.0478
FD
2.046
0.2667
FDI
-0.274
0.9183
GDP
-0.238
0.9236
GNS
-1.2044
0.6607
For testing the unit root using the Phillips Perron test, the criteria for judging was the p values, at 1%, 5% and 10% significance, it showed that the variables are not stationary .Thus we do not reject the hypothesis that the variables have unit roots.
Unit root using the KPSS test (Ghana)
VARIABLES
TEST STATISTICS
DOP
0.311
FD
0.589
FDI
0.619
GDP
0.586
GNS
0.619
The result I got from the kpss test goes with the result I got from the Augmented Dickey Fuller and Phillip Perron test. This means that the conclusion is robust.
The next step was to check if the variables were co integrated and using the Engel Granger co integration technique, I found out that there was co integration, which means the variables tend to move together in the short run. This means I can go further to run my regression and carry out my analysis.
Firstly, I am going to start with analysing the result for Nigeria, using ordinary least square regression
The next analysis will be on Ghana's result.
LogGDP=α0+α1LogGNSt+α2LogDOPt+α3LogFDt+ LogFDIt
LogGDP= -16.01 +0.407 LogGNS- 0.091 LogDOP- 0.107 LogFD + 0.102 LogFDI
(1.11) (0.08) (0.03) (0.05) (0.03)
All the variables are statistically significant. From the regression output, it can be shown that Gross National Savings had a positive impact or influence on Gross Domestic Product, this result conforms to the findings of McKinnon (1973) and Shaw (1973). It is commonly believed that since saving is a source of funding for investment which will increase Gross Domestic product, any policy that is designed to stimulate saving, will also stimulate Gross Domestic Product. Degree Of openness which measures financial liberalization had a negative impact on Gross Domestic Product; this can be as a result of weak financial system. The result for financial deepening has a negative impact on Gross Domestic Product, but foreign direct investment showed a positive relationship which shows that despite the fact the economy is opened and with it economic instability , there are still foreign investments and this has had a positive impact in the economy.
LogGDP=α0+α1LogGNSt+α2LogDOPt+α3LogFDt+ LogFDIt + Ut
GDP= 7.8298-.0405 LogGNS + 0.7660 LogDOP - 0.0143 LogFD + 0.2693 LogFDI
(1.39) (0.05) (0.35) ((0.262) (3.610)
R2= 0.90, F-Stat=60.34076
The results of the equation above shows that all the variables are significant except Gross national savings and Financial Deepening, and I did not want to discard the variables, because I used them for my analysis for Ghana's economy. So I tested the null hypothesis that the parameters on these two variables are jointly zero using an F -test. The resulting F test follows F (2, 27) distribution as there are two restrictions. The F statistics value was 63.4379 with P- value 0.000, which means that we can reject the null hypothesis at 5% level of significance and they are highly significant, that means the variables will be retained.
The result above shows that Degree of Openness which is a measure for financial liberalization has a positive impact on Gross Domestic Product, which means that as financial liberalization positively increases the growth of Gross Domestic Product. This in effect is in conformity with the finding of Gallego and Loauza (2002) and also Okpara (2009). The elasticity of coefficient of Degree Of Openness was 0.766, implying that a 1% increase in Gross Domestic Product will lead to a 0.7666% increase in Gross Domestic Product other variables being constant.
The next variable which is the Gross National Savings is negatively related to Gross Domestic Product, which means that national saving as a result of financial liberalization has had a negative impact on Gross Domestic Product. The elasticity coefficient of Gross National Savings was 0.405, implying that a 1% increase in savings will lead to a reduction in Gross Domestic Product by 0.405%. The result for Gross National Savings is in line with the findings of Bennett et al (2001), Ostry and Levy (1995) and Bayoumi (1993) but contradicts that of McKinnon (1973) and Shaw (1973).
Financial deepening has a negative relationship with Gross Domestic Product despite reform policy of the economy, which implies the inability of financial institutions to effectively mobilise savings for investment purposes. A 1% increase in Financial Deepening will lead to a 0.014% decrease in Gross Domestic Product. Financial deepening has been low for Nigeria (Nzotta and Okereke, 2009, pp52). The main features of the financial deepening aggregates during the 22 year period, as evidenced from Table 2 were as presented below. Financial deepening moved from 35.9 in 1986 down to 24.2 in 1992 and increased to 29.7 by 1994. This declined further to 15.3 by 1997 before rising to 32.0 by 2004. The aggregate moved down to 18.0. The trend above clearly shows that the financial deepening index did not experience any dramatic changes during the period. This is despite the various reforms introduced from 1986 which should have a positive effect on financial deepening in Nigeria. The problem with financial deepening was also as a result of the difficulties encountered after several attempts of liberalization in Nigeria , partly because the government abruptly withdrew public sector deposits from the banking system(Aryeetey et all,1997,pp200) Foreign direct investment from the regression output is seen to have had a negative positive influence on Gross Domestic Product, which means that as a result of the openness of the economy, more foreign investment has been encouraged which has resulted in a positive impact tin the economy. This result contradicts with the findings of Okereke (2009), which showed that Foreign Direct Investment exhibited a negative influence on Gross Domestic Product.
PANEL DATA.
LogGDP=α0 - α1LogDOPt + α2LogFDt - α3LogFDIt + Ut
LogGDP= 10.67 - 0.616 LogDOP + 0.695 LogFD - 0.120 LogFDI
Std. Error (0.544) (0.074) (0.113) (0.04) R2=0.9
T-stat (19.59) (8.27) (6.11) (2.44) F-stat= 889 p(F-stat)=0.0000
All the variables were significant at 1%, 5% and 10% level of significance. And also using the f-test to check the overall significance of the model, the f-stat showed that it was significant at 5%, 1% and 10% level of significance, also the p-value is less than 0.05 which also shows that it is significant. The coefficient of determination was 0.90 and the adjusted R- Squared was 0.89 which shows that over 89% systematic variation of Gross Domestic Product can be explained by all the dependent variables. This therefore is surely a good fit as only about 10% of systematic variation is left unaccounted for in the model. The remaining influence must reflect some combination of measurement errors, random fluctuations, etc. All the variables were logged. There is a positive relationship between financial deepening and Gross domestic product, which signifies that they are linearly related the elasticity coefficient of financial deepening is 0.69 implying that a 1% increase will increase Gross Domestic Product by 0.69% when other variables are kept constant. Since the elasticity value of 0.69 is less than one (1), showing that it is inelastic, this means that a unit change in Financial Deepening will bring about a less than proportionate change in Gross Domestic Product. There is a negative relationship between Degree Of Openness and Gross Domestic Product, the elasticity coefficient of Degree Of Openness is -0.61, implying that a 1 increase in Gross Domestic Product.
For foreign direct investment, the elasticity coefficient is -0.12, implying that a 1% increase in Foreign Direct Investment will decrease Gross Domestic Product by 0.120 when other variables are kept constant. since the value of -0.120 is less than one (1)in absolute terms.
From the result above, the only variable that has had a positive impact on the Gross Domestic Product as a result of the liberalization of the financial sector is financial deepening. Degree Of openness and Foreign Direct Investment has had a negative influence on the Gross Domestic Product. This means that the extent of the degree of openness has actually not worked in sub-Sahara African countries. Degree Of Openness negative impact does not mean that liberalization will never work in African countries, or rather does not work the major reason behind this problem is the fact that the financial system is weak, but as (Pill and Pradhan, 1997 p7) puts it; financial liberalization is only one component of a flourishing development strategy. Suitable macroeconomic policy, institutional development, and structural reform must go along with financial liberalization and create the stable context vital for it to succeed, which is not present in some African countries.
