International Journal of Economics and Financial Issues

4.6. Heteroscedasticity Test
H
0
: No Heteroscedasticity exists.
Heteroskedasticity test: Breusch-Pagan-Godfrey.
Table
10 tests the presence of Heteroscedasticity in the data 
adopted for the model of the study. Decision rule; we accept the 
null hypothesis that there is no Heteroscedasticity in the data of the 
model when the result is not statistically significant. It is important 
that the data has no Heteroscedasticity as this also serves as the 
condition of Homoskedasticity. This shows the data is good or fit 
for estimation.
4.7. Serial Correlation Test
Table 11 contains the result on autocorrelation test. The null 
hypothesis of autocorrelation states that ‘no autocorrelation exists 
in the data used’. It is therefore important that the result is not 
statistically significant. We then accept the null hypothesis that 
there is no autocorrelation among the data of the variables adopted.
Figure 6:
 Impulse response function
Table 10: Heteroscedasticity test results
F-statistic
0.989568
Prob. F (4,30)
0.4283
Obs.* R-squared
4.079700
Prob. Chi-square (4)
0.3953
Scaled explained SS
8.979433
Prob. Chi-square (4)
0.2616
Source: Author’s computation, 2017
Table 11: Breusch-Godfrey serial correlation LM test
F-statistic
1.361626
Prob. F (2.28)
0.2727
Obs.* R-squared
3.102335
Prob. Chi-square (2)
0.2120
Source: Author’s computation, 2017

---

Rafindadi and Musa: An Empirical Analysis of the Impact of Public Debt Management Strategies on Nigeria’s Debt Profile
International Journal of Economics and Financial Issues | 
Vol 9 • Issue 2 • 2019
135
5. CONCLUSION, RECOMMENDATIONS