A Multiple Regression Analysis Of Contributions Of Various Sources Of Tax To The Total Tax Revenue In Nigeria:- Usoro, Anthony E.

ABSTRACT

This work sought to carry out a multiple regression analysis, which involved a dependent and five independent variables. The data for the analysis were obtained from Federal Inland Revenue, Uyo Zonal Office, Akwa Ibom State.The variables Ci, C2, C3, C4, C5 and Cg, which represented Total Budgeted Tax Revenue (TBTR), Petroleum Profit Tax (PPT), Company Income Tax (CIT), Education Tax (ET), Personal Income Tax (PIT) and Value Added Tax (VAT) respectively were all used for the analysis. A multiple linear model was obtained as; Ci = 20.017-0.342C2 -1.115C3 17.633C4 +17.171C5 +0.51C6. The t-test result of table 4.2 revealed that Petroleum Profit Tax and Education Tax are statistically significant while Company Income Tax, Personal Income 8 Tax and Value Added Tax agreed with the null hypothesis of no significant contribution to the Total Tax Revenue. Farrar and Glauber test for detection of multicollinearity showed presence of multicollinearity amongst the explanatory variables. Further tests carried out indicated that the presence of multicollinearity was as a result of inter-correlation between variables C2 and C5. This . multicollinearity attracted further study to finding solution through Principal Component Method. The Principal Component Analysis was actually carried out and this finally yielded higher estimates for Total Tax Revenue.