Application of Multivariate Time Series Analysis to Modelling of Total Tax Revenue and Some of Its Components (Published)
The study applies Multivariate Time Series analysis to model Total Tax revenue and some of its components, examining the dynamic relationships between various tax revenue streams. Using a Vector Autoregressive (VAR) framework, the analysis fits the VAR model and also explores the interactions and causal relationships between different Tax components, such as P.A.Y.E, Stamp Duties, Direct Assessment, Road Taxes as well as Other Taxes. The findings provide insights into the complex dynamics of tax revenue generation, informing policy decisions and forecasting strategies for sustainable revenue mobilization. The findings placed P.A.Y.E on a high premium as one of tax components not only as a good predictive factor to total tax revenue and other tax components in this work, but as a major driver of the economy in Nigeria. The evidence of this is shown in the granger causality test where, at 5% critical value of F level of significance, with 2 and 77 degrees of freedom gives, 3.316 causing the null hypothesis H0 to be rejected.
Keywords: Causality, bayesian informationakaike information criteria, vector autoregression (VAR)