Dynamics of Effective VAT and Per Capita Household Consumption in Nigeria: Evidence from ARDL Cointegration Approach, 1994Q1-2024Q4

Abstract

The study examines the dynamics of effective VAT and per capita consumption in Nigeria: Evidence from ARDL Coiintegration Approach, 1994Q1 -2024Q4. Quarterly data obtained via Denton-Cholette interpolation were analyzed using the ARDL bounds test, VECM Granger causality, Impulse Response Function, and Variance Decomposition, with dummies for structural breaks in 2016Q2 and 2020Q2. ADF and Phillips-Perron tests show that LNPHCE, EFFVAT, LNRGDPPC, LNDISYPC, LNINVPC, and LNPOP are all I(1), while INF is I(0). The ARDL bounds test confirms a long-run relationship with F-statistic of 7.2, exceeding the 1% upper bound of 4.68. Long-run results show that a 1% point increase in EFFVAT reduces household consumption per capita by 0.091%. Real GDP per capita [0.651], disposable income [0.284], and investment [0.113] have positive effects, while inflation [-0.009] and population [-0.420] are negative. In the short-run, the ECT is -0.451, indicating 45.1% adjustment per quarter. EFFVAT has an immediate effect of -0.041 and lagged effect of -0.028%, totaling -0.069%, GDP [0.385] and disposable income [0.165] are positive while inflation [-0.0042] and COVID-2020 dummy [-0.062] are negative. Granger Causality runs unidirectionally from EFFVAT to consumption [Chi-sq. = 8.42, ρ=0.015]. Variance Decomposition at Q12 shows that 11.25% of consumption variation is explained by EFFVAT shocks, second only to GDP at 18.50%. Impulse Response confirms a persistent negative effect of EFFVAT lasting 10 quarters. The study concludes that EFFVAT significantly reduces household welfare. Policy should focus on targeted exemptions and phased VAT implementation.