Tax Administration Transformation and Tax Revenue Collection in Kenya: An ARDL Bounds-Testing Approach

Abstract

This paper evaluates the joint effect of the Kenya Revenue Authority’s (KRA) three-pronged tax administration transformation signaled by iTax system usage, tax base expansion and data-driven compliance, on domestic tax revenue collection. No prior study has jointly evaluated these three reforms using national-level time-series data, a gap this paper addresses. Quarterly secondary data covering January 2014 to March 2023 (N = 37) were obtained from KRA and analyzed using a correlational design. Based on Augmented Dickey-Fuller and Phillips-Perron unit-root tests confirmation of mixed order integration among the series, an autoregressive distributed lag (ARDL) bounds-testing approach to cointegration was applied. This was complemented by an error-correction model (ECM) and pairwise Granger causality tests. iTax usage correlated weakly with revenue (r = 0.29), tax base expansion (r = 0.84) and compliance (r = 0.95) correlated strongly. The bounds F-statistic (11.91) confirmed long-run cointegration. In the long run, compliance raised revenue by 114 per cent (p = 0.000), while tax base expansion (–12 per cent, p = 0.033) and iTax usage (–8 per cent, p = 0.152) were associated with revenue declines. The error-correction term adjusted at 209 per cent per quarter, and the model explained 88 per cent of revenue variance, passing all residual diagnostics. Causality ran bidirectionally between iTax usage and revenue, and unidirectionally from base expansion and from compliance to revenue. The findings suggested that Kenya’s revenue reforms should be sequenced. Investment in data-driven compliance and enforcement be prioritized, followed by deeper iTax system integration. Base-expansion drives should be paired always with monitoring of the existing taxpayer base.
Keywords: iTax, tax base expansion, data-driven compliance, ARDL bounds testing, tax revenue collection