ROBOTIC PROCESS AUTOMATION AND THE ACCELERATION OF TAX ASSESSMENT PROCEDURES: EVIDENCE ON THE MEDIATED PATHWAY TO TAX REVENUE COLLECTION

Abstract

Fiscal losses in transitional economies arise not exclusively from tax evasion but also from administrative delays between return submission and the issuance of legally binding assessments. This study investigates whether Robotic Process Automation (RPA) reduces such delays and whether improvements in assessment efficiency translate into stronger revenue collection. Integrating technology acceptance, task–technology fit and public value perspectives, the proposed three-construct framework examines the relationship between RPA adoption and Tax Revenue Collection (TRC), with Tax Assessment Speed (TAS) serving as a mediating mechanism. Data obtained from 284 tax inspectors, public financial auditors, revenue system administrators and departmental directors across revenue authorities and audit directorates, representing an 81.1% response rate, were examined through Partial Least Squares Structural Equation Modelling (PLS-SEM) using SmartPLS 4. All measurement-model reliability and validity criteria were satisfied, while the structural model accounted for 46.8% of the variation in TAS and 57.3% in TRC. The findings revealed a positive association between RPA adoption and TAS (β = .684), alongside a substantial relationship between TAS and TRC (β = .528). RPA adoption also maintained a direct influence on TRC (β = .298). The significant indirect relationship (β = .361) confirmed partial mediation, indicating that accelerated assessments accounted for 54.8% of the total effect. These findings establish procedural delays as a consequential determinant of fiscal performance and underscore the importance of strategically deploying automation bots, strengthening inter-agency data interoperability and establishing statutory recognition of automated tax assessments.