Technological innovation and carbon dioxide emissions in Nigeria using an EKC STIRPAT VECM framework with artificial neural network robustness

Abstract Developing economies such as Nigeria face mounting pressure to sustain economic growth while reducing carbon dioxide (CO₂) emissions, yet empirical evidence on whether technological innovation can meaningfully decouple growth from emissions in African contexts remains limited. This study addresses this gap by examining the growth–innovation–emissions nexus in Nigeria using an integrated Environmental Kuznets Curve (EKC)–STIRPAT–Vector Error Correction Model (VECM) framework, complemented by artificial neural network (ANN) validation. Annual data for 1989–2023 on CO₂ emissions, gross domestic product, per capita income, and technological innovation are analysed following unit-root and cointegration testing. The results confirm a stable long-run equilibrium relationship and provide strong evidence of EKC-type nonlinearity, with emissions declining beyond income thresholds. Technological innovation exerts both direct nonlinear effects and a statistically significant moderating influence, weakening the emissions-increasing impact of economic growth. VECM estimates indicate gradual convergence toward equilibrium, with approximately 14% of disequilibrium corrected annually. ANN results corroborate the econometric findings and confirm nonlinear predictive consistency. The study’s originality lies in integrating EKC nonlinearity, STIRPAT elasticities, innovation moderation, dynamic cointegration, and machine-learning validation within a unified national framework. Policy implications emphasise innovation-centred growth strategies, sustained R&D investment, and clean-technology diffusion as critical pathways for achieving low-carbon and sustainable development in Nigeria and comparable emerging economies.