344 KB0 KB200920262026Adebanjo, SeunOjedokun, Olatunji Daudaletnik:28številka:iss. 3, article 1str. 176-191DOI:10.15458/2335-4216.1374ISSN:2335-4216COBISSID_HOST:289740547URN:URN:NBN:SI:doc-3EYWGFK7enEkonomska fakultetaEconomic and business reviewdavkilegislationNigeriaNigerijataxeszakonodajaNigeria's 2025 Tax Act in comparative perspective| tax rates, compliance, and fiscal space with econometric and machine-learning evidence from Nigeria, South Africa, the United Kingdom, and the United States|This study compares Nigeria's 2025 Tax Act with regimes in South Africa, the United Kingdom, and the United States using quarterly panel data (2000–2025) and complementary econometric and machine-learning methods. We estimate fixed-effects difference-in-differences, autoregressive distributed lag/cross-sectionally augmented autoregressive distributed lag error-correction models, dynamic panel generalised method of moments, and local projections to identify short- and long-run effects on revenue efficiency, compliance behaviour, and fiscal space. Predictions for VAT/corporate income tax revenues and gaps are generated with Elastic Net, Random Forest, XGBoost, and LightGBM, and heterogeneity is assessed via generalised random forests and doubly robust learners. Results show a significant, durable postreform rise in Nigeria's revenue efficiency, with the strongest gains where enforcement intensity and digitalisation are high; VAT/corporate income tax gaps shrink, and long-run elasticities confirm cointegration with policy fundamentals. Machine-learning models achieve high out-of-sample accuracy and highlight digitalisation, enforcement, VAT productivity, and tax capacity as dominant predictors. Policy implications emphasise pairing rate design with administrative capacity buildingTEXTznanstveno časopisjejournalsSlovenian National E-content AggregatorNational and University Library of SloveniaUniverza v Ljubljani, Ekonomska fakulteta