AI-Powered Predictive Analytics and Its Relationship with Budgeting Efficiency and Financial Accountability in Nigeria's Federal Public Sector

Budgets are the plans against which public financial reports are judged, and weak forecasting undermines both budget credibility and accountability. This study examines whether documented adoption of artificial intelligence (AI)-powered predictive analytics is associated with budgeting efficiency and financial accountability in Nigeria's federal public sector. A balanced panel of five federal institutions, the Federal Ministry of Finance, the Office of the Accountant-General of the Federation, the Federal Inland Revenue Service, the Central Bank of Nigeria and the Bureau of Public Procurement, is observed over 2015–2024. Documentary evidence from annual reports, audit-related documents and AI-implementation documents is scored on a five-point (0–4) rubric and analysed with pooled, fixed-effects, trend-adjusted, first-difference and lagged regressions, interpreted through agency theory. Predictive analytics adoption is positively and significantly associated with budgeting efficiency and accountability (pooled β = 0.694, R² = 0.853; fixed-effects β = 0.712, within R² = 0.845). The association remains significant after adjusting for a common trend (β = 0.639), in first differences (β = 0.727) and with a one-year lag (β = 0.585). The lagged estimate is the smallest of the specifications, consistent with, though not proof of, a maturation period during which models learn from data and users learn to rely on forecasts. Because the design is observational, the measures are documentary indices, and the sample is small and purposively selected, the results indicate association rather than proven causation and are interpreted for the five institutions examined rather than for all federal ministries, departments and agencies. The results support investment in forecasting capability and data infrastructure, with realistic expectations about time-to-benefit, accompanied by data-quality programmes, staff training, publication of forecast errors and expert review of model outputs.