Rethinking decentralization: how fiscal autonomy shapes local government efficiency through nonlinear institutional dynamics

Purpose This study aims to examine whether fiscal autonomy improves capital expenditure efficiency in decentralized systems by addressing the overlooked possibility that its effects are nonlinear and context-dependent. Focusing on Ghana’s 261 Metropolitan, Municipal and District Assemblies (MMDAs), the study investigates whether fiscal autonomy enhances efficiency uniformly or only beyond certain institutional thresholds. Design/methodology/approach This study uses a balanced panel data set from 2018 to 2024 to detect continuous nonlinear and regime-specific effects using fixed-effects estimation with a quadratic specification and threshold-based robustness analysis. The data were obtained from Ghana Audit Service-certified financial reports and Ghana Statistical Service demographic indicators. Findings Fiscal autonomy is associated with a U-shaped relationship with capital expenditure efficiency, although the nonlinear effect is modest and only weakly statistically supported. At low levels of autonomy, increases in internally generated funds are associated with lower capital expenditure shares, while beyond an estimated threshold of approximately 37% of total revenue, the relationship becomes positive. Intergovernmental transfers complement local fiscal capacity, whereas population density and urban classification are not significant predictors in the regression models. Practical implications The findings indicate gradual capacity-sensitive decentralization strategies. Enhanced fiscal autonomy in the absence of institutional development may reduce expenditure efficiency. Policymakers should prioritize administrative capacity, revenue systems and accountability before increasing budgetary discretion. Originality/value This study provides panel-based evidence consistent with a U-shaped fiscal autonomy–efficiency relationship in African local governments. By highlighting threshold effects and institutional conditioning, it advances fiscal decentralization research and contributes to the application of nonlinear modeling in public finance.