Inflation upswings, disinflation and threshold effects of monetary tightening on economic growth in Nigeria

This study examines the asymmetric associations between inflation upswings, disinflation and economic growth in Nigeria using annual data spanning 1990–2025. A nonlinear autoregressive distributed lag (NARDL) model is employed to decompose inflation into positive and negative partial sums, allowing asymmetric short-run and long-run relationships to be examined. The NARDL provides the primary assessment of directional asymmetry, while a threshold interaction model evaluates whether these associations vary across inflation environments and a moderation model examines whether the disinflation–growth relationship is conditional on monetary-policy tightness. Local projections and quantile regression provide supplementary evidence on temporal and distributional heterogeneity, respectively, whereas machine learning is limited to an exploratory assessment of the importance of nonlinear predictors. The NARDL results indicate a long-run relationship among the variables and provide formal evidence that the estimated associations of inflation upswings and disinflation are statistically distinguishable in both the short run and the long run. Both inflation components are negatively associated with long-run economic growth. The threshold interaction estimates indicate different fitted inflation–growth associations across inflation environments, but bootstrap inference does not establish a statistically significant threshold effect. The moderation analysis indicates that the disinflation–growth relationship varies with monetary-policy tightness rather than remaining constant across policy environments. Local projections show that the estimated disinflation–growth association varies across horizons, while quantile regression indicates variation in the inflation coefficients across the conditional growth distribution. Exploratory machine-learning results identify inflation upswings and exchange-rate depreciation as comparatively influential nonlinear predictors. Taken together, the evidence distinguishes directional asymmetry, inflation-environment variation, monetary-policy conditioning, and temporal and distributional heterogeneity. The contribution, therefore, lies in showing how analytically distinct dimensions of the inflation–growth relationship emerges from complementary specifications, rather than treating methodological multiplicity as independent confirmation of a single result. The findings support a context-sensitive interpretation of inflation stabilisation while remaining subject to the limitations of annual observational data.