Energy Consumption, Carbon Emissions and Economic Growth in Algeria: Conditional Evidence from ARDL and NARDL Models (1990–2022)

Abstract

This study asks whether the Environmental Kuznets Curve (EKC) describes the income–emissions relationship in Algeria, a hydrocarbon-dependent economy, over 1990–2022. Per-capita CO₂ emissions are modelled as a function of per-capita income, its square, energy use and trade openness using a parsimonious conditional ARDL specification, with symmetry and integration-robust causality tests as supporting evidence. Three findings emerge. First, the evidence for a stable long-run relationship is weak: the bounds test is marginal at the 10% level and insignificant at 5%, and the Johansen tests do not support cointegration, so the level estimates are reported as conditional descriptions rather than long-run elasticities. Second, the conventional inverted-U EKC is not supported; the fitted level terms imply a U-shaped profile, but severe collinearity between income and energy use, together with a significant functional-form test, precludes any structural or policy-threshold reading of the implied turning point. Third, energy use has the strongest short-run association with emissions and the largest share of forecast-error variance, and its omission destabilizes the income coefficients. For Algeria the policy-relevant margin is therefore the carbon intensity of energy supply and use rather than an income threshold at which emissions would fall automatically. All conclusions are conditional on the specification and on a 33-observation sample, whose identification limits are reported rather than assumed away.