Abstract

This study examines whether renewable electricity is associated with industrial combustion-related carbon dioxide emissions in Azerbaijan over 1995–2024. Five aggregate-share autoregressive distributed lag specifications and three pre-specified component models are estimated with 29 effective annual observations. In the baseline model, the renewable coefficient is negative, but significance depends on covariance choice (conventional p = 0.098; White robust p = 0.033; Newey–West p = 0.038). The hydropower replacement is negative and robust, whereas the non-hydropower replacement is imprecise. In the joint component equation, the hydropower coefficient remains negative (-0.0283; White robust p = 0.004; Newey–West p = 0.0025) and the non-hydropower coefficient remains statistically unsupported (White robust p = 0.423; Newey–West p = 0.3862); the finite-sample bounds statistic rejects no levels relationship at 5%, and the error-correction coefficient is -0.6386. Error-correction terms are negative across all eight equations, although collinearity, specification, serial-correlation, and endpoint cautions remain. The evidence indicates conditional dynamic associations, not causal effects or verified fossil-generation displacement. Policy evaluation should link renewable expansion to measured fossil displacement, industrial electrification, hydrological conditions, and sector-specific emissions.The HAC-robust Wald equality test does not reject H0: HYDRO = NONHYDRO (F(1, 22) = 1.5891, p = 0.2207; chi-square(1) = 1.5891, p = 0.2075), so the paper does not claim statistically distinct component effects.