Industrial energy use and oil dependence as conditional drivers of carbon emissions in Saudi Arabia - evidence from quantile regression.

Saudi Arabia's per capita carbon emissions have risen persistently since the 1970s, driven by energy-intensive industrialisation and a structural reliance on oil. Existing studies have largely treated energy consumption as a direct, additive driver of emissions, overlooking how its environmental impact depends on the industrial structure that absorbs it. This study proposes a conditional emissions framework and tests it on Saudi annual data for 1970-2024, combining an autoregressive distributed lag (ARDL) model with an explicit energy-manufacturing interaction term, quantile regression across τ = 0.1-0.9, and Toda-Yamamoto causality analysis. In the long-run ARDL specification, the standalone coefficient on energy consumption is negative; however, the implied marginal effect at the sample-mean manufacturing share is positive (≈ 0.43), and rises monotonically with the manufacturing share of GDP, confirming that energy becomes emissions-intensive only within industrial production (interaction coefficient = 2.49, p < 0.01). The quantile regression results refine rather than overturn this picture: the energy-manufacturing interaction is most evident in the lower-emission regimes and attenuates toward the upper tail. At the same time, oil dependence exerts a consistent positive effect across most of the distribution. A 1984 structural break, linked to the oil-price collapse and the policy pivot toward subsidised industrial energy, is significant across all emission regimes, providing empirical support for carbon lock-in theory. These findings imply that effective decarbonisation under Saudi Arabia's Vision 2030 requires restructuring industrial energy use rather than simply reducing its aggregate volume.