Structural instability and emerging energy–GDP decoupling in Saudi Arabia: exploratory evidence and forecasting implications

Energy-demand relationships in resource-dependent economies may change when macroeconomic shocks, policy reforms, and structural transformation alter consumption patterns. This study provides an exploratory assessment of structural instability, energy–GDP decoupling, and energy-demand forecasting in Saudi Arabia using annual data from 1990 to 2023. Bai–Perron tests are applied under alternative specifications to identify endogenous changes in per capita energy consumption, while a policy-motivated test at 2016 evaluates whether the year associated with the launch of Vision 2030 is linked to a change in the energy–GDP relationship. The results indicate substantial structural instability in Saudi Arabia’s per capita energy-consumption trajectory. Major breaks are identified in the mid-1990s or early 2000s and around 2009, depending on the structural-break specification, with the 2009 breakpoint remaining robust across alternative models. In contrast, 2016 is not selected endogenously as a dominant breakpoint. A policy-motivated test at 2016 suggests a possible change in the energy–GDP relationship under conventional inference, although the evidence weakens after accounting for serial correlation. Break-adjusted cointegration results are interpreted as exploratory because of the short annual sample and uncertain integration properties. The decoupling analysis shows a descriptive change after 2016. During 2017–2023, real GDP per capita increased while per capita energy consumption declined, indicating strong decoupling. By contrast, the earlier period was characterised by strong negative decoupling. Forecasting comparisons indicate that models incorporating more recent observations produced lower forecast errors than the model estimated exclusively from pre-2016 data. Predictions incorporating realised real GDP per capita and Brent oil prices are interpreted as conditional ex post forecasts rather than fully operational ex ante forecasts. Overall, the findings provide exploratory evidence of substantial structural instability and an emerging energy–GDP decoupling pattern in Saudi Arabia. They do not establish a causal effect of Vision 2030 or demonstrate a permanent new energy regime. The results emphasise the importance of cautious interpretation, transparent data and code, and regularly updated break-aware models for energy-demand forecasting in resource-dependent economies.