Short-Run Statistical Interactions Between Nuclear and Renewable Energy Production in the EU27: A Bivariate VAR Analysis (1990–2022)

This study examines the temporal evolution of low-carbon energy production in the European Union (EU27) using annual data for 1990–2022, focusing on the dynamic interaction between nuclear, renewable, and biofuel production at the EU aggregate level. After evaluating stochastic properties via Augmented Dickey–Fuller (ADF) tests and assessing long-run cointegration through the Johansen framework, short-run interactions are modeled using a Vector Autoregression (VAR) of order one. Dynamic responses and innovation variances are analyzed using impulse response functions (IRFs) and forecast error variance decomposition (FEVD). The Augmented Dickey–Fuller (ADF) results suggest both series are I(1). The Johansen test fails to reject the null of no cointegration, implying that there is no stable long-run equilibrium relationship between the two series over 1990–2022. VAR-based IRFs show small, short-lived cross-responses that dissipate within a few years. FEVD results indicate that variance shares are horizon-dependent and sensitive to the Cholesky ordering. Granger causality tests provide limited evidence of short-run directional predictability. A Zivot–Andrews test does not reject the unit-root-with-break null. These findings suggest that nuclear and renewables follow largely independent dynamics in the EU27 aggregate. A key limitation is that EU27 aggregation masks cross-country heterogeneity (e.g., Germany vs. France) and excludes policy variables, prices, and demand-side drivers. The estimated VAR(1) satisfies the stability condition: all eigenvalues of the companion matrix lie inside the unit circle (modulus < 1), confirming that the system is dynamically stable.