Short-Run and Long-Run Dynamics of Nordic Electricity Prices

This study investigates the dynamics of Nordic electricity system prices by applying a Markov regime-switching state-space model to daily data from 2012 to 2020. The model decomposes log-prices into a permanent random walk component and a transitory first-order autoregressive (AR(1)) component with regime-dependent variance (low, medium, high). Estimation is performed via Gibbs sampling. Results indicate that price movements are predominantly driven by short-term transitory shocks, as the permanent component exhibits low variance. The high-volatility regime demonstrates a variance over 100 times that of the low-volatility regime and is highly persistent. Transitory shocks decay with a half-life of approximately three days. Monte Carlo simulations reveal that standard Augmented Dickey-Fuller (ADF) tests often misidentify stationarity in the presence of regime-switching volatility. Forecast comparisons against a Seasonal Autoregressive Integrated Moving Average (SARIMA) benchmark show that the proposed model achieves a lower Mean Absolute Error (MAE), although Root Mean Squared Error (RMSE) is slightly higher. While both models struggle with extreme price spikes, the regime-switching model offers interpretability and insights into underlying market dynamics. However, its univariate nature and the assumption of normally distributed shocks are limiting factors.