Electricity spot prices are characterized by sudden large movements, followed a few days later by an equally large movement in the opposite direction. These phenomena are called spikes (upward movements) and drops (downward movements). Recent research has suggested that the dynamics of the electricity spot prices can be accurately described by hidden Markov Regime Switching (MRS) models. Regime switch models separate the ordinary dependence and the extreme (spike or drop) dependence. This is a crucial point since it is the extreme dependence that is of interest when computing risks. We fit a large number of MRS models to six European electricity markets (EEX, PowerNext, APX Power UK & NL, Nord Pool System & Sweden) in order to analyze the nature of the prices. The estimated regimes are then used to analyze the probability of an extreme event and the conditional probability for one market to experience an extreme event, conditional that another market is experiencing the same event. We find that the frequency of extreme events is positively related to the amount of renewable energy sources in the power system. It can be argued that the dependence results in this paper give an indication of what future dependence (for even more integrated markets with more renewable energy) will be. The integration of markets will lead to efficient sharing of reserves while the additional renewable energy will increase the need for reserves. The results are of interest when computing financial risks (e.g. VaR or expected shortfall), when designing reserves but also as an indication of the degree of integration between markets.
- Sannolikhetsteori och statistik