Impact of a Single Intraday Coupling of European Electricity Markets on the Intraday Prices Predictability

The main objective of this dissertation thesis is to examine the change of predictability of intraday electricity prices/returns in Czechia traded on the Czech intraday market organized by market operator OTE, a.s., due to the Single Intraday Coupling (SIDC), which rapidly expanded the volume of intraday electricity. As a model-independent predictability quantities, we estimated the Hurst exponent and the Sample entropy in its multiscale augmentation - the Multiscale entropy. By performing a linear regression and intervention analysis we were able to determine that the impact of the SIDC was associated, in general, with the slight increase of the predictability. However, there were many other factors after the connection to the SIDC, which had the influence on the predictability such as weather and the coronavirus pandemic that started in spring 2020. Both linear regression and the intervention analysis of the Hurst exponent on short temporal scales suggests, that the SIDC was associated with swtiching from weakly mean-reverting to persistent behavior. On the contrary, the situation on the long temporal scales was more obscure pointing out that the mean-reversion behavior was either unchanged according to the intervention analysis, or increased in the pre-covid period according to the linear regression, which implies larger predictability. The similar results were generated using the deseasonalized returns. Moreover, results of the Multiscale entropy display that the irregularity right after the SIDC became smaller indicating more predictable patterns in time series for all evaluated coarse-graining scales, which is in agreement with the Hurst analysis on the short temporal scales. These calculations were also performed with the deseasonalized returns yielding the similar results. Since this study is the first one to analyze the relationship between the SIDC and the predictability of the intraday prices/returns in terms of long range dependence and irregularity of time series, it reveals that the SIDC impact on predictability faded out when the Covid-19 pandemic began in Czechia, and in general, the SIDC did not remarkably change the predictability patterns compared to the predictability pattern of the temperature, which is probably the most important factor that affects the predictability of the intraday prices/returns. These results have various implications for traders on intraday market and contribute to the academic knowledge of the intraday prices/returns behavior.