A semi-noparametric approach to risk analysis in electricity markets

Resumen

This thesis complements the existing literature on risk analysis in electricity markets, proposing the use of a flexible tool that allows capturing skewness, kurtosis and higher order moments in the uncertain components of variables that affect electricity markets. To do so, it is proposed a generalization of the normality assumption, traditionally used to describe the uncertainty of these markets, by means of probability density functions defined in terms of a finite Gram-Charlier expansion. This 'semi-nonparametric approach' (SNP) has been successfully implemented to extend the non-normality of different phenomena, but never before to the description of electricity markets. For this purpose, three applications are proposed to solve risk problems typical of the electricity markets and that will be developed through the three chapters of the thesis: (i) "Uncertainty in Electricity Markets from a semi-nonparametric approach", (ii) “Modeling the electricity spot price with switching regime semi-nonparametric distributions” and (iii) “Hedging in Electricity Markets under skewness and leptokurtosis”. In all three cases, it is found that by means of the SNP approach it is possible not only to better describe the behavior of the variables linked to the electricity markets, but also to use their estimation for issues such as efficient pricing, risk management policies or optimal hedging strategies. From a theoretical point of view, generalizations of SNP distributions are proposed to capture switching regimes and to the multivariate modeling of simultaneous energy price and quantity functions. Thus, establish energy policy recommendations that help companies, investors and regulators to make decisions in the presence of risk and uncertainty in electricity markets.