European option pricing on day-ahead electricity prices: the mexican wholesale electricity market case

The present research proposes a novel European option pricing model with the day-ahead electricity price as an underlying asset which could be implemented as the first day-ahead electricity price hedging financial instrument in the Wholesale Electricity Market (MEM). Therefore, this work represents an essential contribution to the MEM's development since, according to Roy and Basu (2020), MEM should be considered an emerging electricity market owing to its small number of participants, and hedging financial instruments, such as futures or options, cannot be acquired. Hence having an instrument of this kind would allow market participants to implement better risk management strategies to hedge day-ahead electricity price volatility to prevent financial losses. This work is divided into five chapters; each concerns a different component of the proposed model. In Chapter 1, the main characteristics of the MEM, as well as a review of the operating rules that are most closely related to the design of the proposed financial instrument, as well as a general context of the MEM and the growth initiatives proposed by the Mexican government, are described. In Chapter 2 an in-depth review of the probability theory necessary for a complete understanding of the proposed model, starting with basic probability concepts and moving on to the Normal Inverse Gaussian and Multivariate Normal Inverse Gaussian probability distributions, as well as the valuation of a European Option by Monte Carlo valuation is provided. In Chapter 3, two topics are addressed; first, a statistical analysis is performed to confirm that well-known LMP stylized facts, such as seasonality, volatility, and autocorrelation, are observable on MEM's day-ahead electricity prices. Second, Normal Inverse Gaussian (NIG) distribution capability to fit LMP logarithmic returns (Series Returns) is shown as follows: the Seasonal and Trend Decomposition Model (STL), NIG parameter estimation by Maximum Likelihood Estimation (MLE) of Series Returns, simulated NIG series generation from obtained parameters, and goodness-of-fit tests are performed to demonstrate NIG's distribution capabilities to fit and simulate electricity returns series. In Chapter 4, the European option pricing model employing Multivariate Normal Inverse Gaussian (MNIG) is proposed. In order to obtain the European option price for 28 days ahead on an hourly basis (672 hours ahead) by applying this model, each week hour is assumed to be a single independent asset, which produces 168 series for a single week. Four lagged log-prices for each hour are then obtained to be modeled employing MNIG distribution to perform Monte Carlo simulations and generate electricity lagged log-prices trajectories which then are employed to estimate the European option price for the 672 hours ahead by applying the European option pricing methodology. Results show that by applying this valuation model, electricity price correlation and seasonality are modeled by the employment of MNIG distribution, which simplifies modeling complexity and MNIG makes it possible to obtain a correct European option price valuation for each of the forecast values. Finally, the research conclusions are presented in Chapter 5.

https://orcid.org/0000-0001-7070-6154