Electricity Demand Forecasting of Value-at-Risk and Expected Shortfall: The South African Context
DOI:
https://doi.org/10.32479/ijeep.16995Keywords:
Extreme Value Theory, Electricity Demand Load, Generalised Extreme Value, Volatility, South AfricaAbstract
The energy landscape, particularly in the electricity sector, is characterised by a complex interplay of various factors that contribute to its inherent volatility. Electricity participants, including investors, regulators, consumers, and policymakers, are constantly seeking methods to better understand and manage the associated risks. The generalised extreme value (GEV) distribution with block minima is applied to model extreme losses on daily electricity demand in South Africa for the period of 1 April 2019 to February 13, 2024. The results of the estimated GEV gave a negative shape parameter implying that both winter and no-winter seasons extremes are correctly model a type III GEV distribution known as the Weibull distribution. When computing the VaR and ES, we found that VaR went as low as 8.74% while for ES had the lowest as 9.57%. Finally, the backtesting procedures further proved that the estimated risk measures are reliable as both the Kupiec and Chrisoffersen tests failed to reject the null hypothesis. In conclusion, the fitted GEV showed some reliance in capturing extremes losses for winter and non-winter returns. Lastly due to reliability of the models, risk analysts together with investors interested in the electricity sector should therefore adopt the procedures used to know the risk of their investment.Downloads
Download data is not yet available.
Downloads
Published
2024-12-22
How to Cite
Masilo, B. M., & Makatjane, K. (2024). Electricity Demand Forecasting of Value-at-Risk and Expected Shortfall: The South African Context. International Journal of Energy Economics and Policy, 15(1), 481–489. https://doi.org/10.32479/ijeep.16995
Issue
Section
Articles