Backtesting energy portfolio with copula dependence structure

Masala, G.

2021-01-01

---

Abstract

The energy markets play a crucial role in the economic development of every country. These markets are characterized by high volatilities due to sizeable price fluctuations. The correct development of risk measures to quantify the inherent risk market is then a challenging task for the risk management systems. We consider in this survey a portfolio composed of two energy assets: crude oil and natural gas. We adopt, as a risk measure, the value-at-risk and the expected shortfall. In order to estimate these risk measures efficiently, we model the tails of each marginal with extreme value theory and we adopt a general dependence structure between the two assets given by a t-copula. The performance of the model is then validated with backtesting technique. To this end, we use a database ranging from years 1997 to 2017. We have then highlighted that the backtesting based on the value-at-risk and the Expected Shortfall passed the most common tests.