Essays on Financial Risks and Derivatives with Applications to Electricity Markets and Credit Markets

Sammanfattning: Contracts traded on international financial and commodity markets are associated with complex risk structures. In this dissertation we are concerned with two specific types of risks; market risks and credit risks. The first chapter investigates market risks in the context of the Nordic electricity market. The paper performs in- and out-of-sample backtesting of a VaR model based on GARCH volatility with NIG innovations. Furthermore, the Cornish-Fisher expansion is applied to get analytical approximations of the NIG based VaR estimates. Backtesting shows that the model is a promising alternative to the well known Gaussian GARCH. Results also show that the Cornish-Fisher approximation gives reasonable outcomes for the less extreme quantiles, especially when the return distribution is close to normality. The second chapter continues to explore market risks on the Nordic electricity market. A mean-reverting jump diffusion stochastic volatility model for the electricity spot price is suggested. The model is estimated with a combination of standard statistical methods and a Markov Chain Monte Carlo (MCMC) algorithm. Results indicate that the model captures large parts of the trajectorial and statistical properties of the spot price. Hence, the model is a candidate for applications in risk management and derivative pricing. The third chapter is again concerned with market risks. The spot price model from chapter two is applied to pricing of the forward curve in the Nordic electricity market. A semi-closed solution for the forward curve is derived, and then calibrated to market data. Results show that the model outperforms its benchmark. The final chapter of the dissertation goes in the direction of credit risks. The paper extends Merton’s classical corporate bond credit risk model to an international setting with stochastic domestic and foreign interest rates. In an extensive comparative static analysis we study how the credit spread in the model depends on the currency and related variables.