Risk management in energy markets

This dissertation concentrates on issues of risk management for corporations with a focus on energy quantity and price exposure hedging. In commodity markets in general, and energy markets in particular, the model corporation produces and/or consumes in future time a random quantity of a commodity. Using combinations of several types of contracts, the firm seeks to reduce its downside risk while maximizing profits. Different type and combinations of contracts are considered. Since the focus is on the energy markets I consider hedging both with such popular and liquid contracts as options and forwards as well as with new types of contracts that are just starting to be used in energy risk management. The properties of options and forwards are well studied in finance. However in the case of corporate risk management the trade-offs caused by the non-linear nature of options are not very well understood. Another difference from financial markets is that the market price of risk, an important parameter when considering trade-offs between maximizing profits and reducing risk, can be positive as well as negative. The sign of the market price of risk significantly influences the qualitative nature of optimal hedges. To address this concern the dissertation contains an empirical analysis designed to estimate the sign of the market price of risk for energy. Although such standard financial instruments as forwards and options are used with great success in energy markets, they cannot address a very important property of electricity price behavior – very sharp spikes. Since one of the major reasons for spikes is inelasticity of demand, interruptible contracts, which effectively increase demand response, are gaining popularity among energy retailers. In the dissertation I analyze optimal static as well as dynamic hedging using forwards, options, and interruptible contracts in various settings (i.e., reduced and structural price models). The analysis leads to several nonlinear problems which I address using both analytical and numerical methods. The static hedging problems result in the standard stochastic programming problem which, in some simple cases, can be solved analytically, and otherwise is solved numerically using well established stochastic programming methods.Computational Science, Engineering, and MathematicsComputational and Applied Mathematic