| Natural Gas is commodity like few others. Excluding electricity, it is the most volatile commodity traded. The price of gas is dependent primarily on weather, with local price shocks felt at other geographic locations to the extent that locations are connected by pipelines with spare capacity.;This paper takes a new approach to modelling natural gas. Instead of modelling the commodity at one location, an approach is developed whereby the natural connections between locations are incorporated.;Furthermore, as gas prices can exhibit both contango and backwardation, a stochastic convenience yield is included in the model as well as stochastic interest rates.;This term structure approach is not unknown in financial modelling; however, incorporating multiple risk factors that correspond to various locations is a new perspective. This paper also empirically tests the data from gas forward prices at Chicago, NYMEX and AECO to understand the statistically properties at each location and to ensure the proposed model is robust enough to include these properties.;This thesis also investigates the time series property of the difference of two locations (the basis) and notes that these empirical properties are consistent with the model properties.;Finally, this paper derives closed-form option solutions for call options of forward contracts and call options on forward basis. The options are calibrated and compared to other models. The thesis concludes with directions for future research. |
