Supply chain management and economic valuation of real options in the natural gas and liquefied natural gas industry

My dissertation concentrates on several aspects of supply chain management and economic valuation of real options in the natural gas and liquefied natural gas (LNG) industry, including gas pipeline transportations, ocean LNG shipping logistics, and downstream storage.;Chapter 1 briefly introduces the natural gas and LNG industries, and the topics studied in this thesis.;Chapter 2 studies how to value U.S. natural gas pipeline network transport contracts as real options. It is common for natural gas shippers to value and manage contracts by simple adaptations of financial spread option formulas that do not fully account for the implications of the capacity limits and the network structure that distinguish these contracts. In contrast, we show that these operational features can be fully captured and integrated with financial considerations in a fairly easy and managerially significant manner by a model that combines linear programming and simulation. We derive pathwise estimators for the so called deltas and structurally characterize them. We interpret them in a novel fashion as discounted expectations, under a specific weighing distribution, of the amounts of natural gas to be procured/marketed when optimally using pipeline capacity. Based on the actual prices of traded natural gas futures and basis swaps, we show that an enhanced version of the common approach employed in practice can significantly underestimate the true value of natural gas pipeline network capacity. Our model also exhibits promising financial (delta) hedging performance. Thus, this model emerges as an easy to use and useful tool that natural gas shippers can employ to support their valuation and delta hedging decisions concerning natural gas pipeline network transport capacity contracts. Moreover, the insights that follow from our data analysis have broader significance and implications in terms of the management of real options beyond our specific application.;Motivated by current developments in the LNG industry, Chapter 3 studies the operations of LNG supply chains facing both supply and price risk. To model the supply uncertainty, we employ a closed-queuing-network (CQN) model to represent upstream LNG production and shipping, via special oceans-going tankers, to a downstream re-gasification facility in the U.S, which sells natural gas into the wholesale spot market. The CQN shipping model analytically generates the unloaded amount probability distribution. Price uncertainty is captured by the spot price, which experiences both volatility and significant seasonality, i.e., higher prices in winter. We use a trinomial lattice to model the price uncertainty, and calibrate to the extended forward curves. Taking the outputs from the CQN model and the spot price model as stochastic inputs, we formulate a real option inventory-release model to study the benefit of optimally managing a downstream LNG storage facility. This allows characterization of the structure of the optimal inventory management policy. An interesting finding is that when it is optimal to sell, it is not necessarily optimal to sell the entire available inventory. The model can be used by LNG players to value and manage the real option to store LNG at a re-gasification facility, and is easy to be implemented. For example, this model is particularly useful to value leasing contracts for portions of the facility capacity. Real data is used to assess the value of the real option to store LNG at the downstream re-gasification facility, and, contrary to what has been claimed by some practitioners, we find that it has significant value (several million dollars).;Chapter 4 studies the importance of modeling the shipping variability when valuing and managing a downstream LNG storage facility. The shipping model presented in Chapter 3 uses a "rolling forward" method to generate the independent and identically distributed (i.i.d.) unloaded amount in each decision period. We study the merit of the i.i.d. assumption by using simulation and developing an upper bound. We show that the model, which uses the i.i.d. unloaded amount, provides a good estimation of the storage value, and yields a near optimal inventory control policy. We also test the performance of a model that uses constant throughput to determine the inventory release policy. This model performs worse than the model of Chapter 3 for storage valuation purposes, but can be used to suggest the optimal inventory control policy, especially when the ratio of flow rate to storage size is high, i.e., storage is scarce.;Chapter 5 summarizes the contributions of this thesis.