Spatial and intertemporal arbitrage in the California natural gas transportation and storage network

Intertemporal and spatial price differentials should provide the necessary signals to allocate a commodity efficiently inside a network. This dissertation investigates the extent to which decisions in the California natural gas transportation and storage system are taken with an eye on arbitrage opportunities. Daily data about flows into and out of storage facilities in California over 2002-2006 and daily spreads on the NYMEX futures market are used to investigate whether the injection profile is consistent with the "supply-of-storage" curve first observed by Working for wheat. Spatial price differentials between California and producing regions fluctuate throughout the year, even though spot prices at trading hubs across North America are highly correlated. In an analysis of "residual supply", gas volumes directed to California are examined for the influence of those fluctuations in locational differentials.; Daily storage decisions in California do seem to be influenced by a daily price signal that combines the intertemporal spread and the locational basis between California and the Henry Hub, in addition to strong seasonal and weekly cycles. The timing and magnitude of the response differs across storage facilities depending on the regulatory requirements they face and the type of customers they serve. In contrast, deviations in spatial price differentials from the levels dictated by relative seasonality in California versus competing regions do not trigger significant reallocations of flows into California.; Available data for estimation of both the supply-of-storage and residual-supply curves aggregate the behavior of many individuals whose motivations and attentiveness to prices vary. The resulting inventory and flow profiles differ from those that a social planner would choose to minimize operating costs throughout the network. Such optimal allocation is deduced from a quadratic programming model, calibrated to 2004-2005, that acknowledges relative seasonality in demand, trade-offs between transportation and storage costs, infrastructure configuration and regulatory requirements. A comparison of the simulated equilibrium with observed behavior identifies where the arbitrage opportunities lie. Moreover, scenario analysis of such as a LNG terminal or additional storage capacity in California reveals the considerable indirect network effects brought about by changes at any node or arc.