Emergency Government Interventions: Case Study of Natural Gas Shortages

Abstract

We present a framework, decision models, and supporting methods to improve government's decisions on a class of intervention problems. The example used is on natural gas shortages. We provide three decision models that move progressively from regulated market mechanisms as a means of gas allocation to nonprice-based directives as shortages become more serious. A critical aspect of this framework is government's timing in switching decision models and changing levels of decision variables. We provide new results on chance constraints that enable time series forecasting as the basis for triggering intervention decisions. In particular, we show that the forecasted regression quantile function for the underlying stochastic process of a chance constraint is the deterministic equivalent for the constraint. Lastly, we evaluate some government decisions on the 1976/77 natural gas shortage in Ohio by applying our methods in a simulation of the conditions of that time. One finding is that some of the emergency actions taken were probably unnecessary.