The video begins at 2:51.

Abstract: Coal-fired electricity plants account for over 50 percent of the nation’s electricity. These plants can purchase coal from a large number of different locations and, often, can have a number of different transportation options. Normally, however, from the array of different options, they use only a handful.  We frame the model as that of a cost-minimization with a large number of input choices, characterized by standard Kuhn-Tucker conditions. Empirically, we estimate a system of input decisions that contain both zero and non-zero levels, using a Multiple Discrete/Continuous Extreme Value model. The payoffs from each choice are a function of costs, coal attributes, and unobserved modal attributes, as well as the increased regulation under the Clean Air Amendment Act of 1990.