2.3 The hybrid approach to demand modeling

This paper presents an engineering-economic approach for developing energy demand models. The approach combines the economic theory of production with engineering process data. We review the economic theory of production, the Constant Elasticity of Substitution (CES) function, and the Logit function. Using the Oak Ridge Industrial Model (ORIM) as a case study, we show how the Logit function can be used to specify a fuel share model and how the CES function can be used to specify an energy conservation model. Shephard's lemma is a key result of the economic theory of production. The lemma can be used to derive input-output coefficients from a cost function. We use CES cost functions because they have good global properties. To obtain more flexibility, we use nested CES cost functions. To identify the parameters in a nested CES structure, we assume that the multilevel cost function corresponds to a multistage production process. When input-output coefficients are derived from a cost function, the tradeoffs between the factors satisfy the law of diminishing returns. Energy conservation occurs when capital, labor, or materials are substituted for energy service. In most cases, the substitution obeys the law of diminishing returns. Thus, a CES cost function can be used to specify an energy conservation model. Normally, perfect substitutes do not obey the law of diminishing returns. Since we include all of the costs for each option (the options use fuel and provide energy service) in ORIM, the options are perfect substitutes and do not satisfy the law of diminishing returns. Since the options are perfect substitutes, we use the Logit function to specify the ORIM fuel share model.