Currently, the project is assessing NEMS sectoral demand relationships for selected fuels.

Steps

(1) Identify important explanatory variables using NEMS documentation.

(2) Configure and run NEMS scenarios with selected explanatory variables varied individually and extract the appropriate solution series.

(3) Test results for nominal sensitivities (relative to the AEO2007 Base Case). Example of sensitivity test.

(4) Set regression specification.

(5) Configure and run NEMS scenarios for Base/High/Low values for the selected explanatory variables based upon a (sometimes modified) Latin hypercube design. Altogether, including the AEO2007 Base Case, ninety-five NEMS solutions were created to support the modeling of demand for ten fuels.

(6) Inspect "apparent" linearity by comparing the sum of impacts for each explanatory variable changed individually with the impact of changing all explanatory variables at the same time, the "total impact". To the degree that the sum of the individual impacts is the same as the total impact the underlying NEMS structure appears linear.

(7) Pool data and run regressions. The regressions are based on data pooled for the solution years 2010-2015.