Data Sets

There are multiple data sets to work with.  The difference between these data sets has to do with how uncertainties are calculated.  Read about each set individually below.

Elasticity Sweep

The elasticity sweep refers to a set of model runs (approximately 5,000) performed to explore the sensitivity of forecasts to uncertainty in a set of static economic parameters called substitution elasticities. These elasticities control the ability of firms and consumers to respond to variations in price signals by substituting out of demand for one commodity and into another. This study was performed on the CIM-EARTH v01 prototype by varying 71 unique substitution elasticity parameters: 16 independent parameters (one for each commodity producing sector) control firms’ abilities to substitute between capital and labor as the relative price of the two factors changes, 16 independent parameters (Armington elasticities; one for each imported commodity) control the ability of importers to substitute import demand of a particular commodity between different trade partners, 16 independent parameters (one for each imported commodity) control the substitutability for firm and consumer demand between the domestic and imported versions of each of the 16 different commodities, and 23 other elasticities that control the ability of different sectors of the economy to substitute between various types of production inputs—types of fossil energy inputs, aggregate energy inputs vs. the aggregate of capital and labor, and raw resource inputs (land or raw fossil fuel resources) vs. non-resource inputs. A detailed explanation of the study is available in the paper, Propagation of data error and parametric sensitivity in computable general equilibrium model forecasts

Share Sweep

The share sweep refers to a set of model runs (approximately 10,000) performed to explore the sensitivity of forecasts to uncertainty in the underlying expenditure data which is used to calibrate the models share parameters. These parameters are calibrated so that the model reproduces the known data for a single base year (or set of years) accurately. In this case the model is calibrated to 2004 using version 7 of the GTAP expenditure database. This study was performed on the CIM-EARTH v0.1 prototype by varying 1600 unique expenditure values from a version of the GTAP 7 expenditure dataset aggregated to 16 regions and 16 sectors and 4 primary factors per region. A detailed explanation of the study is available in the paper, Propagation of data error and parametric sensitivity in computable general equilibrium model forecasts