Land-Use Changes and CO2 Emissions Due to US Corn Ethanol Production 543 Feed composite Livestock Processed feed Crops Sugar crops …… Other agriculture Other grains Intermediate inputs from livestock and processed livestock Energy-protein DDGS-coarse grains Coarse grains DDGS Oilseed-meal Oilseeds Meals Figure Structure of the nested demand for feed in livestock industry Table Cost shares of major feed items in the US livestock industries in 2001 and 2006a Feed items Coarse grains Other crops DDGS Oilseeds meals a 2001 2006 Dairy Meat ruminant Nonruminant Dairy Meat ruminant Nonruminant 67.6 6.4 5.6 20.3 68.4 10.4 6.4 14.9 82.9 2.9 1.1 13.1 61.0 6.0 12.9 20.1 57.5 9.8 17.6 15.0 82.5 2.7 2.1 12.7 Processed feed is dropped from this table to highlight shares of items listed in the table regions, they applied a relatively high elasticity of substitution, 20, between these two feed materials for all types of animal species Following the literature, they used values of 25, 30, and 20 for the elasticities of substitution between coarse grains and DDGS in the dairy farms, other ruminant, and nonruminant feed structure, respectively They also applied a small value, 0.3, for the elasticity of substitution between the energy and protein feedstuffs because DDGS could displace a portion of meals in some feed rations, as shown in Arora et al (2008) and Fabiosa (2009) In the composite of other crops and composite of processed livestock inputs, they applied elasticities of substitution of 1.5 for all types of livestock industry Finally, following Keeney and Hertel (2005), they used elasticity of substitution of 0.9 at the higher level of the feed demand nest This article uses some GTAP simulation results to show how these elasticities shape the cost structure of the livestock industry To accomplish this task, it uses the results obtained from the simulations introduced in its next section In particular, it uses the results of the first simulation of the second group of experiments This particular simulation replicates transition of the global economy from 2001 to 2006 The results of this simulation predict that the cost shares of coarse grain, other crops, and meals in the US livestock industries declined during the time period 2001–2006, whereas the cost share of DDGS increased The largest substitution is DDGS for coarse grains, but there is also substitution for other crops and oilseed meals, depending on the livestock species Note that processed feed was dropped from the list of animal feeds to highlight the changes in the shares of crops, DDGS, and meals during this time period (Table 1) Global markets (including agriculture) are complex and highly linked This complexity and these linkages make it difficult to accurately estimate how a market perturbation, such as the use of sufficient US corn to produce 15 BGs of ethanol, would impact land clearing and other aspects of landuse change around the world This discussion indicates how a CGE model of the global agricultural economy has been structured to allow it to make such a prediction In particular, the GTAP-BIO-ADV model and its database described earlier in this section and in the associated references have the following features: It covers production, consumption, and trade of three types of biofuels: ethanol from crops and from sugarcane, and biodiesel from crude vegetable oil Byproducts are DDGS and oilseeds meals The crude vegetable oil industry uses oilseeds and produces crude vegetable oil and oilseed meals The biodiesel industry uses crude vegetable oil to produce biodiesel The demand for feedstuffs follows a three-level nesting structure The land module handles two new land categories of unused cropland and cropland pasture Although the model could trace changes in these two groups of land