Calibration of a Land Cover Supply Function Using Transition Probabilities

An important question that arises frequently in the economic analysis of environmental and energy policies is, how does the supply of land across various uses change in response to policies? The GTAP Modeling framework of Hertel (1997) addresses this question by determining the supply of land across different uses through a Constant Elasticity of Transformation (CET) supply function. In the standard GTAP Model, the only type of land explicitly modeled is agricultural land, and this is distributed across uses with a one-level, CET function. However, in the GTAP-AEZ framework, data are available on a wider range of activities, including forestry (Sohngen et al., forthcoming). Given the recent availability of data on land cover (Ramankutty et al., 2007) and harvested cropland (Monfreda et al., forthcoming) the land supply decision is more naturally divided into the allocation of land cover across forestry, grazing and crops, followed by the allocation of harvested area across cropping activities. In this Research Memorandum, we focus on the former problem, namely the allocation of land cover between these three competing commercial uses. Naturally, the quality of the land cover responses produced by GTAP-based simulations is contingent on the value of the CET parameter. And the value of this parameter is likely to depend on the length of run for the analysis in question. This research memorandum describes the empirically based calibration strategy used to determine the value for the CET parameter, based on recent research in the United States. The following section reviews the theory of the CET parameter and how it is relevant for modeling land supply. Sections 3 and 4 describe the calibration methodology and data, while the final section discusses the calibration results.