Estimating the spatially heterogeneous elasticities of land supply to U.S. crop agriculture

Sustainable long term agricultural production is the primary issue that US agriculture is currently facing. This is due to increased stress on land and water resources resulting from rising agricultural production. For example, decreasing water quality due to nitrogen discharge into groundwater, and expanding croplands encroaching on ecologically vulnerable land pose threats to agricultural sustainability. Various models have been used to assess long run sustainability of US agriculture. However, an important limitation of the existing research is the limited spatial resolution of the global to national scale modelling. This restricts their usage in understanding local environmental impacts and sustainability stresses since land supply responses vary by locality due to agro-ecological, economic and institutional factors. The key indicator of land supply response is the land supply elasticity which is the percentage change in cropland due to a percentage change in land rents accruing to agriculture. The primary objective of this paper is to estimate geographically explicit land supply elasticities to changes in agricultural cash rents, at a resolution of 5 arc minutes or between 5500 and 7600 hectares depending on the latitude, for the contiguous US using a panel data on land use and controlling for land quality and accessibility attributes, types of crops produced, and indicators of climatic factors conducive to crop growth, along with a number of other factors. We also test for the predictive power of our estimates using both in-sample and out-of-sample validation.