Supply from many: studies on heterogeneous US land use decisions at the extensive and intensive margins

Abstract

Price changes affect the profitability of agricultural land use at the intensive margin (i.e. crop choice) and the extensive margin (i.e. land devoted to crop production). Understanding how prices impact localized land use decisions is important for predicting how production and its allocation across producers change with prices. Due to its wide expanse and diverse geography, the productivity US land differs across space and uses. Understanding the drivers of land use decisions while accounting for such diversity is essential for accurately modeling supply response at the regional and national level. This dissertation contains two studies that provide insight into how price changes impact land use decisions at the extensive and intensive margins. In the first chapter examine the corn supply-price relationship in the United States. I perform this analysis using field-level data across the contiguous US (CONUS). This study is unique in that it incorporates micro-level data from over 3 million fields to estimate region-specific supply response and then aggregates results to the national level. The dataset used in this study is nearly comprehensive, representing field-level decisions across fields that accounted for over 88% of national corn production between 2009 and 2016. The findings from this study illustrate the importance of incorporating heterogeneity in supply response models. Supply response to price differed substantially across regions with high supply sensitivity in the north-central US and Mississippi River Delta, moderate sensitivity in Corn Belt states, andlow sensitivity in the western and Gulf Coast states. The relative importance of corn production in the in the Corn Belt states of Iowa, Illinois, Indiana, and Nebraska meant that it was far less sensitive and, in the long-run, more stable to price changes than national corn supply as a whole. Including heterogeneity in supply response also provided policy relevant context to supply response studies. Overall supply response was negatively correlated with area yields. This meant that price changes have a larger effect on planted corn acres and a smaller effect quantity of corn itself. In the last chapter I examine the impact that ethanol plant capacity has on local land use at the extensive margin. The Renewable Fuel Standard (RFS) has been one of the most influential agricultural policies in the past 20 years, increasing general US crop prices by over 20% and inducing a substantial in US ethanol production capacity (Carter et al., 2016; Roberts and Schlenker, 2013). Its effect on cropland extensification was a concern before it was passed since the policy includes a stipulation forbidding ethanol production on cropland converted after 2007. Lands at the extensive margin tend to be less productive and more environmentally sensitive. Extensive transitions also tend to be less frequent than transitory breaks in crop rotations making their impacts longer-lasting. The goal of this final analysis is to isolate the impact of ethanol expansion on cropland transitions from the general price changes. The concurrent increase in general crop prices and ethanol construction from the RFS complicates the estimation of plants’ effects. I isolate these effects using difference- in-differences (DID) which removes impact from common price trends between the treatment and control group. The standard DID approach results show significant pre-treatment effects stemming from non-random ethanol plant construction. Treatment is likely non-random since ethanol plants lo- cate in areas that provide better returns. Factors that impact the returns to plants confound the analysis since they likely also impact cropland transition decisions. To address this, I use propensity score matching to ensure these confounding factors are identically distributed between the treatment and control groups. Under the matched DID models, the expansion of ethanol plants tended to increase cropland retainment and reduce lands transitioning from non-cropland to cropland. While these results seem contradictory, they are consistent with the findings in recent literature. These impacts are thought to arise due to higher program retention in the major US cropland retirement program CRP due to changes that disproportionately impacted major ethanol production areas.