Utilizing geo-referenced and “big-ag” data to improve US agricultural policy
Date
Authors
Journal Title
Journal ISSN
Volume Title
Publisher
Abstract
Study 1: Utilizing Topographic and Soil Features to Improve Rating for Farm-level Insurance Products Previous studies have shown a strong correlation between topographic/soil features and agricultural production; however, linkages between these features and agricultural insurance products are scarce. Agricultural insurance is an ever-growing means of governmental support for producers globally. However, failure to set insurance premiums that accurately reflect risk exposure can lead to low participation rates and/or adverse selection. The U.S. federal crop insurance program partly guards against this at the farm-level by inducing pricing heterogeneity via a rate multiplier curve, which does not consider topographic/soil information. We develop a method for econometrically incorporating this information into existing rating procedures used by the Risk Management Agency (RMA). The empirical application leverages 149,267 farm-level observations of Kansas producers across four dryland crops (corn, soybean, sorghum, and wheat), spanning 46 years, and matched to fine-scale topographic/soil features. The results suggest that incorporating these features does improve the prediction accuracy of yield losses and can, in general, improve rating performance. However, these improvements are specific to farms with limited yield histories, as there are no improvements for farms with the commonly used yield history of ten years. This suggests substantial rating improvements for new farms or those with limited histories for a particular crop, but more general improvements for the program are not likely to occur given a large number of current participants with a full ten-year yield history.
Study 2: Tradeoffs Between Production-History-Based and Index-Based Insurance for Field Crops Agricultural insurance products based on Actual Production History (APH) typically suffer from adverse selection, moral hazard, and high program costs associated with pricing, loss assessment, and monitoring. On the contrary, Index-based insurance offers the opportunity of reducing, and even sometimes eliminating, some of these concerns; however, by design, they cannot guarantee that indemnities will be paid when producers experience losses. This concern is commonly referred to as basis risk and is the biggest limiting factor in the potential expansion of Index insurance programs. An extensive body of literature has shown that basis risk could be reduced to an appreciable extent by improving product design. Nonetheless, a knowledge gap on farm-level tradeoffs between APH- and Index-based insurance exists because observable data is limited. The novelty of this study is that it overcomes these limitations and extends the literature by providing ex-post simulated evidence of the tradeoffs between Index-based and APH-based insurance at the farm level. Using a sample of 5,428 corn, soybean, sorghum, and wheat KS farms from 1973-2018 the study shows that economically significant tradeoffs do exist between APH- and Index-based insurance and that different types of index products are associated with differing levels of basis risk. Index-based insurance that protects against killing-degree-days (i.e., degree-days >30 °C) accumulation generates the most significant gains in economic rents and is associated with relatively low basis risk.
Study 3: The Potential Significance of “Big Ag Data” in Corn Futures Markets The advent of precision agriculture technologies has left researchers to grapple with how to best-use its associated “Big Ag-Data”. While the wealth of information output from precision equipment can easily be aggregated to a higher level in real-time, this poses an interesting question of whether aggregated real-time data will be relevant vis-à-vis periodic information from public sources. To this end, this study utilized advances in event study and yield projection methodologies to test the potential market value of simulated live streamed yield monitor data vis-à-vis USDA report yields. The results shows that the market for corn exhibits only semi-strong form efficiency, as the “news” provided by the monthly Crop Production and World Agricultural Supply and Demand Estimates reports is incorporated into prices in at most two days after the release. As expected, an increase in corn yields relative to what was publicly known, elicits a futures price decrease. On the contrary, live-streamed yield information does not significantly correlate with historic market reactions. Nonetheless, this study advances the market-price event-study methodology by utilizing sources of information not previously considered. Second, the study provides policy implications centered around the ongoing debate about the economic significance of USDA reports in the presence of growing information availability in the private sector.