Files
Abstract
Crop yield distribution estimation has long been a major focus of agricultural policy, insurance, and risk management research. Yet, explicit incorporation of variety or genetic data in the estimation of yield distributions remains elusive in large scale contexts, both within agricultural economics as well as the broader crop sciences. The purpose of this study is to investigate methods for integrating crop variety data into crop yield distribution models. Through a unique data collection effort, we have assembled a large scale dataset of yield, weather, soil type, and varietal data for planted acreage for the last four and a half decades for the U.S. rice market. This dataset represents virtually all rice grown in the southern U.S. since 1970 (approximately 125 million acres). This research is the first to our knowledge that attempts to model the relationships between varieties, yield, soil, and weather using such large scale market data. Significant variation was found among crop variety performance, and on average varieties have improved through time as it regards yield impacts. We find that about half of all technology gains in the U.S. rice market can be explained by the introduction of, and shifting towards, new rice varieties and adaptation related to adoption of better performing varieties, suggesting that the market actively adapts. This is perhaps not surprising given the investments in these markets, the rational actions of economic agents in choosing varieties, as well as advancements in breeding programs.