In this dissertation, I examine three issues related to agricultural and catastrophic risk management. In my first essay, "Modeling Yield Distribution in High Risk Counties: Application to Texas Upland Cotton", I develop models of yield distributions for crops and regions whose yields exhibit irregular behavior. Specifically, I undertake a statistical case study of Texas upland cotton, which in recent years has exhibited poor actuarial performance under the U.S. crop insurance program, and propose an alternative mixture distribution model in which the conditional distribution of yield depends upon an observable drought index. The results indicate that the mixture distribution model provides a better fit to the data than conventional parametric distributions and, for more than two-thirds of the Texas counties examined, implies Group Risk Plan insurance premium rates that are greater than those currently published by the U.S. Department of Agricultural Risk Management Agency.

In this dissertation, I present three essays studying agricultural insurance in both developing and developed countries. Farmers face numerous risks including production, price, financial, etc. In response, farmers may seek agricultural insurance to cover some of these risks. Different questions of rainfall insurance and crop insurance are considered in this dissertation, and natural experiment and Propensity Score Matching method are applied in measuring the impact a “treatment” on insurance participation rates for farmers. Chapter 1 is an extensive review of China’s agricultural insurance market. I first summarize the history of the market for agricultural insurance in China, which between 2007 and 2013 grew an average of 30% annually and is now the second largest market in the world. I then review major agricultural insurance programs introduced by the Chinese government during this period. I also identify the major challenges that have to be overcome to further develop agricultural insurance in China, and discuss reforms that have been proposed to address them. In Chapter 2, I estimate farmers’ willingness to pay (WTP) for a hypothetical rainfall index insurance product, with a special interest in whether recent experience with a natural disaster increases the WTP. My findings are based on a survey of small farm households in China’s Heilongjiang Province, which suffered a large flood in the summer of 2013. I find a significantly higher propensity to purchase rainfall index insurance for farmers who experienced the flood in 2013 than farmers who did not. I also find a much higher WTP for rainfall index insurance for farmers affected by the flood. My findings imply that past experience in major disasters could alter farmers’ risk perception and lead to a higher demand for risk management tools such as crop insurance. In Chapter 3, I examine whether U.S. farmers using fixed-cash lease are more likely to purchase federal crop insurance (FCI). Farmland rental lease choice plays a significant role in farm management decision making. Fixed-cash agreement has become the most widely used leasing contract in the United States, despite the fact that it is riskier than other types of farmland leasing contracts. A propensity score matching method is used to compare FCI participation rate between farmers under fixed-cash rental contract and farmers under other types of leasing arrangement. By analyzing a farm-level national survey data, the results show that using fixed-cash rental agreement significantly increases FCI participation rate, providing evidence that FCI is adopted by farmers as an effective risk management tool to reduce production risks.

The general theme of this dissertation is risk and uncertainty in agriculture, with each chapter addressing a specific topic related to agricultural risk and uncertainty. Chapter 2 examines the effects of production uncertainty on the types of contract structures used in specialty grain markets. A theoretical model of a contractual relationship between a monopsonistic processor and risk-neutral producers is presented. Two common contract structures, and their resulting effects on the sharing of production risk between buyer and seller, are compared. The spatial structure of yields and farm-level yield volatility are shown to have significant impacts on the processor's preferred choice of contract structure and expected profits of both the processor and farmers in the resulting equilibrium. Chapter 3 provides a critical look at a classic definition regarding the relationship between input use and risk, and attempts to reconcile an apparent paradox in the production literature. Experimental corn yield response data is used to estimate a stochastic production relationship between applied fertilizer, soil nutrient availability, and corn output. Optimal fertilizer application rates for risk-averse and risk-neutral producers are found using numerical methods. In addition to the empirical analysis, primary data collected through a farmer survey instrument, designed to elicit information from farmers regarding their risk attitudes and subjective beliefs regarding the relationship between risk and fertilizer use, is presented and compared with the results of the empirical analysis. Chapter 4 turns to the opportunities for managing weather risk using weather derivative markets. Developing regions are areas in which weather based risk management tools show significant potential. However, the success and long-term viability of insurance programs depends heavily on the availability of accurate and reliable historical data. The lack of this type of historical data for developing regions is one of the largest obstacles to insurance program development in these regions. A framework which utilizes statistical methods to estimate unbiased rainfall histories from sparse data is developed. To validate the methodology's usefulness, a drought insurance example is presented using a rich data set of historical rainfall at weather stations across the state of Iowa.