Coble, Keith H.

September, 2022

Estimating Crop Yield Densities for Counties with Missing Data

By: Park, Eunchun ; Harri, Ardian ; Coble, Keith H.
Crop yield densities are often estimated at the county level. However, county-level yield data providers often omit county records due to low participation or other reasons. The data omission can undermine insurance premiumsÕ credibility and thereby lead to restrictions on the provision of area insurance products in specific locations. To address this problem, we propose a novel Bayesian spatial interpolation method to estimate crop yield densities for counties with missing data. Empirical results indicate that our approach is consistently superior to the benchmark approaches. Importantly, our approach offers noticeable estimation accuracy even at a significant level of data omission.
Download Full Article

September, 2021

Divvying Up the Pie: U.S. Adults’ Preferences for USDA Expenditures in 2018

By: Collart, Alba J.; Ishee, Shea G.; Coble, Keith H.
The U.S. Department of Agriculture (USDA) spends roughly U.S.$140 billion yearly in public funds on farm, nutrition, conservation, and other programs, yet scarce research has elicited the preferences of the U.S. public regarding USDA spending. We survey a representative sample of U.S. adults to examine preferences for USDA spending and find respondents would spend less on nutrition, about the same on farm programs, and more on conservation and other programs. However, respondents' allocation toward nutrition increases after receiving information on the USDA's 2018 budget. These results provide insights into the state and malleability of public support for policy options.
Download Full Article

May, 2015

Accounting for Weather Probabilities in Crop Insurance Rating

By: Rejesus, Roderick M.; Coble, Keith H.; Miller, Mary France; Boyles, Ryan; Goodwin, Barry K; Knight, Thomas O.
This article develops a procedure for weighting historical loss cost experience based on longer time-series weather information. Using a fractional logit model and out-of-sample competitions, weather variables are selected to construct an index that allows proper assessment of the relative probability of weather events that drive production losses and to construct proper “weather weights” that are used in averaging historical loss cost data. A variable-width binning approach with equal probabilities is determined as the best approach for classifying each year in the shorter historical loss cost data used for rating. When the weather-weighting approach described above is applied, we find that the weather-weighted average loss costs at the national level are different from the average loss costs without weather weighting for all crops examined.
Download Full Article

December, 2011

The SURE Program and Its Interaction with Other Federal Farm Programs

By: Ubilava, David; Barnett, Barry J.; Coble, Keith H.; Harri, Ardian
We investigate potential effects of the Supplemental Revenue Assistance Payments (SURE) program introduced in the 2008 Farm Bill. Results suggest little impact on optimal crop insurance purchase decisions, though the SURE program does seem to provide an incentive for mid-level insurance coverage. For producers in the price counter-cyclical payment (PCCP) program, SURE payments are actually higher (lower) when commodity prices are high (low). This is not the case for producers in the Average Crop Revenue Election (ACRE) program.
Download Full Article

April, 2009

Techniques for Multivariate Simulation from Mixed Marginal Distributions with Application to Whole-Farm Revenue Simulation

By: Anderson, John D.; Harri, Ardian; Coble, Keith H.
Alternative techniques for representing dependencies among variables in multivariate simulation are discussed and compared in the context of rating a whole-farm insurance product. A procedure by lman and Conover (IC) that is common in actuarial applications is compared to a new technique detailed by Phoon, Quek, and Huang (PQHl. Results suggest that rates derived from the IC procedure may be inaccurate because the procedure produces biased estimates of correlation between simulated variables. This situation is improved with the PQH procedure.
Download Full Article

April, 2005

Target Markets for Grain and Cotton Marketing Consultants and Market Information Systems

By: Vergara, Oscar; Coble, Keith H.; Hudson, Darren; Knight, Thomas O.; Patrick, George F.; Baquet, Alan E.
This paper examines the use of market consultants and market information systems by grain and cotton producers. A model of producer demand for marketing information and consultants is proposed that decomposes price received into exogenous and endogenous components. The analysis is based on a survey of over 1,600 producers. The results suggest that expenditures on market information systems and market consultants are not independent and, more specifically, expenditures on marketing consultants substitute for expenditures on market information systems.
Download Full Article

July, 2001

DEVELOPING AND PRICING PRECIPITATION INSURANCE

By: Martin, Steven W.; Barnett, Barry J.; Coble, Keith H.
Production agriculture and agribusiness are exposed to many weather-related risks. Recent years have seen the emergence of an increased interest in weather-based derivatives as mechanisms for sharing risks due to weather phenomena. In this study, a unique precipitation derivative is proposed that allows the purchaser to specify the parameters of the idemnity function. Pricing methods are presented in the context of a cotton harvest example from Mississippi. Our findings show a potential for weather derivatives to serve niche markets within U.S. agriculture.
Download Full Article

December, 2000

IMPLICATIONS OF CROP YIELD AND REVENUE INSURANCE FOR PRODUCER HEDGING

By: Coble, Keith H.; Heifner, Richard G.; Zuniga, Manuel
New types of crop insurance have expanded the tools from which crop producers may choose to manage risk. Little is known regarding how these products interact with futures and options. This analysis examines optimal futures and put ratios in the presence of four alternative insurance coverages. An analytical model investigates the comparative statics of the relationship between hedging and insurance. Additional numerical analysis is conducted which incorporates futures price, basis, and yield variability. Yield insurance is found to have a positive effect on hedging levels. Revenue insurance tends to result in slightly lower hedging demand than would occur given the same level of yield insurance coverage.
Download Full Article

July, 2000

MEASUREMENT OF PRICE RISK IN REVENUE INSURANCE: IMPLICATIONS OF DISTRIBUTIONAL ASSUMPTIONS

By: Goodwin, Barry K.; Roberts, Matthew C.; Coble, Keith H.
A variety of crop revenue insurance programs have recently been introduced. A critical component of revenue insurance contracts is quantifying the risk associated with stochastic prices. Forward-looking, market-based measures of price risk which are often available in form of options premia are preferable. Because such measures are not available for every crop, some current revenue insurance programs alternatively utilize historical price data to construct measures of price risk. This study evaluates the distributional implications of alternative methods for estimating price risk and deriving insurance premium rates. A variety of specification tests are employed to evaluate distributional assumptions. Conditional heteroskedasticity models are used to determine the extent to which price distributions may be characterized by nonconstant variances. In addition, these models are used to identify variables which may be used for conditioning distributions for rating purposes. Discrete mixtures of normals provide flexible parametric specifications capable of recognizing the skewness and kurtosis present in commodity prices
Download Full Article

December, 1998

ON CHOOSING A BASE COVERAGE LEVEL FOR MULTIPLE PERIL CROP INSURANCE CONTRACTS

By: Ker, Alan P.; Coble, Keith H.
For multiple peril crop insurance, the U.S. Department of Agriculture's Risk Management Agency estimates the premium rate for a base coverage level and then uses multiplicative adjustment factors to recover rates at other coverage levels. Given this methodology, accurate estimation of the base coverage level from 65% to 50%. The purpose of this analysis was to provide some insight into whether such a change should or should not be carried out. Not surprisingly, our findings indicate that the higher coverage level should be maintained as the base.
Download Full Article