Feedlot cattle crush: Joint distribution model development, evaluation, and application.

摘要

Successful marketing, investment, and risk management endeavors rely heavily on the ability to assess the risk facing a firm and to incorporate all relevant, available, and affordable information in making managerial decisions. Identifying sound ways to make managerial decisions that reduce uncertainty and/or increase mean returns over time is critical to maintaining the viability of any business, and in particular to agricultural producers who face numerous uncertainties with production and government policy that are not nearly as inherent in other businesses. This dissertation attempts to improve the way in which cattle feedlot managerial decision making is modeled. In particular, this dissertation develops a model that better represents the price risks and opportunities actually faced by feedlot investors and managers. This is accomplished by recognizing the multivariate situation that feedlots operate in and employing appropriate multivariate models, rather than univariate examinations that currently dominate the literature.; The main objective of this research is to develop a practical method to use in providing a joint distribution of live cattle, feeder cattle, and corn cash prices at a future time in a manner from which the price forecasts can be of actual realistic use to cattle feedlot managers and investors in assessing different opportunities and market situations. In accomplishing this main objective, a set of additional contributions to the literature are made, including (1) an examination of the ability of implied volatility, historical volatility, and composite approaches to forecast volatility realized in subsequent periods of live cattle, feeder cattle, and corn prices analyzed using univariate techniques, (2) an evaluation of the appropriateness of the volatility forecasting techniques as inputs in a joint distribution simulation model, and (3) a demonstration of the economic value of employing our joint distribution model in subsequent simulated models representative of situations faced by practitioners.; Results of this dissertation suggest that properly modeling the cattle crush as a multivariate set of prices significantly improves the accuracy of forecasting future cattle crush values realized in the cash market. In addition, the developed joint distribution model is shown to provide distributional information that is currently lacking in the literature which relies heavily on futures market price implied techniques of forecasting the cattle crush.