Specification issues in agricultural supply /
| Main Author: | |
|---|---|
| Other Authors: | , , |
| Format: | Thesis Book |
| Language: | English |
| Published: |
1987.
|
| Subjects: | |
| Online Access: | ProQuest, Abstract Link to OAKTrust copy |
| Abstract: | The unobservable nature of some important explanatory variables in economic analyses make it necessary to employ proxy variables in empirical work. Two areas where this is common in the analysis of agricultural supply response are in the specification of output price expectations and government policy variables. This study empirically addresses the question of choosing between proxies for these unobserved variables. Microcomputer software for the calculation and evaluation of Bayesian composite forecasts using matrix beta priors was developed. Prior subjective outperformance probabilities from four experts were used to form Bayesian composite hog price forecasts. The composite forecasts achieved a lower mean squared forecast error than the individual forecasts over the period studied. An indirect test was applied to examine the significance of the improvements in mean squared forecast errors. Three price expectation series were constructed for crops produced under government programs in Iowa and Texas. They included: (a) quasi-rational expectations based on cash prices and (b) futures-based expectations based on futures market information. These two series were used to form a composite expectation using the outperformance criterion. The composite series generally achieved a lower mean squared forecast error than the individual series. The market price expectation series were each used alone and in two different combinations with effective support prices to form nine different specifications for each state. These specifications were sequentially imposed on a dual restricted profit function model to examine the appropriateness of each specification. A non-nested hypothesis test was used in conjunction with out-of-sample predictive accuracy and theoretical consistency as measures of appropriateness. The results support the use of composite price expectations and document the importance of government support prices in modeling agricultural supply. Using the composite price expectations for the program crops, nine alternative specifications of government policy information were developed and examined. The specifications were sequentially imposed on the dual model of agricultural supply. Non-nested tests of hypotheses, out-of-sample predictive accuracy and theoretical consistency were examined. The results document that agricultural producers in individual states respond differently to government programs. They also suggest caution in using non-nested hypothesis tests alone as a model selection criterion. |
|---|---|
| Item Description: | Typescript (photocopy). Vita. "Major subject: Agricultural Economics." |
| Physical Description: | xiv, 253 leaves ; 29 cm |
| Bibliography: | Includes bibliographical references (leaves 81-84). |