Three Essays on Performance and Behavior of Agricultural Market Advisory Services in Corn and Soybeans

This dissertation investigates trading behavior and performance of agricultural market advisory services. The data employed consist of advisory programs tracked by the AgMAS project at the University of Illinois for 1995 through 2004 corn and soybean crops. The first essay develops measures of marketing style for advisory programs and estimates the relationship between style characteristics and pricing performance. Style is measured by the intensity of futures and options use, degree of activeness in marketing, and seasonality of sales. Results indicate that active programs making large bets on price movements obtained a higher average price than more conservative programs. This is consistent with active advisors possessing superior information and/or analytical skills rather than being overconfident. However, estimates of the relationship between pricing performance and activeness are sensitive to the inclusion of a single high-performing program. The second essay employs a Bayesian hierarchical approach to estimate individual expected performance of market advisory programs. Three versions of the model are estimated. The first combines information across the entire sample, while the second includes skeptical beliefs based on the efficient market hypothesis. The third divides programs into two groups based on the degree of activeness in marketing recommendations. Results indicate that even when skeptical beliefs are incorporated into the model a few programs in corn and several programs in soybeans appear to be better marketing alternatives compared to a naive strategy that mimics the market benchmark. However, expected gains are of small magnitude. In the last study, concepts from behavioral finance are employed to study the process that market advisory services employ to develop marketing recommendations. Specifically, this study evaluates whether advisory service recommendations in corn and soybean markets exhibit evidence of the representativeness bias and loss realization aversion. Estimated relationships between trading recommendations and recent price history do not provide evidence of advisory programs extrapolating past price movements. Also, there is no relationship between the time that futures positions remain open and whether the positions represent gains or losses. Therefore, results do not imply behavioral biases in trading recommendations.

Name thesis:degree_name

Ph.D.

Level thesis:degree_level

Dissertation

Discipline thesis:degree_discipline

Agricultural and Consumer Economics

Grantor

University of Illinois at Urbana-Champaign

Year dc:date

2015

Author dc:creator

• Cabrini De Colonna, Silvina Maria

Contributors dc:contributor

• Irwin, Scott H.

• Business Administration, Marketing

Language dc:language

spa

Identifier

(MiAaPQ)AAI3250215

OAI identifier oai:identifier

oai:www.ideals.illinois.edu:2142/82983