Advancing Statistical Methodology for USDA Surveys and Their Analysis

• Activity Number: 489
• Type: Invited
• Date/Time: Wednesday, August 4, 2010 : 10:30 AM to 12:20 PM
• Sponsor: ENAR
• Author(s): Jianqiang (Jay) Wang*+ and Scott Holan and Balgobin Nandram and Criselda Toto and Wendy J. Barboza and Edwin Anderson
• Companies: National Institute of Statistical Sciences and University of Missouri and Worcester Polytechnic Institute and Worcester Polytechnic Institute and National Agricultural Statistics Service and National Agricultural Statistics Service
• Keywords: Bayesian hierarchical models ; Forecasting corn yield ; Composite estimation ; Model comparison

Abstract

Forecasting the end-of-year crop yield is critical for agricultural decision-making. Historically, a panel of specialists known as the Agricultural Statistics Board meets regularly to set estimates based on expert review of a combination of survey data and administrative/auxiliary information. To make this process less subjective and more repeatable, a Bayesian hierarchical model (BHM) is developed that produces superior yield forecasts/estimates, while quantifying all sources of uncertainty. The proposed BHM naturally combines information from multiple monthly surveys, including a field measurement survey and two farmer opinion surveys. The dependence between the monthly updated surveys and serial dependence of annual yield are incorporated at different stages in the hierarchical model. The effectiveness of the model is demonstrated through simulation and application to USDA data.