EUVIMEDEUVIMED entdecken
Crossref

A Horse Race Comparison of County-Level Crop Yield Prediction Methods

Junkan Li, Francis Tsiboe

AgEcon Search (University of Minnesota, USA) · 01.01.2025

DOI: 10.22004/ag.econ.391348 ↗

Kurz gefasst

Fachartikel in der Zeitschrift AgEcon Search (University of Minnesota, USA), erschienen am 01.01.2025. Verfasst von Junkan Li und 1 weiteren Autor:innen.

AbstractSchlagwörterBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

Accurate county-level crop yield prediction is essential for agricultural outlooks and risk management, yet the predictive value of complex models remains uncertain. This study conducts a horse race comparison of alternative yield prediction methods for corn, soybeans, and cotton using USDA NASS yield data and PRISM weather data. Models range from simple historical averages to specifications incorporating spatial dependence, time dynamics, and weather variables. Evaluated using out-of-sample forecasts from 2015 to 2024, results show that simple models based on recent county-level yield averages consistently outperform more complex approaches. The findings highlight the robustness and practical value of parsimonious benchmarks for operational yield forecasting.

Schlagwörter

crop yield predictionmaizesoybeancottonspatial dependenceout-of-sample forecast evaluationagricultural outlookagricultural risk management

Zitieren

APA

Li, J., & Tsiboe, F. (2025). A Horse Race Comparison of County-Level Crop Yield Prediction Methods. AgEcon Search (University of Minnesota, USA). https://doi.org/10.22004/ag.econ.391348

Vancouver

Li J, Tsiboe F. A Horse Race Comparison of County-Level Crop Yield Prediction Methods. AgEcon Search (University of Minnesota, USA). 2025. doi:10.22004/ag.econ.391348

BibTeX
@article{Li2025,
  author = {Li, Junkan and Tsiboe, Francis},
  title = {A Horse Race Comparison of County-Level Crop Yield Prediction Methods},
  journal = {AgEcon Search (University of Minnesota, USA)},
  year = {2025},
  doi = {10.22004/ag.econ.391348}
}
Kostenlos starten →