Abstract
Farm managers possess a broad spectrum of capabilities, ranging from tech-savvy, strategically focused managers to those grappling with operational challenges. Our study examines survey data from 403 US commercial producers in order to identify subsets of producers who differ in terms of resilience, management practices, producer sentiment, and other farm characteristics. Utilizing various supervised and unsupervised machine learning techniques, we uncover key farm characteristics that capture the most pronounced variations in survey responses. We then cluster the dataset using these key variables, maximizing separation between clusters and minimizing differences within clusters using Ward’s hierarchical clustering. Fisher’s exact tests are implemented to determine the statistical significance of differences in farm characteristics across the constructed clusters. Results confirm that resilience to strategic risk, managerial ability, producer sentiment, technology adoption, and demographics all vary significantly among commercial farms. In particular, we observe a trade-off among farms in regard to operators’ management abilities and their farms’ resilience. Farms with the highest resilience levels tend to show slightly lower managerial abilities. Conversely, farms with the strongest managerial abilities exhibit somewhat poorer farm resilience. The third group of farms, which encompasses 49% of our sample, displays the lowest levels of farm resilience, precision agriculture technology adoption, and managerial abilities and has the weakest growth expectations.
Recommended Citation
Lippsmeyer, Margaret; Langemeier, Michael; Mintert, James R.; and Thompson, Nathan
(2024)
"Farm Resilience, Management Practices, and Producer Sentiment: Segmenting US Farms Using Machine Learning Algorithms,"
Journal of Applied Farm Economics: Vol. 7
:
Iss.
1,
Article 4.
DOI: 10.7771/2331-9151.1080
Available at:
https://docs.lib.purdue.edu/jafe/vol7/iss1/4