Farm to Fridge: Digital Traceability and Quality Upgrading in Kenyan Dairy Value Chain

How to cite

Xu, Guanghong. 2024. "Farm to Fridge: Digital Traceability and Quality Upgrading in Kenyan Dairy Value Chain." AEA RCT Registry. AEARCTR-0014255. https://doi.org/10.1257/rct.14255-1.0

Download: BibTeX · RIS — for reference managers without a browser translator. Zotero and most others pick up this page's citation metadata directly.

Guanghong Xu*

* Primary investigator

Registry ID
AEARCTR-0014255
DOI
https://doi.org/10.1257/rct.14255-1.0
First registered
Last updated
Status
On going
Trial dates
2024-04-01 – 2025-01-31
Countries
Private
JEL codes
O12, O13, Q12, Q13, D82, C11
Registry record
http://www.socialscienceregistry.org/trials/14255

Abstract

Disorganized agricultural value chains often prevent the transmission of quality incentives to upstream farmers, especially when quality is unobserved at the farm gate. When the quality-revealing costs are prohibitively high, and without an effective traceability system throughout the value chain, farmers might not be rewarded for producing high-quality products even if quality incentives exist in the downstream markets. Thus, farmers have low incentives to invest in quality upgrading. In this study, I establish digital traceability systems among Kenyan dairy cooperatives and develop an innovative quality monitoring method based on Bayesian models to overcome the high test costs for individuals’ milk. The model uses the aggregated milk container quality information and traceability data to detect whether individual farmers are producing high- or low-quality milk. The model-predicted quality significantly correlates with the one-time milk test among 940 farmers. I conduct an individual-level randomized controlled trial to provide both cooperatives and farmers with either the model-predicted milk quality or one-time random quality test results. I evaluate the differential effects on cooperatives’ reactions to different types of farmers, farmers’ investment in quality upgrading, farmers’ milk quality, and sales.

Experimental design

I randomly assign farmers to one of the three treatment conditions. Group 1 farmers’ model-predicted quality information is shared with the cooperatives (as well as with the farmers themselves) as periodic reviews. Group 2 farmers’ one-time quality test information is shared as random monitoring. Group 3 serves as the control for the evaluation.

Design details

Randomization method
Randomization was done in office by a computer
Randomization unit
Farmer
Clusters
940 Farmers
Observations
940 Farmers
Arms
Group 1 (model detection): 315 farmers; Group 2 (random tests): 314 farmers; Group 3 (control): 311 farmers.
Primary outcomes
milk quality, sales, farming practices

Cited by

No citation data available for this registration yet.

Keywords