Do Non-Classical Measurement Errors Affect the Effects of Farm Input Subsides on Crop Productivity? Evidence from Malawi
No Thumbnail Available
Date
2026-01
Authors
Martin Limbikani Mwale
Journal Title
Journal ISSN
Volume Title
Publisher
AERC
Abstract
Non-classical measurement errors produce biased and inconsistent estimates
in estimations. Previous studies have examined the effects of farm production
on crop productivity without accounting for non-classical measurement
errors. Using satellite data, which is free from neo-classical errors, this study
sought to establish whether the effects of Malawi’s farm input subsidy on crop
productivity are biased. The study compared results generated using satellite
yields to those generated using yields reported by farmers through a survey.
The study tested the sensitivity of the satellite estimates by changing the
possible yields to fertilizer response rates, and the results remained consistent
with the main findings. The findings revealed that the effect of farm input
subsidies on crop productivity generated through survey data are upward
biased. Farmers over-report yields to demonstrate that they are productive
and retain their status as subsidy beneficiaries. Studies on farm input
subsidies, therefore, need to pay attention to non-classical measurement
errors to provide reliable results and policy advice on farm input subsidies.
Furthermore, policy on farm input subsidies needs to strengthen the targeting
of beneficiaries to evade inclusion and exclusion errors of productive farmers