Survey Measurement Errors and the Assessment of the Relationship between Yields and Inputs inSmallholder Farming Systems : Evidence from Mali

An accurate understanding of how input use affects agricultural productivity in smallholderfarming systems is key to designing policies that can improve productivity, food security, and living standards inrural areas. Studies examining the relationships between agricultural productivity and inputs typically rely on landproductivity measures, such as crop yields, that are informed by self-reported survey data on crop production.This paper leverages unique survey data from Mali to demonstrate that self-reported crop yields, vis-à-vis(objective) crop cut yields, are subject to non-classical measurement error that in turn biases the estimatesof returns to inputs, including land, labor, fertilizer, andseeds. The analysis validates an alternative approach to estimate the relationship between crop yields andagricultural inputs using large-scale surveys, namely a within-survey imputation exercise that derives predicted,otherwise unobserved, objective crop yields that stem from a machine learning model that is estimated with a randomsubsample of plots for which crop cutting and self-reported yields are both available. Using data from a methodologicalsurvey experiment and a nationally representative survey conducted in Mali, the analysis demonstrates that it ispossible to obtain predicted objective sorghum yields with attenuated non-classical measurement error, resulting in aless biased assessment of the relationship between yields and agricultural inputs. The discussion expands on theimplications of the findings for (i) future research on agricultural intensification, and (ii) the design of futuresurveys in which objective data collection could be limited to a subsample to save costs, with the intention to applythe suggested machine learning approach.