Estimating Aflatoxin Spillover Effects at Different Production Stages Using a Spatial Autoregressive Regression Model: The Case of Peanuts in Mali
35 Pages Posted: 9 Oct 2012
Date Written: September 13, 2012
Management practices that reduce the incidence of aflatoxin contamination have focused on decreasing aflatoxin levels on the farm or in the storage unit without taking into account the potential impact of neighboring farms and storage units. In this article I discuss an application of a spatial autoregressive regression model not previously studied in the literature: the estimation of aflatoxin prevalence levels in peanuts in Mali. I estimate this model using a spatial two-stage least squares procedure that produces a consistent estimator of the parameters in the model. I find that the aflatoxin levels of farmer i’s neighbors help explain farmer i’s own aflatoxin level only during the post-harvest stage of production. In contrast, during the pre-harvest stage, the neighbors have no influence on the farmer’s aflatoxin level. This suggests that farmers’ sensitivity to the aflatoxin prevalence levels in their reference group differs according to the different stages of the production process.
Keywords: Aflatoxin, Spillovers, Mali, Spatial autoregressive models
JEL Classification: Q12
Suggested Citation: Suggested Citation