Hierarchical crop yield linear model

Abstract

This research investigates which statistical procedure—panel random effect model or hierarchical linear model accounts for the observed spatial random variation in crop yields. Identification of the statistical procedure is accomplished using Akaike information criteria, covariance test of the spatial random variations and out-of-sample performance using holdout sample. Following the identification of the statistical procedure, normality of crop yield residuals are statistically tested using skewness, kurtosis and omnibus test. An empirical application to United States county yields of 20 crops, grown across 48 states during 1957–2013 suggests the need to account for random variation based on the multi-level hierarchy.