The Effect of Weather Conditions on Fertilizer Applications A Spatial Dynamic Panel Data Analysis

Document pages: 37 pages

Abstract: Given the extreme dependence of agriculture on weather conditions, this paperanalyses the effect of climatic variations on this economic sector, byconsidering both a huge dataset and a flexible spatio-temporal modelspecification. In particular, we study the response of N-fertilizer applicationto abnormal weather conditions, while accounting for other relevant controlvariables. The dataset consists of gridded data spanning over 21 years(1993-2013), while the methodological strategy makes use of a spatial dynamicpanel data (SDPD) model that accounts for both space and time fixed effects,besides dealing with both space and time dependences. Time-invariant short andlong term effects, as well as time-varying marginal effects are also properlydefined, revealing interesting results on the impact of both GDP and weatherconditions on fertilizer utilizations. The analysis considers fourmacro-regions -- Europe, South America, South-East Asia and Africa -- to allowfor comparisons among different socio-economic societies. In addition tofinding both spatial (in the form of knowledge spillover effects) and temporaldependences as well as a good support for the existence of an environmentalKuznets curve for fertilizer application, the paper shows peculiar responses ofN-fertilization to deviations from normal weather conditions of moisture foreach selected region, calling for ad hoc policy interventions.