The Economics of Pesticides and Pest Control299Table 2. Sources of suboptimal use of pesticides and their impactsLevel and cause of impactSource of suboptimalityFarmRegionGlobalPesticide resistanceXXDestruction of beneficial speciesXXXWorker safetyXXFood safetyXGroundwater contaminationXXXDamage to nontarget speciesXXXAir pollutionXXXProperty damageXaway from a population center. Lichtenberget al.(1993) show the damage associatedwith exposure to sprayed chemicals varies between regions that have dry harvest seasonsand wet harvest seasons. Furthermore, the method and timing of application matter.The heterogeneity in pesticide costs and benefits implies optimal use can be achievedby financial incentives only if they are specific to chemicals, uses, localities, applicationmethods, use levels, etc. The difficulty in designing simple tax systems to address pesti-cide externalities may justify the use of direct controls in many cases. The wide array ofpesticide regulations we witness reﬂect, to a large extent, the heterogeneity and multi-plicity of the impacts of pesticide use. Computational limitations and enforcement andimplementation costs have led governments to use direct controls, and in many casesthey present the best available regulatory approach. Economic research can, however, bevaluable in assessing and improving various regulations.First-best pesticide policies must be calibrated to account for this heterogeneity. Poli-cies must induce greater pesticide reductions where the costs are greater. The difficultyof achieving these conditions has led to a focus on second-best policy alternatives.Another obstacle to the implementation of effective pesticide policy is the nonpointsource nature of pesticide pollution. More generally, it is difficult to determine cause andeffect in farm pollution. This is due to the complicated environmental fate processes thatdetermine the delivery of pollution to environmental resources. It is also caused by thediffuse nature of the pollution source. In any given region — a watershed for instance —there are many farms that may contribute to the contamination of water. Determiningexactly how much environmental pollution is attributable to which farms requires costlymonitoring. The monitoring costs are prohibitive, and precise determinations of theorigins of pollution are seldom made. With point-source pollution, it is possible tostop delivery of pesticides to the environment, but the diffuse nonpoint nature of farmpollution renders such instruments largely infeasible.Because it is difficult to assign pesticide pollution to individual sources, one mayattempt to infer pollution from the use of agricultural inputs, which may be more
300Sexton, Lei and Zilbermanreadily observable. However, complicated production processes and uncertainty abouthow pollution is delivered to the environment render input-based mechanisms imprecise.