Perhaps Still the Lesser Evil? The Trouble Investigating Glyphosate Health Effects in the US

Glyphosate, the active ingredient of the herbicide RoundUp, has been subject to extensive scrutiny since the introduction of genetically engineered glyphosate tolerant crops in 1995. However, both a lack of observational studies and continued uncertainty around the toxicity of glyphosate emissions contribute to international political controversy and numerous civil actions at a time when US rural health measures are worsening and youth cancer rates are rising. Recent quasi-experimental studies using genetically engineered crop adoption as a source of exposure variation found glyphosate applications caused deleterious infant health effects in rural US areas, but we are unsure their exposure is truly exogenous. We reassess glyphosate’s effects on infant health using two sources of exposure variation in two-stage least squares instrumental variable approaches for twelve different infant health measures. We first similarly employ genetically engineered crop adoption but using seed sale marketing data, which is a more accurate measure of adoption relative to previous attempts. We then use changes in glyphosate product characteristics from the introduction of generic product types using product sales marking data, which is a novel separate source of variation. We do not find any consistent significant health changes from glyphosate exposure, where estimates are generally precise, pass typical assumption tests, and are robust to alternative specifications. By investigating the association between genetically engineered crop adoption and several rural co-occurring health and agricultural trends, we obtain evidence exposure variation from genetically engineered crop adoption is probably not independent from other agricultural production externalities. While this study provides evidence glyphosate did not contribute to rural health problems from 1995 to 2013, we believe pesticide data limitations, continued glyphosate controversy, and remaining potential for toxic agricultural emissions justifies evaluations of improved environmental monitoring.