Persistent narratives: why is the "Failure of Bt cotton in India" story still with us?

Whose Numbers Count? Persistence of the Bt-cotton-failure story has seemed puzzling to me. The issue seemed settled: Indian farmers had collectively decided that Bt technology in cotton was useful—neither a miracle seed nor a suicide seed. Insect resistance was a valuable trait and offered some respite from the pesticide treadmill that is both financial and biological. How do we account, then, for the persistence of reports in and about India that “Bt cotton has failed”—in the sense of agro-economic catastrophes ending in suicides, deaths of livestock grazing in Bt fields, allergenicity, and so on? Could purely instrumental reporting of failure be the reason? That is, are the reports of catastrophic failure at variance with the dominant global pattern and general Indian experience purposely falsified? At the Ravello conference of 2008, my paper spent some time on the question of how honest—i.e., noninstrumental—errors could be made in measuring yield and income effects of Bt cotton on-farm in India. What had settled in India by 2008 was an empirical consensus about Bt cotton: the technology works as predicted, with predictable results, increasingly well-understood by farmers, and incorporated into their risk-avoidance strategies (Roy, Herring, & Geisler, 2007). One excellent recent summary of studies, including their own, is that of Rao and Dev’s presentation entitled ‘Biotechnology in Indian Agriculture: Evidence from Panel Studies on Bt Cotton’ (2008).1 In what the Bush administration in the United States dismissively refered to as the “realitybased community,” something approaching consensus had emerged from years of partial results and some considerable confusion produced by studies claiming disaster in the Bt fields.