Here's an excellent response to an article recently published in India's 'Economic & Political Weekly' (April 9th 2005) which was authored by Gopal Naik, Matin Qaim, Arjunan Subramanian and David Zilberman. The piece, "Bt Cotton Controversy: Some Paradoxes Explained", attempts to question studies showing problems with Monsanto's Bt cotton in India.

One of the industry's principal doctrinal myths is the overwhelming success of GM crops in the Third World. You can imagine the embarrassment, therefore, when farmers in the first country to approve Monsanto's Bt cotton in Asia - Indonesia - experienced such bad results that the company was actually forced to abandon selling GM seeds in Indonesia.

It became essential that Bt cotton prove beneficial in India. Unfortunately for Monsanto, Bt cotton commercialisation was instead dogged by a whole series of bad reports from farmers, NGOs, independent scientists and even State governments.

Monsanto via its own surveys and via industry-friendly scientists has done its best to paint an entirely different picture. One of the most notorious pieces of research was a paper by Matin Qaim (University of Bonn) and David Zilberman (University of California, Berkeley) published in SCIENCE ("Yield effects of genetically modified crops in developing countries." Science Vol 299, No. 5608, pp. 900-902).

Qaim and Zilberman are 2 of the authors now trying to critique the studies showing problems with GM cotton. Their original claimed outstanding (80%!) yield increases from Monsanto's GM cotton - results they projected as relevant to farmers throughout the developing world.

Qaim and Zilberman's paper derived all of its data from Monsanto and its findings were so at odds with the reports coming from Indian farmers that its publication caused a storm of protest.

Devinder Sharma called Qaim and Zilberman's paper a 'scientific fairytale" while Dr Vandana Shiva - pointing out that Qaim and Zilberman based their findings entirely on data drawn from Monsanto's trials and "not on the basis of the harvest from farmers' fields" - dismissed it as "fabricated data that presents a failure of Bt Cotton as a miracle."

The paper proved an embarrassment even to India's pro-GM lobby. In a piece posted on CS Prakash's AgBioView list, former Syngenta man, Dr Shantu Shantharam, complained that, "This kind of shoddy publication based on meagre and questionable field data in reputed journals like SCIENCE do more harm to science and technology development, perhaps set GMO technology backwards."

Response to EPW, April 9th 2005 article titled:
"Bt Cotton Controversy: Some Paradoxes Explained"

It seems to be the norm with Matin Qaim and David Zilberman to bring in data that they have from the past to explain something in the present. While governments, the company involved and independent agencies are discussing Bt Cotton performance data from Kharif 2004, these authors are undertaking a current discussion based on primary data from 2002. In 2002, when the rest of the interested groups were discussing the performance of the first season of approved commercial cultivation, these two authors chose to portray the excellent potential of Bt Cotton based on company-supplied data from field trials prior to 2002.

Coming to the present article, the authors decide to question studies from all the three years of commercial cultivation by independent agencies as well as government bodies, questioning their projection of Bt Cotton as a failure, by presenting data from their own data from 2002. How is their data more correct than data from other sources? How can the paradoxes of all studies be explained from a study that they have done in one year which according to them shows agronomic benefits, when other studies do not have such findings? In any case the conclusion on heterogeneity amongst farmers and germ plasm effects on production is known and logical and applies to any new seed/technology introduced in agriculture and its subsequent performance. By stating the obvious, the authors cannot wish away the findings from other studies.

How about taking on board other studies which also have considered heterogeneity as a factor and have shown that across heterogeneous groups of farmers also, in even the most resource-intense situation, the technology has failed (the DDS/APCIDD reports, for instance?).

There are three things to be discussed here 1. yields in connection with germ plasm and yields in connection with saving in crop loss due to pests; 3. pest control costs and efficacy of Bt Cotton on what it is purportedly brought in for.

1. Yields: If the whole issue rests so much on germ plasm, why hype up the technology why not give credit to higher yield potential of some existing hybrids and teach farmers how not to lose these yields to pests by approaching Non Pesticidal, sustainable practices? Or even allow for a small margin of yield loss, by using a high-yielding conventional hybrid to begin with. We all knew that a large factor of yield is of course the germ plasm why did the company not say so too in the first year? Why insert the gene in non-performing varieties, charge higher seed prices and realise its mistake at the expense of farmers’ incomes, even if we were to assume that the pest control technology works?

The authors need not teach others about things like market prices being affected by staple length and so on (which are of course understood by other researchers). They should instead question the marketing gimmicks of the companies involved which promise a uniform thing across varieties to a heterogeneous group of farmers who purchase Bt Cotton seeds but obviously fail to deliver longer staple length and higher yields were indeed promised in the company’s propaganda without making a distinction between varieties or without explaining what "increased yields" would mean here. Why should protection against crop loss be mis-represented as "Increased Yields"? In a year where such crop loss is minimal, farmers and their supporters are indeed right to question the undelivered promise of Increased Yields.

While talking about yield increases over-compensating the higher expenditure on Bt cotton, the authors deliberately chose to ignore even official data that talks about stress intolerance of Bt Cotton varieties, where no yields have been obtained by thousands of farmers in some locations for two of the three years. Why are there no references to such data? Why is this brushed off as something not connectable to the technology?

2. Pest Control costs: In any case, the decrease in pesticide sprays has not been dramatic since the incidence of secondary pests not only remains but grows as farmers' experience collected through many other studies shows, as university and agriculture department data in some states shows and as experience elsewhere shows. Why is such data not taken on board?

On the other hand, we wonder if they would have included the extremely questionable data put out by a company-commissioned report from 2004, if their EPW article was to have come a week later? Monsanto-Mahyco Biotech chose to commission a study on Bt Cotton performance in 2004 through IMRB, just as they had done with A C Nielson the earlier year.

MMB seems to be under great pressure to show that Bt Cotton is an effective pest management strategy and that pesticide sprays are progressively decreasing in Bt Cotton, probably to counter the experiences of farmers that bollworm resistance is in fact building up in their observation. The IMRB survey has put the pesticide costs of non-Bt hybrid users at around Rs. 1412/- in spite of an average use of 6.24 average sprays per acre. This finding seems highly inconsistent with data from other sources on the matter, and given that farmers use high-value, low-volume pesticides these days . Such an inconsistency is also reflected in the average pesticide cost of a meagre Rs. 275/- per acre for 1.73 sprays per acre on Bt cotton. This data is further inconsistent with MMB's own survey of last year (the AC Nielsen study). In AP, in the second year (Kharif 2003), just bollworm control costs in Bt Cotton were reported to be Rs. 1369/- per acre as per AC Nielsen/ORG-MARG. For Kharif 2004, it is inexplicable how the total cost of pesticides in Bt Cotton works out to Rs. 283/acre.

Other questions to the authors:

ß Why did the authors not show us their original data analysis on heterogeneity amongst farmers before arriving at Table 5? What was the profile of farmers that they studied in terms of capturing heterogeneity? For what kind of farmers are they finally recommending Bt Cotton as an answer? How many such farmers exist in India, according to the authors?

ß What do the authors want to make of the high Standard Deviation values of Bt Cotton and its Gross Margin of Rs. 5294/-?

ß The authors further show from their data that Bt Cotton adopters in Karnataka and Tamil Nadu realised the biggest net benefits. If that is the case, why is the extent of Bt Cotton in Tamil Nadu coming down (from 19000 acres in Kharif 2003 to 13000 acres in 2004)?

ß If increasing demand is an indicator of a technology’s effectiveness or performance, how do the authors explain the increasing demand for pesticides? Does increasing demand automatically imply higher efficacy, successful experience and so on? Or does it also have elements of hyped up propaganda? Does increasing demand also justify the use at the expense of other concerns?


-An Article in Economic & Political Weekly, April 9th 2005, under the "Perspectives" section, authored by Gopal Naik, Matin Qaim, Arjunan Subramanian and David Zilberman

-Farmers usually use two pesticides by the brand names of "Tracer" (spinosad pesticide) and "Avaunt" (Indoxacarb) to control heliothis. For 75 ml of Tracer, which can be used for one spray over one acre, the cost is Rs. 850/- per container. The effect is supposed to last for 15-20 days; Avaunt, another pesticide in a 200 ml container, costs Rs. 680/-. A 200 ml container is supposed to be used for one spray over one acre to control heliothis. This is supposed to be effective for 7-10 days.