‘Studies’ that hype Bt cotton benefits, don’t show how Bt technology contributed to yields: a brief critique

The following is a brief critique in the context of two ‘studies’ that claim great benefits, especially to smallholders, from Bt cotton in India.



Has Bt technology improved cotton yields in India?

Though studies after studies are being brought out by the biotech proponents and industry, claiming that they have findings to show that Bt technology has improved cotton yields in India, and therefore net returns of farmers (increase that more than makes up for the increased cost of cultivation, is the claim) and from those net returns, everything from women’s empowerment to better quality education is being claimed, nobody has been able to give a convincing answer to a fairly common sensical question, based on something that everyone accepts – that bollworm incidence itself has been low since 2002 – if pest pressure is low, how did the Bt technology end up contributing to yield increases?

It is interesting that despite some senior cotton scientists themselves pointing out that this cause and effect relationship being attributed to Bt technology and cotton yields is questionable, the hype of the biotech proponents continues. It is worrisome that the government then chooses to pick up such false hype to argue a case around Bt cotton bringing about changes in the cotton scenario in India on the floor of the Parliament!

Here, we look at two recent studies that are getting much publicity, also because an expensive international PR agency was engaged in the case of one “study”.

Kathage & Qaim (2012): “Economic Impacts and Impact Dynamics of Bt (Bacillus thuringiensis) cotton in India”, PNAS (www.pnas.org/cgi/doi/10.1073/pnas.1203647109 )

Qaim has been an author in other controversial papers related to Bt cotton in India. For instance, it is well-known that data supplied by the crop developer from the field trials was sought to be passed off as the author’s findings in 2002. Incidentally, in a 2005 paper, he was one of the authors who sought to explain away “paradoxes” in Bt cotton by arguing that it is the host germplasm that matters while everything’s fine with the Bt technology.

It is interesting to note that in the present paper under discussion, the authors say in passing, while claiming higher yield per acre with Bt cotton in both time periods being reported about, “this finding is not because of higher yield potentials of Bt hybrids, but because of more effective pest control and thus lower crop losses”. WHILE SUCH A TALL CLAIM IS MADE, NOT A SINGLE PIECE OF EVIDENCE IS PROVIDED TO SUBSTANTIATE THIS. For instance, no analysis is apparent about what the hybrids in the study and control groups were. More importantly, nothing’s shown on whether non-Bt cotton had hybrids or even varietal seed sources.

What is also missing in this paper is any information on pest incidence in the first instance, to make any claims on effective pest control. Anyone following the Indian cotton story closely, and also interacting with farmers would know that bollworm incidence has been low since 2002, when Bt cotton area was negligible. The authors show their clear bias by claiming that widespread adoption of Bt cotton has led to area-wide suppression of bollworm populations, to explain away the fact that both conventional and Bt cotton farmers were spending the same amounts as pesticide costs on an average (with the standard deviation in fact being lesser in the case of conventional farmers). If this is indeed true, what explains the yield increases because Bt cotton can lead to yield increases only if there is pest pressure? An additional question is what explains low pest incidence of bollworms from 2002 itself?

The authors should read the AICCIP reports to get more information on this, the official confirmation of no great bollworm pest pressure on cotton crop, even in years that Bt cotton had not pervaded most of the cotton cultivation. Official documents are clearly pointing out that it is weather conditions that have not allowed the pest load to be high.

We want to draw the readers’ attention to another major problem with the analysis the authors have come up with, to claim that confounding factors and possible non-random selection bias have been controlled for. They have not accounted for the real confounders, which are clearly showing up as the major factors for cotton yield increases in India: shift from varietal seed sources to hybrids and more irrigation being provided to cotton cultivation now, especially in more-productive cotton cultivating states like Gujarat (states which have seen major shifts in these two factors are the ones which are also posted some yield improvements in the past decade, though yields are on the decline in the recent past. There is nothing much shared about how these factors has been controlled for, while the study makes education, age, landholding etc., appear to be the major confounding variables. A simple example should illustrate this: official data from Gujarat shows that despite rapid spread of Bt cotton in the state (official data may not reflect this since unapproved Bt cotton spread more than approved till recently), the average productivity in irrigated cotton is 689 kilos of lint per hectare whereas it is as low as 247 kilos per hectare in unirrigated conditions. This works out to just about one quintal of lint per acre.

This study of the German researchers does not account for this, even as the irrigation is shown to be higher in the Bt cotton plots in the supplementary tables (Table S1). Also, there has been no controlling for higher fertilizer use which is reflected in the same table, with the statistical significance of the difference being at the 1% value in both the time periods analysed. Similarly, a senior cotton scientist in India has been pointing out that a new seed treatment chemical has contributed to yield increases – however, this study does not account for a factor like that.

The study makes a far-fetched claim that increased returns with Bt cotton have also resulted in raised consumption expenditure, a common measure of household living standard. Let us not forget that household consumption expenditure increasing could also be due to various confounding factors including increase in wage rates in India and probably better incomes in non-cotton crops that the farmers were growing etc. No mention is made of what the other crops were and what the other income sources were.

It is also important to note that in absolute numbers, the consumption expenditure for the Bt farmers actually declined from the earlier reporting period (2002-04) to the recent period (2006-08)! Further, consumption between Bt and non-bt households in 2006-08 period is not significantly different. If this is indeed a measure of living standards, then the study findings are actually pointing to something other than what is being interpreted. It is also not clear if has been normalized for family size. Further, without a breakup of consumption expenditure, it cannot be taken as a measure of raised living standard, as the authors have claimed. Such a breakup has not been provided, however. As an economist is pointing out, “they refer to Deaton for arguing in favour of consumer expenditure but do not take his suggestion of using consumer data for generating prices that can be used to deflate”.

Further, to examine the ‘non-random selection bias in technology adoption’, data of the reported households/plots as well as those not reported need to be put out in the public domain for any analysis.

Most importantly, the authors seek to get away with inaccurate reporting on a very important matter. This is by calling their sample farmers as smallholders. According to the Indian government’s definition, farmers classified as small and marginal landholders are those who have a land area of 2 hectares [5 acres approx] or less. They form the majority of farmers in the country. Authors have deliberately misled the readers by calling their sample farmers as smallholders. His sample farmers during all sample periods are above 4 hectares and these are not India’s smallholders.

This study fails to reflect several macro-level trends, including yield declines in Indian cotton. The study also has a major failure in not being able to compare Bt cotton with other successful pest management alternatives like NPM.


A “study” ostensibly done by Council for Social Development, commissioned by Bharat Krishak Samaj (BKS) is being publicized in press conferences around India, by Bharat Krishak Samaj’s Ajay Vir Jakhar, with the help of one of the more expensive PR agencies Hanmer MSL.

While these press conferences are continuing, nobody has seen the study report yet to understand what the study is all about, or analyse its methodology and findings. The Coalition for a GM-Free India would have ideally liked to have put out this critique after going through the report, which is not being shared by CSD and BKS despite repeated requests.

There are a few problems with this publicity blitzkrieg by BKS: (1) The said report on the basis of which these claims are being made is neither READY nor RELEASED (2) CSD’s Dr.Haque has asked Mr.Jakhar to refrain from publicizing the findings and that too in a misleading manner (3) From whatever is known from the summary released, the study is based on discredited methodology (4) the study conclusions being projected to the media seem to have no relation to the reality on the ground and (5) a joint conference of CEE/CSA/CSD 3 weeks back arrived at the conclusion that Bt cotton needs a thorough review. Below are the details:

The Press Releases around this study quote Dr Haque, Director of CSD as such:“It is very evident from the study that hybrid Bt cotton technology is scale-neutral and farmers’ responses have underlined significant improvement in livelihoods, better yields and higher returns. The country has also benefitted immensely as India has become a net exporter from being a net importer in recent years.”

However, an undated letter that was shared on June 11th 2012 by CSD, with some members of the Coalition, with the text of a letter that was reportedly sent by Dr Haque of CSD to Ajay Jakhar of BKS, contradicts the fairytale being projected and has the CSD’s Dr Haque pointing out to Ajay Jakhar the following:

‘In fact, even before the conference started, we told you clearly that we are not sharing the report with anybody at this stage as some editorial changes have to be made’…..

‘I shall appreciate if you kindly issue a corrigendum to the press, mentioning exactly what I said at the press conference which is as follows: “ The impact of Bt cotton varies from region to region and farmers in some cotton growing areas are in distress because of high cost of production, low yield, low returns, indebtedness etc. This is particularly true about Vidarbha in Maharashtra and some parts of Tamil Nadu and Andhra Pradesh, while there was appreciable yield improvement in Punjab and Gujarat. Even at the national level, there is a decelerating trend in the yield in the past few years for which there is need for further technological improvement”’.

Elsewhere in the letter, Dr Haque says, “Also the pesticide consumption continued to be high because of various sucking pest attacks in several places”.


It is apparent that the study methodology is completely discreditable, from what little information appears in the Farmers’ Forum website and what was presented in a National Conference organized by CSD/CSA/CEE to review a decade of Bt cotton in India. It is based on RECALL METHOD, where farmers were asked to remember their experience in the past and experience now. This is a totally unreliable method and to claim amazing findings based on such a methodology is ludicrous at best.

In fact, the first press invitation to Delhi media on 7th June gives away the total bias of BKS in doing this : It says that the study was conducted by CSD to ascertain and validate their own belief of farmers’ experiences with hybrid Bt cotton seeds across the nine cotton-growing states in India (our emphasis).

The study has amazing findings like 375% increase in net returns per hectare from pre-Bt cotton period to Bt cotton, translating to Rs. 65,307.82/- per hectare. These are average net returns, please note. Let us for a moment assume that Bt cotton farmers do not incur any cost of cultivation (though the increased use of fertilizers, increasing labour costs etc., are mentioned by the CSD-BKS report), and everything that they produce translates into net returns. Let us assume that the average price fetched per quintal was Rs. 4000/- (somewhere above the MSP of Rs. 3300/- and below the CSD’s survey figure of an average of Rs. 4300/quintal) – this means 16.33 quintals yield per hectare.

How come India’s yield figures do not show this reality? How does one explain the dismal one quintal per acre yield in rainfed cotton in Gujarat, even after all these years of Bt cotton, as per official presentations from that state? How does one explain the average 1.4 quintals of yield per acre from Maharashtra (irrigated and unirrigated averaged out) as admitted by Union Minister for Agriculture on the floor of the Parliament if things were this hunky dory with Bt cotton?

How many cotton farmers today can claim to have earned Rs. 26,000/- net income from each acre of Bt cotton they cultivated? If they are doing so, what explains the farm suicides of thousands of Bt cotton farmers?

The study also associates India’s cotton exports and imports with Bt cotton when it is well-known that policy decisions are what drive exports and imports.

The study also irresponsibly comes up with “findings” on socio-economic impacts of Bt cotton and connects it to better quality education, more intake of high value and nutritious food etc. etc. All this is egregious misrepresentation and we are deeply affronted by this as any conscionable civil society group should be. The “study” had not accounted for any confounding variables when coming up with such conclusions and findings and this is simply irresponsible and does not behoove the word socio-economic research.

In the national conference on June 11th and 12th, it was apparent that the study’s findings are highly discountable and contestable and this was a critique raised clearly in front of Ajay Jakhar, including by very senior scientists of this country. However, BKS is choosing to still publicise these findings. If they respect scientific understanding on such issues at all, they would immediately refrain from publicizing these findings any more.

A conference that CSD, CEE & CSA organised recently, came up with other findings (press release and annexure) – most importantly, this Conference brought to the fore a need for a thorough review of Bt cotton in India along with a review of the cotton scenario itself.

Leave a Reply

Your email address will not be published. Required fields are marked *

Back To Top