Untitled Document

The 2005 Bt Cotton Craze Presentation version

Shortened paper available; explanation in ppt

<india>

In the global struggles over GM crops, there are few places where the stakes are higher than India -- which released its first GM crop in 2002

It was cotton, genetically modified with an insecticidal gene from the bacterium Bt.

By last year, Bt cotton sales climbed to 1.3 million packs, but by this year's crop there were 20 Bt seeds for sale, and Monsanto (which provided the gene construct to all) was pleased to announce that that sales jumped to over 3 million packs.

Actually in some areas Bt sales climbed more sharply than that. In Warangal District of Andhra Pradesh, whjere I have been working since 2000, sales went through the roof.

This is a key cotton area that became notorious in 1998 due to hundreds of suicides linked to cotton farming -- a tragedy used in the rhetoric of both Monsanto and its chief antagonist, Vandana Shiva. So warangal has been a much-watcdhed area.

My survey of Warangal seed vendors shows that in the last 3 yrs, the market share held by Bt hybrids climbed from 1% to 20% to 62%. In Gudeppad village, where they're used to me asking about cotton, one farmer said with a laugh that my work would be easy this summer: the village was "mutham Bt!" (all Bt). It was true: 90% of Gudeppad's cotton choices this year were Bt seeds. This wasn't just a Tipping Point; it was a stampede.

What are we to make of such a phenomenal stampede to the first GM crop in this cruciual area of this crucial country? Monsanto predictably sees it as a vote of confidence by the farmers, but I want to look at the larger cultural context. The first results of my recent study show that this cotton craze actually fits into a disquieting history of local cotton crazes, which reflect agricultural deskilling.

Let us first note that Bt cotton should fit classic innovation- diffusion theory, which stresses smallplot experiments and information spread through social networks.

Cultural evolutionary theorists have stressed a firmer distinction between crop evaluation (or environmental learning) and cultural transmission (or social learning). It has been suggested that when the accuracy of ewnvironmental learning drops, farmers rely more on social learning. This is a point worth coming back to.

I have been less interested in the adoption of discrete traits than on the process of integrating technologies and practices into agricultural performance, a process known as skilling. I have described specific impediments to that lead to "agricultural deskilling."

Biotech seed firms attribute the spread of Bt cotton to environmental learning. Monsanto's cites smalplot experimentation, and recognition of consistent results. And other commentators salute the integrity of farmers' evaluations.

That social dimension to adoption is claimed to be a problem....

Although seed companies actually use it in marketing -- for instance, setting up demonstration plots

and giving free seeds to "pedda ryutulu" -- "big farmers" -- like this one.

Farmers turn out to be remarkably quick to plant whatever seed their neighbor is planting, even if they know little about the seed's traits. A major reason these farmers have become generally non-analaytic is the market dominance by hybrids.

Southern India is unusual in that cotton farming is based on hybrids. Hybrid maize had been blamed for deskilling of US farmers, although I think this is only one component of a larger problem.

In Warangal, it's not just hybrids, but a remarkably unstable market for them -- 125 different seeds from 61 companies in the last 3 yrs. And of the 78 seeds on sale this year, only 24 had been around since 2003.

Seeds often vary from pack to pack, and there are often flawed seeds on the market -- known as "spurious seed," these are a bane for farmers AND vendors, who have been closed down for selling a seed that turned out to be spurious. And some different seed brands are in fact identical.

All this militates against skilling, as do several unpredictable external factors like insect outbreaks and drought. So there is remarkably little empirical evaluation of seeds.

So what patterns in cotton seed choices emerge in this context?

Cotton Crazes

I have just finished a census of 6 villages in 2004, and 9 villages in 2005. Both surveys samples were randomized to avoid clustering. Let's look at cotton choices -- which of the ,many seeds farmers bought.,

Taken collectively, there is nothing unusually volatile in the trends except for the precipitous rise of one seed: RCH2-Bt. But when we look at individual villages we see a volatile pattern of crazes.

<Gudeppad-graph>

To show this I have ranked cotton choices for each year and each village. (don't have to know cotton brands for test) But nore that in Gudeppad, Brahma and Ganesh were local favorites in 2003 but had almost disappeared by 2005. Chitra (purple line) went from negligible, to town favorite, before also disappearing in 2005, when while RCH2-Bt purchases have literally gone off the chart.

In Kalleda, Brahma was a runaway favorite in 2003 before dropping sharply; here too there was a strong 1-year favorite but it was Gemini. Here too RCH2-Bt takes off dramatically.

In Tekumatla, there was an even more pronounced craze in 2004, for Durga, 40% of the town's cotton choices before being abandoned the next year, while RCH2-Bt took off.

In Pathipally, there is more brand stability, but also a craze -- Dyna, the local favorite in 2004 which then dropping into obscurity.

Neither I, nor any of the vendors or farmers I've interviewed, offer any agro-ecological rationale for any of these short-term, localized cotton seed crazes.

Naive and Experimental Planting

Since these crazes are hard to reconcile with the idea of experimentation and evaluation, let us look at the extent of "naive planting" -- defined as the planting of a type of seed for the first time.

Of all cotton choices for 2005, 79% were naive plantings.

But maybe farmers are testing new seeds on small plots?

Small plantings -- under an acre -- comprised just 4% of the cotton plantings in 2003-04. The median household cotton acreage here is just 2 acres, so this is clearly not experimental small- planting.

<boll>

The absence of seed evaluation is shown by this analysis of farmer knowledge of the one trait farmers claim to value most highly -- boll size -- in the seed that is the runaway 2005 favorite -- RCH2-Bt.

Note that the Rasi Co., which produces the seed, says RCH2-Bt has a medium-sized boll.

Planters were asked if they knew the boll size, and this chart shows only those 280 cases where the farmer planted RCH2-Bt AND SAID HE DID KNOW ITS BOLL SIZE.

But there was no agreement. Bhandarupally and Pathipally are neighboring towns but in one village most feel the boll is small while in the other no one thinks so. Neighboring Oorugunda and Gudeppad are in general agreement on medium, while Kalleda and Ravuru are completely split.

So the market offers a fickle roster of seeds; many of the key determinants of a good crop are unpredictable; villages show ephemeral fads; most cotton plantings are na‹ve and non- experimental; and farmers do almost no environmental learning, even for the one trait they value the most. So what is driving cotton seed choices then?

One answer is corporate promotion.

Cotton advertising is everywhere in Warangal -- altho rice seed, which is selected more on environmental learning, is rarely advertised.

The other obvious driver of cotton selection is recommendations and emulation. I interviewed partcipants in those crazes for Gemini in Kalleda and Chitra in Gudeppad.

Gemini was a new hybrid in 2003

Many Kalleda farmers buy their seeds from 1 Warangal dealer who happened to be the sole distributor for Gemini, and he recommended it.

Gemini had a marketing campaign in Kalleda; Farmers who made advance purchases got scratch cards for prizes.

An owner of Gemini Seeds is a relative of the mandal president, and he too recommended it

One of the 2003 adopters happened to get a good crop (although other early adopters didn't).

Many farmers adopted Gemini simply because "other farmers around here were planting it."

The Chitra craze was also driven by marketing and by emulation of a single local farmer. It was also a new seed, and a company marketeer knew Jaipal Reddy to be a pedda rytu, and gave him a free box in 2003 -- then Enough farmers emulated Jaipal to make Chitra sales soar before crashing in 2005.

For a more theoretical understanding of these crazes, let's return to the idea that farmer reliance on social learning is a function of the cost and accuracy of individual/env learning.

This makes sense, but it also assumes that social learning is adaptive because it "exploits already paid costs of individual learning needed to generate novel adaptations." But what if payoffs are highly unpredictable, and so the accuracy of individual learning is extremely low?

The skilling process is so impeded that there is little individual learning for social learning to exploit. In this situation, pure social learning generates crazes with little environmental basis.

Michael Dove suggests that our rural development concepts follow a life cycle, in which they are eventually appropriated by the interests they were developed to oppose. This is what has happened with farmer seed wisdom in Warangal cotton.

So there was an important and surprising cultural context to the big year for GM cotton in this corner of India. On the surface, the news seems decidedly good for Monsanto. But whether this stampede signals a sustained change in agricultural practice is much less certain, as it fits into a prior pattern of crazes without agroecological bases. And if my analysis is correct, the news is decidedly bad for the cotton farmers who will surely be appearing in industry media thius coming year.