Bt Cotton In India

  • Uploaded by: Akshay
  • 0
  • 0
  • May 2020
  • PDF

This document was uploaded by user and they confirmed that they have the permission to share it. If you are author or own the copyright of this book, please report to us by using this DMCA report form. Report DMCA


Overview

Download & View Bt Cotton In India as PDF for free.

More details

  • Words: 4,123
  • Pages: 15
Bt cotton: the crop that kills?  Examining the relationship between farmer  suicides and genetically modified cotton    Akshay Birla  6/10/2009     

There is some controversy as to whether the introduction of genetically modified cotton has led to an increase in farmer suicides because it increases the amount of debt that famers carry to buy Bt cotton seeds. I examine the science behind genetically modified crops and test this accusation. I find that at a 90% confidence interval, controlling for external technological and weather shocks by using cereal productivity, the introduction of Bt cotton has actually led to a decrease in farmer suicides. I also find that increase in crop yield in general leads to a decrease in farmer suicides.  

Introduction

A gene sequence is a segment of deoxyribonucleic acid (DNA) molecules that contains the necessary instructions to make protein. DNA is composed of nucleotides—four in kind—which are “read” three nucleotides at a time by the enzymes that make protein. It is the permutation and combination of these nucleotides that is at the core of the existence and diversity of life. In the last century it has became clear that there lies a great potential in manipulating genes to suit a certain set of goals in living organisms. One of the more practical applications of such manipulation has been the development of genetically modified (GM) crops with traits that are desirable to humans that cultivate them. Most often, genetically modified crops are created by inserting a gene segment into an existing gene sequence. To do this, scientists must first locate and isolate a gene of interest from another organism and create a new gene out of it. This gene may be what can give tobacco a glowing color, or what gives rice extra vitamin A, or what allows the starch in corn to break down into ethanol more easily. It must contain a marker which allows scientists to test whether or not the gene is effective. The marker may be an unusual color, or a trait such as resistance to some drug which would allow easy identification of the drug. For example, if a plant is being tested for tolerance of a chemical and the new plants are sprayed with that chemical, all unsuccessful mutations will die leaving only successful mutations as choice options. Then the gene is inserted into a bacteria cell where it can be replicated easily. The bacteria then transfers the DNA into the chromosomes of the cells, which is grown as callus in a solution called Agrobacterium. The solution in which they are grown can be manipulated so that an entire plant can grow from the mutated cells. Another method to insert genes, instead of using the plasmid in bacteria, is to coat minute metal bullets with pieces of DNA and fire them into the cell that the scientists are trying

to mutate. This long and complicated process is what makes GM crops, GM crops. Some firms that sell GM food argue that methods of selection and cross breeding that have been used in agriculture for centuries are also means of genetic modification and so the new techniques are nothing new, per se. However, there is a fundamental difference—without the new technology it would be impossible to insert the gene from an organism belonging to one species or kingdom into an organism from another. These processes simply do not happen naturally. Traditionally, genes were manipulated primarily to produce breeds that had characteristics of the two parent plants. While these sorts of manipulations did make new and better plants, the scope of manipulation was limited. Using GM technology scientists can today make plants that have different mechanisms of increasing quality, consistency, and yield. Plants are made resistant to insects by inserting a gene called Bacillus thuringiensis (Bt) into them. Bt contains toxins that are specific to insects—it contains cryprotein, given the name because it has spores that are enclosed in crystals. These crystals can be broken down by enzymes that are specific to the insect gut, releasing toxins that kill the insect. Other plants are genetically modified to give them genes which make them herbicide tolerant (HT), so when fields are sprayed with herbicides, only the weeds that are soaking up nutrition from the soil and producing no yield, die, leaving the crops intact. These genes in these plants need to be activated by soaking them in certain solutions, and these solutions need to be bought each season. HT technology reduces the need for tilling, a process that consumes time and energy, and leads to losses of nutrients from the soil. The commercial introduction of genetically modified crops—where it has been permitted—has been controversial because of its alleged adverse effects on health, the environments, and on rural economies. Some social activists, most notably Vandana Shiva her fellow anti-globalization

activists claim that genetically modified crops are evil. In particular, they claim that the introduction of Bt cotton, a type of genetically modified cotton, has lead to an increase in farmer suicides. They reason that multinational corporations like Monsanto have coerced farmers to buying seeds of GM cotton. These purchases have put households in immense debt. Farmers are unable to pay these debts because GM cotton gives low yields and so they commit suicide. I attempt to examine this claim rigorously. I begin by explaining the science behind genetically modified crops, give a brief overview of the policy issues surrounding Bt cotton, and then use two-stage least squares regression to determine causation. Bt cotton was approved for commercial use in six states in the year 2002 1 . However, there was F

F

already a thriving market for illegal GM cotton seeds in the state of Gujarat by this time. Reports indicate that farmers from other states in India were coming to Gujarat to buy a popular brand of illegal GM seeds, Navbharat 151, by 2002 2 , and there were test-trails by Monsanto in the midF

F

90s. This contradicts Shiva’s claim that farmers were coerced into buying GM seeds—they were actually incurring significant costs to procure GM seeds. On the other hand, a study often cited to claim the success of GM cotton, by Richard Bennett et al the University of Reading, is deeply flawed. Bennett et al compare farms that adopt GM cotton with those that do not, and randomly select farmers to interview, but this isn’t actually a randomized experiment because of the selection bias: farmers that are willing to test new seed technology are likely to be the farmers that use new technology for other aspects of their farms already—tractors instead of bullocks, for example—and so it would hardly be surprising to find them to have better yield.

                                                             1

“ Background Information on Bt Cotton” http://www.envfor.nic.in/divisions/csurv/geac/bgnote.htm 2 “Bt cotton through the back door” http://www.grain.org/seedling/?id=151 

I try to answer whether the introduction of Bt cotton actually increases suicides through debt. Here I ignore the farmers-being-coerced story because it seems implausible and I don’t have the data to test it anyway. The effect of Bt cotton on suicide, if any, has major implications for agriculture policy. I seek to understand whether the use of genetically modified cotton increases rates of suicide in farmers because of its effect on farmer debt. It seems plausible that farmers would need more credit if they were to buy Bt cotton seeds instead of regular seeds, since the investment required to plant Bt cotton seeds is higher—even after accounting for the lower costs for pesticides, the cost of growing Bt cotton is higher than that of growing regular cotton. But farmer’s willingness to pay for Bt cotton indicates that it must have a higher, not lower, yield. I was unable to find, however, reports comparing the yields of the two after the 2002-2003 periods, which leaves open the question as to whether these costs change in a non-linear function. If it turns out to be the case that costs increased unexpectedly for every following time period, then it may require farmers to take on more credit than they anticipated. At the same time, if costs were increasing every year, cost increases would no longer be unexpected.

Cost (Rupees/acre) Cotton seed Pest sprays Bollworm pest sprays Total

Season 2002 Non-Bt Bt 460 1527 634 568 984 280 2078 2375 Table 1

Season 2003 Non-Bt Bt 471 1491 520 529 1166 2157

195 2215

Model I use instrumental variable (IV) regression to examine the claim that the introduction of GM seeds has increased the number of farmer suicides. I use indebtedness as an instrument, thereby making two assumptions—that debt and suicide are correlated, and that the use of Bt cotton does not affect suicides except through indebtedness. That indebtedness and suicides are correlated is fairly intuitive: given the inability to pay back debts, farmers chose to kill themselves and their families rather than incurring the social and financial costs of the debt 3 . The second assumption, F

F

that using Bt cotton does not affect suicides except through debt is also reasonable—the use of technology changes only the initial amount of credit required to buy Bt cotton seeds. If farming Bt-cotton is found to be hazardous in that produces chemicals that induce suicidality, this argument will fall apart. A key thing to note is that it is very difficult to determine a time at which Bt cotton was introduced. However, it can be ascertained for sure that the proportion of the cotton crop that is Bt cotton has increased each year since 1998. Official data for this, however, is not very clear. Some states where illegal Bt cotton was being planted stopped publishing estimates of the quantities and yields of their cotton crop in their quarterly economic bulletins. So what I measure by using yield is the amount of crop that is genetically modified. This in and of itself would be absurd, since yield is a reflection of the weather, agricultural policies, and general agricultural technology. But these can be controlled for if I use the yield of cereals or pulses as a control. That is, any effect of weather or policy or technology that is non-cotton specific would be captured in the coefficients for my controls. To make sure that I am controlling correctly, I use                                                              3

Why is it that some farmers live in debt while others kill themselves? A future theoretical study can examine the implications of the existence of something that is the equivalent of a default line at which farmers “default” by killing themselves.

three different sets of controls—yield of cereal-crops, yield of pulses, and yield of sugarcane. I choose these because there was no exogenous shock to the production function in either cereal or pulses—no genetically modified cereal crops have been introduced. Golden rice, which contains additional Vitamin-A had been proposed but worldwide protests ensured that it was a non-starter. Sugarcane is not such a good instrument because domestic politic ensures constant intervention in the market, but I use it anyway, since it might make for an interesting comparison. I measure suicides of farmers using a proxy, the number of deaths from pesticide consumption. This seems to be the primary method of committing suicide in rural India—it is simpler to execute than hanging or shooting oneself in the head, for example, and since insecticides are easily available at a low price. The results using this data would be distorted if the means of committing suicide has changed over the time period I examine, but there is no reason to believe that suicide technology has changed. I can also cross-check these results using another proxy that measures suicides in agricultural families that own the land they work, but these are not reported in the this paper. Finally, I use savings in rural cooperatives as a proxy for household debt. If more households fall in debt, there should be a reduction in the rupee amount of savings that rural cooperatives hold. This should also be consistent over time because interest rates in rural cooperatives are often higher than those offered by private banks, so it is unlikely that money would leak from the cooperatives to other sources. If, however, it turns out to be the case that there is a general trend that saving patterns are changing—if, for example, farmers are saving in bullion instead of cash, this argument will fall apart. I will not investigate this particular question in this study, however.

The instrumental variable regression model is:

However, applying the log transformation provides a more intuitive interpretation—the percentage change in suicide as a result of the percentage change in percentage of the total cotton crop that is genetically modified. This also makes sense because the percentage change in debt has a more meaningful interpretation than debt in and of itself. These equations are 

Here β2 measures the effect of cotton yield on suicide after controlling for weather and general technological advancement. If the claim of Shiva et al is correct, this should be positive. That is, as the proportion of the cotton crop that is genetically modified increases, the number of suicides increases. If the researchers at Reading University, who are incidentally on the payroll of Monsanto, are correct, this should be zero—they don’t claim anything about suicides, only about material well being. But presumably increased material well being leads to a decrease in suicides so I’ll give them credit even if my results turn out to be negative and statistically significant.

Data Collecting data to test the hypothesis is challenging. While sufficient data was collected to be able to test a hypothesis, much of this data was for proxies for what the model really was testing, which makes the validity of my results is dubious. The time-period for which data for suicides was available was extremely limited—the newest data is that for 2003 4 . This limited my dataset F

F

severely, and to compensate for this, I used data from different states—Andhra Pradesh, Maharashtra, Gujarat, Punjab, and Tamil Nadu. All the data used is aggregated at the state-level. The “Accidental Deaths and Suicide” reports from the Ministry of Home Affairs in India provided the number of suicides from the consumption of pesticides and for suicides in landowning farmer households. To find data of the areas under cultivation, the production, and the yield of each of the crops that I use, I used economic bulletins from the states. These have been listed in the Bibliography. In some instances data on pulses was not available. Here, data for grams (a kind of pulse) was used. Where an aggregate data for cereal was not available, data for the predominant cereal crop was used—in the case of Punjab this is wheat, while in Gujarat it is rice. However, since yield is of importance as an indicator of technology, the difference due to area under cultivation and of one particular crop should not be a problem. The same set of economic reports was used to estimate money deposited in rural banks. These are aggregates of various mechanisms for farmers to save money—fixed deposits (CDs) that pay a higher than market interest rate, savings accounts at the post office, rural credit cooperatives and other farmer-specific financial institutions. These serve as a proxy for debt, and I use them because the only alternatives would have been to use some measure of consumption which                                                              4

The data for 2003-2007 has been requested from the Home Ministry using the Right to Information Act of 2005, but has yet to deliver results.

would have tremendous difficulty with reverse causality. The change in savings rate can provide a reasonable estimate in the movements in indebtedness. If famers are in general more indebted the net deposits to these financial institutions should decrease, and if farmers are less indebted, the net deposits should increase, assuming that there isn’t a big shift in the ways in which people save to bullion or something not captured in my data. Results The model is tested at the 90% confidence interval. At the 90% confidence interval, when controlling for general technological advancement and weather using the yield of pulses, there is a statistically significant result that can be interpreted to mean that an increase in the proportion of the crop that is genetically modified leads to a decrease in the percentage change in farmer suicides. That is, the rate of farmer suicides decreases and more Bt cotton is planted. Using the yield of cereals and sugarcane, I fail to find anything conclusive.

Source | SS df MS Number of obs = 28 ------------+-----------------------------F( 2, 25) = 8.04 Model | 12.2801736 2 6.14008682 Prob > F = 0.0020 Residual | 25.2113443 25 1.00845377 R-squared = 0.3275 ------------+-----------------------------Adj R-squared = 0.2737 Total | 37.4915179 27 1.38857474 Root MSE = 1.0042 ----------------------------------------------------------------------------lnp | Coef. Std. Err. t P>|t| [90% Conf. Interval] ------------+---------------------------------------------------------------lnsav | -1.183763 .6553851 -1.81 0.083 -2.303253 -.0642733 lncereal | -2.465354 .6594769 -3.74 0.001 -3.591833 -1.338874 _cons | 18.68259 5.698948 3.28 0.003 8.947982 28.41719 ----------------------------------------------------------------------------Instrumented: lnsav Instruments: lncereal lncotton -----------------------------------------------------------------------------

Source | SS df MS Number of obs = 28 -------------+-----------------------------F( 2, 25) = 5.12 Model | 8.28315075 2 4.14157537 Prob > F = 0.0136 Residual | 29.2083672 25 1.16833469 R-squared = 0.2209 -------------+-----------------------------Adj R-squared = 0.1586 Total | 37.4915179 27 1.38857474 Root MSE = 1.0809 ----------------------------------------------------------------------------lnp | Coef. Std. Err. t P>|t| [90% Conf.Interval] -------------+--------------------------------------------------------------lnsav | -.3610605 .5492963 -0.66 0.517 -1.299336 .5772149 lnpulses | -2.379418 .778267 -3.06 0.005 -3.708807 -1.050028 _cons | 8.669034 4.541638 1.91 0.068 .911277 16.42679 ----------------------------------------------------------------------Instrumented: lnsav Instruments: lnpulses lncotton -----------------------------------------------------------------------------

Source | SS df MS Number of obs = 28 ------------+-----------------------------F( 2, 25) = 0.16 Model | -2.24126243 2 -1.12063121 Prob > F = 0.8557 Residual | 39.7327803 25 1.58931121 R-squared = . ------------+-----------------------------Adj R-squared = . Total | 37.4915179 27 1.38857474 Root MSE = 1.2607 ----------------------------------------------------------------------------lnp | Coef. Std. Err. t P>|t| [90% Conf. Interval] -------------+--------------------------------------------------------------lnsav | -.328963 .6041388 -0.54 0.591 -1.360917 .7029911 lnsugar | -.0985501 .4218661 -0.23 0.817 -.8191568 .6220566 _cons | 10.33873 6.180565 1.67 0.107 -.2185493 20.896 ----------------------------------------------------------------------------Instrumented: lnsav Instruments: lnsugar lncotton -----------------------------------------------------------------------------

Source | SS df MS Number of obs = 28 ------------+-----------------------------F( 1, 26) = 0.29 Model | -2.52738381 1 -2.52738381 Prob > F = 0.5921 Residual | 40.0189017 26 1.53918853 R-squared = . ------------+-----------------------------Adj R-squared = . Total | 37.4915179 27 1.38857474 Root MSE = 1.2406 ----------------------------------------------------------------------------lnp | Coef. Std. Err. t P>|t| [90% Conf. Interval] ------------+---------------------------------------------------------------lnsav | -.3424379 .6311734 -0.54 0.592 -1.418979 .7341028 _cons | 10.03047 5.134315 1.95 0.062 1.273286 18.78765 ----------------------------------------------------------------------------Instrumented: lnsav Instruments: lncotton

----------------------------------------------------------------------------

Source | SS df MS Number of obs = 28 ------------+-----------------------------F( 2, 25) = 5.54 Model | 7377532.62 2 3688766.31 Prob > F = 0.0102 Residual | 49151911.1 25 1966076.44 R-squared = 0.1305 ------------+-----------------------------Adj R-squared = 0.0609 Total | 56529443.7 27 2093683.1 Root MSE = 1402.2 ----------------------------------------------------------------------------p | Coef. Std. Err. t P>|t| [90% Conf. Interval] ------------+---------------------------------------------------------------lnsav | -1833.809 915.1004 -2.00 0.056 -3396.929 -270.6883 lncereal | -3009.929 920.8137 -3.27 0.003 -4582.808 -1437.049 _cons | 19220.84 7957.322 2.42 0.023 5628.616 32813.07 ----------------------------------------------------------------------------Instrumented: lnsav Instruments: lncereal lncotton -----------------------------------------------------------------------------

The final regression table suggests that a 1% increase in the amount of crop that is genetically modified results in a 1800 less deaths. This is statistically significant at the 90% confidence interval. Even though the coefficients for savings do not exclude zero from the 90% confidence interval, each of the coefficients is negative. Furthermore, in each of the controls, while controlling for cotton yield, an increase in yield decreases farmer suicide. These are statistically significant for every single test at the 90% confidence level. This suggests that even if my rationale behind using the yield of one crop to control for all technological progress in the field of agriculture turns out to be deeply flawed—and that is an empirical question that would require measuring the yield of different crops and knowing the percentage of the crop that is genetically modified—there does indeed exist a relationship between yield and suicide that acts through debt. Several measures can be taken to improve this study so that the estimates received show more clearly whether a relationship exists. In particular, if data with greater granularity—that is data at

the district level instead of the state level—can be found and compared against data describing which districts Bt cotton was planted in, this would allow for an extremely clear answer for the picture. Since much of the Bt cotton-planting was illegal, it is unlikely that such a study can be conducted, we can design a panel study to collect data going forward. This could also contain other measures that would be useful for our study. The Townsend Thai Survey provides a good example of the kind of survey that can be designed. A randomized study might also be extremely useful, but it would be difficult to design if the connection to suicide is to be ascertained. A study limited to examining the effect of Bt cotton on yield at a more granular level than has currently been published would be very welcome. Since the mechanism through which Bt cotton is alleged to kill farmers is the credit market (defaulting by pesticide consumption), randomized studies allowing farmers access to different levels of credit would be very useful 5 . F

F

There are important policy recommendations that follow from the results of this paper. How seriously they are to be taken depends on how seriously the reader views the results from this study. This study shows that yield has a tremendous impact on farmer suicides. This operates through the savings/debt mechanism—that is, an increase in yield leads to an increase in savings and consequently to a decrease in suicides. Furthermore, this study shows that as the use of Bt technology increase, after controlling for the increases in general agricultural productivity, farmer suicides actually drop, not increase as some environmentalists would have us believe. This calls for a reevaluation of the ban on genetically modified crops and the arguments that are presented in favor of the ban, and demonstrates that policy makers and voters desperately need to be more educated about genetically modified food. Finally, this suggests that the data on which                                                              5

 If we find at the end of this recession that enough economists are unemployed, we could

probably go over to India and do some planting and screening ourselves. Our desire for truth over our preconceived notions of Bt cotton will make us unbiased farmers; our suicides will be fodder for future economists.  

Dr Shiva and other environmentalists have based their accusations need to be examined. It would help if these researchers released public versions of their data to begin with (repeated inquiries got me no data from Dr Shiva). This is not to carry out an inquisition but because these researchers have played a significant role in shaping agricultural policy as it exists today. This also suggests that a way to raise the standard of living for rural households would be to provide farmers with greater access to credit to buy more productive seeds and other materials for farming. Tentatively, I can answer the question I started with: genetically modified crops are not killers; those that prevent them from entering the market, however…

Working Bibliography Andhra Pradesh (India). Directorate of Economics and Statistics, Government of Andhra Pradesh. Andhra Pradesh Economics and Statistical Bulletin. Vol. 36-43, 49-51. Hyderabad, 1992-2004. Andhra Pradesh (India). Directorate of Economics and Statistics, Government of Andhra Pradesh. Handbook of Statistics: Andhra Pradesh 1998-1999. Hyderabad, 1999. Andhra Pradesh (India). Directorate of Economics and Statistics, Government of Andhra Pradesh. Statistical Abstract. Hyderabad, 1993-2001. "Background Note on Bt Cultivation in India." Ministry of Environment and Forests. 9 May 2009 . "Bt Cotton through the back door." Seedling 4 Dec. 2001. Grain. 4 Dec. 2001. 9 May 2009 . Fedoroff, Nina, and Nancy Marie Brown. Mendel in the Kitchen A Scientist's View of Genetically Modified Food. New York: Joseph Henry P, 2004. Grada, Cormac O. Famine: a short history. Princeton, N.J: Princeton UP, 2009. "Guj worst-hit by illegal Bt cotton production." Business Standard. 22 Apr. 2008. 9 May 2009 . "Gujarat Asked to Control Spurious BT Cotton Seeds." www.business-standard.com 22 Apr. 2008. Business Standard. 9 May 2009. . Gujarat (India). Directorate of Economics and Statistics, Government of Gujarat. Statistical Abstract of Gujarat State. Gandhinagar, 1998-2004. India. Ministry of Home Affairs. National Crime Records Bureau. Accidental Deaths & Suicides in India. New Delhi, 1996-2003. Maharashtra (India). Directorate of Economics and Statistics, Government of Maharashtra. Quarterly Bulletin of Economics and Statistics. Vol. 35,36, 41-45. Mumbai, 1995-2005. Punjab (India). Economic and Statistical Organization, Government of Punjab. Chandigarh: Economic Adviser to Government, 2003. Tamil Nadu (India). Department of Economics & Statistics, Government of Tamil Nadu. Annual Statistical Abstract of Tamil Nadu: 2003-2004. Chennai, 2004. Tamil Nadu (India). Department of Evaluation & Applied Research, Government of Tamil Nadu. Tamil Nadu: An Economic Appraisal. Chennai, 1995-2002.    

Related Documents

Bt Cotton In India
May 2020 15
Cotton
June 2020 12
Cotton
June 2020 8
Cotton
June 2020 8
Cotton
November 2019 16

More Documents from ""