arm level impact of Bt cotton in South Africa
Yousouf Ismael, Richard Bennett and Stephen Morse
Keywords:  South Africa, Bacilllus thuringiensis (Bt), Plant production, Disease/pest resistance, Cotton, Socioeconomic impacts.
Correct citation: Ismael, Y.; Bennett, R. and Morse, S.. (2001), "Farm level impact of Bt cotton in South Africa." Biotechnology and Development Monitor, No. 48, p. 15-19.

This paper describes the method and findings of the first independent survey of smallholder farmers in the Republic of South Africa designed to explore the economic benefits of their adoption of Bt cotton. The study found that the Bt variety generally resulted in a per hectare increase in yields, value of output and reduction of pesticide costs which outweighed the increase in seed costs to give a substantial increase in gross margins. There are several surveys being carried out at the moment on different aspects of the Makhathini experience. The Monitor will be reporting on their results as these become available.

The development of agricultural biotechnology offers new opportunities to increase agricultural productivity. Some of the potential gains that can be achieved with this technology in the developed countries include higher yields, enhanced product quality and lower pesticide and labour costs. Such gains are particular important in developing countries, especially in the fight to alleviate poverty. So far there has been little evidence as to whether agricultural biotechnology can benefit farmers in developing countries. The aim of the study outlined here was to provide some early evidence on the economic implications of adopting a genetically modified crop, Bt cotton, at farm level in a developing country - the Makhathini Flats region of the Republic of South Africa. The study concentrated on the uptake of the technology by smallholder farmers and on the economic gains, if any, that they derived from adopting Bt cotton.

Bt cotton

The worldwide area under commercially grown genetically modified (GM) crops has been rapidly increasing since they were first introduced in 1996. Insecticide and herbicide resistant traits account for more than 60 per cent of the GM crops grown worldwide. In South Africa, the commercial release of insect resistant Bt cotton and Bt maize was made possible by the Genetic Modified Organism Act (GMO Act 15) of 1997. Bt cotton was the first commercial release of a GM crop variety in Sub-Saharan Africa. In 1999/2000 a total area of 100,000 hectare of Bt cotton was grown in South Africa by 1530 commercial farmers and 3000 small-scale farmers mostly under dryland conditions in the Northern Province, with some in KwaZulu-Natal and the Free State. Since 1998, smallholder farmers in the Makhathini Flats (KwaZulu-Natal province) have been adopting the genetically modified cottonseed variety NuCOTN 37-B with Bollgard™. 60 per cent of these farmers have plots of between 10 and 20 hectare. Seed companies estimate that in the 2000/2001 growing season, around 95 per cent of the 4000 smallholder farmers in the Makhathini region will have adopted this Bt cotton variety.

Agriculture is an important source of income in the Makhathini area, and rural households typically occupy between 1 and 3 hectare of land allocated to them by their tribal chiefs. The major crops in the area are beans, maize and cotton. The latter usually occupies most of the farm and is grown as a commercial crop. Planting takes place from mid-October to mid-December and harvesting from mid-May to mid-June. One of the main reasons for growing cotton is that the crop needs less intensive management than maize or beans and can survive fluctuating weather conditions. The main constraints on market-orientated agriculture in the region are a labour force depleted because of male migration to towns, low capitalization because of a heavy reliance on credit, the climatic constraint of erratic rainfall, and pest attack.

Vunisa Cotton is at the heart of the farming system in Makhathini. Vunisa Cotton is a private, commercial organization supplying seed, fertilizer, pesticide, credit and information to smallholder farmers in the region. It also then buys their cotton after harvest. Seed companies such as Delta Pineland, Clark Cotton and OTK, and agrochemical companies such as VenChem supply their products to Vunisa Cotton, which then retails them to the small farmers. In the case of Bt cotton, Monsanto owns the Bt gene that Delta Pineland has used to develop the NuCOTN 37-B with Bollgard™ variety. The Landbank of South Africa provides the finance (credit) and Vunisa Cotton is responsible for allocating finance to farmers following stringent assessment that takes into account their assets, personal finance, experience, repayment history, and yield assessment over time. Information on cotton is disseminated to the farmers via extension personnel employed by Vunisa Cotton. Each of the extension staff is responsible for a specific area and regularly visits the farmers and holds meetings. Although there is a government extensionist next door to Vunisa, they are not responsible for cotton production. This is left the hands of the Vunisa. There has been no research done so far, as they are private extensionists. All the farmers deliver their cotton to Vunisa Cotton for weighing and grading before payment is made. Vunisa Cotton is the sole supplier of cotton inputs and buyer of cotton output in the region in the northeast of KwaZulu Natal. As such, it plays a pivotal and vital role in the farming system. Vunisa, however, does not provide inputs, credit and marketing for other crops grown in Makhathini.

Survey data

Our study was conducted in November 2000. It is based on a questionnaire survey of 100 smallholder farmers in the Makhathini Flats. A stratified sample was taken of smallholder farmers according to whether they had adopted the Bt cotton variety. The questionnaire was completed during face-to-face interviews in the field and at Vunisa Cotton premises. The questionnaire was designed to obtain information on the physical characteristics of the farm; characteristics that defined the farmer; cropping patterns; input use and costs; cotton output and revenue; and other income sources and assets.

The aim of the study was to examine the factors involved in the adoption of Bt cotton as well as its impact on yields, gross margins and technical efficiency - notably pesticide use. Our analysis compares Bt cotton farmers (adopters) with non-Bt cotton farmers (non-adopters). The survey covered two growing seasons: 1998/1999 (first year) and 1999/2000 (second year). In the 1998/1999 season, only 10 per cent of the 4000 farmers in the region had adopted the new Bt cotton variety, while by 1999/2000 this had risen to 40 per cent.

The average farm size of respondents was 6 ha, although 62% of the farms included in the survey were less than 5 hectare. Approximately 73 per cent of 100 respondents owned livestock such as hens, sheep, goats and 25 per cent had non-farm income sources. The average is 6 hectare for the sample, and most farmers own between 0.5 to 3 hectare. This is very typical for the Makhathini region where land is allocated to farmers by the chief. No one actually owns the land, but the chief has control over it. If a farmers does not used the land allocated to him, someone else can to do so. They farm small areas of cotton because of the lack of labour and credit, and cotton is only a cash crop. Some migrants work in nearby towns and most of them move to Johannesburg or Durban in search of work.

Farmers faced several problems. Amongst the agronomic problems, 57 hectare of farmers consider pests to be their biggest problem and 62 hectare of these categorized the bollworm as the major pest. Too much rain was a more distant second at 24 hectare, followed by weeds, which was ranked highest by only 11%. Among the non-agronomic problems, 82 hectare of farmers cited lack of credit as their major non-agronomic constraint, as compared with 14 hectare who rated land scarcity as their biggest problem and only 4 hectare whose greatest difficulty was lack of labour.

Adoption of Bt cotton

The stratified sample consists of 19 adopters and 81 non-adopters in the first year and 60 adopters and 40 non-adopters in the second year. All 19 farmers who grew Bt cotton in the first year carried on growing it in the following year, suggesting that they were satisfied with the variety. Generally, the adopters of Bt cotton were more experienced and had larger farms. Soil differences are negligible; the whole area has similar soil and used to be a flat riverbed before the construction of the huge Jozini Dam in 1974. None of the 4000 smallholders use irrigation in their farming. They all plough using tractors and carry out planting, weeding and harvesting manually. There is a variation between farmers in the amount of labour and seed used. The larger farmers use more hired labour and are probably more efficient in using resources having spent more years in farming than the non-adopters. A combination of these two characteristics ensured that these farmers were specifically targeted by Vunisa Cotton agents for the promotion of Bt cotton in 1998/99 - the first year of release - and were therefore more likely to be exposed to the technology and receive credit. The key factors affecting early adoption of Bt cotton were thus the availability of credit, or other means of purchasing inputs, such as non-farm income, and pressure from the Vunisa Cotton personnel.

When questioned in 2000, after one season of experience, about why they adopted Bt cotton or might consider adopting it in the future, 44 per cent of respondents directly cited savings on the cost of insecticide as the main reason, while 24 per cent cited expected increases in yield. Approximately 10 per cent said that the expectation of less time spent spraying Bt cotton was critical in their decision to adopt it. Most of the farmers surveyed did not identify any problems with Bt cotton. Until recently they were not given any training on how to grow it. However, now information is being made available, for example, on the subject of refuges. Almost 90 per cent of the non-adopters in 1999/2000 were willing to adopt the technology in the future, but cited cost of the seed as the main reason for not adopting it at the moment. As already noted the vast majority of respondents felt themselves to be constrained by a lack of capital, and Vunisa officials targeted the larger, more creditworthy farmers, hence it is perhaps not surprising that credit has played such a major role in the adoption of Bt cotton in Makhathini. All cotton is for domestic use. South Africa is not self-sufficient in cotton and still relies on imports.

Table 1. Average per hectare cost and returns based on adoption and growing season

1998/1999 season 1999/2000 season
Variable non-adopters adopters difference (%) non-adopters adopters difference (%}
Yield (kg/ha) 434 (49) 511 (82) 18 261 (43) 417 (30) 60
Value of output (Rand/ha) [1] 944 (107) 1,111 (178) 18 568 (94) 907 (65) 60
Seed cost (Rand/ha) 100 (8) 202 (14) 102 91 (17) 197 (12) 116
Pesticide cost (Rand/ha) 112 (8) 98 (14) 13 116 (11) 72 (8) 38
Gross Margin (Rand/ha) 732 (91) 811 (150) 11 361 (66) 638 (45) 77
Numbers of farmers sampled 74 17 32 59

[1] Price per kg is SAR 2.175
( )standard deviation

Economic impact of Bt cotton

Table 1 summarizes the average yield performance and economic impact of Bt cotton. For this analysis some nine farmers in the sample were excluded due to incomplete information. Bt cotton gave higher yields per hectare than other non-Bt varieties. However, there was a marked seasonal effect. In 1998/99 the average yield increase of Bt over non-Bt was nearly 18%, and in the 1999/2000 season this rose to 60 per cent. These increases were despite the fact that the seeding rate of Bt adopters was only 45 per cent of the recommended rate (the equivalent rate for non-adopters was 55 per cent of the recommended rate). Thus, Bt cotton gave even higher yields per kg of seed planted than the non-Bt crop. The shift in yield differential in the two seasons was probably related to rainfall, which can affect the incidence of bollworm or growth of the crop or even the planting time. The 1999/2000 season was a bad one due to unusually heavy rainfall, and average cotton yields fell. However, the Bt adopters suffered a fall in yields between the two seasons (of 18 per cent) much less than those who did not adopt (40 per cent). The rain could have affected the incidence of disease and pest such as bollworm and the length of the growing season. It meant a delay in planting and also affected the germination rate.

The use of Bt cotton increased seed cost per hectare by over 100 per cent in both seasons. The increase in price to Rand 464 per bag includes the technology fee of Rand 240. The price is fixed and does not vary with demand or supply. This high seed cost was only partially offset by the fact that Bt cotton reduced pesticide costs for Bt adopters in both growing seasons. On average, in the first season pesticide costs for Bt adopters were reduced by nearly 13 per cent, while in the second season the reduction was 38 per cent. It could be that in the first season some producers continued to spray either as a risk reduction strategy in case the Bt variety proved not to be as resistant as claimed, or because they received little information on the nature of the Bt variety and its pesticide attributes.

Fifty-four per cent of heads of farm household in this area are male. However, there seems to be little difference in terms of income generated between male and female headed households.

As a result of increases in yield and value of output, and reduced pesticide costs (despite seed cost increases), the average gross margin per hectare (value of output minus the cost of seed and pesticide) for the Bt crop was higher than for the non-Bt. In the first season the average gross margin for adopters was 11 per cent higher than that of non-adopters. In the second season, adopters had an average gross margin 77 per cent higher than that of non-adopters. Input cost differential was the same for both seasons, but the main factor at play in this increase in gross margin differential was the change in yield. The second season saw a much greater yield advantage for Bt adopters relative to non-adopters.

The average figures in Table 1 hide considerable variation and some interesting differences related to farm size. Table 2 shows a breakdown of results according to four categories of cotton area plantings. Not surprisingly, the smaller holdings grew more intensively (with higher seed and pesticide costs per hectare) and achieved higher per hectare yields on average than the larger farms. Indeed, in the first year of adoption some producers with plantings of greater than 5 hectare of Bt cotton suffered yield and gross margin reductions compared to those growing conventional cotton varieties, while the smaller producers growing Bt cotton have substantial increases both in yields and gross margins. By the second year of adoption, all holding sizes appear to benefit from the Bt variety, although the greatest benefit, in terms of gross margin per hectare, still accrues to the smaller producers. Judging by the data we collected there appears to be no widening inequality between Bt and non-Bt farmers. The increase in yield if sustained would not make a difference if the price of cotton were to fall. We have not as yet considered the effect of Bt on other crops. This is an early study; we are now in the process of expanding the work, finance permitting.

Table 2. Average per ha cost and returns based on adoption and growing season for different categories of cotton area planting

1998/1999 season 1999/2000 season
Cotton area <2.5 ha >2.5-5 ha >5-10 ha >10 ha <2.5 ha >2.5-5 ha >5-10 ha >10 ha
Yield kg/ha Adopters 839 598 354 326 565 480 319 273
Non-Adopters 576 447 366 468 330 327 178 228
Value SAR/ha (2.175/kg) Adopters 1825 1301 770 709 1229 1044 694 594
Non-adopters 1253 972 796 1018 718 711 387 496
Seed cost/ha Adopters 232 285 115 152 286 231 139 128
Non-Adopters 149 116 80 73 149 129 58 95
Insecticide cost/ha Adopters 108 140 48 50 131 70 52 31
Non-Adopters 189 114 63 105 156 125 77 192
Gross margin [Output value - (seed cost + insecticide cost)]
Gross margin SAR/ha Adopters 1485 876 607 507 812 743 503 435
Non-Adopters 915 742 653 840 413 457 252 209
Number (N) Adopters 2 6 5 4 17 22 12 8
Non-Adopters 27 29 15 3 11 15 5 1

Discussion and conclusion

Despite the variation between the two seasons, the results of this survey provide considerable cause for cautious optimism regarding the impacts of Bt cotton at farm level. Although it was not so apparent in the first year, by the second year the adopters were clearly gaining economically in terms of increased yield, lower insecticide costs and hence a higher gross margin. This perhaps suggests that the advantages of Bt cotton become particularly apparent in times of environmental stress, since the second year was very wet and such conditions favour the bollworm. For risk-prone, resource-poor farmers this is an important consideration. Moreover, although our analyses are based on small sample sizes, it would appear that the smaller holdings could benefit most from growing the Bt variety. This may be a particularly important element of the technology if the aim is to reduce poverty among small farmers.

There may be additional benefits from Bt cotton that our study has not addressed. First, the analysis of savings in pesticide costs from adopting the Bt variety do not include labour costs or time saved from reduced spraying, which would be an additional benefit unaccounted for in our analysis. Moreover a reduction in the quantity of insecticide applied would be beneficial in terms of the health of farmers and others in the area (there are numerous cases of poisonings due to pesticides) and the environment (pesticides often find their way into water courses).

In a system such as that of cotton production on the Makhathini Flats, with only one supplier of inputs and credit which is at the same time the only market for cotton, the main dynamics at play in adoption of Bt cotton are perhaps not surprising. The tendency was for the older, more experienced farmers and those with larger farms to have higher percentages of adopters especially in the first year of adoption. These farmers were particularly targeted by Vunisa Cotton staff and made aware of the Bt cotton and its presumed advantages over other varieties. They were also more likely to be granted credit or be able to finance the higher seed costs from savings or other income sources. Almost all smallholders in the sample said they would adopt Bt cotton if they had the financial resources to do so and if they had been more aware of its availability and advantages in the first season.

However, given the short time period of our study (two years), no definitive conclusions can be drawn about the adoption dynamics in the region. Some farmers may decide to return to non-Bt varieties in the long run, if the seed suppliers decide to appropriate a greater share of the benefits by raising their seed prices. If this were to happen, it would be unfortunate, as this study shows that there are potential gains at the farm level. Finally, further years of data are required for a larger sample of farmers, together with more detailed data on the labour and other aspects of adoption before final judgement of the benefits of Bt cotton to smallholders is possible. Moreover, it would be interesting to gauge the relative benefits of uptake of the crop for larger, commercial farmers compared to smallholders. This research is currently being undertaken.

Yousouf Ismael, Richard Bennett & Stephen Morse

Department of Agriculture and Food Economics,The University of Reading, Reading, UK.
Phone (+44) 118 9875123; E-mail y.ismael@rdg.ac.uk

Betz, F. S., Hammond, B. G. and Fuchs, R. L. (2000), "Safety and advantages of Bacillus thuringiensis-protected plants to control insect pests." Regulatory Toxicology and Pharmacology , No. 32 (2), p. 156-173.

Fernandez-Cornejo, J. and Klotz-Ingram, C. (1998), Economic, environmental and policy impact of using genetically modified crops for pest management. . Selected Papers presented at the 1998 NEREA Meetings. Ithaca, NY 22-23 June 1998.

Fernandez-Cornejo, J., Klotz-Ingram, C., Jans, S. and McBride, W. (1999), Farm level production effects related to adoption of genetically modified cotton for pest management. Available at http://www.agbioforum.org/vol2no2/klotz.html

Kock, M. (2000), Case study: the introduction of gm cotton in south africa. paper presented at the ICGEB Workshop: Biosafety 2: Advanced research and procedures: case studies for designated experts..

Pray, C.E., Ma, D., Huang, J. and Qiao, F. (2000), Impact of Bt cotton in China. Paper presented at the Agricultural Economics Society annual conference, Manchester, April 2000; forthcoming in World Development.

Rother, H. A. (2000), "Influences of pesticide risk perception on the health of rural South African women and children" African Newsletter , 2/2000. Available at http://www.occuphealth.fi/e/info/anl/200/pesticide05.htm

Contributions to the Biotechnology and Development Monitor are not covered by any copyright. Exerpts may be translated or reproduced without prior permission (with exception of parts reproduced from third sources), with  acknowledgement of source.


back to top
monitor homepage
index of this issue
contact us