Agriculture and Human Values

, Volume 29, Issue 1, pp 3–15

Genetically-engineered crops and their effects on varietal diversity: a case of Bt eggplant in India


    • Research Collaborator, International Food Policy Research Institute and Visiting FellowCornell University
  • William Lesser
    • Dyson School of Applied Economics and ManagementCornell University

DOI: 10.1007/s10460-011-9320-3

Cite this article as:
Kolady, D.E. & Lesser, W. Agric Hum Values (2012) 29: 3. doi:10.1007/s10460-011-9320-3


Building on the evidence from the impact of hybrid technology on varietal diversity loss, this paper explores ex ante the possible effects of introduction of Bt eggplant on on-farm varietal diversity of eggplant. The public–private partnership involved in the development and introduction of Bt eggplant provides a great opportunity to develop locally-adapted Bt open-pollinated varieties (OPVs) instead of having a limited number of generic hybrid varieties. The study shows that introduction of multiple Bt OPVs by public institutions will reduce the rate of replacement of OPVs by hybrids and thus help in conserving varietal diversity. However, the cost of developing multiple Bt OPVs is high; hence policy makers need to look at alternative measures to maintain the varietal diversity of crops such as eggplant in its centers of diversity.


Bt transgenic cropsGenetically engineered cropsPublic–private partnershipVarietal diversityIndia



Bacillus thuringiensis


Contingent valuation


Eggplant shoot and fruit borer


Genetically engineered


Genetic use restriction technology


Multinomial logit


Open-pollinated variety


Research and development


United States Agency for International Development


Willingness to pay


Technological advances in agriculture, especially the introduction of modern varieties, have contributed significantly to production and productivity. However, there have also been real and perceived costs associated with these gains. One of those costs is the replacement of traditional varieties by “modern” or “improved” types, which some contend has led to a narrower genetic base for major crops; hence to greater vulnerability of crops to biotic and abiotic stressors and large-scale disasters (Cooper et al. 1992). This vulnerability is especially serious if the loss of crop varieties from centers of diversity leads to an overall loss of genetic resources that be needed in the future by breeding programs and farmers worldwide. While maintaining diversity by ex situ conservation, where genetic resources are sustained in gene banks, botanical gardens, and agricultural research stations, may help to conserve the genetic resources themselves, managing the diversity on farms is economically important because it provides small-scale, resource-poor farmers with an important mechanism for risk management. Such farmers value diversity within and between their crops because of different production conditions, risk factors, and market demand. Moreover, resource-poor farmers often rely on diversity on other nearby farms and communities to provide new seeds when a crop fails (Louette et al. 1997).

The oftentimes repeated example of losses caused by resistance-based crop seed stocks is that of the 1970 southern corn leaf blight in the US. At that time, southern corn leaf blight—began in Florida and swept through the US Corn Belt—eventually reaching into Canada. Losses were estimated at 15% of the US crop, or somewhere between $500 million and $1 billion (Walsh 1981). The causal organism was subsequently identified as a new fungus which attacked the cytoplasm; and 80% of the US corn crop at the time contained T-cytoplasm, which was vulnerable to the fungus. T-cytoplasm was intentionally bred into the line, and caused male sterility, which explains why it was used so extensively across varieties. This example highlights the risk of depending on a limited number of closely-related modern varieties for crop production and the importance of in situ genetic diversity as a risk management tool for farmers.

The commercialization of genetically engineered (GE) crops has further raised the concern that a limited number of GE varieties will supplant numerous local varieties leading to further diversity loss (Steinbrecher 1996; Wolfenbarger and Phifer 2000). However, some evidence suggests that the availability of a greater number of GE varieties could avert the issue of increased genetic uniformity and genetic vulnerability (Bowman et al. 2003; Sneller 2003; Qaim et al. 2005; Raney and Pingali 2005). Further, proponents of the technology claim that the adoption of GE crops help to achieve nutritional and food security especially in developing countries (Reiss and Straughan 2001; Qaim et al. 2005). Indirectly, more productive GE crops reduce pressure on scarce land resources, an environmental benefit. There is an abundant literature, both ex post and ex ante, examining the economic and welfare effects of GE crops in developed and developing countries (Falck-Zepeda et al. 2000; Huang et al. 2002; Qaim and Zilberman 2003; Shankar and Thirtle 2005; Kolady and Lesser 2008a; Krishna and Qaim 2008; Subramanian and Qaim 2009).

Some of those studies have examined the factors affecting farmers’ varietal choices and farmers’ perceptions of potential environmental impacts in the context of GE crops in developing countries (Soleri et al. 2005; Briol et al. 2007). However, those studies generally focused on field crops, suggesting that empirical studies which analyzed the potential impacts of the introduction of GE technology on on-farm varietal diversity of non-field crops were limited. In this study, we begin to fill that gap by analyzing the possible effects of the introduction of Bt eggplant (a non-field crop) on varietal diversity of eggplant in India, which importantly is one of the vegetable’s centers of origin.1 The paper also presents policy options worth pursuing as the range of technological opportunities in agriculture expands to other horticultural and field crops.

Recognizing that the requisite careful taxonomy of local varieties required for a systematic study of local diversity is a protracted scientific process beyond the scope of this study, we instead utilized a simplified approach which provided initial indications of pending declines in crop genetic diversity. Bt eggplant is yet to be released commercially in India, our geographical focus, so that the analysis is ex ante, which has the advantage of allowing estimates of crop diversity prior to the loss of traditional varieties. This in turn permits identifying alternative approaches to genetic resource conservation which can be implemented prior to any losses.

This report provides the following sections (a) description of the characteristics of eggplant production in India, and the economic importance of Bt eggplant; (b) briefly reviews prior studies examining the environmental effects of modern and GE varieties2; (c) describes the data used for the empirical analysis and associated methodology (d) presents the results of the analysis, and finally (e) offers conclusions and policy implications.

Eggplant production in India and significance of GE eggplant

Eggplant—Solanum melongena—is an economically important non-seasonal vegetable crop cultivated in countries of South and South East Asia, and Africa. Eggplant (known as brinjal locally) occupies 9.4% of the vegetable cropland in India. In 2007–2008, area and production of eggplant in India were 0.56 million hectares and 9.6 million metric tons, respectively (NHB 2008). Many varieties of eggplant differing in shape, color, taste, texture, and size are grown in most of the agro-climatic zones in India.

Farmers grow open pollinated varieties (OPVs) and/or hybrids of eggplant in India. OPVs are propagated by natural (e.g., uncontrolled) pollination. Farmers can save and use the seeds of OPVs without much loss in attributes of the breeding line. Because of the seed saving attribute of OPVs, appropriability (i.e., recouping returns from investments) of research and development (R&D) investments toward improved OPVs is limited. Hence public sector is the major provider of OPV seeds.

Hybrids (F-1) are the first generation progenies of a cross between two genetically dissimilar parents, one designated male and the other female, and the seeds are produced by controlled pollination. The Indian Institute of Horticulture Research developed the first hybrid in eggplant ‘Arka Navaneet’ in 1981 (Sidhu 1998). Unlike OPVs, F-1 hybrids suffer significant declines in yields in subsequent generations, and because of this farmers tend to purchase hybrid seeds annually (Ramaswami 2002). However, because of the increased appropriability through hybridization, the private sector holds the major share of the hybrid seed market.3 Proprietary hybrid seeds are marketed on average at $1.91/10 g, well above OPV seed prices of $0.08/10 g charged by the public sector. The high seed replacement rate of 63% for eggplant in India suggests that many eggplant growers are purchasing seeds annually (MANAGE 2002).

Compared to other vegetable crops in India such as tomato, the adoption of hybrid technology by eggplant growers at 30% is low (Kataria 2005). Eggplant farmers in eastern states grow mainly OPVs while those from southern and western states grow mainly hybrids. Data on total numbers of hybrids and OPVs of eggplant grown nationally or the number of companies engaged in eggplant breeding are difficult to collect. Available information is provided in “Appendix”.

Eggplant production in India is seriously affected by eggplant shoot and fruit borer (ESFB)—Leucinodes orbonalis—with reported yield losses up to 70% (Dhandapani et al. 2003). Because of the feeding nature of the pest, which—once entered into a shoot or fruit—feeds internally during the larval stage; and further, when matured, descends into the soil for pupation, chemical and manual control of the pest (such as by removing wilted shoots and damaged fruits) are not very effective. Frequently applied toxic insecticides threaten the health of producers and consumers, pollute the environment, and raise costs. In this context, Bt eggplant assumes importance because it provides inbuilt protection against the targeted pest, ESFB (Chaudhary 2009).

Bt eggplant containing the “cry1Ac” gene transferred from the soil bacterium Bacillus thuringiensis expresses Bt protein in all parts of the plant throughout its life cycle. To become activated and exhibit insecticidal properties, Bt protein must be ingested by ESFB larvae. When the FSB larvae feed on Bt eggplant plant they ingest the Bt protein along with the plant tissue. In the insect gut, this protein is solubilized and activated by gut proteases which generate a toxic fragment which in turn causes the death of the larvae (Chaudhary 2009).

Mahyco, the developer of hybrid Bt eggplant in India, initiated the research on Bt eggplant in 2000. Even though the Genetic Engineering Approval Committee, the apex regulatory body for GE crops in India, gave its regulatory approval for commercialization of Bt eggplant (five Bt hybrids) in October 2009, a moratorium on the commercialization of Bt eggplant was introduced, based on safety grounds, by the central Ministry of Environment and Forestry in February 2010.4 Another important aspect of the development of Bt eggplant in India is that Mahyco donated the Bt gene for developing Bt open pollinated varieties to selected public breeding institutions such as Tamil Nadu Agricultural University, University of Agricultural Sciences, Dharwad, and Indian Institute of Vegetable Research in India. The donation was facilitated by the Agricultural Biotechnology Support Program funded by the United States Agency for International Development (USAID) and was aimed at making Bt seeds available at affordable prices for resource-poor farmers. Currently in n India, 22 Bt OPVs are awaiting regulatory approval.

From the private sector’s point of view, this public–private partnership is an expression of corporate social responsibility. Even though concerns have been raised about the implications of partnerships, especially for smallholders in developing countries (Glover 2007), studies by Kolady and Lesser (2006) and Kolady and Lesser (2008b) found that because of the differences in production practices between hybrid and OPV growers of eggplant in India, market segmentation is possible for Bt eggplant in India, which in turn would sustain the economic feasibility of the partnership. Further studies by Krishna and Qaim (2008) and Kolady and Lesser (2008a) highlight the welfare implications of the partnership in the Indian context. The partnership involved in the development and introduction of Bt eggplant provides a great opportunity to develop locally-adapted Bt varieties instead of relying on a limited number of Bt hybrids. This opportunity is important because the question of how transgenic crops will influence the diversity of plant types in farmers’ fields is understood as largely dependent on the forces shaping agricultural research, variety development, and adoption. However, the effects of such partnership on the varietal diversity on farm has been studied less from an empirical perspective.

Literature review

Many studies have analyzed the effect of the Green Revolution on genetic diversity. Two major propositions about the Green Revolution are: the Green Revolution caused genetic erosion, and the Green Revolution increased genetic vulnerability (Cooper et al. 1992). Each of these propositions was examined by Smale (1997) in the context of wheat genetic diversity of the developing world. The study concluded that because of the difficulties in defining and measuring genetic erosion and proving causality with multiple intervening factors, a causal relationship between the Green Revolution and genetic erosion in bread wheat varieties could not be established. However, the study noted that the adoption of modern cereal varieties has been characterized first by concentration on a few varieties followed by diversification as more varieties became available.

Studies by Brush (1991) and Bellon (1996) illustrated that present levels of diversity are the outcome of farmers’ needs and preferences on the one hand, and the increasing availability of a limited number of high yielding varieties on the other. Results from these studies suggest that supply (R&D) and demand (varietal preference by farmers) factors are important in deciding the extent of on-farm varietal diversity.

As in the case of the Green Revolution, conflicting findings are reported on the effect of GE crops on genetic diversity (Sianesi and Ulph 1998; Ulph and O’Shea 2002). Sianesi and Ulph (1998) examined the potential threat of GE crops on species diversity and argued that diversity is directly related to the supply of non-GE crops grown, and that the growing of non-GE crops generates external benefits in terms of species diversity. Based on the findings from the study, the authors recommended that whether GE technology is available or not, a subsidy for non-GE crops would be necessary to deliver both an optimal mix of GE and non-GE crops, and achieve the optimal level of R&D. In a related study, Ulph and O’Shea (2002) showed that it is necessary to go beyond intervening in the growing of GE crops, but also to redirect R&D through a tax on the adoption of new GE technology. According to the authors, without intervention at both levels, the cultivation of GE crops and the rate of innovation of GE technologies will exceed their socially optimal levels.

Soleri et al. (2005) evaluated farmers’ perceptions of potential environmental impacts of Bt maize in Cuba, Gautemala, and Mexico. The study concluded that even though perceptions varied between farming communities, farmers saw many potential consequences such as pests developing resistance against the Bt transgene as harmful. The study also highlighted the importance of having a participatory approach, including farmers’ feedback, in building risk management strategies. In a study analyzing farmers’ preferences for milpa diversity and GE maize in Mexico, Briol et al. (2007) found that farmers with the greatest private value from landrace conservation due to various reasons such as incomplete input or output markets were more likely to continue with landrace cultivation.5

However, studies by Sneller (2003) and Bowman et al. (2003) found that because transgenic technology was incorporated into multiple cultivars, its introduction has not affected diversity of cultivars of soybean and cotton in the US Qaim et al. (2005) reaffirm these findings by arguing that biotechnology can preserve crop genetic diversity mainly because it allows for separation between the act of developing novel crop traits and the process of breeding plant varieties. A study by Cleveland et al. (2006) on gene flow effects of Bt maize on landraces highlighted the complex nature of empirical exercises examining the environmental effects of GE crops; that is, conflicting findings could be found depending on the sampling framework and methodology used. Perusal of the above literature shows that empirical studies analyzing the impact of GE technology on varietal diversity of crops other than field crops are limited and prior studies present mixed results. We now examine primary field level data on Indian eggplant producers to project whether and what form (hybrid or OPV) Bt technology will be adopted when available, and use those estimates to determine the potential loss of traditional OPVs as they are replaced by Bt variants.


Field data collection on eggplant production in Maharashtra was conducted during 2004–2005. Maharashtra is one of the major eggplant growing states in India, accounting for 5.2% of area and 5 of eggplant production. In the year 2007–2008, area and production of eggplant were 0.029 million hectares and 0.48 million MT, respectively (National Horticulture Board (NHB) 2008). According to the data available from the state agricultural marketing board, 60% of the area in the state under eggplant is planted with hybrids, making it a good case example for estimating the impact of hybrid technology on varietal diversity. Further, when developing Bt hybrids, private companies initially target hybrid-growing farmers in states such as Maharashtra.

The survey team followed a multi-stage stratified sampling procedure to select the 290 households included in the survey from four districts (Jalgaon, Nagpur, Ahmad Nagar, and Nanded). These districts were chosen to represent the four major geographical zones and to collect information on different market segments of eggplant. The survey covered 20 talukas (an administrative division smaller than a district) and 38 villages known to include farmers producing substantial amounts of eggplant. The research team randomly selected farmers from the list of eggplant farmers or from the list of all farmers provided by the village administrative authorities (Panchayat) (Table 1). In line with the state average, about 60% of the eggplant farmers in our survey were hybrid growers. Because the survey included all major agro-climatic zones in the state, different market segments, and different classes of eggplant farmers, it is considered broadly representative of the state and results are interpreted based on this assumption. A village-level questionnaire was used to collect village-specific information on eggplant production.
Table 1

Classification of farmers participated in the survey



Hybrid eggplant growers


OPV/traditional variety growers


Non-eggplant vegetable farmers




Source: Survey data

The research team used separate questionnaires to interview eggplant growers, non-eggplant growers, and village administrative authorities. The questionnaire for eggplant growers had three sections. The first part included questions on general cropping patterns, years of growing eggplant, adoption details of hybrid seeds, and detailed cultivation practices for eggplant. Questions about farmers’ knowledge of and perceptions towards Bt technology, their willingness to adopt Bt hybrid seeds, their preference towards Bt OPV seeds, and questions exploring their willingness to pay (WTP) for Bt technology were included in the second part. Income, land ownership, and demographic details were included in the last part of the questionnaire.

In order to explain the technology to farmers, we relied on field trial data and inputs from subject experts from the field. Since Mahyco is the developer of the technology, field trial data were available only from the company. According to the field trial data it would be expected that there would be complete elimination of sprayings against ESFB in the Bt fields. Yields from Bt plots would be 117% higher than non-Bt counterparts, 108% higher than a popular hybrid check, and 192% higher than a popular OPV check. Even though performance of the technology on farmers’ fields may differ from experimental results and vary from field to field, these were the best available indicators of the potential performance of the technology available.

Based on the data from the field trials of Bt hybrid eggplant conducted by Mahyco, and on interactions with the subject experts in the field, farmers were told that adoption of Bt hybrid might cause a reduction in insecticide use against ESFB by 70–75% and a yield increase of about 30% over conventional hybrids.6 Farmers were also informed about the possible increase in seed prices for Bt hybrids. Because of the prior introduction of Bt cotton in the state, most farmers were aware of Bt technology and its high price. According to Mahyco scientists, the behavior of the Bt gene is likely to be similar in both Bt hybrid and Bt OPV eggplant. Hence the same benefits as from Bt hybrids were attributed to Bt OPV eggplant; moreover farmers were reminded that once purchased, the Bt OPV seeds could be saved and used as crop seeds for subsequent years.7

A modified version of double-bounded dichotomous choice Contingent Valuation (CV) approach was followed to elicit the information on adoption and willingness to pay (WTP) for Bt hybrid technology. For the WTP question for Bt hybrids, the first bid offered was Rs 400/10 g packet, and if the response was “no” from the farmer, a lower bid was offered. The lower bids offered were: Rs 350, Rs 300, Rs 250, Rs 200, and Rs 150 each for a 10 g packet. The bid ranges were chosen to cover what we perceived to be the likely range of retail prices, and WTP for Bt hybrid seeds. During the pre-testing of the survey, we identified farmers’ difficulties in responding to a double-bounded CV framework for Bt OPV. This may be due to the fact that OPV seeds are marketed at a cheaper price and that not all farmers replace seeds of OPVs annually. Hence, an open-ended CV format where farmers were requested to state their WTP for Bt OPV eggplant was used.

To capture the expected time lag between the introduction of Bt hybrids and Bt OPVs and to assess the effects of the introduction of Bt OPVs by the public sector on expected adoption of Bt hybrids, we used two different hypothetical scenarios. In the first scenario, farmers were requested to express their willingness to adopt Bt hybrid as one of the three options: Bt hybrid, hybrid, and OPV. After collecting information on farmers’ willingness to adopt Bt hybrid, a second scenario was presented to farmers whereby they were told about the public initiative to develop Bt OPVs, and requested again to state their willingness to adopt Bt OPV. In order to capture any possible shift in their stated preference due to the additional information on publicly-developed Bt OPVs, each of the surveyed farmers was again asked to state his/her willingness to adopt Bt technology (Bt hybrid, Bt OPV, both Bt hybrid and Bt OPV, hybrid, and OPV).


Each variety represents a unique combination of genes, some of which might be rare. The extinction of a variety then means the loss of that unique combination and possibly of selected gene(s) of significant potential value, even if the need for those genes may not be known for years. In our study, varietal diversity refers to the number of varieties within a particular crop species, in this case eggplant, recognized by farmers as distinct on the basis of morphological and agronomic criteria. To simplify our analysis we made the following assumptions regarding varietal diversity:
  1. (1)

    Each variety has a unique name and varieties with the same names are morphologically and genetically similar while those with different names are morphologically and genetically dissimilar.

  2. (2)

    If all eggplant farmers in a village switch from variety to hybrids it is more likely that farmers growing the same variety in the neighboring villages, subjected to same economic pressures and opportunities, would switch to hybrids as well. Hence we consider any variety completely replaced by a hybrid in a village as lost, and;

  3. (3)

    Varietal diversity of a cultivated species on farm is a source of genetic diversity, and conservation of on-farm varietal diversity leads to conservation of genetic diversity.


To examine the link between genetic use restriction technology (GURT) (i.e., a technology developed to address the issue of intellectual property rights protection and appropriability), and the rate of diffusion of yield gains to developing countries, Goeschl and Swanson (2000) drew on the experience of hybrid technology- a technology which provided its developers a certain degree of intellectual property right protection and appropriability over OPVs. On the basis of the experiences with hybrid crops, the authors argued that if GURT use parallels that of hybrids that the situation will negatively impact the rate of diffusion of yield gains to developing countries. Because of the lack of data on the performance of GURT, drawing from the experience of a similar technology was justified.

In this study, we used a similar approach to examine the potential effect of Bt eggplant on on-farm varietal diversity. Building on the evidence from the impact of hybrid technology on varietal diversity, we examined the possible effects of Bt eggplant on on-farm varietal diversity. Since Bt eggplant is not yet commercialized, our analysis is based on farmers stated preference and a stated shift in farmer’s willingness to adopt when Bt OPVs are available. In the real world, these responses could vary and hence the impact on varietal diversity would be different from our predictions. However, we justify our approach by noting that an effective policy response to minimizing biodiversity loss should be identified prior to that loss, as is done here.

Currently eggplant growers in India can choose either hybrid or OPV or both. In this analysis we assumed that the social, economic, and institutional factors shaping individual decisions at farm-level influence the varietal diversity at aggregate levels. With the introduction of Bt eggplant, farmers have three choices to select from (OPV, hybrid and Bt hybrid) in scenario 1 (where only private Bt hybrids are available) and four choices to select from (OPV, hybrid, Bt hybrid, and Bt OPV) in scenario 2 (where both private Bt hybrids and public Bt OPVs are available).

Because on-farm varietal diversity at aggregate levels (village, district, regional, or national) depends largely on farmers’ decisions at farm-level, extrapolating results from the analysis of farmers’ adoption behavior at the farm-level to aggregate levels may shed some light on the possible impacts of introduction of Bt eggplant on on-farm varietal diversity. Because of the nominal nature of the categorical dependent variable we used multinomial logit model (MNL) to examine factors affecting farmers’ expected adoption in the (hypothetical) context of Bt eggplant and to predict probabilities of different possible outcomes of technology choices at the farm-level. We analyzed the potential impact of Bt eggplant on on-farm varietal diversity in three steps. First, using assumptions (1–3) stated earlier, we estimated the varietal diversity loss by counting the number of villages completely shifted to hybrids. Second, a multinomial logit model is used to analyze farmers’ technology choice decisions under the two hypothetical scenarios of introduction of Bt eggplant. Third, we extrapolate the results at the farm-level on technology choice to the village level to classify the villages based on type of eggplant grown by the farmers. Comparing the outcomes in step 1 and step 3 provides the estimate of variety diversity loss under the two hypothetical scenarios of introduction of Bt eggplant in India.


Impact of hybrids on varietal diversity

Extrapolating farmer-level information gathered through farm-level surveys to the village level, the villages included in the survey were classified into three categories: the villages growing only hybrids; villages growing hybrids and OPVs; and villages growing only OPVs (Table 2).8 Using the assumptions above, we found that 55% of the 38 villages in the survey had switched completely to hybrids, while another 26% is under transition. As expected, the percentage of villages cultivating only OPVs is the smallest (18%).
Table 2

Classification of villages based on hybrid adoption




Villages growing only hybrid



Villages growing OPV and hybrid



Villages growing only OPV



Source: Survey data

Using information from the survey respondents and cross checking with village heads, we calculated the variety loss (number of varieties lost) due to the introduction of hybrids as 14 among the 38 sampled villages. Since another 26% of the villages are under transition (growing hybrid and OPV), the number of villages growing only hybrids is likely to increase over time (assumption 2). Even though 17 villages in our survey grow OPVs, we identified only 12 OPVs. Assuming the sample is representative of the state, this gives an estimate of 0.7 varieties per OPV growing village.

Impact of Bt eggplant on varietal diversity

Current production practices of farmers, especially hybrid versus OPVs, reflect farmers’ revealed preferences towards a particular type of variety which could be shaped by socio-economic characteristics at the farm level or by more regional characteristics such as population density, agro-ecology, or development of commercial markets. Farmers’ stated preferences towards Bt eggplant gathered through the survey were used to examine farmers’ expected adoption behavior in the context of Bt eggplant, and the factors contributing to that adoption behavior. Descriptions of the variables used in the regression analyses (MNL model) are presented in Table 3. The explanatory variables included in the regression analysis could be broadly classified into household characteristics and contextual characteristics. The household characteristics include both farmer and farm attributes.
Table 3

Description of the variables used in the multinomial logit model



Hybrid growers

OPV growers

District 1

1 if Jalgaon, else 0

.09 (.29)

.54 (.50)*

District 2

1 if Nagpur, else 0

.28 (.45)

.28 (.45)

District 3

1 if Ahmad Nagar, else 0

.4 (.5)


District 4

1 if Nanded, else 0

.24 (.4)

.2 (.4)

Family size

# of family members

7.3 (4.9)

6.6 (3.4)


Age of the head in years

44.4 (11.6)

48.5 (12.8)*


1 if yes, else 0

0.87 (.34)

.82 (.39)


1 if good access to banks

.87 (.34)

.76 (.43)


Years of growing eggplant

7.8 (4.4)

12.8 (8.4)**

Total land

Total land in acres

10.6 (10.3)

8.9 (6.5)*

Expenses (Rs/farm)

To control ESFB

9,307 (9,970)

2,098 (2,154)**


# of irrigations/farm

28.7 (25)

17.1 (11.6)**

Season of growing

1 if kharif, else 0

.36 (.48)

.77 (.42)**


To eggplant market (Km)

76.3 (91.7)

49.7 (87.5)**

Preference of variety

1 if yield attribute weigh most

.33 (.47)

.14 (.35)**

Diversity loss

1 if concerned about variety loss, else 0

.31 (.52)

.35 (.66)

Source: Survey data

Values in parentheses are standard deviations. Based on the survey, the average yields of hybrid and OPV eggplants are 18.5 MT/ha and 12.6 MT/ha, respectively. Also, crops grown with the onset of first rains in the May–June (South-West monsoon) period are known as kharif crops

*  Mean values are statistically different at the 10% level from hybrid growers

**  Mean values are statistically different at the 5% level from hybrid growers

Contextual characteristics such as district dummies are expected to capture the agro-climatic conditions and infrastructure variations which may influence farmers’ technology choice decisions. In addition to the socio-economic conditions of the farmers, infrastructure and institutional set ups in the district also might contribute to farmers’ adoption decisions. Among the districts selected for the study, Jalgaon is the least developed district and borders with Madhya Pradesh, where OPV eggplant is more popular than hybrids. Ahmad Nagar is a well developed district with better infrastructure and almost all of the eggplant farmers in the district grow hybrids, whereas hybrids and OPVs are equally popular in Nanded and Nagpur. Distance to the market (eggplant market) is included as a proxy for market access. It is evident from Table 3 that the hybrid production system is more intensive (i.e. uses more inputs and has higher yields) compared to the OPV production system.

The MNL regression analysis helps predict the probabilities of different possible outcomes of categorical dependent variable at the farm-level given the independent variables. Based on farmers’ responses to the willingness to adopt Bt technology, the dependent variable ‘alternative’ was constructed. This dependent variable takes three discrete values in scenario 1, recalling farmers have three choices: hybrid, OPV, and Bt hybrid. The ‘alternative’ OPV eggplant was taken as the reference point in the regressions.

Regression results from the MNL model analyzing farmers’ adoption decision in the context of introducing Bt hybrid eggplant (scenario 1) are presented in Table 4. Examination of the significance of coefficients of the variable included in the MNL model suggests that contextual characteristics such as district dummies and farmers’ socio-economic characteristics such as access to credit and literacy status, for instance, are significant in farmers’ decisions to adopt hybrid or Bt hybrid eggplant. The more or less similar coefficients for significant variables for hybrid and Bt hybrid alternatives in Table 4 might suggest that an initial shift to Bt hybrids could occur from current hybrid growers.
Table 4

Estimated coefficients from the multinomial logit model based on scenario 1


Hybrid (SE)

Bt Hybrid growers (SE)

District 1 (Jalgaon)

−3.2 (.97)**

−2.6 (.75)**

District 2 (Nagpur)

−.14 (.77)

−.87 (.66)

Family size

−.03 (.09)

−.08 (.07)

Age of head

−.003 (.02)

−.01 (.02)


−.13 (.72)

1.06 (.63)*

Access to banks

1.3 (.68)**

1.9 (.59)**

Total land

−.002 (.05)

.05 (.04)


−.004 (.004)

−.0002 (.003)


−.21 (.07)**

−.11 (.04)**


.0004 (.0001)**

.0005 (.0001)**


.02 (.03)

.04 (.02)*


.6 (.73)

.06 (.6)


.45 (.65)

.39 (.54)


.03 (.54)

.11 (.4)


−.64 (1.6)

−1.5 (1.3)

In scenario 1, eggplant farmers have three technology choices available to them: Bt hybrid, hybrid and OPVs. OPV is taken as the reference group. Standard errors in parentheses. Dummy variable for district 3 (Ahmad Nagar) was dropped from the analysis due to high collinearity with hybrid and Bt hybrid adoption. District 4 was the reference district in the analysis

* Significance at 10% level

** Significance at 5% level

From the results in Table 4, farmers’ expected adoption behavior is as follows: 63% are likely to adopt Bt hybrid, 4% are likely to adopt hybrid, and 33% are percent likely to adopt OPV.9 The villages included in the survey were classified into various eggplant growing categories based on farmers’ expected choices. As presented in Table 5, the number of villages expected to grow only hybrid or Bt hybrid is 14, same as the estimated diversity loss presented above due to hybrid technology. However, compared to Table 2, results presented in Table 5 show a 57% decrease (from 7 to 3) in the number of villages expected to grow only OPV eggplant. This implies that with the introduction of Bt hybrids, some farmers presently cultivating OPV eggplant and located in villages cultivating only OPVs are likely to shift from OPVs to hybrids or Bt hybrids. The three villages expected to grow only OPVs are located in Jalgaon, the most remote of the districts surveyed. This penetration of Bt hybrids may lead to the loss of OPVs, as previously happened with the introduction of hybrid technology. Using the previous estimate of 0.7 varieties per OPV growing village, the number of varieties that could be lost by hybrids or Bt hybrids is about 17. However, it should be noted here that since Bt eggplant contains more valuable attributes than simple hybrids, estimating varietal diversity loss this way may be a conservative one.
Table 5

Classification of villages based on scenario 1




Only OPV



Only hybrid



Only Bt hybrid



OPV and hybrid



OPV and Bt hybrid



Hybrid and Bt hybrid



OPV, hybrid, and Bt hybrid



Source: Survey data and author’s calculations

In scenario 1, eggplant farmers have three technology choices available to them: Bt hybrid, hybrid and OPVs

Under scenario 2, seven farmers expressed their willingness to adopt both Bt hybrid and Bt OPV. However, since their number was too small to include as a separate category of dependent variable (i.e., technology choice), those seven farmers were excluded from the analysis. Results from the MNL model under scenario 2 are presented in Table 6. The estimated coefficients in Table 6 were used to identify farmers’ choice among the four alternatives, and this information was extrapolated to the village level to classify villages into different categories based on farmers’ expected adoption decision. Classification of villages using estimated coefficients under scenario 2 is presented in Table 7. The classification of farmers based on their expected adoption behavior is as follows: 54% of the farmers are likely to grow Bt hybrids, 28% is likely to grow Bt OPVs, 6% is likely to grow hybrids, and 12% is likely to grow OPVs (please see footnote 9 for details on this calculation). As expected, most OPV farmers prefer Bt OPV eggplant.
Table 6

Estimated coefficients from the multinomial logit model based on scenario 2


Hybrid (SE)


Bt hybrid (SE)

District 1 (Jalgaon)

−3.4 (1.4)**

−1.8 (1.4)

−5.4 (1.4)**

District 2 (Nagpur)

−2.3 (1.3)*

−2.2 (1.3)*

−2.2 (1.3)*

Family size

−.08 (.1)

−.06 (.09)

.02 (.09)

Age of head

−.03 (.03)

−.02 (.02)

−.03 (.03)


.18 (1.009)

−.15 (.85)

.51 (.94)

Access to banks

.47 (1.1)

−1.3 (.92)

−.31 (.97)

Total land

.09 (.06)

.05 (.06)

.06 (.06)


−.02 (.009)**

−.009 (.004)**

−.001 (.004)


−.09 (.06)

−.009 (.04)

−.16 (.05)**


.0002 (.0001)*

−.0001 (.0001)

.0002 (.0001)*


.04 (.04)

.006 (.03)

.05 (.03)


−.12 (1.04)

−.95 (.96)

−1.1 (.96)


1.9 (.94)**

1.5 (.87)*

2.1 (.92)**


.09 (.48)

−.68 (.39)*

−1.3 (.51)**


2.9 (2.6)

6.1 (2.2)

4.9 (2.4)

In scenario 2, eggplant farmers have four technology choices available to them: Bt hybrid, Bt OPVs, hybrid and OPVs. OPV is taken as the reference group. Standard errors in parentheses. Dummy variable for district 3 (Ahmad Nagar) was dropped from the analysis due to high collinearity with hybrid and Bt hybrid adoption. District 4 was the reference district in the analysis

* Significance at 10% level

** Significance at 5% level

Table 7

Classification of villages based on scenario 2




Only OPV



Only Bt hybrid



Only Bt OPV



Bt hybrid and Bt OPV



OPV and Bt hybrid



OPV and Bt OPV



Hybrid and Bt hybrid



OPV, Bt hybrid, and Bt OPV



OPV, hybrid, and Bt OPV



OPV, hybrid, and Bt hybrid



Hybrid, Bt OPV, Bt hybrid



OPV, hybrid, Bt OPV, and Bt OPV



Only hybrid



Source: Survey data and author’s calculations

In scenario 2, eggplant farmers have four technology choices available to them: Bt hybrid, Bt OPVs, hybrid and OPVs

Our results show that 12 villages are likely to switch completely to Bt hybrid or hybrid eggplant, which is 14% less than the number from scenario 1. The number of villages expected to grow only OPVs (OPV and Bt OPV eggplant) is six, which is 100% higher than that in scenario 1. In scenario 2, the percentage reduction in the number of villages expected to grow only OPVs (OPV and Bt OPV eggplant) due to Bt technology was lower (14% as opposed to 57% in the scenario 1) than that in scenario 1. The number of villages expected to grow non-Bt OPVs is 20 (compared to 24 in scenario 1) which implies that the number of varieties that could be lost with the introduction of Bt eggplant is 14 in scenario 2, 18% lower than that in scenario 1.

As is evident from the above analysis, in addition to farm or farmer characteristics there are contextual characteristics that influence farmers’ choice of varieties. Introduction of Bt OPVs provides an opportunity for OPV growers to continue with OPV cultivation, and thus reduces the rate at which OPVs will be replaced with hybrids or Bt hybrids. Introduction of novel crop traits into both OPVs and hybrids may also help to address the concern that a limited number of Bt hybrids will supplant numerous OPVs leading to diversity loss on farm. Findings from the study indicate that the public sector R&D investments in agricultural biotechnology, especially in crop segments where the private sector has no incentive to invest, can not only provide access to technology for the resource-poor farmers, but also can help in addressing the dynamic biological problems posed by the technological advances in agriculture.

Cost of developing Bt OPVs

When determining which aspect of agricultural biodiversity to conserve, one can value the benefits, the costs, or ideally both. Here the focus is on the costs of conservation which, compared to available public budgets in India, provides an indication of the economic feasibility of the in situ methods for reducing crop biodiversity loss following the introduction of Bt eggplant.

The prior analysis assumes that all OPV varieties currently grown by farmers will be available as Bt OPV varieties. In our study we identified 12 OPVs from 17 villages in Maharashtra, and this indicates the difficult task faced by the public sector in developing Bt OPVs satisfying farmers’ varied needs and preferences. Based on cost data collected from Tamil Nadu Agricultural University, the average cost for developing a single Bt OPV variety using royalty-free Bt technology from Mahyco is about US $89,000, which includes the cost of development (back crossing) and field trials (for agronomic performance).10 Thus the estimated cost of developing 12 Bt OPVs from the 12 varieties identified in the survey is about one million US dollars.

Public breeding programs could generate revenues from Bt OPV seed sales, which might be used to fund the development of Bt OPVs. However, the expected revenues from seed sales would depend largely on the demand and seed prices charged. In order to examine the economic feasibility of developing a large number of Bt OPVs by the state, we estimated the expected sales revenue from Bt OPV seeds in two steps. First, assuming that all of the eggplant growers in the state would grow Bt OPV variety in 100% of the current area under eggplant, we estimated the demand for Bt OPV seeds as 3,233 kg in the state. In the second step, the estimated value of average WTP for Bt OPV eggplant by expected adopters of Bt OPV eggplant (Rs 115/50 g) was used as a proxy for seed price to calculate the expected sales revenue. Under these assumptions, the sales revenue is expected to be US $ 0.2 million, much lower than the cost of developing the 12 varieties identified in the survey. These costs apply only to the 38 surveyed villages so that the cost of transforming all varieties in the state would be substantially higher. It should be noted here that the estimated values of costs and sales revenue are sensitive to the assumptions we used. However, overall the analysis indicates that large amount of public sector investment would be required to develop large number of Bt OPVs in the state.

Conclusions and policy implications

On-farm varietal diversity is one available risk management tool, especially for resource-poor farmers. While the introduction of hybrid eggplant seeds contributed to productivity enhancements, their large scale adoption resulted in a decline of on-farm varietal diversity in states such as Maharashtra, India. This decrease in diversity attributable to hybrid adoption can be expected to accelerate with the introduction of a limited number of Bt hybrids which farmers are predicted to prize mainly because of the clear benefits by means of pesticide cost reductions and damage abatement. In this study, building on the evidence from the impact of hybrid technology on varietal diversity, we examined the possible effects of the introduction of Bt eggplant on on-farm varietal diversity under two scenarios. In scenario one, Bt technology is available only in hybrids, whereas in scenario two Bt technology is available in both hybrids and OPVs. Since Bt eggplant is as yet not commercialized, our ex ante analysis was based on farmers’ revealed preferences for hybrid eggplant and a stated willingness to adopt Bt when available in both hybrid and OPVs forms.

Findings from our study suggest that the proposed introduction of Bt OPVs by the public sector could reduce the rate of replacement of OPVs by Bt hybrids which in turn maintains a greater level of genetic diversity than is likely if the only Bt cultivars available were hybrids. Thus, this example of research and development of Bt eggplant in India indicates that investment by the public sector in agricultural biotechnology R&D, especially in crop segments or markets where the private sector has no incentive to invest, could address some of the dynamic problems posed by technological advances in agriculture. However, the need for significant public sector investment for developing multiple Bt OPVs limits the potential of this option for on-farm varietal conservation. As the range of technological opportunities expands, it is important to examine the economic feasibility of other supply and demand-related policy interventions to maintain on-farm varietal diversity. Some of the policy options worth pursuing are briefly discussed below.

India is considered one of the centers of diversity of eggplant, and the National Bureau of Plant Genetic Resources has identified 134 diversity-rich districts in the country. As is shown in this study, due to different contextual characteristics, eggplant growers in some states/districts may be less likely than others to replace traditional varieties with hybrids. Identifying those farmers and targeting such sites may provide an economically efficient mechanism for on-farm varietal conservation. However, in locals where on-farm conservation is not economically efficient, both from the farmers’ and the government’s perspectives, efforts must be made for ex situ conservation. The timely identification of those areas will allow for the collection and conservation of the threatened local varieties before they have become extinct.

One supply-related policy intervention for in situ conservation is to provide subsidies for growing OPV eggplant by a limited number of farmers in diversity-rich villages, for example by creating community gene banks. However, the identification and selection of farmers and effective monitoring of such a program may be difficult. The social opportunity cost and welfare implications of such an intervention could be a topic of future research.

Demand-related mechanisms such as labeling to identify a unique attribute of a traditional variety (taste, flavor etc.) might provide market incentives for maintaining on-farm varietal diversity. The economic feasibility of such programs depends largely on consumers’ willingness to pay a premium for traditional varieties and farmers’ ability to build a niche market for those local varieties, including a possible need for public subsidies to develop those markets. However, it should be noted here that, even in the best case scenario, only a limited number of varieties could reasonably be sustained by niche markets.

These are but a few examples of the kinds of policy responses to biodiversity loss which must be identified and implemented prior to any loss. More research needs to be conducted to examine the economic feasibility of various alternative supply and demand-related policy instruments for the conservation of crop genetic diversity, not only for eggplant but also for other crops in the GE technology pipeline. Results from such research could lead to evidence-based policy responses to pending declines in crop genetic diversity. Now is none too early a time to begin.


Bt eggplant, developed through genetic modification, contains the “cry1Ac” gene transferred from the soil bacterium, Bacillus thuringiensis. Bt eggplant provides resistance against eggplant shoot and fruit borer (ESFB), one of the major pests of the crop (see below for more details).


In this study “modern” varieties refer to those varieties developed by human intervention using techniques of plant breeding. GE varieties are those varieties developed using techniques of ag-biotechnology, especially, genetic modification.


Because of tight regulations, public sector was the major player in Indian seed industry until late 1980s. After the 1980s, private companies were allowed to obtain breeder seeds directly from public research institutions. However, the “New Policy on Seed Development” introduced in 1988 permitted import of germplasm for research and import of commercial vegetable seeds, which enhanced private research in vegetable seed development (Ramaswami 2002).


Subsequent to the moratorium, the name of the apex regulatory authority has been changed from Genetic Engineering Approval Committee to Genetic Engineering Appraisal Committee.


A landrace is a local variety of a domesticated plant species which has developed largely by natural processes, by adaptation to the natural and cultural environment in which it is grown.


To draw a comparison, field trial data of Bt cotton in India suggested 70% reduction in insecticide use and 80% increase in yield (compared to non-Bt counterpart) due to use of Bt technology (Qaim and Zilberman 2003). Based on panel survey data in Maharashtra, Karnataka, Andhra Pradesh, and Tamil Nadu, Subramanian and Qaim (2009) reported that insecticide use in Bt plots was less than that in conventional plots by 50, 5 and 21% in years 2002–2003, 2004–2005, and 2006–2007, respectively. Similarly, yields from Bt plots were higher by 34, 35 and 43% in years 2002–2003, 2004–2005, and 2006–2007, respectively.


Stakeholders in the public sector expect that farmers would purchase Bt OPV seeds annually to ensure seed quality. This might be a realistic assumption given the high seed replacement rate of eggplant in India. However, there are concerns about the performance of the Bt gene in varieties over subsequent generations.


Using results from a probit model of hybrid adoption and extrapolating the same to village level gave the same distribution of villages as that based on farmers’ revealed preference gathered through the survey instrument.


Farmers’ expected adoption behavior is calculated as follows. After running the MNL regression, predicted probabilities for each of the alternatives were estimated for each of the farmers in the sample. The alternative with the highest probability was selected as the expected choice. For example, if the predicted probabilities for the first observation were 0.6, 0.3, and 0.1 for Bt hybrid, hybrid, and OPV alternatives, respectively, then Bt hybrid was taken as the expected choice for the first observation. Once the expected choice for each of the observations was calculated, frequencies for each of the alternatives were calculated for the whole sample. Tabular analysis was used to classify villages based on farmers’ expected adoption behavior.


This information is based on personal communication with Mr. Vijayaraghavan K, the coordinator of Agricultural Biotechnology Support Project II in India.



Authors acknowledge the financial support of the Agricultural Support Project II (ABSP II) for the study. Special thanks to the farmers in Maharashtra, who actively participated in the farm-household survey conducted by the research team. Authors are grateful to Dr. Usha Zehr of Mahyco, for sharing with us the field trial data of Bt eggplant.

Copyright information

© Springer Science+Business Media B.V. 2011