1

Ex-ante assessment of the potentialimpact of genetically modified bananaresistant to Xanthomonas wilt in the

Great Lakes region of Africa

International Institute of Tropical Agriculture

Report prepared by:Ainembabazi John Herbert (IITA – Uganda),Tripathi Leena (IITA – Kenya),Rusike Joseph (IITA – Tanzania/ AGRA – Kenya)

August 2014

2

Contents

Acknowledgement ------------------------------------------------------------------------------------------------------ 3Abstract --------------------------------------------------------------------------------------------------------------------- 4Introduction --------------------------------------------------------------------------------------------------------------- 5Conceptual framework and methods --------------------------------------------------------------------------- 7

The data sources --------------------------------------------------------------------------------------------- 7The analytical approach ----------------------------------------------------------------------------------- 9

The results ---------------------------------------------------------------------------------------------------------------- 11BXW awareness, spread and impact on banana production ----------------------------- 11Awareness of genetically modified crops and development of GMB-BXW ----------- 13Potential adoption of GMB-BXW --------------------------------------------------------------- 15Consumption preferences: local versus improved banana varieties ------------------- 18Potential consumers of GM banana resistant to BXW------------------------------------- 19Geographical areas and banana varieties for potential intervention with GMB-BXW---------------------------------------------------------------------------------------------------------- 21Economic evaluation of GMB-BXW development------------------------------------------- 22

Conclusions --------------------------------------------------------------------------------------------------------------- 25

List of TablesTable 1. Characteristics of respondents (percentages and means) ---------------------------------- 9Table 2. Control methods of BXW and awareness of genetically modified crops------------- 15Table 3. Willingness to pay and reasons for adoption of GMB-BXW----------------------------- 16Table 4. Preferences and consumption attributes between improved and local bananavarieties -------------------------------------------------------------------------------------------------------- 19Table 5. Parameter values used to estimate the economic benefits of GMB-BXW using ESM------------------------------------------------------------------------------------------------------------------- 23Table 6. Potential benefits of developing GMB-BXW in the GLA region ------------------------ 24

List of figuresFigure 1. Farmers’ awareness of BXW on target countries----------------------------------------- 12Figure 2. Banana production loss due to BXW incidence ------------------------------------------- 12Figure 3. Willingness to adopt GMB-BXW-------------------------------------------------------------- 16Figure 4. Projected adoption rate of GM BXW resistant banana --------------------------------- 17Figure 5. Left pie chart shows farmers’ consumption preference of banana varieties. Rightpie chart shows consumers’ inquiry about banana before purchase from either farmers ortraders. --------------------------------------------------------------------------------------------------------- 19Figure 6. Potential consumers of GM banana--------------------------------------------------------- 20Figure 7. Frequency of reasons for consumer preference of GM banana (%) ----------------- 21

3

AcknowledgmentWe greatly appreciate interest, guidance and participation of national scientists/researchersfrom Burundi (Emmanuel Njukwe), Democratic Republic of Congo (Dowiya Nzawere), Kenya(Margret Onyango), Rwanda, Tanzania (Mgenzi Byabachwezi and Rony Swennen), andUganda (Jerome Kubiriba). We acknowledge the information given by the governmentofficials, private tissue culture laboratory operators of Kenya, extension agents, bananatraders and farmers. Lastly, we appreciate technical advice given by Tahirou Abdoulaye andcomments from Victor Manyong.

4

AbstractThis study evaluates the potential economic impacts of a genetically modified (GM) banana

cultivar resistant to banana Xanthomonas wilt disease. The ex-ante findings show that the

development of GM banana is viable and has large expected economic impacts. The main

beneficiaries of this technology development are consumers and farmers. The magnitudes

of economic impacts vary substantially across the target countries, being highest in

countries where disease incidence and production losses are high. The report presents

results highlighting the important triggers of successful adoption and consumption of new

technologies that have policy implications for the successful scaling up of GM banana in

the target countries.

5

1. Introduction

The Great Lakes region of Africa (GLA) comprising of Kenya, Tanzania, Uganda, Burundi,

Rwanda and Democratic Republic of Congo (DRC), is the largest producer of bananas in

Africa. The metrics from Food and Agriculture Organization (FAO) show that GLA contributes

over 60% of the total area under banana cultivation across Africa, producing 63% of Africa’s

and 21% of the world’s banana supply (FAO, 2014). At the same time, GLA is the highest

consumer of its bananas in the world with 3 – 22% of total calorie consumption per capita,

and 147 kcal daily consumption per person, which is 15 times the world’s average and 6

times Africa’s average (FAO, 2014)1. Banana is not only a staple food in the GLA region but

also an important cash income source. Increasing the productivity and profitability of

banana production and marketing systems is important for achieving household and

national food security and for driving down the real price of food in order to accelerate

economic growth, increase sustainable management of natural resources, improve nutrition

and health and reduce poverty.

However, the banana production is being seriously threatened by the outbreak and

spreading of Banana Xanthomonas wilt (BXW) (Tripathi et al. 2009). This disease is caused

by Xanthomonas campestris pv. musacearum (Xcm). BXW affects both production and

quality. Crop losses can be as high as 100%. Currently two major approaches exist to control

the disease. First, the use of cultural practices to identify and manage the disease. This

involves removing the male bud (to prevent infection from insects), using sterilized farm

tools and destroying single infected stems (or the whole mat). However, the level of BXW

control by cultural practices can be inconsistent due to noncompliance in implementation

by the value chain actors, especially the farmers and traders.

The second approach is to use natural host plant resistance to prevent the disruptive

impacts of BXW. This natural host plant resistance is non-existent in cultivated banana

germplasm. In the absence of known host plant resistance among cultivated banana

cultivars, the International Institute of Tropical Agriculture (IITA) in partnership with the

National Agricultural Research Organization (NARO) of Uganda and African Agricultural

Technology Foundation (AATF) has successfully developed a genetically modified (GM)

1 The figures are own calculations from FAOSTAT database.

6

banana germplam resistant to BXW. The germplasm was developed through constitutively

expressing Hypersensitive Response Assisting Protein (Hrap) or plant ferredoxin like protein

(Pflp) gene from sweet pepper (Capsicum annuum) (Tripathi et al. 2010; Namukwaya et al.

2012). The wilt-resistance genes from pepper; Pflp and Hrap were obtained under an

agreement from Academia Sinica in Taiwan and the proteins they confer are not listed as

being a potential allergen in AllergenOnline and should be safe for human consumption (and

are widely distributed across a broad range of plant species including rice and fruits that are

eaten raw). Even so, the GM bananas will be tested for food and environmental safety in

compliance with biosafety protocols and regulations. The risk of gene escape to other crops

is unlikely since most edible bananas are sterile and the process involves clonal propagation

which limits gene flow to other crops.

The GM banana plants were evaluated for resistance against Xcm in the laboratory,

screen house, and confined field trial in Uganda (Tripathi et al. 2013). Based on results from

mother plants and their first ratoon plants, 11 lines were identified showing absolute

resistance. The transgenic lines showed flowering and yield (bunch weight and fruit size)

characteristics comparable to non-transgenic varieties. Currently, these lines are under

evaluation for durability of disease resistance and agronomic performance in the second

trial in Uganda. The identified lines will be further evaluated in multi-locations in Uganda for

environmental and food safety in compliance with Uganda biosafety regulations. The best

lines will then follow risk assessment and management, and seed registration and release

procedures. The GM banana is likely to be released for multiplication, distribution and

commercialization in 2020. While evidence exists on the technical performance of the GM

banana under development in controlling BXW as explained above, no empirical data have

documented yet the economic profitability and potential acceptability of this new

technology.

This study assesses the potential impact of the GM banana varieties prior to their

release, dissemination and commercial sale to farmers. Specifically, the study objectives are

three-fold. First, to analyze production and consumption patterns of bananas, map out

geographical areas for intervention to overcome impacts of BXW and estimate a baseline

scenario for production and marketing of bananas and forecast the likely changes with

adoption of GM bananas. Second, to evaluate the potential benefits and costs of GM

7

bananas to producers and consumers. Third, to identify institutional and policy innovations

that are needed for value chain actors to fully capture the benefits and minimize the costs.

2. Conceptual framework and methods

The study used rapid appraisal value chain analysis and ex-ante benefit-cost framework

using an economic surplus model (ESM). The value chain framework was used to collect and

analyze data on production, consumption of bananas and identification of geographical

areas and banana varieties for potential intervention. The economic surplus model was used

to estimate the economic benefits of developing GM bananas resistant to BXW (GMB-BXW).

Because the GM bananas have not yet been commercially released, it is not possible to

measure the impacts under farmers’ conditions and acceptability to consumers. However,

the ESM, described in detail below, has the ability to quantify the potential benefits and

costs of GMB-BXW accruing to consumers and farmers.

2.1 The data sources

The ex-ante assessment of the potential impacts of GMB-BXW on economic benefits was

conducted in six countries in 2013: Burundi, DRC, Kenya, Rwanda, Tanzania, and Uganda.

The selection of study areas in each country was based on major banana producing areas

and the incidence of BXW. Three districts (or an equivalent administrative unit depending

on the country) were selected in each country. Each of the selected districts represented

the regional block in each country. The selection of districts was done with the help of

resource people knowledgeable on banana production at the national research institutions.

In each district, a key informant working with the government agricultural department or

research institution was identified. The key informant served at least one purpose. First, to

identify banana farmers, traders both in major towns and rural markets, and agricultural

extension agents (both from government and agricultural based development

organizations). Second, to participate in the study by providing information on banana

production, adoption of new varieties, marketing and local institutional policies on

development of agricultural technologies.

Table 1 reports the background information of the participants in the study. The

data were collected in 2013 in the 6 countries. Farm level data on banana production, BXW

8

incidence and its effects on production, potential adoption of GMB-BXW, banana varieties,

consumption and marketing of banana were collected from at least 6 farmers per country,

making a total of 37 farmers of whom 65% represented smallholder farmers and 35%

represented progressive or model farmers. Out of the 6 farmers selected in each country, at

least two farmers would be progressive farmers. Progressive farmers were defined as those

considered by the community to be the largest banana producers compared to others and

largely using improved agricultural practices. Smallholder farmers were leaning largely on

subsistence production, while progressive farmers largely produced bananas for sale.

Market information on consumer preferences and attributes for bananas was collected

from at least 4 banana retailers and 2 wholesalers per country, making a total of 47 traders

(29 retailers and 18 wholesalers). Retailers included banana traders in district towns and

itinerant traders in villages where selected farmers lived. Wholesalers were mainly those

supplying banana to major district towns. Extension agents provided broad overview of the

incidence of BXW, its impacts on banana production, and potential adoption of GMB-BXW.

Majority of extension agents (66.7%) interviewed worked either for international or local

non-governmental organizations. In addition to information similar to that collected from

extension agents, key informants provided more information on the spread of BXW

countrywide and biotechnology policy related issues. At least 3 extension agents and one

key informant were interviewed from each country. The seven key informants included 4

senior agricultural government officials, 2 banana breeders, and 1 private tissue culture

laboratory official.

9

Table 1. Characteristics of respondents (percentages and means)Farmers (N=37)% of smallholder farmers (versus progressive) 64.9% of male respondents 75.7Average age (years) 55.4Education (years of schooling) 7.7Experience in banana production (years) 26.7% of crop income from bananas 62.2Traders (N=47)% of banana retail traders (versus wholesalers) 61.7Experience in agricultural produce trade (years) 14.1Experience in marketing of bananas (years) 12.2Number of banana bunches sold on a market day 86.3% of business income from banana trade 57.1Agricultural extension agents (N=21)% of extension agents working with government institutions 33.3Experience in area of academic training (years) 11.5Experience in banana extension work (years) 10.1Key informants (N=7)% of key informants working with government institutions 71.4Experience in area of academic training (years) 16.2Experience in banana extension work (years) 14.6

2.2 The analytical approach

There are several approaches to evaluate the ex-ante impact of agricultural technologies

including economic surplus model (ESM), benefit-cost analysis and econometric models

(Alston et al. 1998). For this study, an ESM was used. Despite its drawbacks to control for

measurement error, general equilibrium effects, transaction costs, and externalities, it is

still preferred to other approaches when acceptable assumptions are used. Unlike benefit-

cost and econometric analyses, the ESM does not assume perfectly elastic or inelastic

demand and supply, and it controls for both international prices and distributional effects

(Alston et al. 1998).

For ex-ante analysis in this study, the ESM model for the closed economy was used

to estimate the economic benefits for each country from the change in banana supply due

to introduction of GMB-BXW. Alston et al. (1998) provide details of the formulation to

10

estimate ex-ante economic benefits (producer and consumer surpluses) from a given

technology, and these details are not repeated here. The release and the eventual adoption

of GMB-BXW are expected to increase banana yields substantially from the baseline of the

BXW presence scenario (i.e., a downward shift in the banana supply curve), while the

demand level remains unchanged. This will lead to a fall in banana prices. As a result,

consumers gain through paying less for more and farmers benefit through large supplies to

the market. In other words, both the consumer and producer surpluses increase and so are

the total economic surpluses (benefits) resulting from the release and adoption of GMB-

BXW.

The changes in economic surpluses are measured for the research period from 2013

– 2038 for which tangible research outputs are, and will be, realized. By 2013, researchers

had successively developed and tested transgenic banana cultivar lines resistant to BXW.

This means that research costs prior to 2013 are considered as unrecoverable costs since

these costs were invested in acquisition of equipments and building knowledge base, which

are more or less fixed. The period from 2013 to 2020, the expected release date, is

associated with the investments in multi-location fields and biosafety testing. The annual

supply changes are then estimated based on the projected adoption rate for GMB-BXW for

the period from 2020 – 2038. Costs associated with the projected adoption rate would

represent the extension costs for the period following the official release of GMB-BXW. For

each country i at time t, the net present value (NPV) is used to evaluate the changes in

economic surplus (ES) for period 2013 – 2038 as:

2038 2038, ,

2013 2013(1 ) (1 )i t i t

t tt t

ES CNPV

r r

(1)

where r is the real discount rate set at 10% per annum, and Ct is the total research cost.

To measure returns on investment into development of GMB-BXW, the internal rate

of return (IRR) was computed. The IRR is the discount rate that equates the NPV in equation

(1) to zero. The benefit-cost (BC) ratio was also calculated to measure returns from each

dollar invested in development of GMB-BXW. The BC ratio is computed as the ratio of NPV

of benefits to the net present value of research and extension costs.

11

3. The results

This section presents results based on the information elicited from the survey participants

described in section 2.1. Results are presented on the spread of BXW on both the farms of

farmers who participated in the study and those in the study region. Then awareness of

genetically modified crops is discussed as well as their potential benefits, adoption and

consumption.

3.1. BXW awareness, spread and impact on banana production

The awareness of BXW among the participating farmers dates as back as 2001, but the

majority of them became aware it as recent as 2010 (Figure 1). Although farmers in the

sample first heard about BXW around 2001, the first occurrence of BXW on the farmers’

banana plantations in study areas generally started in 2005 with majority of them

experiencing it in 2010 through 2011. The impact of BXW on banana production in the study

areas is devastating. The average banana production losses due to BXW are highest in DRC

(83%) and Uganda (71%), and ranging from 39 – 51% in other countries (Figure 2). These

figures were obtained as averages from farmers, extension agents and key informants.

These losses are within the same range reported in earlier studies in the region. For

example, Karamura et al. (2010) estimated 65 – 80% banana production loss due to BXW in

Uganda. The implication is that the existing control methods of BXW appear somewhat

ineffective in combating the effects of BXW. In Uganda, for example, the first occurrence of

BXW was observed in 2001, but production losses have remained high despite the

application of control methods and wide mass media campaign by NARO. Results in Table 2

show that more than half of the farmers (54%) used a combination of control methods

including removing male buds and infected plants, using sterilized tools, while only 3% of

them avoided introduction of suckers from unknown locations. This potentially leads to a vicious

cycle of BXW occurrence on farmers’ fields – in the sense that after removing all infected plants or

mats, the source of new planting materials remains a potential source of BXW.

12

Figure 1. Farmers’ awareness of BXW on target countries

Figure 2. Banana production loss due to BXW incidence

13

3.2. Awareness of genetically modified crops and development of GMB-BXW

While GM crops have been widely grown for more than two decades and considered to be

safe, it was necessary to gain an understanding of the stakeholders’ perception about GM

banana in the target countries. Key informants, extension agents and banana traders were

asked to define what GM banana is. The focus was on these key stakeholders because they

have the capacity to cause a behavioral change among farmers through capacity building

(change agents) and increased demand (market agents). Of the respondents who thought

they knew the meaning of a GM banana, about 35% understood it as the “banana which has

been bred to resist pathogens that cause diseases”, and 41% said it is “an improved banana

which has been developed with different genes from other sources”, while others, largely

traders, defined the GM banana as “a variety with different attributes from the traditional

varieties”. Some of the attributes mentioned included nice looking banana figures, not tasty

and have long term health effects on humans. These definitions reflect a knowledge gap in

information sharing between research scientists in the breeding profession and the end

users of their products. This suggests that successful adoption of GM crops, may partly, be

driven by the breeders’ effort to develop and disseminate simple-user friendly manuals

describing the process of the GM crop development and its biosafety effects.

Regarding the awareness of the research development on GMB-BXW, nearly a third

of the respondents had heard about GMB-BXW through seminars and fellow friends. To

enrich the understanding of the respondents about GM crops, an explanation was done for

each respondent about the process of developing GMB-BXW. The majority of respondents

(85%) supported the development, while 4% where undecided and 13% did not support the

development. However, majority of them (95%) reported a number of potential benefits

associated with the development of GMB-BXW including: increased banana yields and

income, stable and continued supply, and increased banana consumption.

Majority of the respondents (60%) did not perceive any potential disadvantages

associated with GMB-BXW development, but 40% reported otherwise. The most important

potential disadvantages included: the GM banana may lead to outbreak of new banana

diseases through mutation, which may lead to major losses of local varieties. This was

followed by the costly establishment of new banana plantations. This is likely to slow down

early adoption of GM banana by farmers who have already experienced significant income

14

losses due to BXW and may be unable to establish new plantations or even afford the likely

expensive BXW-resistant banana planting materials. The perception that GMOs are not

good for human health was frequently mentioned. Other anticipated disadvantages

included undesirable attributes, which may affect GM banana market demand negatively,

and the failure of GM banana to adapt to local environmental conditions, which may lead to

loss of resistance to BXW within a short-term time. For successful adoption of GM banana

resistant to BXW, all these potential disadvantages need to be addressed during its

dissemination.

15

Table 2. Control methods of BXW and awareness of genetically modified crops% of farmers using control methods for BXW (N=37)

Removing the male bud 10.8Removing infected plants 21.6

Uprooting whole infected mat 18.9Avoid introduction of suckers from unknown locations 2.7Combination of control methods 54.1

% aware of the development of BXW resistant banana varieties (N=111) 27.9% of respondents aware of the meaning of GM banana (N=75) 36.0% defining GM banana as … (N=29)

a banana which has been bred to resist diseases 34.48an improved banana with integrated gene(s) from other sources 41.38a banana variety with different properties from local varieties (good eye appealbut tasteless, has long term health effects)

24.14

% of respondents’ opinion on the development of GMB-BXW (N=112)Support the development 83.0Indifferent 3.6Do not support the development 13.4

% reporting potential benefits of GM banana resistant to BXW (N=111) 94.59Perceived potential benefits (N=217)

Increase in yields 28.11Increase in income 38.25Credit to research institutions 3.69Stable supply 12.44Increase in banana consumption 17.51

% reporting potential disadvantages of GM banana resistant to BXW (N=110) 40.00Perceived potential disadvantages (N=54)

High cost of establishing new plantations 24.07Undesirable consumption attributes 14.81Outbreak of new diseases and loss of local varieties 27.78Health problem concerns 24.07Failure of GM banana to adapt to local conditions 9.26

Note: in this table, the number of observations is the sum all respondents in the study (farmers, traders,extension agents and key informants). The number of observations varies for some questions due to missingresponses.

3.3 Potential adoption of GMB-BXW

Figure 3 reports the potential adoption of GMB-BXW among the sample farmers. The

majority of the farmers (65%) reported immediate adoption upon release of GMB-BXW in

order to limit the spread of BXW, restore the destroyed plantations, and improve declining

crop income through increased banana yields (Table 3). Other farmers (19%) indicated that

16

they would delay adoption in order to first learn about the effects and performance of

GMB-BXW. Interestingly, the majority of the farmers (90%) were willing to pay for GMB-

BXW planting materials if they decided to adopt it. The minimum they would be willing to

pay per banana plantlet was about US$ 71 cents and a maximum of about US$ 1.4 per

plantlet. The current market price is US$ 80 cents on average. About 2 acres, on average,

would be allocated to GMB-BXW for cultivation. Part of this land would come from cutting

down existing banana plantation already affected as reported by 27% of the farmers, while

the majority of them (73%) reported that they would establish new plantations by land re-

allocation from other crops.

Figure 3. Willingness to adopt GMB-BXW

Table 3. Willingness to pay and reasons for adoption of GMB-BXWReasons for early adoption of GMB-BXW (N=33)

To limit the spread of BXW 36.4To restore plantations destroyed by BXW 36.4To improve yields and income 24.2To be among early adopters 3.0

Reasons for late adoption of GMB-BXW (N=9)To limit the spread of BXW 22.2Need to first learn from early adopters 66.7Negative attitude on improved varieties 11.1

% of farmers willingness to pay GMB-BXW (N=31) 90.3Minimum price farmer is willing to pay (US$) 0.71Maximum price farmer is willing to pay (US$) 1.43Area expected to be allocated to GMB-BXW (acres) 2.2% reporting to cut down existing banana plantation and replace GMB-BXW 26.9

17

In addition to farmers, extension agents and key informants were asked to estimate

potential adoption rate of GMB-BXW based on their experience with the adoption of

improved banana varieties. The survey results show that if GMB-BXW is released in 2020,

the minimum adoption rates are impressively high but vary across countries ranging from

21% in Kenya to 70% in DRC (Table 5). The maximum adoption rates of up to 100% would be

attained in 2 to 10 years from the time of adoption. Figure 4 reports an adoption path for

each country predicted over 25 years using analytical approach suggested by Alston et al.

(1998). The first segment of zero adoption indicates field trials and biosafety tests of GMB-

BXW from 2013 up to 2020, expected year of release.

Figure 4. Projected adoption rate of GM BXW resistant banana

18

3.4. Consumption preferences: local versus improved banana varieties

The adoption of new varieties is largely, among other factors, driven by the

consumer demand – both at the beginning of the supply chain (farmers) and at the

end of supply chain (net consumers). The potential consumption of GMB-BXW was

tested against existing experience of both farmers and traders with improved

varieties. Figure 5 shows that 41% of the sample farmers prefer local banana

varieties to improved ones because the former are tastier and easier to cook than

the latter. On the flip side, for the same reasons, more than a third of the farmers

(35%) preferred improved banana varieties to local ones, while nearly a quarter of

them find no difference between improved and local varieties in terms of taste and

cooking time (Table 4).

The key question however is: what attributes do consumers look for when

purchasing bananas from either farmers or traders? Both farmers and traders

reported that more than half (60%) of their banana buyers ask about the type of

the banana variety before purchase, while 40% of them do not. However, both

traders and farmers stressed that the type of banana (local or improved) does not

matter as much in influencing the decision to buy, but the size of the bunch and its

fingers, the quality and taste matter a lot (Table 4). These attributes are, thus,

important to consider when developing new banana varieties.

19

35%

24%

41% Improved varieties

Indifferent

Local varieties

Farmers' variety preferences (N=37)

40%

60%

No

Yes

Consumers ask about variety type (N=84)

Figure 5. Left pie chart shows farmers’ consumption preference of banana varieties. Right pie chartshows consumers’ inquiry about banana before purchase from either farmers or traders.

Table 4. Preferences and consumption attributes between improved and local banana varieties

Reasons for preferring local varieties to improved ones (N=19)Local varieties are tasty 84.2Local varieties cook fast 15.8

Reasons for being indifferent between consuming local and improved varieties (N=9)Both improved and local varieties are tasty 55.6Both varieties cook fast 44.4

Attributes considered during purchase of bananas (N=146)Quality and taste of banana 37.0Size of the bunch and fingers 50.7Banana prices 5.5Good eye appeal 5.5Long shelf life 1.4

3.5. Potential consumers of GM banana resistant to BXW

Against the background in preceding section, all respondents in the survey were asked to

identify the potential consumers of GMB-BXW. More than half (56%) of the respondents

perceive all consumers not to select against GMB-BXW because of the high market demand

and limited choices, and that consumers are more concerned about consumption attributes

20

than the type of the variety (Figures 6 and 7). Comparing urban and rural (farmers)

consumers, the former would be the main consumers (24% against 13%) as the price of the

banana is more important to them than the banana variety. The minority (7%) who would

not consume GMB-BXW, they would be limited to consumption of local varieties due to

perceived negative attitudes toward GM crops. These results further underscore the

importance of preserving or enhancing quality and taste of banana when breeding for

disease and pest resistance.

Figure 6. Potential consumers of GM banana (N=101)

21

Figure 7. Frequency of reasons for consumer preference of GM banana (%)

3.6. Geographical areas and banana varieties for potential intervention with GMB-BXW

The target areas for potential intervention in each country were prioritized by key

informants. The prioritization was based on BXW incidence levels and the contribution of

the area’s share to national banana production. The following regions or districts in each

country were prioritized (Appendix A1): Uganda (South West, Central and Eastern regions),

Kenya (Siaya, Busia, Bungoma, and Vihiga), Tanzania (Kagera, Kigoma, Tarime, Ukerewe), DR

Congo (South-Kivu, North-Kivu, and Oriental Provinces), Burundi (Kirimiro, Buyenzi, Bututsi,

Bweru, Bugesera, Buragane, Buyogoma, Imbo, Moso, and Bweru).

For the regions identified as priority for dissemination of GMB-BXW, we asked the

key informants to prioritize banana varieties for transformation based on the area

cultivated under given variety relative to the total banana area in the country (Appendix

A1). Interestingly, a number of identified banana varieties are similar to the farmer-

preferred varieties (Appendix A2). The top three varieties prioritized by both key informants

and farmers in each country are: Uganda (Nakitembe, Kayinja and Kisansa), Kenya

22

(Cavendish cultivars, Ekegusi and Ngombe), Tanzania (Enshakara, Enyoya and FHIA 17),

Burundi (Igisahira, Igitsiri and Kayinja), DR Congo (Nshikazi, Nguma and Kamaramasengi),

and Rwanda (Injagi, Intuntu and FHIA 17).

3.7. Economic evaluation of GMB-BXW development

Although few studies have evaluated the economic impacts of BXW on banana production

(Abele and Pillay 2007; Smale and Tushemereirwe 2007; Karamura et al. 2010; Kikulwe et al.

2011), hardly anyone has evaluated the potential adoption of GMB-BXW and its impact on

both the consumers and producers. Kalyebara et al. (2006) show that, in a period of 10 -15

years, banana farmers in Uganda would lose 5.6 US$ billion if BXW was not controlled and

3.1 US$ billion if controlled through removing the male bud and infected plants, using

sterilized farm tools, and uprooting whole infected mat. These control methods exclude

resistant banana cultivars to BXW. In a similar line, Abele and Pillay (2007) indicated that

uncontrolled BXW would benefit farmers in early years only and lose out in later stages of

BXW, whereas consumers would be worse-off: measured against the baseline, farmers

would gain 13 millions US$ due to increased prices by 50%, while consumers would incur

losses worth 160 million US$. Kayobyo et al. (2005) predicted that uncontrolled BXW

spreads at annual infection rate of 8% translating into an annual production loss of 2 million

tonnes. This report complements these studies by going step farther to estimate the

economic benefits of GMB-BXW based on the survey data collected from key stakeholders.

Table 5 reports the survey results and assumptions to measure ex-ante economic

benefits of GMB-BXW using the ESM described in section 2.2. The results show that farmers

expect to increase area allocated to banana cultivation by 10 – 50%, incur increased input

costs especially on acquiring GMB-BXW plantlets and establishing new plantations by 20 –

50%, and increase yields by 25 – 70% across target countries. The change in input costs also

accounts for labor cost savings from avoided practices to control the spread of BXW. A

question was asked if there would be reductions in labor hours used in management

practices to control for BXW, but farmers indicated this labor requirement would be

negligible in places where BXW incidences are low. This is because some of these

management practices such as removing suckers (sick plants) and male buds are part of the

routine activities.

23

Other variables included in ESM analysis are the minimum and ceiling adoption rates

and the time taken to attain the ceiling adoption. A uniform lag of 10 years was used across

all countries to attain the ceiling adoption levels from reported minimum adoption levels.

This is because the diffusion of agricultural technologies often takes relatively long time due

to resource limitations for dissemination. Secondly, the minimal variation may also be

attributed to the small sample of respondents.

Table 5. Parameter values used to estimate the economic benefits of GMB-BXW usingESM

Burundi DRC Kenya Rwanda Tanzania UgandaAverage loss of banana production due toBXW (%)

51 83 39 45.9 43.5 71.4

Expected initial adoption rate of BXWresistant banana (%)

30 70 20.9 48 24 35.3

Expected ceiling of adoption rate of BXWresistant banana (%)

65.5 100 60.4 70 77.5 73.8

Expected number of years to attainmaximum adoption

3 - 5 3 - 5 5 - 10 2 - 5 5 - 10 5 - 10

Average % increase in area allocated tobanana after adoption

23.1 50 30.6 26 10.3 40

Average % increase in input per ha afteradoption

27.5 50 28 35 20 33

Average % yield gain after adoption per ha 56.7 70 39.4 30 25 53.8% banana production from study areas withrespect to whole country

15 13 27 38 33 28

Production (‘000 tonnes) 1,184 832 1,424 3,220 3,260 9,770Banana area (ha) 178 361 61 349 727 1,830Price per ton (US$) 267 255 198 205 203 207Research costs (million US$) 3.2 3.2 3.2 3.2 3.2 3.2

To obtain economic benefits of GMB-BXW at national level also requires information

on the national banana cultivation area and production data. This information was obtained

from FAOSTAT (FAO, 2014). The farm gate banana prices were obtained from farmer survey

as an average price for each country. The total research costs of US$ 3,233,800 were based

on the GMB-BXW project proposal budget for activities from 2013 to 2020. The costs

include: laboratory and screen house testing and confined field trials and food and

environmental safety studies. The extension service costs during dissemination were

24

assumed at US$ 80 per hectare. The values for other parameters used in ESM were based

on available literature including, among others, Alston et al., 1998. These parameters

included price elasticity of demand (0.5%) and price elasticity of supply (1.0%). Table 6

reports the ESM results.

Table 6. Potential benefits of developing GMB-BXW in the GLA regionCountry Burundi DRC Kenya Rwanda Tanzania UgandaAdoption ceiling area (‘000 ha) 117 361 36 244 567 1,354NPV of gross consumer surplus (millions US$) 110 119 42 19 61 658NPV of gross producer surplus (millions US$) 55 60 21 9 31 329NPV of net total benefits (millions US$) 161 168 60 20 76 953Internal rate of return (%) 56.48 57.79 43.21 29.6 43.14 85.64Benefit – Cost ratio 33.85 17.13 20.71 3.62 6.11 30.05

Net Present Values (NPV) computed using a real interest rate of 10%.

Table 3 reports measures of the ex-ante economic impacts of GMB-BXW in the GLA. An ex-

ante analysis of a 25-year period, including research and extension costs, and assuming a

10% discount rate, the net present value (NPV) of the net benefits ranges from US$20 – 953

millions in the target countries. The benefits are highest in Uganda (US$ 953 millions), DRC

(US$ 168 millions) and Burundi (US$ 161 millions). This is not surprising given that these

same countries have, and are, still facing high yield losses due to BXW. This suggests that

investment in development of GMB-BXW is not only essential but also economically viable.

Certainly, returns on investments in research and extension are highest (56 – 86%) in these

three countries compared to other countries in the region (30 – 43%). Correspondingly, the

benefit-cost ratio estimates are high (17 – 34:1) in Uganda, Burundi and DRC compared to

the rest of the target countries (4 – 21:1), indicating that investments in research and

extension programs on GMB-BXW are feasible and have a great potential to generate a

stream of benefits in excess of costs. For each dollar invested in the development and

dissemination of GMB-BXW generates the highest amount of dollars in Burundi (US$ 34)

followed by Uganda (US$30), Kenya (US$ 21), DRC (US$ 17), Tanzania (US$ 6) and Rwanda

(US$ 4).

Across all countries, the development and production of GMB-BXW would benefit

consumers twice as much the benefits accruing to farmers. For example, in Burundi the

25

consumer surplus is US$ 110 million which is twice the producer surplus of US$ 55 million.

This distribution of benefits are similar to those of ex-ante analysis of GM ring spot virus–

resistant papaya in Thailand (Napasintuwong and Traxler, 2009) and Philippines (Yorobe

and Laude, 2007) that consumers benefit twice as much as farmers.

4. Conclusions

Banana production has been greatly affected by BXW in the Great Lakes region of Africa

(GLA). The existing management practices to control BXW only reduce the effects to a small

extent. If unchecked, BXW remains one of the biggest threats to banana production, which

provides food security and income to millions of farm households in the GLA. To overcome

the threatening effects of BXW, scientists have successfully developed a GM banana

resistant to BXW. Although GM crops have been developed and proved to be economically

beneficial in other countries, the development of GM crops in GLA is still in its infant stages.

On the one hand, the government policies legalizing the release and commercialization of

the GM crops are still under formulation although some countries have allowed research on

GM crops to take place. Only Kenya in GLA have a biosafety law, and a national Biotechnology

and Biosafety Policy in place whereas Uganda has national Biotechnology and Biosafety Policy

in place and the formulation of biosafey laws are in advanced stages of the policy cycle. The

discussion of biosafety laws and policies in other countries in the GLA region have not yet

gained enough momentum. On the other hand, studies to assess the ex-ante adoption and

potential benefits are still limited in the region. This study provides evidence on the

potential economic importance of yet to be released GM BXW-resistant banana in the GLA.

This study provides evidence based on data collected from farmers, extension agents and

key informants from Burundi, Democratic Republic of Congo (DRC), Kenya, Rwanda ,

Tanzania, and Uganda.

The results show that farmers would be willing to adopt the GM BXW-resistant

banana when released. The initial adoption rate ranges from 21 – 70% and would reach the

adoption ceiling of up to 100% in 2 – 10 years. The findings further show that if GM banana

is successfully adopted in the GLA, both consumers and producers would benefit. However,

the former would benefit as much as twice the latter due to price reduction from excess,

stable and continuous supply of banana. The magnitudes of benefits vary considerably

26

across countries. Largest benefits would accrue to consumers and farmers in countries that

have and are experiencing large production losses due to BXW. Uganda, Burundi and DRC

would receive the largest benefits ranging from 161 – 953 million US$ compared to 20 – 76

million US$ in the rest of the countries.

The study has implications for the research scientists and extension agents, as they

are frequently interested to know the important triggers of successful adoption and

consumption of new technologies. The results in this report not only demonstrate that

investment in the development of GM banana resistant to BXW is viable, but they also

provide a basis on which to focus research and extension efforts on the attributes that

would significantly enhance adoption of new banana varieties.

27

References

AATF [African Agricultural Technology Foundation]. 2009. Feasibility Study on Technologiesfor Improving Banana for Resistance Against Bacterial Wilt in Sub-Saharan Africa.Nairobi, Kenya: African Agricultural Technology Foundation.

Abele, S., and Pillay, M. 2007. Bacterial wilt and drought stresses in banana production andtheir impact on economic welfare in Uganda: Implications for banana research in EastAfrican highlands. Journal of Crop Improvement, 19(1-2): 173-191.

Alston, J. M., Norton, G., and Pardey, P. G., 1998. Science under scarcity: Principles andpractice for agricultural research evaluation and priority setting (2nd Edition).Wallingford, UK: CAB International.

FAO. 2014. Statistical database of the Food and Agricultural Organization of the UnitedNations, FAO. http://faostat.fao.org, accessed April, 2014.

Kalyebara, M.R., Ragama, P.E., Kagezi, G.H., Kubiriba, J., Bagamba, F., Nankinga, K.C.,Tushemereirwe, W.K. 2006. Economic importance of the banana bacterial wilt inUganda. African Crop Science Journal, 14: 93–103.

Karamura, E., Kayobyo, G., Tushmereirwe, W., Benin, S., Blomme, G., Eden Green, S.,Markham, R. 2010. Assessing the impacts of banana bacterial wilt disease on banana(Musa spp.) productivity and livelihoods of Ugandan farm households. In: Dubois, T.,Hauser, S., Staver, C., Coyne, D. (eds) Proceedings of the International Conference onBanana and Plantain in Africa: Harnessing International Partnerships to increaseResearch Impact, 5-9 October, 2008, Mombasa, Kenya. Acta Horticulture, 879: 749-756.

Kayobyo, G., Aliguma, L., Omiat, G., Mugisha, J., and Benin, S. 2005. Impact of BXW onhousehold livelihoods in Uganda. Assessing the impact of the banana bacterial wilt(Xanthomonas campestris pv. musacearum) on household livelihoods in East Africa,workshop held on Dec. 20, 2005, Kampala, Uganda.

Kikulwe, E. M., Birol, E., and Wesseler, J. and Falck-Zepeda, J. 2011. A latent class approachto investigating demand for genetically modified banana in Uganda. AgriculturalEconomics, 42: 547–560.

Namukwaya B., Tripathi L., Tripathi J.N., Arinaitwe G., Mukasa S.B., Tushemereirwe W.K.(2012) Transgenic banana expressing Pflp gene confers enhanced resistance toXanthomonas Wilt Disease. Transgenic Research 12: 855-865.

Napasintuwong, O. and Traxler, G. 2009. Ex-ante impact assessment of GM papaya adoptionin Thailand. AgBioForum, 12(2): 209-217.

Smale, M. and Tushemereirwe, W.K. (Eds.), 2007. An economic assessment of bananagenetic improvement and innovation in the lake Victoria region of Uganda andTanzania. Research Report No. 155. International Food Policy Research Institute,Washington, D.C.

28

Tripathi L., Mwaka H., Tripathi J.N., Tushemereirwe W.K. (2010) Expression of sweet pepperhrap gene in banana enhances resistance to Xanthomonas campestris pv. musacearum.Molecular Plant Pathology, 11: 721-731.

Tripathi, L., Mwangi, M., Abele, S., Aritua, V., Tushemereirwe, W.K. and Bandyopadhyay, R.2009. Xanthomonas wilt a threat to banana production in East and Central Africa. PlantDisease 93(5), 440 – 451.

Tripathi L., Tripathi J.N., Tushemereirwe W.K., Arinaitwe G., Kiggundu A. (2013) Transgenicbananas with enhanced resistance against Xanthomonas wilt disease. Acta Horticulture,974: 81-90.

Yorobe, J. M. Jr., and Laude, T. P., 2007. Costs and benefits of PRSV-resistant technology inthe Philippines (ISAAA and Agricultural Biotechnology Support Program II [ABSP-II]Impact Study). Manila, The Philippines: International Service

29

Appendix A.

Appendix A1. Prioritization bannana varieties and regions for intervention with GMB-BXWCounty Cooking/

Plantain% areacoveragerelativeto totalbananaarea

Beer % areacoveragerelativeto totalbananaarea

Dessert % areacoveragerelativeto totalbananaarea

Priority areas

UgandaNakitembe 12 Kayinja 9 Gros Michel 4 South West,

Central, EasternKisansa 8Mpologoma 5M9 0.5

KenyaEkegusi 13 Cavendish

cultivars30 Siaya, Busia,

Bungoma, VihigaNgombe 12 Gros Michel 7Other AAA-EA types 10 Sukali Ndiizi 3Nusu-Ngombe 5 FHIA 17 2Sialamule 2 FHIA 1 1

TanzaniaEnshakara 20 FHIA 23 10 FHIA 17 10 Kagera, Kigoma,

Tarime, UkereweEnyoya 15 Nay Poovan 8Enshansha 8 Gros Michel 8Entobe 8Mshare 8Plantain (Enkonjwa) 5

DR CongoNguma 25 Nshikazi 42 Kamaramasengi 14.9 South-Kivu, North-

Kivu, OrientalProvinces

Vulambya/Barabesha 9.3 Ndundu/Tuntu

8.3 Gros Michel 9.1

Musheba 8.3 Kisubi 7.5 Semi-dwarfcavendish

8.1

Yangambi 5.7Burundi

Igisahira 25 Kayinja 15 FHIAs ( 17,21, 25) 12 Kirimiro, Buyenzi,Bututsi,Bweru,Bugesera,Buragane,Buyogoma, Imbo,Moso, Bweru

Igitsiri 25 FIA 25 Yangambi Km5 8Incakara 5 Kamaramasenge 6Ikigurube 1 Gros Michel 1Sohokunkorere 1Mbwazirume 1

Rwanda **Barabeshya Mazizi Apple bananaInjagi Umukora Gros MichelIngaju Intutu CavendishIncakara Kayinja

**Data were not available for Rwanda, the varieties reported are the preferred varieties according to AATF(2009)

30

Table A2. Frequency of banana varieties of farmers’ fieldsUganda N=70 Kenya N=40 Rwanda N=34Entaragaza 10.0 Kampala 12.5 Injagi 17.65Mbwazilume 8.6 FHIA 17 7.5 Intuntu 14.71Nakabululu 5.7 Cavendish 7.5 FHIA 17 11.76Bogoya/Mbogoya 5.7 Bogoya/Mbogoya 5 Inkazikamumwa 11.76Nakitembe 4.3 Mulure/Mutule 5 Wakubimba/Mpologoma 8.82Nsakala/Musakala/Mushakara 4.3 Kiganda 5 FHIA 25 5.88Kibuzi 4.3 Ngombe 5 Kamara 5.88Enyeru 4.3 Golden finger 5 Intutsi 5.88Kufuba 4.3 Naombe 5 Gonja/Plantain 2.94Nabusa 2.9 Sirembe-2 5 Igisubi/Ibisubi 2.94Musa 2.9 FHIA 23 2.5 Gros Michel 2.94Gonja/Plantain 2.9 GT 2.5 Kayinja/Mugirwozi 2.94Fiya 2.9 Munaru 2.5 Inyumba 2.94Mujuba 2.9 Valene 2.5 Intokatoke 2.94Entukura 2.9 Nyaluo 2.5 Burundi N=34Mulure/Mutule 2.9 Udhigo 2.5 Igisahira 26.09Nasara/Nansara/Nakati 2.9 Namayeye 2.5 Mugomozi 17.39Muvubo 1.4 Sirembe-1 2.5 FHIA 17 13.04Kabula 1.4 Muhirak 2.5 Igitsiri/Igitsiru 8.7Nsowe 1.4 Gerogero 2.5 Igisubi/Ibisubi 8.7Kisubi 1.4 Ngolo 2.5 FHIA 25 8.7Ndizi/Sukari Ndizi 1.4 Israel 2.5 Ikingurube 4.35Kivuvu 1.4 Tissue-culture 2.5 FHIA 23 4.35Enjagata 1.4 FIA 18820 2.5 FHIA 3 4.35

Rwakashita 1.4East African highlandbanana 2.5 KM5/Yagambi 4.35

Mpologoma 1.4 Tanzania N=27 DRC N=19

Sira 1.4 Nshakara 51.85East African highlandbanana 36.84

Lisindalo 1.4 Nyoya 18.52 Musa/Camera 26.32Mbalala 1.4 Nchoncho/Enchoncho 18.52 Gonja/Plantain 21.05Mudwale 1.4 Ntobe/Entobe 7.41 Gros Michel 15.79Nabuholi 1.4 Nshansha/Enshansha 3.7Liyase/Limwa 1.4Kamakako/Likago 1.4Namboko/Wamboko 1.4

The number of observations (N) represents the frequency of responses by farmers not the number ofrespondents.