MR 33 - Kenya Impact Assessment: Baseline Research Report

Assessing the Impact of Kenya BDS
and
Horticulture Development Center
Projects in the Tree Fruit Value
Chain in Kenya
Baseline Research Report
microREPORT #33
JULY 2005
This publication was produced for review by the United States Agency for International Development. It was prepared by
Don Snodgrass and Jennefer Sebstad with Action for Enterprise under the Accelerated Microenterprise Advancement Project
(AMAP) Business Development Services (BDS) Kenya STTA Task Order. Action for Enterprise is a subcontractor to
ACDI/VOCA under the AMAP BDS IQC.
DISCLAIMER
The author’s views expressed in this publication do not necessarily reflect the views of the United States Agency for Interna-
tional Development or the United States Government.
Accelerated Microenterprise Advancement Pro-
ject (AMAP) is a four-year contracting facility that
USAID/Washington and Missions can use to acquire
technical services to design, implement, or evaluate
microenterprise development, which is an important
tool for economic growth and poverty alleviation.
For more information on AMAP and related publica-

tions, please visit www.microLINKS.org.
Accelerated Microenterprise Advancement Pro-

ject
Contract Number: GEG-I-00-02-00016-00
Task Order: Kenya STTA
Task Order Number: GEG-I-800-02-0016-00
Contractor: ACDI/VOCA
Olaf Kula, Program Manager
Tel: (202) 879-0213
E-mail: OKula@acdivoca.org
Donald Snodgrass, Ph.D., Institute Fellow Emerit-

us at Harvard University, is a scholar of economic de-
velopment and an advisor in the areas of small en-
terprise development, microfinance, education and
human resource development, and poverty and eco-
nomic growth in developing countries.
Jennefer Sebstad is a consultant working on devel-

opment programs that focus on expanding income,
employment, and asset building opportunities for
poor people in developing countries.
Action for Enterprise (AFE) is a non-profit organiz-

ation dedicated to private sector/enterprise develop-
ment based in Arlington, Virginia.
1
ACDI/VOCA is a private, non-profit international de-
velopment organization based in Washington, DC.
2
Assessing the Impact of Kenya
BDS and Horticulture Develop-
ment Center Projects in the
Tree Fruit Value Chain in Kenya
Baseline Research Report

microREPORT #33
Don Snodgrass and Jennefer Sebstad
2005
3
CONTENTS
LIST OF TABLES AND FIGURES..............................5
LIST OF ACRONYMS AND ABBREVIATIONS.........7
EXECUTIVE SUMMARY............................................8
I. INTRODUCTIONS DEVELOPMENT ...................12

A. Main Features of Program Environment..........12
B. Program Descriptions......................................17
II. DESIGN OF THE IMPACT..................................22

A. Key Questions.................................................22
B. The Causal Model............................................22
C. Hypothesis......................................................22
D. Framework of Analysis....................................23
E. Data Collection Strategy.................................26
III. BASELINE RESEARCH FINDINGS...................30

A. Tree Fruit Value Chains....................................30
B. Smallholder Producers of Avocados, Mangos,
and Passion Fruit.................................................43
IV. CONCLUSIONS OF THE BASELINE STUDY AND

IMPLICATIONS FOR ROUND TWO OF IMPACT
ASSESSMENT.....................................................62
REFERENCES..........................................................63
ANNEX A. DESCRIPTION OF KENYA BDS AND

FINTRAC HDC PROJECT ACTIVITIES ................65
ANNEX B. ISSUES TO FOLLOW UP IN ROUND

TWO....................................................................69
ANNEX C. CALCULATION OF THE ASSET SCORE

GROUPINGS.......................................................71
4
LIST OF TABLES AND FIGURES
TABLES
Table 1: Value of Horticulture Exports

Table 2: Baseline Sample Size
Table 3: Geographic Location of Sample
Table 4: Export Statistics for Avocado, Mango, and Passion Fruit,
Kenya, 2003
Table 5: Survey Respondents by Location
Table 6: Number of Trees per SME by Site: Average and Distribu-
tion
Table 7: Pieces of Avocado and Mango Harvested in Past Year: Average
and Distribution
Table 8: Kilograms of Passion Fruit Harvested in Past Year: Average
and Distribution.
Table 9: Pieces of Avocado and Mango Harvested in Past Year per
Producing Tree/Vine
Table 10: Kilograms of Passion Fruit Harvested in Past Year per Producing
Tree/Vine
Table 11: Earnings from Sales of Fruit in Past Year (Kenya shillings)
Table 12: Sales by Type of Customer and Intervention
Table 13: Percentage of Contract Sales
Table 14: Distribution of Respondents by Amount of Hired Labor Used
Table 15: Changes in Cultivation Methods and Planting of Fruit Trees
Table 16: Average Expenditure on Fertilizer Used on Targeted Fruit Trees
in Past Year
Table 17: Average Expenditure on Sprays Used on Targeted Fruit Trees in
Past Year
Table 18: Percentage of Respondents who are Members of a Producer
Group
Table 19: Selected Enterprise Data by Asset Score Group (Entire Sample)
Table 20: Average Number of Household Members, Earning Members, and
Earner/Dependent Ratios by Intervention
Table 21: Respondent Households by Children in School and Inter-
vention
Table 22: Respondent Households by Members with Salaried Employment
and
Intervention
Table 23: Percent Distribution of households by number of income
sources
Table 24: Average Number of Household Income Sources by Inter-
vention
Table 25: Sources of household income (Total sample)
Table 26: Percent of Households Ranking Tree Fruit Income #1 or
#2
5
Table 27: Proportion of Household Income from Tree Fruit (Estim-
ated)
Table 28: Proportion of Household Income from Tree Fruit (Estim-
ated) by Intervention
Table 29: Average Monthly Consumption Expenditure per Capita
by Intervention (Ksh)
Table 30: Distribution of Respondent Households by Monthly Consumption
Expenditure per Capita
Table 31: Distribution of Respondent Households by Asset Score
Table 32: Distribution of Respondent Households by Asset Score
and Intervention
Table 33: Distribution of Respondent Households by Size of Land-
holding
Table 34: Distribution of Respondent Households by Size of Land-
holding by Intervention
Table 35: Percentage of Woman-headed Households in Study
Sample
Table 36: Selected Household Data by Gender of Household Head
Table 37: Selected household data by asset score group (total
sample)
FIGURES
Figure 1: Activities under BDS Tree Fruit Project

Figure 2: Activities under HDC Tree Fruit Project
Figure 3: Causal Model for Kenya BDS and Fintrac HDC Projects
Figure 4: Framework for Studying Impacts
Figure 5: Kenya Tree Fruit Value Chain.
Figure 6: Governance Continuum
6
LIST OF ACRONYMS AND ABBREVIATIONS
AGOA African Growth and Opportunity Act
BDS Business Development Services
EAGA East Africa Growers’ Association
EPZ Export Processing Zone
EU European Union
EUREPGAP Euro-Retailers Produce Good Agricultural Practices
FPEAK Fresh Produce Exporters Association of Kenya
GOK Government of Kenya
HCDA Horticultural Crop Development Authority
HDC Horticulture Development Centre (Fintrac)
HIV/AIDS Human Immunodeficiency Virus/Autoimmune Deficiency Syndrome
IR Intermediate Result
KACE Kenya Agricultural Commodities Exchange
KADI Kamurugu Agricultural Development Initiatives
KARI Kenya Agricultural Research Institute
KHE Kenya Horticulture Exporters
Ksh Kenya Shilling
KWETU Swahili for “our home”, name of a local non-governmental organization
MOA Ministry of Agriculture
MOU Memorandum of Understanding
MSEs Micro and Small-Scale Enterprises
NGO Non-Governmental Organization
SITE Strengthening Informal Sector Training and Enterprise
SMS Short Message Service
SO Strategic Objective
USAID United States Agency for International Development
7
XECUTIVE SUMMARY
THE PROJECTS AND THE the project issued tenders and holder associations, and two
IMPACT ASSESSMENT awarded contracts to eight small businesses producing
private sector and NGO part- plant stock. In the future, it in-
This study assesses1 the im- ners. The contracts were de- tends to work through input
pacts of two USAID/Kenya fun- signed to facilitate the devel- suppliers as well. This five-
ded projects that focus on de- opment of sustainable solu- year project began in late
veloping sustainable solutions tions/services that provide ma- 2003 and was in its first year
to constraints facing busi- terial inputs (agro-chemicals of operation at the time of the
nesses in targeted industries and seed varieties), appropri- baseline survey.
and the degree to which these ate technology to upgrade
solutions impact: products and production pro- The study is longitudinal, with
 the competitiveness of the cesses, business and skills a baseline study including
mango, passion fruit, and training, and extension and in- both quantitative and qualitat-
avocado value chains; formation services. Several ive research. This will be fol-
contracts promote market link- lowed up in two years with a
 the integration of micro
ages between smallholder pro- resurvey of the same respond-
and small enterprises
ducers and lead firm exporters ents as well as further qualitat-
(farmers and others) into
through supply contracts and ive research. The major find-
these value chains so that
lead firm provision of embed- ings on the impact of the two
they contribute to and be-
ded services, and encourage projects will emerge after this
nefit from the tree fruit in-
inter-firm cooperation through second stage of research.
dustry’s increased com-
organization of producer
petitiveness
groups and provision of em- This report presents the find-
bedded services. ings of a baseline study that
The projects are the Kenya
featured a survey of 1,947
Business Development Ser-
The HDC project focuses on a smallholder farmers who grow
vices (BDS) project implemen-
wide range of horticulture avocado, mangos, or passion
ted by the Emerging Markets
products. We study only their fruit in Central, Eastern, and
Group and Fintrac’s Horticul-
passion fruit work, which in- Rift Valley provinces. The
ture Development Centre
cludes plans to: (1) introduce sample included farmers who
(HDC) project. Both projects
new varieties of passion fruit are participating in the two
support USAID/Kenya’s stra-
for fresh export; (2) improve projects as well as a control
tegic objective of increasing
agricultural practices of local group of non-participants. The
rural household incomes in
producers; (3) expand local survey was complemented by
Kenya (SO 7). They seek to
processing capabilities for loc- qualitative research (in-depth
raise smallholder productivity,
al market products; and (4) interviews and focus group
widen market outlets, facilitate
strengthen the farm-to-market discussions) with over 60 indi-
vertical and horizontal link-
value chain, inclusive of busi- viduals involved in the tree
ages, and promote the sus-
ness services to small farmers. fruit value chain, including
tainable development of busi-
Unlike Kenya BDS, the HDC farmers, farmer producer
ness services for rural MSEs
project does not operate group leaders, input suppliers,
through contracts but carries extension workers, brokers,
Kenya BDS, a five-year project,
out activities directly through exporters, and Kenya BDS and
started in 2002 and worked
project staff based in Nairobi Fintrac project directors and
initially on the tree fruit value
and agronomists based in four staff.
chain. In its first two years,
field offices. It works with and
1
“Value chain” and “sub-sector” are used through cooperating partners,
synonymously throughout this report. including KARI, existing small-
8
ASSESSING THE IMPACT OF KENYA BDS AND HDC PROJECTS IN THE TREE FRUIT VALUE CHAIN: BASELINE RESEARCH REPORT
The study design is based on a Hypotheses 3: Greater integ- local supermarkets, and ex-
causal model of impact that ration of smallholder MSEs into porters. Exporters buy fruit
shows how project facilitation productive value chains con- produced by smallholders and
activities to promote sustain- tributes to improved competit- medium-scale farmers and
able solutions can address iveness and growth of the tar- also produce some fruit on
constraints to smallholder par- geted value chains. their own plantations. Export-
ticipation and the competitive- ers have just recently begun to
ness of the tree fruit value buy directly from producer
chain2. These activities in turn FINDINGS OF THE groups and to provide embed-
lead to sustained access to the BASELINE STUDY ded services to smallholders
solutions, smallholder upgrad- through these groups.
ing, increased smallholder MSE
Tree Fruit Market
profits from tree fruit activit- Constraints to smallholder par-
Avocados, mangos, and pas-
ies, increased rural household ticipation in the tree fruit value
sion fruit are among the most
incomes, and overall sector chain include:
common fruit crops in Kenya.
growth and competitiveness  Lack of information and
Most of the fruit produced is
within the value chain. knowledge of the markets
sold in the domestic market,
but all three fruits are import-  Limited access to inputs
The study tests three hypo-  Limited access to re-
ant and growing export crops.
theses about the impact of sources for, and/or weak
In Europe, Kenyan fruit has a
donor interventions in opening incentives for, upgrading
competitive advantage based
up opportunities for smallhold- Weak vertical and hori-
not on volume, quality, or 
er MSEs in local, regional, and zontal linkages within the
price, but rather on seasonal-
global markets and in improv- value chain
ity. Avocados, mangos, and
ing the competitiveness of the
passion fruit each have a ‘win-  Lack of trust among pro-
overall value chain:
dow’ when these crops are ducers, brokers, and ex-
less available from other sup- porters
Hypothesis 1: Project activit-
pliers. Kenya is also better set
ies to promote sustainable
up to meet certification stand- Governance in the tree fruit
solutions in the tree fruit value
ards than other countries. value chain is characterized by
chain contribute to better in-
a mix of market and network
tegration of smallholder MSEs
The main marketing outlets for relationships (see page ___ for
into the value chain.
tree fruit producers are traders more detail). Smallholders
and brokers, who in turn sell to have traditionally sold their
Hypothesis 2: Better integra-
both domestic and export mar- fruit to brokers on a spot basis;
tion of smallholder MSEs into
kets. In the domestic market, contractual relationships have
the tree fruit value chain con-
they sell to wholesalers, fresh been marked by distrust. With
tributes to enterprise upgrad-
fruit retailers, and small retail increasing concentration
ing, improved performance,
shops. Producers also sell dir- among European buyers and
and enhanced household well-
ectly to fresh fruit retailers and rising standards in end mar-
being.
the Horticulture Crop Develop- kets (especially Europe, but
ment Authority (HCDA). also in other international, re-
2 Brokers and traders are the gional, and domestic markets),
Sustainable solutions here refer to more the power of the retailing
main conduits for smallholders
than business services from third party
to formal and informal pro- groups to impose governance
providers - they also include sustainable
access to markets, business relationships, cessing plants and to export- rules on the value chain is in-
TA provided in an embedded fashion from ers who buy tree fruits. Medi- creasing. Horizontal linkages in
one firm to another, improved business en- um-scale growers often link the form of farmers’ associ-
vironment, capacity of industry represent- directly to processing plants, ations exist but need strength-
atives to influence policy, etc.
9
ening. The horticulture value accepted for farmers selling household income sources
chain has had limited govern- avocados to EAGA under the suggest an active working
ment involvement and private Kenya BDS project and those population among respond-
firms have generally been left selling passion fruit in the Fin- ents. There are no major differ-
free to organize the trade. This trac areas; remaining groups ences in earner-dependent ra-
differs from the pattern that sold their fruit predominantly tios between men and women
characterizes some other com- in spot markets. headed households or by
modity value chains in Kenya wealth level (as indicated by
-- for example, coffee, tea, and Hired labor was used fairly ex- asset scores), suggesting that
pyrethrum, for which official tensively by richer farmers, this may not be a major de-
marketing boards still control while poorer farmers relied terminant of vulnerability for
procurement and prices. As primarily on family labor. Wo- households in the sample.
producer groups form to link to man-managed farms tended to
inputs and markets, and as ex- hire more labor than compar- The asset scores and con-
porters form associations, the able farms managed by men. sumption expenditure data
patterns are shifting more to- show a significant number of
ward network relationships. Producer group membership poor households in the
was almost universal among sample, in both the participant
Tree Fruit Enterprises program participants, both and control groups. This sug-
We surveyed five interventions male and female. Moreover, gests the projects are in-
intended to promote upgrad- nearly all of the farmers who volving poor households and,
ing and raise productivity and belonged to producer groups thus, have potential for direct
income from tree fruit among characterized them as either impact on their income from
smallholder producers of avo- very or fairly useful. tree fruits. The sample also in-
cado, mango, and passion cludes non-poor households,
fruit. The MSEs included in the Few farmers had access to ir- which should provide a good
survey cultivated varying num- rigation and less than one-half basis for comparing impacts
bers of trees/vines, with avo- purchased fertilizer for use on across poverty groups at the
cado holdings the smallest on their fruit trees. A larger num- end line.
average and passion fruit the ber said they had bought
largest. For each fruit, the pesticide or fungicide sprays. Households are quite diversi-
range of holding sizes was fied in their sources of income
wide. With one exception, pro- Considerable numbers of re- and tree fruits are an import-
duction and productivity were spondents had instituted im- ant source. While these figures
higher for program parti- proved cultivation or market- may reflect an upward bias in
cipants than for controls, dif- ing methods in the past two some respondents who associ-
ferences that may reflect se- years. Large numbers in some ated the study with the tree
lection bias and/or early im- areas had planted fruit trees in fruit projects, it suggests the
pacts of program participation. the past year. Farmers looked importance of relatively small
Between the two passion fruit to a wide range of sources for amounts of cash income for
sites, Fintrac works with larger useful technical advice, in- rural households.
farmers. formation, or training.
Gender differences in the divi-
Nearly all the farms surveyed Tree Fruit Households sion of labor related to tree
sell tree fruit, primarily Household size in the sample fruit production, the control of
through traders of different is large relative to the total tree fruit income, and access
sorts, but most earn only small population, but about average to productive resources are
amounts from these sales. for poor rural households. The likely to play out in the impact
Contract sales have become number of earning members in of the projects. Producer
dominant and relatively well households and the number of groups appear to be an effect-
10
ive means of reaching women It is too soon to tell whether
and poorer tree fruit farmers. To varying degrees, the five in- the project activities will result
terventions included in the in "sustainable solutions" to
Role of the Projects baseline study succeed in the recurrent needs of tree
Both projects facilitate activit- reaching low-income farmers. fruit producers. This includes
ies to promote upgrading of This means that there is po- both embedded and stand-
tree fruits, primarily product, tential for direct impact by alone solutions/services that
process, and inter-chain up- raising rural household in- provide inputs, TA, or market
grading. The aim is to improve comes through the projects. access. In some cases changes
the capacity of smallholders to might take place due to direct
respond to changing market Building stronger horizontal provision by the projects but it
demand and increase rural in- linkages by grouping produ- remains to be seen if embed-
comes. The baseline research cers and achieving economies ded service arrangements, the
identified specific forms of up- of scale is an important part of commercialization of nursery
grading in the tree fruit value this potential because it helps and extension services, or the
chain and polled the views of poor farmers link to export “network broker” concept of
producers and other actors in markets – something they EAGA and Kenya BDS will last
the value chain on the incent- have very little opportunity to once the project activities end.
ives and disincentives to up- do by other means. The pro-
grade. All the sub-projects in jects have been instrumental While scrupulous efforts were
the study promote the forma- in organizing and strengthen- made to select control group
tion of producer groups as part ing tree fruit producer groups. samples for the baseline sur-
of their strategy to link small- vey that were comparable to
holders to input, service, and Vertical links to higher-value the participant samples, at the
product markets. At the time markets provide critical incent- time of the survey the parti-
of the qualitative research, ives for tree fruit producers to cipants as a group were signi-
producers groups had been upgrade. So far, only one of ficantly better-off and more
formed, but most of them (ex- the interventions studied – the productive than the controls.
cept the avocado groups) were EAGA avocado intervention – When each group is resur-
still at an early stage in their has begun to realize this po- veyed two years hence, care
actual activities. tential by forging a direct link will need to be taken in analyz-
from farmers to the European ing the results to ensure that
Baseline Research Conclu- market. This linkage has in- differences in household
sions and Implications for volved the provision of embed- wealth and other mediating
Round Two ded spraying services by the variables are taken into ac-
Smallholders are part of the exporter and negotiated MOUs count in determining the im-
tree fruit value chain, but they between producer groups and pact of the programs.
occupy a low position within the exporter. The process has
that chain. They are numerous required considerable “hand In the second round it will be
and active producers, but their holding” by Kenya BDS and crucial to review and docu-
productivity is low and they other support from USAID to ment the interventions care-
sell much of their produce un- help prepare smallholders to fully. The activities are very
der unfavorable conditions. In- meet EUREPGAP standards. different and the scopes of
come from tree fruits plays an their activities and the ap-
important role as a source of Brokers remain alive and well proaches they take are likely
household income, especially in all three fruit value chains to evolve over time. Finally, it
for the poorer farmers, but in- and continue to be important will be important to analyze
come from tree fruit and total marketing channels for many the commercialization issue,
household income are both farmers. including a careful look at the
very low in most cases. specific services/solutions pro-
11
moted during the course of the
projects.
12
I. INTRODUCTION
cipants in this activity, and A. MAIN FEATURES OF
This report presents the find- provide some preliminary in- THE PROGRAM
ings from a baseline study of dications of what impacts the ENVIRONMENT
the impact of two projects to projects may be having.
develop tree fruit value chains 1. DEVELOPMENT IN
in Kenya. The study featured a
KENYA3
survey of 1,947 smallholder
farmers who grow avocado,
Kenya achieved independence
mangos, or passion fruit in
from Great Britain in 1963 fol-
three provinces of Kenya –
lowing a nationalist struggle.
Central, Eastern, and Rift Val-
Significant economic growth
ley. The sample included farm-
was achieved through the
ers who are participating in
1970s, but growth slowed in
the Kenya Business Develop-
the 1980s and per capita in-
ment Services project (imple-
come declined in the 1990s
mented by the Emerging Mar-
under the dictatorial rule of
kets Group, formerly known as
Daniel Arap Moi. Hopes of na-
Deloitte, Touche, Tohmatsu)
tional revival were raised by
and Fintrac’s Horticulture De-
the free election of 2002,
velopment Centre project, as
which brought to power a gov-
well as a control group of non-
ernment headed by Mwai
participants. The survey was
Kibaki. Yet the economic
complemented by qualitative
growth rate remained low and
research involving in-depth in-
per capita income in 2003 was
terviews and focus group dis-
still below the 1990 level. Pre-
cussions with over 60 individu-
dicted economic growth rates
als involved in the tree fruit
have recently been revised up-
value chain, including farmers,
ward to 3.3 percent in 2005
farmer producer group lead-
and four percent in 2006,
ers, input suppliers, extension
based on accelerated dis-
workers, brokers, exporters,
bursement of donor funds,
and Kenya BDS and Fintrac
strong performance by cash
HDC project directors and
crops and tourism, and rising
staff.
garment exports to the U.S.
under AGOA. These rates re-
This baseline study will be fol-
flect improvement on past per-
lowed up in two years with a
formance but remain far from
second survey of the same re-
the six percent annual growth
spondents as well as further
that the World Bank believes
qualitative research. The ma-
to be achievable if planned re-
jor findings on the impact of
forms are implemented in full.
the two projects will emerge
after this second stage of re-
Slow economic growth has
search. In the meantime, the
combined with HIV/AIDS to
present report will describe
smallholder tree fruit cultiva- 3
References for this section include the
tion in Kenya, note some of CIA World Factbook, the Economist Intel-
the characteristics of parti- ligence Unit, USAID, and the World Bank.
13
cause poverty to rise and  Rapid population growth age of 1.8 hectares in the
health conditions to worsen.  HIV/AIDS rainy season. 5 Just over one-
The poverty headcount in-  Low levels of investment half of farmers have a deed to
creased from 49 percent of the  Inefficient and dilapidated their land, while another one-
population in 1990 to more infrastructure third own the land but have no
than 56 percent in 2003. Life formal title. Many farms lack
 Vulnerability to drought
expectancy fell from 57 years good access to markets. Close
 Threats to Kenya’s ex-
in 1986 to 45 years in 2004 to one-half are located within
traordinary environment4
while the infant mortality rate five kilometers of a paved
rose from 63 per live births in road. The average farm house-
The government that came to
1990 to 78 in 2002. The estim- hold has 6.8 members and is
power in 2002 pledged to ac-
ated HIV/AID prevalence rate headed by a 53-year old. Male
celerate economic growth and
is currently 6.7 percent. household heads (86 percent
reduce poverty. To this end, it
of the total) average six years
formulated a poverty reduction
High fertility and rapidly in- of schooling, woman house-
strategy, known as the Eco-
creasing population have com- hold heads four years.
nomic Recovery Strategy for
pounded Kenya’s economic
Wealth and Employment Cre-
problem. The population grew
ation, and committed itself to
from 9.4 million at the time of
shift public expenditure to-
independence to 31.9 million
wards programs that benefit
in 2003, averaging more than
the poor, notably a free
three percent per annum.
primary education program.
However, the total fertility rate
Subsequently, however, polit-
has now declined to 3.3 births
ical in-fighting over constitu-
per woman and the population
tional reform and other issues,
growth rate in 2004 was only
together with signs of reluct-
1.1 percent.
ance to tackle high-level graft,
have raised doubts about the
Kenya has many natural ad-
government’s ability to reform
vantages as well as the largest
and shake off the bad habits of
and most diversified economy
the past. Kenya is heavily de-
in the East Africa. According to
pendent on donor funding,
USAID/Kenya, the country’s
which was withheld during the
perennial failure to achieve
1990s and is currently
sustained economic growth is
threatened once more by the
attributable to several factors:
governance issue.
 Governance issues: lack of
democracy; over-concen-
Poverty and inequality remain
tration of power in the ex-
severe, largely because most
ecutive branch with inad-
Kenyans are still low-pro-
equate checks and bal- 5
Data cited in this paragraph derive from
ductivity farmers. Agriculture
ances the Rural Household Survey carried out by
absorbs 75 percent of the
 Corruption that pervades labor force but produces less Egerton College, Tegemeo Institute, and
public administration than 20 percent of GDP. Farms
Michigan State University in 2000, as re-
 Inconsistency in policies, ported in Nicholas Minot and Margaret
are small on average. Most Ngigi, “Are Horticultural Exports a Rep-
laws, and regulations that farmers are heavily dependent licable Success Story? Evidence from
adds significantly to costs on rainfall and plant an aver- Kenya and Cote d’Ivoire.” Paper presented
of doing business and dis- at the InWEnt, IFPRI, NEPAD, CTA con-
courages investment 4
USAID/Kenya. 2000. Integrated Stra- ference, “Successes in African Agriculture,
 Low productivity tegic Plan 2001-2005, pp. ii-v. Pretoria, December 1-3, 2003.
14
2. USAID KENYA’S STRA- fore heavily dependent on (USAID/Kenya 2000, pp.
TEGIC OBJECTIVES its agricultural productiv- 65, 83)
ity. Over the past decade,
Based on U.S. foreign policy in- however, agricultural pro- Kenyan agriculture is primarily
ductivity has declined and organized in smallholdings and
terests and Kenya’s develop-
poverty has increased. is almost exclusively rain-fed.
ment constraints, USAID/Kenya
identified four strategic object- According to the Mission’s ana-
Over the same period,
ives and one special objective drought has plagued lysis, factors contributing to
for its programming in 2001- Kenya on an increasingly low and falling agricultural pro-
2005: frequent basis, affecting ductivity include HIV/AIDS, a
 Strategic Objective 6: ‘traditionally’ drought- confused policy environment,
Sustainable reforms and prone areas, as well as the survival of marketing
accountable governance many other agro-ecologic- boards for a few key commod-
al zones (AEZs) of the ities (coffee, tea, and pyr-
strengthened to improve
country. While poverty is ethrum), poor access to credit
the balance of power found in both urban and
among the institutions of and extension services, and
rural areas, 75 percent of
governance weak smallholder organiza-
the poor are in rural areas.
 Strategic Objective 7: USAID/Kenya will, there-
tions.
Increased rural household fore, focus on increasing (USAID/Kenya 2000, pp. 68-75)
incomes the incomes of rural
households in selected The results framework adop-
 Strategic Objective 3:
high and medium poten- ted by USAID/Kenya for SO 7
Reduce fertility and the
tial and arid and semi-arid includes four high-level inter-
risk of HIV/AIDS transmis- lands, most of which mediate results (IRs). IR 7.1
sion through sustainable, already rely on a combina- calls for increased productivity
integrated family planning tion of on- and off-farm in three targeted agricultural
and health services activities...
sub-sectors: dairy; horticul-
 Strategic Objective 5:
Increasing rural household
ture; and maize. IR 7.2 aims to
Improved natural resource
incomes is essential to increase the volume and value
management in targeted
achieving a prosperous of traded agricultural commod-
biodiverse areas by and
and democratic Kenya. To ities, especially dairy and hor-
for stakeholders
sustain and improve pub- ticultural products. IR 7.3
 Special Objective 4: lic services and build seeks increased access to
Critical needs met for democratic institutions, business support services
Kenyans affected by the Kenyans must have higher (credit and savings; appropri-
bombing of the Nairobi incomes. A population ate technology; skills, and
Embassy in 1998 and ca- with higher incomes is a business training) for micro
pacity built to address fu- population with higher ex-
and small enterprises. IR7.4
ture disasters pectations for its future
and the future of its chil- targets increased effective-
dren. When people are ness of smallholder organiza-
The activities covered in this tions in providing business ser-
able to pay for health and
report fall under SO 7. The vices to members and repres-
education services, these
Mission justifies this strategic services can be sustained enting their business interests.
objective as follows: and improved. Likewise, Below these four IRs, 15 sub-
economic growth will cre- IRs are specified.
Since 80 percent of the ate financial stability and
Kenyan population lives in allow Kenyans to take a
rural areas, and 75 per- As discussed in the following
more constructive interest
cent are somehow in- section, tree fruit cultivation
in the political environ-
volved in agriculture, ment that affects their and other forms of horticulture
Kenya’s economy is there- economic well-being. are important activities for
15
Kenyan smallholders. Raising production in Kenya include a lackluster economy (Minot and
productivity in these activities climate that allows for year- Ngigi 2003, pp. 3-8). Domestic
and the revenue earned from round cultivation, fertile soils, sales through traditional retail
them should therefore contrib- and a competitive labor force and public wholesale outlets,
ute significantly to the in- with good education and tech- by far, dominate the market.
creases in average rural nical background. While there are two large su-
household income that are permarket chains, they com-
sought by the GOK and According to the 2000 Rural prised less than five percent of
USAID/Kenya. Household Survey carried out domestic market horticulture
by Egerton College, Tegemeo sales in 2003. Much of what is
Curiously, given the emphasis Institute, and Michigan State sold in these supermarkets is
on raising average rural University, almost all farmers procured directly from pre-
household incomes, little in- in Kenya (98 percent) grew ferred growers – mostly com-
formation seems to be avail- some fruits and vegetables mercial farmers and a small
able on the actual levels of and 35 percent of fruit and ve- number of organized small-
these incomes. Tegemeo Insti- getable production was sold in holders (Tschirley et al 2004).
tute, on behalf of the market. Overall, fruits and Only two percent of farmers
USAID/Kenya, does track annu- vegetables contributed 18 per- currently produce for export
al movements using a proxy cent of average household in- markets. The Horticulture
method that it developed in come. Over 90 percent of Crops Development Authority
partnership with Michigan households across income (HCDA) estimates that 40 per-
State University.6 groups grow fruits and veget- cent of exported fruit is pro-
ables, although richer house- duced by smallholders (cited
3. BACKGROUND OF THE holds market a larger share of by Minot and Ngigi 2003, pp.
HORTICULTURE SECTOR their output and account for a 10-11), with the remaining 60
large proportion of total sales. percent produced by commer-
Kenya’s tropical and temper- According to a study by the In- cial farms.
ate climate zones favor cultiv- stitute of Development Studies
ation of a wide range of horti- at the University of Sussex, Horticulture Exports from
cultural crops. In the coastal households involved in the Kenya
lowlands, farmers grow man- production or processing of ex- Over the past two decades,
gos, citrus fruits, cashews, ba- ported horticultural crops export horticulture in Kenya
nanas, hot peppers, brinjals, earned higher incomes than has grown in importance, al-
and melons. In the middle alti- households that are not, other most tripling in value between
tudes, crops include bananas, things being equal. This sug- 1996 and 2001 (Table 1). Hor-
mango, avocado, pineapple, gests that enabling more ticulture (comprising fresh
grapes, passion fruit, pawpaw, households to participate in fruits and vegetables and cut
citrus, flowers, onions, garlic, the sector could reduce flowers) has become the na-
tomatoes, kale, cucumbers, poverty substantially in both tion’s third most important for-
pepper, okra, and French rural and urban areas (McCul- eign exchange earner after
beans. At high altitudes, avo- loch and Ota). tourism and tea.
cado, pears, apples, plums,
carrots, cabbages, peas, pota- While horticulture products Kenyan horticulture products
toes, and flowers are grown. have long been grown for are exported primarily to
Factors that favor horticulture home consumption, production Europe and the Middle East7
for sale in domestic and export
6 7
See David Tschirley and Mary Math- markets began in the early Regional exports, especially to neighbor-
enge. 2003. “Developing Income Proxy 20th century and has recently ing Tanzania and Uganda, are minimal.
Models for Use by the USAID Mission in Overall, Kenya is a net importer of horti-
become one of the few suc-
Kenya: A Technical Report.” Tegemeo culture from these countries (Tshirley, et al
Working Paper No. 7.
cess stories in an otherwise 2004).
16
Table 1: Value of Horticulture Exports (in millions of Kenya Shillings)
Year Fruits Vegetables Cut Total Horticulture
Flowers
1996 770 2,577 4,366 7,713
1997 805 3,116 4,888 8,809
1998 820 4,025 1,856 9,728
1999 1,256 5,713 7,235 14,204
2000 1,098 5,293 7,166 13,557
2001 1,560 8,035 10,627 20,221
Source: Cited in Dolan and Sullivan
where they compete with pro- are the main vegetables sup- getables and fruit are sold on
ducers from EU countries as plied to Europe. The leading export contracts that specify
well as from other African, destinations for fresh fruit ex- quantities and prices. British
Middle Eastern, and Southern ports (mango, avocado, and supermarkets took an increas-
European countries. Consign- passion fruit) are France, ing role in the vegetable trade
ments of fresh cut flowers, Dubai, the Netherlands, and during the 1990s as a way of
fruits, and vegetables are air the UK. Overall, nearly 90 per- ensuring the quantities,
freighted daily to various des- cent of Kenyan horticultural safety, and qualities that they
tinations from Kenya’s two in- exports go to Europe. The wanted. This shifted the trade
ternational airports. Some Middle East is a significant from Kenyan wholesale mar-
bulky produce is shipped from market for mangoes. Fruit ex- kets, where Asian traders are
the port of Mombassa. The ports grew rapidly from 1996 active, to contracts with large
European Union is the princip- to 2001 but remained much exporters that obtain their pro-
al importer of Kenya’s fresh smaller in value than either duce primarily from their own
produce. The bulk of flower ex- cut flowers or vegetables. The farms and large contract
ports go to the Netherlands for official figures are shown in farms. The move hurt small
sale by auction. By 1999, Table 1. out-growers. Pre-packs for the
Kenya had become the leading supermarkets and Asian veget-
supplier of flowers to the EU, The demand for horticulture ables became increasingly im-
followed by Israel, Costa Rica, products in the European mar- portant products during the
Colombia, the USA, Ecuador, kets is increasingly concen- 1990s.
and Zimbabwe. Britain, France, trated on fresh produce distri-
the Netherlands, and Germany bution channels in supermar- Some 10-15 major export-
are the major importers of ve- kets. Another important factor ing companies dominate
getables. Kenya has been de- influencing demand is increas- the sector. These compan-
scribed as one of the worlds’ ing importance among con- ies are very well organ-
sumers of food safety and the ized, often with an integ-
most successful exporters of
rated system of produc-
fresh vegetables to EU coun- environmental and social di-
tion/processing/
tries; in 2002 it ranked second mension of the food supply transport/marketing.
among non-members in the chain. As a result, the regulat- There is also a quite well
value of fresh vegetables (Jaf- ory environment is becoming developed small/medium
fee 2003). 8 Beans and peas more stringent, raising the bar size exporter sector who
for new entrants and posing are well organized on pro-
8
The $6 billion annual fresh vegetable new challenges for existing duction/transporting level,
market in the EU was supplied largely by suppliers (Jaffee 2003). but less on processing and
EU producers. Among the $950,000 mil- marketing due to their
lion of vegetables imported from non-EU Export marketing systems for size of operation. There is
members, Kenyan exports account for a third level of exporters
horticulture differ by crop.
$100 million (Jaffee 2003). Similar data who still perform more or
While cut flowers are sent for less in an ad hoc manner,
to show the position of Kenyan fresh auction in the Netherlands, ve- and rely on the prevalent
fruits was not found.
17
market situation and by increasing employment in supply chain. In response,
brokers for their existence. production, transport, input the Fresh Produce Exporters
However, the latter group supply, processing, sorting; by Association of Kenya (FPEAK)9
has almost disappeared increasing jobs for unskilled adopted a Code of Practice
from the flower export
workers, especially women; by for growers in 1999. The
sector in the last five
years, and will, most prob- increasing employment on Code includes a 14-step doc-
ably, decline also in the large farms and plantations; umentation procedure for en-
vegetable sector in the and by building new know- suring the traceability of pro-
next five years due to the ledge and technology that is duce handled by the export-
effects of the Code of valuable in producing and er. “This is an important step
Practice to be implemen- marketing other high value in establishing a common set
ted. However, brokers products. of standards regarding safe
make out an essential part handling of fresh fruits and
of the fruit export sector Constraints to Horticulture vegetables and disseminat-
and will continue to be im-
Exports ing the information. However,
portant if Kenya is going
to remain a fruit exporting
some aspects of the Code im-
country in the future. Demand Side Constraints: ply significant costs and
(FKAB Feldt Consulting According to Minot and Ngigi there are currently no en-
2001, p. 8) (2003, pp.9-10), the trans- forcement mechanisms.”
formation of food retailing and (Minot and Ngigi 2003, p.10)
According to a sector study changes in the structure of More recently, EUREPGAP has
contracted by USAID/Nairobi consumer demand in Europe significantly raised the stand-
(FKAB Feldt Consulting 2001), are serious challenges for ard that Kenyan produce
Kenya has several competitive Kenya and other horticultural must meet to enter the
advantages in export horticul- exporters: European market, as well as
ture:  The rise of supermarkets: the cost of compliance.
 A strong and well organ- The share of fresh fruits and  Competition from other sup-
ized private sector vegetables sold by supermar- pliers: Kenyan horticulture
 A variety of suitable cli- kets in the UK rose from 33 enjoys duty-free access to
mates for different species percent in 1989 to 70 percent European markets. If and
 A rather good main road by 1997. Increasingly, super- when this preference is ter-
infrastructure and good market chains bypass whole- minated, Kenya will face in-
local supplies of inputs salers and buy directly from creased competition from
and implements exporters in Kenya and other countries such as Egypt,
 Access to good air cargo countries. To protect their South Africa, Chile, Brazil,
handing facilities and air- reputations, the chains im- and Thailand. Even without
port services with ad- pose new restrictions and such a change, horticultural
equate cargo space to ma- even organize production in markets are highly competit-
jor destinations developing countries. ive and subject to rapid shifts
 Rather simple export doc-  Increasing concern over food
umentation procedures safety: The demand for horti- 9
FPEAK is an organization that represents
 Incentives for exporters culture products in the than 140 members who are active export-
European markets is shifting ers and other interest groups. Besides for-
(VAT reimbursement and
with consumers increasingly mulating and implementing a Code of
duty-free imports of most Practice to ensure quality produce grown
inputs and implements) aware of the health con-
and shipped in an ecology- and worker-
sequences of pesticide
friendly environment, FPEAK maintains a
Horticulture production for ex- residues and placing more database of local products and suppliers
port has potential to benefit importance on food safety and provides market leads and contacts to
poor people in several ways: and the environmental and members. Its secretariat is assisted by
social dimension of the food USAID.
18
in export competitiveness.  A general shortage of skilled ley et al 2004; Muendo, Tschir-
Kenya lost the European mar- labor and qualified manage- ley, and Weber 2004).
ket for fresh pineapple to ment staff Moreover, recent data suggest
Cote d’Ivoire in the 1980s,  High air freight rates and a a downward trend in the share
was squeezed out of avocado need for more cargo capacity of smallholder production in
exports to Europe by higher to London, Paris, and Frank- these markets. Smallholders’
quality products from Israel furt share in export horticulture
and South Africa, and also has fallen from 75 percent in
 Inadequate communications,
lost the market for several the early 1990s to perhaps 45
power supply, and rural feed-
temperate vegetables. It re- percent today, indicating a
er roads. Failure to exempt
sponded by finding new mar- “clear decline and rough chal-
contract farmers and out-
kets and expanding exports lenges ahead” (Muendo,
growers from VAT (because
of French beans, Asian veget- Tschirley, and Weber). Because
their products are exported
ables, and cut flowers. Export exports have soared, this does
through a third party)
competitiveness evolves con- not necessarily imply an abso-
tinuously in response to lute decline in the quantities
The industry has perceived
changes in markets, techno- that smallholders supply to the
threats to its prosperity from
logy, and competitors (Ibid). export market, but it does sug-
both the Kenyan government
gest limitations on new oppor-
and the EU. Recently there
Supply side constraints: En- tunities.
was a general fear that the
hancing the capacity of the government might raise taxes
Kenyan horticulture industry to In the context of this dualistic
and fees that impact export-
respond to changes in market market, smallholders particip-
ers. There was also a move to
demand is critical to remain ate primarily in traditional
increase government control
competitive in export markets. markets, which at present are
of horticulture by broadening
Small farmers need to become not competitive even on a re-
the role of the Horticulture
more competitive, not only gional basis. There are relat-
Crops Development Authority10
today but also tomorrow. Pro- ively few regional exports,
– from being a facilitator to a
jects such as those reviewed largely because of high trans-
more active role in buying and
here need not only to create portation costs. Kenya is in
selling commodities like a mar-
competitive advantage but fact a net importer of horticul-
keting board -- but this seems
also to sustain it. At present, tural products from Uganda
to have been withdrawn fol-
however, challenging con- and Tanzania. While Kenyan
lowing the change in govern-
straints exist on the supply exports have been competitive
ment. (Minot and Ngugi, 2003,
side. Among the most critical in international markets, the
p. 5)
are shortages of the seedling stringent quality standards
varieties needed for participa- that are being introduced in
Constraints to Smallholder
tion in exports and the lack EU and other export markets
Participation in Horticul-
(on the part of smallholders) of are likely to further raise the
ture Export Markets
the knowledge, skills, and fin- bar for small scale producers –
Despite the potential of horti-
ance needed to grow fruit in thus further limiting their par-
culture sales to increase
ways that will safeguard qual- ticipation in these markets. As
household incomes and reduce
ity and protect them from dis- stated in an article from The
poverty, a large majority of
ease. The same study identi- Financial Times:
Kenya’s smallholder horticul-
fied several important con- ture producers remain outside
straints and areas requiring In the wake of mad cow
the more lucrative export and disease and other scares,
improvement: supermarket segments (Tshir- European authorities de-
 A shortage of irrigation water
mand ever tighter food
in many areas 10
The HCDA was formed in 1967 and car- quality controls. A bewil-
ries out a variety of promotional activities.
19
dering array of these  Increase outreach and sus- cess and incentives for small-
already apply. There are tainability of solutions/ser- holders and enhancement of
more than a dozen quality vices offered by multiple pro- their capacity to respond to
standards across the EU, price incentives through em-
viders to large number of mi-
usually set up and mon-
croenterprise clients bedded services provided by
itored by the trade…For
poor countries like Kenya,  Foster a better-skilled and lead firms and input stockists.
the question is whether more competitive MSE sector The Fintrac project includes a
the regulations, or non- technical component (im-
tariff barriers, are becom- The projects seek to: proved planting stock and cul-
ing incompatible with the  Raise productivity through tivation methods) and tries to
vision of development that market intervention by pro- improve the international en-
sees small-scale crop promoting the production of abling environment through its
duction of export crops as work with European retailers
higher grade, better quality
central to poverty reduc- to shape their standards so
tion. (Wallis)
fruit by facilitating access to
improved stock and seed- that Kenyan farmers can meet
lings, productive inputs, them and by helping produ-
As the Financial Times article
training, extension and in- cers meet the retailer’s rising
suggests, large producers and
formation services quality standards.
exporters find it easier and
 Increase market outlets in
cheaper to comply with such
regulations than do small and selected areas by facilitating
medium firms because large direct links between small- 1. KENYA BDS PROGRESS
firms can spread the cost of holder producers and lead TO DATE
compliance, which is substan- firms involved in fruit export
tial, over a larger volume of and processing and promot- Kenya BDS, a five-year project
sales. ing the formation of producer that started in 2002, was in-
groups tended to work in three sub-
 Facilitate inter-firm cooper- sectors; tree fruit was the first
ation and organization within sub-sector selected.11 During
B. PROGRAM DESCRIP-
the overall value chain, its first two years, the project
TIONS between producers, input issued tenders and awarded
suppliers, producers and buy- contracts to eight private sec-
The two projects covered by ers, by organizing and build- tor and NGO partners active in
this assessment are designed ing the capacity of tree fruit the production and marketing
to promote growth in Kenya’s producer groups, linking of tree fruit. The contracts
tree fruit agriculture and en- smallholder MSEs to lead were designed to facilitate the
courage smallholder participa- firms that provide embedded development of sustainable
tion in the tree fruit value services, and facilitating oth- business solutions that provide
chain. USAID/Kenya funds er business arrangements material inputs (agro-chemic-
both in support of their stra- and relationships als and seed varieties), appro-
tegic objective to increase rur-  Promote the development of priate technology to upgrade
al household incomes in Kenya sustainable business solu- products and production pro-
(SO 7). tions/services for rural cesses, business and skills
MSEs training, and extension and in-
The overall goals of the Kenya formation services. Several
BDS and Fintrac HDC projects contracts promote market link-
The Kenya BDS project focuses
are to: ages between smallholder pro-
on vertical linkages, especially
 Increase small farmer and the link connecting farmers to 11
household incomes The second sub-sector chosen was Lake
lead firms. It is essentially a Victoria Fish; the third sub-sector has not
 Promote growth in final sales business model, which em- yet been identified but is likely to be a
in selected commodities phasizes improved market ac- non-agricultural activity.
20
ducers and lead firm exporters standing with East Africa  Contracts were drawn up
through supply contracts and Growers Association (EAGA), a between the producer groups
lead firm provision of embed- large horticulture export firm and EAGA. The producer
ded services, and encourage in Kenya, to link with avocado groups agreed to sell exclus-
inter firm cooperation through producer groups in two loca- ively to EAGA, to upgrade
organization of producer tions in Central Province. In their avocado production,
groups and provision of em- the first year, Kenya BDS mo- and to follow a good agricul-
bedded services. Embedded bilized 803 avocado farmers tural practices protocol.
services are who organized producer
products/services/solutions groups and, with the help of a
that are provided on a non-fee facilitator hired by Kenya BDS,
basis by one firm to another as negotiated a contract with
part of their commercial trans- EAGA to supply avocados that
actions. Examples include: 1) meet agreed upon standards.
buyers/exporters who offer EAGA provides embedded
pre-financing, technical ad- spraying services, grades the
vice, or inputs to their produ- fruit, and transports it to their
cers in order to ensure a qual- warehouses. Group members
ity product that meets market have been trained in the ap-
standards; 2) input suppliers plication of manure and fertil-
who provide training/technical izers, pruning and orchard hy-
advice to MSEs in the use of giene to upgrade the quality of
the product they sell in order their fruits. Near the end of
to ensure correct/successful the first year, Kenya BDS mo-
usage of the product. bilized an additional 283 avo-
cado farmers in 10 groups in
The box below shows the several new locations. These
range of activities undertaken producer groups are working
under the Kenya BDS tree fruit with another lead firm, Kenya
contracts. Annex A details Horticultural Exporters (KHE).
activities by fruit, partner, and Farmers began basic pruning,
location. and spraying and had plans to
negotiate contractual arrange-
Kenya BDS initiated on-the- ments with KHE. Kenya BDS
ground project activities in also has plans to work with an-
2003, almost a year before the other exporter, Indu Farm
first round of data collection (EPZ) Limited in the next
for this impact study. Over the quarter.
course of that year, project
staff observed a number of At the survey site in Kandara,
changes in the tree fruit value EAGA activities with avocado
chain as a result of project groups during the year prior to
activities. In order to capture the baseline survey included
the full impact of the project, it the following:
is important to document  EAGA, with Kenya BDS sup-
these activities and the ob- port, initiated the organiza-
served changes. tion of avocado producer
Avocado groups (the farmers were not
Kenya BDS in 2003 negotiated previously organized).
a memorandum of under-
21
 EAGA purchased avocados at
a negotiated price, provided
spraying and grading ser-
vices for a fee deducted at
the time of sale, transported
fruit to their warehouse in
Nairobi, and paid farmers
through their group accounts.
According to farmers, spray-
ing was not done on time
during the first year, which
reduced the yield of high-
grade fruit.
 Kenya BDS encouraged the
revival of a processing fact-
ory that will buy lower grade
fruit to process avocado oil
and promoted a linkage
between EAGA and this fact-
ory. If this plant becomes op-
erational, EAGA plans to buy
all grades of fruit from produ-
cers and drop off lower grade
fruit at this processing fact-
ory on their way to Nairobi.
During this time, project staff

Figure 1: Activities under BDS Tree Fruit Project observed the following
Facilitate the provision of inputs, changes:
Establish nurseries, Avocado brokers have been
Establish a credit facility link between less active in the area since
agrochemical distributors and stock- the contractual agreement
Input supply 
ists, and was initiated. EAGA met with
Develop a monitoring system to in- the brokers (who they buy
form manufacturers and stockists on from) and asked them not to
consumer trends. buy from group members in
Improve commercial extension ser- this area. At least one other
vices, exporter has entered the area
Create farmer-led extension teams, to compete with EAGA to buy
Launch information campaigns, upgraded fruit. Production
Extension and
 Train agrochemical stockists in advis- and sale of avocados was very
training
ory services and business manage- low prior to the EAGA initiat-
ment, and ive. Since then, avocado sales
Raise farmer awareness on safe use of volumes have increased dra-
chemicals. matically and the prices re-
Facilitate market linkages, ceived by producers have also
Improve market information through increased. Kenya BDS staff cite
SMS technology and trading floors, this as a critical impact to cap-
Market access  Establish collection sites, ture, since avocado income
Facilitate improved transportation,
22
was almost nil at the begin- the groups to input suppli- ern province. These efforts
ning of the project. ers and microfinance ser- include an information
vice providers. While SITE campaign to raise aware-
Kenya BDS handholding activ- was able to get four export- ness of farmers on nursery
ities have promoted informa- ers to establish direct link- development and the bene-
tion flows, built trust, and ages with farmers, only 17 fits of nursery seedlings;
forged linkages between the percent of the mango farm- training of nursery operat-
exporter (EAGA) and producer ers in the groups benefited. ors in mango husbandry
group members, between Nevertheless, through and business management;
training and extension workers these contracts, which of- and training of extension
and producer group members, ten benefit only large farm- service providers on graft-
and among producers them- ers, small farmers were ing, budding, and top work.
selves. able to participate. Export-
ers made payments 4. Coastal Development Au-
Mango through the farmer groups’ thority provides on-farm
Kenya BDS is working with four bank accounts, which training of trainers for un-
partners to develop the mango provide a more secure employed extension of-
value chain. means of payment for the ficers. They also are creat-
farmers. Kenya BDS reports ing Farmer Led Extension
1. SITE is a local NGO suppor- indicate that prices im- Teams (made up of lead
ted by Kenya BDS. As of proved from Ksh 3-6 per farmers and extension
December 2004, when the mango to Ksh 7-12. agents) to provide commer-
quantitative survey was cially viable extension ser-
conducted, SITE was work- 2. Kenya BDS staff observed vices; launching an inform-
ing with groups of farmers that creating longstanding ation campaign to increase
in eight mango production win-win market relation- awareness of value of ex-
clusters in four districts of ships between mango tension services; and es-
Central and Eastern farmers and lead firm buy- tablishing a revolving fund
province. Its work with ers and exporters is chal- to finance adoption of good
farmers focused on building lenging. It requires time agricultural practices
linkages with reliable mar- and considerable ‘hand-
kets and increasing the ac- holding’ to build trust, 5. KWETU, KARI, and Kenya
cessibility of business ser- change attitudes, and instill Gatsby Trust are working
vices to increase the qual- work ethics. In late 2004, with mango farmers in the
ity and productivity of Kenya BDS awarded a Watamu/Msabaha and Ma-
mango orchards. At this second phase of support to lindi areas of Coast
time, 2,461 farmers grow- SITE, focused on strength- province. Through a con-
ing Apple, Ngowe, Tommy, ening producer groups to tract with Kenya BDS, they
Kent, and Van Dyke mango operate as units in dealing are forming producer
varieties had been mobil- with buyers; coordinating groups and facilitating mar-
ized through 83 producer farmer access to material ket linkages to buyers
groups. During the first inputs; and actually estab- through the development
phase of the project, SITE lishing longstanding market of a market information
hired five private service linkages. data base and brokerage
providers to offer extension workshops. They also are
services and strengthen 3. KADI and the Catholic Dio- training private extension
the farmer groups. They ceses of Embu have con- workers in mango hus-
organized training for farm- tracted with Kenya BDS to bandry and business man-
ers in mango as a farming promote extension services agement and launching a
business and helped link for mango farmers in East- campaign to sensitize pro-
23
ducer groups to the value material inputs provided by This approach has led to signi-
of extension services. KARI -- representing a ficant progress during the first
public/private partnership. The year in avocado upgrading and
Passion fruit site provides extension ser- market linkages. A similar ap-
Kenya BDS is working with vices and quality seedlings. proach is underway in
Fineline Systems and Manage- Growing out of this, as of Meru/Embu with passion fruit
ment Limited, who are now co- December 2004, five group growers.
ordinating the whole program nurseries and 48 individual
among the farmers while EAGA nurseries had been estab- Market principles underpin the
is providing both the export lished and farmers had Kenya BDS approach.12 The
and domestic market and em- planted 96,000 passion fruit project does not cover costs
bedded services in terms of in- vines. The plan is to expand to for spraying, other inputs, or
put supplies and extension 80 groups involving 2,400 extension services, or subsid-
services. Just Juice is now left farmers in 2005. ize producers in other ways.
with the role of providing a Nor does it inform producers
demonstration plot and quality Kenya BDS’s Facilitation about the prices at which ex-
seedlings to help passion fruit Role porters sell. While some small-
growers in Meru and Embu dis- The Kenya BDS project is de- holder participants thought
tricts. This work was just be- signed to promote efficiencies that Kenya BDS would be a
ginning to take off in October and growth in tree fruit agri- more active advocate for them
2004 when the baseline re- culture through the develop- in getting better prices, project
search began. It took some ment of business services and staff emphasized that this is
time to get producer groups other broad facilitation activit- the role of the market and that
organized in this area. By ies. While the approach has if they played this role, the in-
December 2004, however, 744 been to award contracts on formation provided would dis-
farmers had been organized the basis of competitive bids, tort the market. They feel that
into 26 producer groups. for its work with avocado farm- the litmus test should be
Twenty-three percent of the ers in Kandara and passion whether the smallholders are
farmers were women. Project fruit farmers in Meru and better off than they were be-
activities include training Embu (which started out as an fore, not the price they re-
group members in nursery es- agreement with Just Juice) ceived in relation to world
tablishment, land preparation Kenya BDS has taken a some- market prices. Kenya BDS
and planting, orchard mainten- what more active “hand hold- wants to play a very business-
ance, and post harvest handing” role in promoting upgrad- like role in the process. Their
ling. They also focus on facilit- ing and market linkages. For main aim is to diversify market
ating direct linkages with ex- avocados, this had involved outlets and promote competi-
porters through supply con- placing a full-time Kenya BDS tion. Kenya BDS staff members
tracts; regular collections of staff person on the ground in do not want to embed them-
fruit; and exporter-led technic- the Kandara area. She has selves in the supply chain.
al advice and spraying for pest played a role in identifying
and disease management. The producer group leaders, form- Several lessons emerged dur-
aim has been to ensure regu- ing producer groups, setting ing the first year of the Kenya
lar payments, stable and guar- up record-keeping systems, BDS project. From the per-
anteed market prices, and and negotiating contracts spective of Kenya BDS leader-
prices that are higher than between the producer groups ship, the approach of contract-
those paid by brokers. A cent- and EAGA. Her roles in linking ing through open bids has
ral demonstration plot and producer groups to EAGA and worked well. The short-term
nursery had been established building trust by keeping regu- nature of their contracts is a
with land donated from Just lar lines of communication 12
Based on interview with Muli Musinga,
Juice, Ltd. and technical and open has been instrumental. David Knopp, and Rose Warui.
24
Figure 2: Activities under HDC Tree Fruit Project EAGA than buying directly
Introducing new varieties of passion fruit from middlemen, it is an
Input supply  Producing plant stock approach that ensures
Establishing commercial nurseries good quality and good sup-
Extension and train- Establishing demonstration plots ply – something middlemen
 cannot always provide.
ing Providing extension services to farmers
Linking smallholder producers to domestic Similar models can be used
fresh fruit markets for snow peas, sugar
Linking smallholder producers to processors snaps, and French beans.
Market access 
of juice concentrate for domestic and export
markets
Training in EUREGAP certification 2. FINTRAC HORTICUL-
TURE DEVELOPMENT
Inter firm coopera- Delivering services through farmer groups CENTRE (HDC) PROJECT
 PROGRESS TO DATE
tion
positive feature of the project broker.” Its follow-on Memor- Fintrac’s HDC project fo-
in that it provides flexibility andum of Understanding with cuses on a wide range of horti-
and responsiveness in their EAGA is addressing this issue. culture products, one of which
approach. In the case of KACE, One question is the potential is passion fruit. This impact
for example, Kenya BDS had a for EAGA or producers groups study focuses only on their
one-year contract to promote to absorb the costs of this passion fruit work, which in-
market information on tree function as an embedded ser- cludes plans to:
fruits through SMS technology. vice.  Develop Kenyan varieties of
They discovered that KACE passion fruit for fresh export
had a number of problems and From the perspective of one  Improve agricultural prac-
a business model that they did lead exporter, forward plan- tices of local producers
not really agree with, and after ning is one of the biggest chal-  Expand local processing cap-
a year they decided not to lenges in the horticulture ex- abilities for local market
continue with this sub-project. port business. Exporters have products
Other USAID projects are typ- forward contracts so they  Strengthen the farm-to-mar-
ically longer (up to five years) must plan ahead for the uplift ket value chain, inclusive of
and provide much less flexibil- of fruit and cannot operate ad business services to small
ity to cut losses. hoc. A challenge in working farmers
with smallholders, from this
Another lesson is the import- exporter’s perspective, is pro- Unlike Kenya BDS, the HDC
ance of Kenya BDS’s “hand jecting a timeframe of produc- project does not operate
holding” role in building trust tion, sales, and returns. They through contracts and MOUs
between producers and the need to establish ground rules but carries out activities dir-
exporters. Kenya BDS staff and work together to develop ectly through project staff
have brokered the relationship a commercialization strategy based in Nairobi and agronom-
and facilitated communication from day one. ists based in four field offices.
to help each party to under- It works with and through co-
stand the needs of the other. EAGA’s experience in linking operating partners, including
In the future, as part of its exit directly to avocado farmers the Kenya Agricultural Re-
strategy, Kenya BDS wants to has had a demonstration ef- search Institute (KARI), exist-
explore the potential for com- fect within the company. While ing smallholder associations,
mercializing the role that its the costs of interacting directly and two small businesses pro-
staff currently plays - referring with groups and providing em- ducing plant stock. In the fu-
to this role as a “network bedded services is more for
25
ture, it intends to work will train 600 producer groups
through other input suppliers (with 20 members each).
as well. This five-year project
began in late 2003 and was in Progress to Date
its first year of operation at Fintrac’s HDC activity with pas-
the time of the baseline sur- sion fruit growers in the El-
vey. doret area prior to the baseline
survey focused primarily on
The project decided to focus promoting input supply and
on passion fruit because it is extension activities. These in-
regarded as a relatively cluded:
friendly crop for smallholders.  Identifying farmer groups in-
Production carries low risk and terested in planting grafted
the market potential is high. passion fruit, which has
Initial project activities related stronger root stock and more
to passion fruit focus primarily disease resistance
on product development by  Linking farmer groups with a
addressing two key constraints nursery operator who pro-
to smallholder production: produces grafted passion fruit
duction technology and farmer plant stock
knowledge. To this end, Fintrac  Coordinating with HCDA and
HDC is cooperating with KARI the MOA in linking farmer
on training and plant produc- groups to public training and
tion, and with various small extension resources, some of
businesses in Eldoret, includ- it related to compliance with
ing input suppliers and nurser- EUREPGAP standards
ies. The HDC project hopes to
establish 30-40 good demon- Activities within the project
stration plots and, through area were just getting under-
them, have a ripple effect on way when the baseline re-
passion fruit production search began in October 2004.
throughout Kenya. It is trying Fintrac HDC had identified pro-
to develop and produce fruit ducer groups and had linked
varieties that will yield more these groups to a nursery op-
juice, including the introduc- erator supplying grafted seed-
tion of the jumbo variety from lings. Producer groups began
Uganda. It also hopes to find a receiving grafted seedlings to
good investor to build a pro- plant in demonstration plots in
cessing plant that would re- October 2004.
quire input of 50 to 100 tons
per week.
Training in EUREPGAP certifica-

tion also will be an important
project activity. They will train
producers and companies on
requirements for export certi-
fication. The goal is certifica-
tion in 2005. They will train
three companies who in turn
26
II. DESIGN OF THE IMPACT ASSESSMENT
The purpose of this study is to high level of smallholder parti- iveness of the value chain.
assess the impacts of the cipation in the value chain. A These activities, in turn, lead
Kenya BDS and Fintrac HDC related question is the poten- to sustained access to solu-
projects on: tial for smallholder involve- tions, increased smallholder
ment in the tree fruit value MSE profits from tree fruit
1. The competitiveness of the chain to contribute to poverty activities, increased rural
mango, passion fruit, and reduction by increasing small- household incomes, and over-
avocado value chains; holder enterprise profits and all value chain growth. (Figure
2. The integration of micro household incomes. 1)
and small enterprises
(farmers and others) into Accordingly, this impact as- C. HYPOTHESES
these value chains in a way sessment examines the impact
that they contribute to and of the project-facilitated inter- We use this causal model to
benefit from the tree fruit ventions in improving the com- test a number of hypotheses
industry’s increased competitiveness of Kenya’s tree about the impact of donor in-
petitiveness; fruit agriculture, integrating terventions in opening up op-
3. The development of sus- smallholders -- including wo- portunities for smallholder
tainable solutions to con- men smallholders and small- MSEs in local, regional, and
straints facing businesses holder households headed by global markets and in improv-
in the targeted industries; women -- into the value chain, ing the competitiveness of the
and and raising household in- overall value chain.
4. Rural household incomes. comes. A related question is
whether project facilitated in- General Hypothesis: Project
The baseline study will be fol- terventions have led to the de- activities can be effective in
lowed up in two years, with velopment of sustainable solu- the development and improve-
the major findings on the im- tions to problems faced in tree ment of sustainable solutions
pact of the two projects emer- fruit agribusiness, which are in the areas of market access,
ging after this second stage of key for sustained impact. extension services, input sup-
research. ply, and inter-firm cooperation
Our research design attempts that result in increased vertical
to address these questions by and horizontal integration of
A. KEY QUESTIONS defining suitable impact vari- MSEs into value chains and
ables and measures. These greater competitiveness of
The Kenya BDS and Fintrac flow from the causal model de- those value chains.13
HDC projects have taken on a scribed in the following sec-
twofold challenge: improving tion. 13
In general, program interventions can be
the competitiveness of Kenya characterized as potentially involving ef-
tree fruit exports in global B. THE CAUSAL MODEL forts to boost product demand, improve
markets and increasing the the business environment, strengthen ver-
tical and horizontal linkages, and/or im-
participation of smallholders in The study design is based on a
prove private sector supply response. The
the tree fruit value chain. In causal model of impact that Kenya BDS and Fintrac projects emphas-
this context, key questions fa- shows how project facilitation ize linkages and supply response. Except
cing both projects is whether activities to promote sustain- for Fintrac interventions around EUREP-
Kenya can stay competitive in able solutions/services can ad- GAP, they seem to do little to improve the
global tree fruit markets and dress constraints to smallhold- business environment for tree fruits in
at the same time maintain a er participation and competit- Kenya (possibly because major constraints
in this area were not identified).
27
is will compare a sample of
Hypothesis 1: Project activit- Hypotheses 3: Greater integ- smallholders who participate
ies to promote sustainable ration of smallholder MSEs into in the Kenya BDS or HDC pro-
solutions in the tree fruit value productive value chains con- ject with a sample of compar-
chain contribute to greater in- tributes to improved competit- able smallholders who do not
tegration of smallholder MSEs iveness and growth of the tar- participate in either of these
into the value chain through: geted value chains. projects. It will study changes
 Strengthened vertical link-  Increased production by the associated with participation in
ages permitting increased value chain as a whole one of these projects in the fol-
market access for smallhold-  Increased average productiv- lowing domains of impact.
er MSEs producers ity
 Improved/increased inter-firm  Increased share of production Smallholder MSE Integra-
cooperation/collaboration marketed tion into Productive Value
(horizontal linkages)  Increased share of production Chains
 Improved supporting markets exported This will involve assessing and
o Increased use of appro- comparing changes over time
 Improved inter firm coopera- in smallholder MSE participa-
priate inputs (agro- tion (horizontal and vertical
chemicals, plant stock, tion in the tree fruit value
coordination and business ar- chain.
and other supplies) rangements)
o Use of higher quality in-
Integration into the tree fruit
puts D. FRAMEWORK OF ANA- value chain will be measured
 Improved/increased quality LYSIS by the volume and percentage
and quantity of extension, of production that is marketed,
advisory, and information As indicated earlier, we will the average price received for
services provided by lead study impacts at four different marketed output, and thus
firms (embedded) and fee- levels: participating smallhold- sales value. The study will fo-
based providers er MSEs; their associated cus on access to and use of
households; the tree fruit market information and sales
Hypothesis 2: Greater integ- value chain; and the provision to different market outlets.
ration of smallholder MSEs into of sustainable business solu- Other issues related to small-
the tree fruit value chain con- tions. At each level of analysis, holder integration into value
tributes to improved enter- we have identified several do- chains will be explored
prise performance and house- mains of impact and indicators through qualitative, in-depth
hold well-being through: of change, as shown in Figure interviews with smallholders,
 Increased production in parti- 2. for example, whether and how
cipating enterprises participation in producer
 Increased revenues in parti- 1. PARTICIPATING TREE groups provides advantages to
cipating enterprises FRUIT SMALLHOLDER MSES smallholders; the extent to
 Increased employment and which access to new market
employee earnings in parti- The study will focus on the outlets changes smallholder
cipating enterprises main intended beneficiaries of relationships with brokers and
 Increased income in particip- the two projects in Eastern, the implications of this over
ating smallholder MSE house- Central, and Rift Provinces, time; the nature of smallholder
holds namely smallholders who grow relationships with lead firms or
mangoes, avocados, or pas- other buyers or suppliers
 Reduced vulnerability
sion fruit for consumption or providing embedded services.
through diversification of in-
sale14. The smallholder analys-
come sources in participating
smallholder MSE households 14
Further details are given in the sample
Participation of lead firms,
design section. brokers, and other buyers in
28
marketing smallholder produc- Enterprise Production Pro- sumption (as a proxy for in-
tion and their experience in cesses and Performance come), changes in sources of
providing embedded services Changes in production pro- household income ranked by
also will be examined through cesses will reflect changes in importance (including tree
qualitative interviews. skills, knowledge, and prac- fruit income) and changes in
tices related to tree fruit pro- household assets. The use of
duction and processing (plant several variables allows for tri-
husbandry, use of agrochemic- angulation in assessing
als, etc.); use of market in- changes in household well be-
formation; use of technologies; ing.
and capital investments (e.g.,
tools and equipment). Meas- Increased household incomes:
ures of enterprise performance The measurement of house-
will include production, pro- hold income
ductivity, employment, and
technologies used. Production
is the total amount of fruit pro-
duced in a season. Productivity
will be measured by quantity
produced per hectare or per
tree (the more appropriate
measure is to be determined)
over a season. Employment
will be measured by the repor-
ted person/days of hired labor
used for tree fruit production,
harvesting, processing or sale
over a season. Technology use
will be measured by the use of
planting stock and inputs, and
the watering system.
2. SMALLHOLDER MSE
HOUSEHOLDS
This part of the impact assess-

ment focuses on the house-
holds of smallholder MSEs par-
ticipating in the projects and
households of smallholder
MSEs not participating in the
projects.15 It studies impacts
on household well being using
a combination of variables:
changes in household con-
15
To the extent possible, we will also fo-
cus on smallholder MSE employees – de-
pending on whether this emerges as a sig-
nificant group among participating small-
holder MSEs.
29
FIGURE 3: CAUSAL MODEL FOR KENYA BDS AND FINTRAC HDC PROJECTS
Pre Intervention
Project Activities Outputs Outcomes Impacts
Activities
Facilitate integration Market Access

Select tree fruit into value chains by: Increase in sus-
sub-sectors – tainable market
mango, passion 1) forming/linking pro- outlets for mango, Sub-sector perform-
fruit, avocado ducer groups with lead passion fruit, and ance
Increased
firms, promoting inter- avocado produ- participation Growth in sales,
Analyze con- Improved house-
firm collaboration, and cers of smallhold- productivity and
straints and op- hold incomes for
strategic alliances (ver- ers in high- trade in overall
portunities in sub- mango, passion
tical and horizontal Training and Ex- value por- mango, passion
sectors fruit, and avocado
linkages) tension tions of fruit, and avocado smallholders [and
Increase in the mango, pas- sub-sectors
Identify priority for MSE employ-
2) upgrading through provision of com- sion fruit,
solutions/services ees in mango, pas-
the promotion of commercially viable
and other needs and avocado Firm level perform- sion fruit, and
mercially viable busi- extension (e.g.,
for mango, pas- value ance avocado sub-sec-
ness services (private training, technical
sion fruit, avocado chains/mar- Increased sales, pro- tors]
extension agents, agro- assistance, advis-
sub-sectors kets ductivity, and trade
chemical stockists, em- ory services, in-
for participating Increased remu-
bedded services by lead formation ser-
Design interven- Improved smallholders in nerative employ-
firms, private nurseries, vices, and new
through tions and compete competitive- mango, passion ment
training and registration technologies) to
the and award tenders in EUREPGAP/SPS) smallholder ness in the fruit, and avocado
mango, passion entire value sub-sectors
Improve enabling envir- fruit, and avocado chain record-keeping; and respondent mis-
onment, especially in producers reporting due to recall error, mis-
end markets come in
under-
Input Supply rural
standing, or
Increase in com- settings
mistrust.
mercially viable poses diffi-
provision of in- cult challenges,
puts (e.g, agro- however: the ex- Consumption is considered by many to be
baseline survey is clearly important. After ist- chemical supplies, ence of multiple a more reliable measure of household
the follow-up survey round, we would like in- planting materi- come sources; economic status than income in contexts
to be in a position to say whether the two the als)
importance of in- like rural Kenya. It is seen to be less sub-
projects helped to raise the house- come in ject to measurement error
hold incomes of project parti- kind; irregularityPhysical,
in income than income, and is a better proxy meas-
cipants. Measurement of household in- Social,flows; lack of
and Economic Context
ure not only of current welfare but long-
run wealth. The study uses the following ownership of selected assets relevant to cioeconomic status are considered
consumption/expenditure measures of rural Kenya that will be used to construct throughout.
household income: consumption in the an asset score. This includes some assets
last seven days of items grown at home; that are likely to be responsive to short
expenditure on education in the last 12 term changes in household income. We
months; and other expenditures over the use the asset score for two purposes: (1)
last four weeks including, for example, to assess the impact of project participa-
vegetables, meat, packaged food, grocer- tion and increases in tree fruit income on
ies, cooking fuel, transport, and commu- household assets and (2) as a proxy to
nication.16 These measures are then com- determine the relative wealth level of
bined to create a proxy for household in- households in the sample.
comes.
To complement the survey data on
Reduced Vulnerability households, the study also includes in-
An important dimension of poverty is vul- depth interviews with a small number of
nerability, which has been defined as the smallholders to explore the role of tree
exposure to and the capacity to manage fruit income in households, and (at the
risk. Diversification of income sources is end line) the implications of additional
one way poor households manage risk by tree fruit income for household well-be-
‘income smoothing,’ or evening out sea- ing. The in-depth interviews explore this
sonal fluctuations of income throughout source of additional income for income
the year. The survey includes a short set smoothing, control and use of tree fruit
of questions about sources of household income within the household, labor alloc-
income and the relative importance of ation related to tree fruit production, and
each source of income, including tree quality of employment issues for those
fruit income. This allows us to assess di- involved in tree fruit production. The in-
versification, income smoothing, and the depth interviews also explore decision
role of tree fruit income in this process. processes and incentives at the house-
Assets also help households to manage hold level related to participation in the
risk by providing a store of wealth to tree fruit value chain. How the broader
draw upon in times of need or opportun- portfolio of household economic activities
ity. An increase in household assets can affects decisions related to expanding
indicate reduced vulnerability. The survey tree fruit production, switching from an-
also includes a set of questions related to other cash crop to fruit trees, or selling
16
tree fruits to new market outlets will be
These indicators have been used in the Kenya Welfare explored. Differences by gender and so-
Monitoring Survey (1999)
3. TREE FRUIT VALUE CHAINS cipants, no valid control group can be be combined with information gathered
constructed for the value chain as a from stockists and other sellers of inputs
The next level of analysis to be ad- whole. Although determining attribution to smallholders. This will include study of
dressed is the tree fruit value chain, com- will thus be difficult at this level, develop- the provision of embedded business ser-
prising all producers of mangoes, avoca- ing the entire value chain is an important vices to smallholders by input suppliers.
dos, and passion fruit in Kenya (or, altern- objective of the two projects and thus
atively, based on data availability, in the cannot be ignored. Some of the growth in Improved/expanded training and ex-
areas covered by the two projects). production, income, and other impact tension, advisory, and information
variables experienced by non-participants services will be measured through ana-
Production processes in the tree fruit within the value chain will in fact be at- lysis of the use of training, extension, ad-
value chain will be gauged by total pro- tributable to the projects through visory, and information services by small-
duction of the three tree fruits, average spillover effects, but this will be difficult holders, how much they pay for these
productivity, employment, and the tech- to identify and measure. services, and the types of agents who
nologies used. provide extension and training on either a
freestanding or an embedded basis:
Integration of MSEs into the value chains 4. PROVISION OF SUSTAINABLE  Private extension officers (offering ex-
will be measured through the value and SOLUTIONS tension services related to plant hus-
volume of sales to export and domestic bandry, application of agrochemicals,
markets. Changing marketing channels The provision and use of sustainable solu- organization of producer groups, and
for both exports and domestic sales will tions will be measured in a variety of business management training)
be examined. To determine what is going ways. Improved market access will be  Farmer-led extensions teams
on in the marketing process, information measured based on increased/improved  Lead firms or suppliers providing em-
will be collected not only from smallhold- market linkages between bedded extension services
ers (to whom do they sell their products, smallholder/MSEs and their buyers. The
at what price, etc.), but also those who  Stockists and other input suppliers
study will assess the extent to which
buy from smallholders as well as from providing embedded extension/training
MSEs currently benefit from market link-
buyers and sellers at higher levels of the services
ages and to what extent those market
marketing chain including lead firms. The linkages are increased or improved over  Agrochemical distributors supplying
inquiry will investigate the nature and ex- the life of the projects. stockists
tent of embedded services as well as  Nursery operators/seedling suppliers
commodity transactions. Improved provision and use of agri-  Financial brokers or financial service
cultural inputs will be measured by providers
While analysis at the MSE and household total usage of agro-chemical inputs, im-  Providers of market information
levels will compare the results achieved proved planting stock, and other sup-  Organizers of producer groups
by program participants with those plies. Information gathered from small-
achieved by a control group of non-parti- holders (about their input purchases) will
The sustainability of services/solutions17 period on the production and sale of adapted to accommodate the specific
will be determined by assessing whether mangos, passion fruit, and avocados. fruit (for example, measures of productiv-
or not commercial relationships are intact ity, descriptions of varieties, or sales out-
at round two. The study will also consider The smallholder survey is quasi-experi- lets). They were translated from English
whether someone is providing a fee- mental in design, with data collected at into one local language – Kikuyu in Cent-
based or embedded solution and whether two points in time, two years apart, on a ral province. In Eastern province and Rift
someone is using it. sample of participating and non-particip- Valley province, Kiswahili was used. The
ating smallholders in tree fruit agricul- questionnaires took approximately 50
Figure 2, below, summarizes levels of ture. This longitudinal quasi-experimental hours to administer.
analysis, domains of impact, indicators of design allows for a comparison of
change, and sources of information. changes over time in enterprise and
household level variables between parti-
cipating and non- participating smallhold-
ers. The difference between participants
E. DATA COLLECTION STRATEGY
and non-participants at the end of the
study will indicate the impact of the pro-
The baseline study included: (1) a quant-
ject on variables studied.
itative component involving survey of
smallholders and review of secondary
A study team, led by Research Interna-
market information and (2) a qualitative
tional Kenya, carried out the first round of
component consisting of focus group dis-
data collection between November 2004
cussions and in-depth interviews with
and January 2005.
actors in the tree fruit value chain.
Questionnaire Design
1. SURVEY OF SMALLHOLDERS
The survey questionnaires asked about
sources of market information, use of
The quantitative component of the study
capital, labor and material inputs, parti-
involves (1) a longitudinal survey of
cipation in training, use of extension ser-
smallholder MSE tree fruit producers with
vices, market linkages, productivity, em-
plans to collect data in two rounds, two
ployment and income. Household level
years apart and (2) a review of secondary
questions focused on changes in house-
market level information during this time
hold income, consumption expenditures,
and assets. A similar set of questions and
17
indicators was used across the three
Sustainability of services/solutions is defined as the fruits and five sub-projects. Where neces-
ability of the services/solutions to be kept going over
sary, some of the questions were slightly
time without donor subsidy.
Figure 4: Framework for Studying Impacts
Levels of Domains of im- Indicators of change Sources of in-
analysis pact formation
Tree fruit Increased integra- Increased sales/market link- Survey
Smallhold- tion of smallholder ages Case studies
er MSEs MSEs into tree fruit Increased price received In-
value chain creased marketing channels
used Increased/improved use
of agricultural inputs
Increased/improved use of ex-
tension services
Improved produc- Skills, knowledge and practices Survey
tion processes Use of market information Case studies
Use of technology
Capital investment (tools and
equipment)
Improved small- Increased revenues Survey
holder MSE per- Increased productivity Case studies
formance Increased employment
Smallhold- Increased incomes Proxy measure of increased Survey
er MSE household income (consump- Case studies
House- tion/expenditure)
holds Higher ranking of tree fruit in-
come as source of household
income
Reduced vulnerab- Diversification of household in- Survey
ility come sources Case studies
Income smoothing
Increased assets
Markets Provision of sus- Improved and sustainable mar- Survey

tainable solutions ket access Secondary market
to recurrent con- Improved and sustainable in- level information
straints of MSEs in put supply Interviews with
the value chain Improved and sustainable ex- buyers (brokers and
tension, advisory, and informa- lead firms), input
tion services suppliers, extension
service providers
Growth of tree fruit Increased production Secondary market
value chain Increased productivity, level information
Increased employment In- Interviews with
creased sales buyers (brokers and
Increased exports lead firms)
Improved inter-firm
Collaboration
Survey Table 2: Baseline Sample Size Ideal Business Link be-
Sample Tree Fruit Test Control Total cause the wide geo-
Design Avocado 250 250 500 graphic dispersion of
The ori- Mango 420 349 769 trainees and indirect link
ginal Passion Fruit 354 324 678 to smallholders made
sample Total 1,024 923 1,947 sampling problematic;
design and CDA because its re-
was based on the total population of mote location in the Tana River district of
smallholders participating in (or targeted Coast province made data collection very
by) the Kenya BDS tree fruit and Fintrac expensive and logistically difficult, espe-
HDC passion fruit activities by degree of cially in the rain.
participation (based on project docu-
ments and discussions with project staff, The original sample design was to include
and discussion with intervention part- 1,380 participants and 1,380 non- parti-
ners) and the population of smallholder cipants, but the final sample size was re-
farmers in similar geographic areas, but duced to 1,024 participants and 923 non-
-- as far as possible -- outside the sphere participants for a total sample size of
of influence of the project. 1,947. The reduced size was the result of
a decision to exclude project participants
From these populations a sample of parti- on the list who did not already have some
cipant and non-participant tree fruit farm- passion fruit growing. (Many intended to
ers was drawn. Test respondents were start growing at some point in the future,
drawn from participant lists provided by but had not taken action to do so.) Those
Fintrac HDC and Kenya BDS. This in- included in the sample were all of the
cluded participants covered by Fintrac farmers on the passion fruit project lists
HDC’s passion fruit work out of their El who already had at least five vines of
Doret office; and eight Kenya BDS sub- passion fruit growing. The size of the con-
components. These lists covered avo- trol group was also reduced because the
cado, mango, and passion fruit farmers in field teams could not find enough passion
Central, Eastern, Rift, and Coast fruit growers in the relevant matched dis-
Provinces. In finalizing the sample design, tricts.
the project team decided to eliminate
four Kenya BDS sub components from the The study sample covered a total of 8
study -- KACE and Kenya Gatsby districts, 33 divisions, and 191 villages in
Trust/KWETU/KARI because they were not Central, Eastern and Rift provinces as fol-
moving forward in their implementation; lows:
For Kenya BDS, the sample includes parti- understand the context of enterprise, questions focused on understanding in-
cipants in four sub-projects. These in- household, and market level impacts and centives and risks for smallholders asso-
cluded: participant avocado farmers in improve understanding of factors that en- ciated with upgrading and accessing new
Kandara Division and non-participant courage or inhibit the integration of markets. It also considered the incentives
farmers in areas of the same division; Kenyan smallholder SMEs into the tree and risks for exporters and input suppli-
participant mango farmers in Makueni, fruit value chain. Interviews were conduc- ers to provide solutions/services to small-
Machakos, and Mbeere and non-parti- ted at the baseline and will also be con- holders. It looked at the nature of cooper-
cipants in divisions of Makueni District ducted in round two. The qualitative ation and coordination among actors
that the project had not reached but baseline findings have been summarized within the value chain as it relates to
where mangos are grown; and passion in analysis tables and these findings will smallholder participation and competit-
fruit growers participating in the project be compared to the findings from qualit- iveness, specifically, the extent to which
in Embu District and non-participating ative research addressing similar ques- lack of trust, power asymmetries, and
farmers in neighboring Kirinyaga District. tions in round two. cultural biases may be affecting small-
Ministry of Agriculture and project offi- holder participation, upgrading, and mar-
cials were consulted in the selection of The baseline qualitative research was ket linkages. The study also explored
non-project, but similar, fruit-growing carried out over seven days in October household level factors that affect small-
areas. For Fintrac HDC, participant farm- 2004, just prior to implementation of the holder participation in the value chain.
ers were selected from Uasin, Gishu, and baseline quantitative survey. It involved
Kieyo Districts, while non-participants in-depth interviews and focus group dis-
were from non-project divisions of the cussions with 50 actors in the avocado,
same districts. Selection of the non-parti- mango and passion fruit value chains.
cipant sample continued until the survey This included 30 smallholder farmers, six
teams could not locate any more eligible producer group leaders, four input suppli-
respondents in these regions. ers/stockists, one nursery operator, four
extension agents, three brokers, and
two exporters. They were all linked to Table 3: Geographic Location of
2. QUALITATIVE STUDY OF THE TREE three sub-projects covered by the Sample
FRUIT VALUE CHAIN quantitative survey: EAGA (avocado), Tree Dis- Divi- Vil-
SITE (mango), and Fintrac (passion Fruit tricts sions lages
The qualitative component of the impact fruit) . The qualitative research also Avo-
18 1 1 30
assessment includes interviews with a involved interviews with the directors cado
small sample of value chain actors: small- and field staff of the two projects. The Mango 3 17 57
holder MSEs, input suppliers, service pro- Pas- 4 15 104
viders, lead firm exporters, other buyers, 18
The brokers interviewed for the qualitative research op- sion
and producer group leaders. The purpose erated in the geographic areas covered by the EAGA sub fruit
project and used to buy avocados from project parti- Total 8 33 191
of the qualitative component is to help
cipants.
III. BASELINE RESEARCH FINDINGS
While mangos have long been grown for recorded by HCDA in 2003, and 14 per-
A. THE TREE FRUIT VALUE CHAINS home consumption, new export varieties cent of all recorded horticulture exports.
were introduced in the 1980s. Avocados Moreover, avocados and tree fruits in
Within the horticulture sector in Kenya, were introduced primarily as a cash crop general are growing in importance in the
avocados, mangos, and passion fruit are for export in the 1990s. Passion fruit has export market.
among the most common fruit crops a long history of cultivation in Kenya. It
grown. Mangoes are grown primarily was introduced in the first half of the 20th Kenya is not among the main suppliers of
along the coast and other low land areas, century but due to disease never took off avocados in the world market. The top
while avocado and passion fruit are cul- as an export crop. It is grown primarily for suppliers in rank order of importance in-
tivated primarily in the highlands. While home consumption, with some limited clude Mexico, Chile, South Africa (the
most of these fruits are produced for sales in domestic, regional, and export main competitor for European and Near
home consumption or sale in traditional markets. East markets), Spain, Israel, and a mix of
markets, recent data suggest their grow- other countries such as the Dominican
ing role among exports and as a source 1. DEMAND Republic and Indonesia. To give a relative
of foreign exchange. These three fruits sense of volumes produced and exported
make up the top three fruit exports from Smallholders are the by far the major pro- in 2003, Mexico, the top supplier, pro-
Kenya – comprising 85% of fruit exports ducers of tree fruits in Kenya. Most of the duced approximately one million metric
in 2003 (Table 4). Hass avocados are ex- avocados, passion, and mangos produced tons and exported 135,000 tons; South
ported primarily to Europe (many are by smallholders continue to be sold in do- Africa produced 70,000 tons and expor-
shipped to Marseilles for sale in France mestic markets. However, all three fruits ted approximately 38,000 tons; and
and Germany); Fuerte avocados to the are important exports crops and they are Kenya exported approximately 20,000
Middle East; mangos to Dubai; and pas- growing in importance.19 Some points metric tons (USDA, 2005)
sion fruit to specialty markets in Europe, with regard to the demand for Kenyan
especially France. On average, growth in tree fruits in the export markets: The export market for avocados is large,
the volumes and earnings for fruits is but competitive. In the past, Kenya has
higher than for horticulture as a whole; Avocado supplied primarily Fuerte avocados, a
tree fruits in particular are seen as an al- Avocados are an important horticultural large, smooth skinned variety with some-
ternative to primary commodities like cof- export crop in Kenya, comprising approx- what limited demand to the Middle East.
fee and tea as export crops, given uncer- imately 75 percent of fresh fruit exports However, Hass avocados, which are a
tainties and trends that place Kenya at 19 smaller, rough skinned variety, are
Between 2002 and 2003, HCDA data shows that the
risk in these markets. volume and value of tree fruit exports in Kenya grew at favored in the European export markets
Exports of mangos and avocados started higher rates (17 percent and 36 percent, respectively) because they are disease resistant and
in the 1980s and in 1990, respectively. than the rates for horticulture exports as a whole (10 per- less risky to ship.
cent and 8 percent, respectively) (HCDA 2003)
Table 4: Export Statistics for Avocado, Mango and Passion Fruit, Kenya 2003
Fruit Volume Value
Kgs. Metric Tons KSH USD*
Avocado 19,020,028 19,020 892,871,043 11,787,076
Mango 2,226,550 2,227 273,611,917 3,612,039
Passion Fruit 1,505,630 1,506 393,783,047 5,198,456
All fruit 60,982,885 60,983 1,752,645,572 23,137,227
All horticulture 133,232,517 133,233 28,839,583,186 381,380,063
Source: GOK, 2003 * Ksh. 75.75 = USD$ 1 as of Dec. 31, 2003
According to one main exporter (EAGA), local investor is re-opening a factory that major player. Kenyan mangos have a
one million boxes of avocados are expor- will process avocado oil in Central somewhat flawed reputation because of
ted from Kenya each year and they sup- Province (in an area not far from Nairobi poor quality. Deliveries of diseased and
ply one quarter of this amount: 250,000 and close to the avocado groups facilit- spoiled fruit in the past have made buy-
boxes per year, including 20% Haas and ated by Kenya BDS). It is anticipated that ers wary. Improved, disease resistant
80% Fuerte varieties. Haas avocados are this firm will buy lower grade avocados varieties were introduced in Kenya about
sold primarily to Europe and Fuerte are not suitable for export or sale in the do- 8 years ago, but are still relatively new.
sold primarily to the Middle East. Accord- mestic fresh fruit market. This will Kenyan mangos have not yet made major
ing to this exporter (EAGA), Kenya avoca- provide an incentive to smallholders to inroads into the European markets and
dos have a tarnished reputation in export grow avocados and increase their sales are sold primarily in the Middle East.
markets due to past shipments of dis- and income. However, these countries place restric-
eased fruit and mistiming of shipments tions on mangos from Kenya due to a his-
that resulted in fruit being either under- The main competitive advantage of tory of weevils. If they find one weevil in
ripe or spoiled. EAGA, a large horticul- Kenyan avocados is their availability in a container, the whole container is
ture export firm in Kenya, sells mostly to late January and February when the sup- thrown out and exporters loose the value
the speculative market and, more re- ply of avocados from other countries is of the entire container.
cently, to wholesalers with links to super- low (counter-seasonal supply). It has a
markets. It is also penetrating fair trade four-week jump on South African avoca- Nevertheless, there is potential for fur-
markets, where the price premium is 12 dos – a main competitor -- in the world ther development of mango exports, es-
percent. markets. pecially for the apple mango. There also
appears to be potential for developing
The local market, to date, has primarily Mango the industry through branding; however,
involved wholesale and retail fresh fruit In the context of a highly competitive these efforts are still very under-
markets. In addition to these markets, a world market for mangos, Kenya is not a developed.
HIV/AIDS. RI staff in Kampala interviewed able from other suppliers. Another ad-
Passion fruit brokers selling Kenyan passion fruit in the vantage that Kenya has in horticulture in
The passion fruit market is unique among local market and found it to be a highly general is that it is better set up to meet
these three fruits in that there is unmet lucrative market. The Ugandan brokers certification standards than other coun-
demand in both domestic and export are seasonal buyers – they come only tries. Considerable efforts have been
markets. At this point, while passion fruit during the off season for passion fruit in made to encourage exporters to work
is grown in countries throughout the Uganda and Rwanda. with smallholders through MOA, HCDA
world, no one country appears to have
the competitive edge with passion fruit. HCDA buys the highest grade passion fruit, which it sells to exporters in Nairobi.
Zimbabwe was a competitor in producing
and exporting passion fruit in previous and a number of bilateral donor-suppor-
years, but land redistribution activities in In terms of domestic demand, producers ted projects.
recent years have disrupted production. historically supplied passion fruit to a
Both fresh and processed passion fruit government-supported passion fruit pro-
(e.g., juice concentrate) are seen to have cessing factory located first in Kitale, 2. SUPPLY CHAINS – PRODUCERS TO
significant, but as yet untapped, potential then in Sotik, and finally in Thika before it MARKETS
in world markets. It is a specialty market, closed down a few years ago in part due
however, and considered to have less to lack of supply. This may in part relate Producers to markets: overview of the
mass based appeal as a fresh fruit than to the fact that prices in export markets tree fruit value chain
as a pulp used in marinades, sherbets, are much higher (Ksh. 50/kg) than local
and juice. Passion fruit has many positive prices for juice grade passion fruit (Ksh. Figure 3 shows key actors in the tree fruit
attributes: it is easy to ship and a high- 7-15/kg). Moreover, passion fruit has not value chain in Kenya and how they are re-
value crop. Most of the fresh passion fruit been actively promoted. lated.
exported from Kenya goes to Europe, es- Tree fruits are produced primarily by
pecially France. In sum, the competitive advantage of smallholders and medium-sized growers.
Kenya tree fruits is not volume, quality, or Small-scale chemical stockists (referred
The research found a regional value chain price but rather seasonality.20 Avocados, to as agro-vets) sell fertilizers, insect-
for passion fruit extending from Kenya to mangos, and passion fruit each have a icides, and other chemicals to producers
Uganda. Kenyan passion fruit is in high ‘window’ when these crops are less avail- through privately owned shops located in
demand in Uganda at certain times of the 20
towns throughout Kenya. Many of these
See Steven Jaffee (2003). “From Challenge to Oppor-
year (October through December). There shops also offer advice to farmers on
tunity: Transforming Kenya’s Fresh Vegetable Trade in
is a ready market in Kampala for fresh the Context of Emerging Food Safety and Other Stand- what chemicals to use as well as their
fruit and juice sold in hotels and restaur- ards in Europe.” Agriculture and Rural Development Pa- proper application and safe handling. Big-
ants. Household consumption is also up per 1. World Bank; Steven Jaffee and Spencer Henson name chemical wholesalers (such as Bay-
in Uganda because it is seen as a healthy (2004). “Standards and Agro-Food Exports from Devel- er) provide periodic training for stockists
drink, especially for people affected by oping Countries: Rebalancing the Debate.” World Bank and farmers to raise awareness of the be-
Policy Research Working Paper 3348.
nefits of using their products. This train- bedded services to smallholders through Historically, smallholders in Kenya have
ing is provided through existing farmer these groups. grown tree fruits largely for home con-
groups and meetings organized by lead sumption or small-scale local trade. They
farmers in local communities. Publicly Key Problems in the Tree Fruit Sup- have not considered tree fruit a major
supported agricultural research centers ply Chain at the Time of Baseline Re- cash crop or a business activity to invest
and government organizations, such as search in. A repeated theme in the qualitative re-
the Kenya Agricultural Research Institute search is that these crops that have not
(KARI), the Horticulture Crops Develop- Constraints to smallholder participation in been taken seriously in the past.
ment Authority (HCDA), and the Ministry the tree fruit value chain include:
of Agriculture (MOA) produce and distrib-  Lack of information and knowledge of Avocado
ute tree fruit plant stock. They link to the markets Avocado production in the Kandara area
farmers through demonstration centers  Limited access to inputs experienced almost total collapse in the
and farmer groups. Private nurseries also  Limited smallholder access to business 1990s as a result of several factors:
have begun to produce and sell tree fruit solutions and services  Disease, specifically anthracnose, which
plant stock.  A long and inefficient supply chain with causes black spots on the skin of avoca-
poor vertical and horizontal linkages. dos
The main marketing outlets for tree fruit Small producers are not well organized  The collapse of the government para-
producers are traders and brokers, who in in terms of access to inputs and mar- statal that bought avocados prior to the
turn sell to both domestic and export kets; buyers face high transaction costs liberalization of the economy
markets. In the domestic market, they when purchasing fruit from dispersed  The related collapse of a private sector
sell to domestic wholesalers, fresh fruit smallholders. plant to process low grade avocados
retailers, and small retail shops. Produ-  Lack of trust among producers, brokers, into oil
cers also sell directly to fresh fruit retail- and exporters, related to problems in  A tarnished reputation in export markets
ers and to HCDA. Brokers and traders are the past in the enforcement of supply due to the poor quality of Kenyan avoca-
the main conduits for smallholders to the contracts; from the perspective of pro- dos and the mistiming of deliveries res-
formal and informal processing plants ducers, buyers are not always depend- ulting in spoiled fruit.
and to exporters who buy tree fruits. Me- able or honest.
dium-scale growers often link directly to
processing plants, local supermarkets, These constraints result in low yields (es-
and exporters. Exporters buy tree fruits pecially of export quality varieties), low
produced by smallholders and medium- quality, low sales volumes, low selling
scale farmers and also produce some prices, high rejection rates, and excessive
fruit on their own plantations.21 Exporters post harvest waste – all of which affect
have just recently begun to buy directly the competitiveness of Kenyan fruits in
from producer groups and to provide em- export markets.
21
We were able to get very little information on this.
Markets Figure 5: Kenya Tree FruitDomestic
Domestic Value Chain22 Export
Brokers continued to purchase avocados Fresh fruit
grade due to lack of oil processing
Processed ards Fresh
(chemicals, protective clothing, dig-
during these years, but in limited facility ging wells, insecticides, sprayers.
volumes and at very low prices (50 cents
Functions  Limited emphasis on inputs and services  Limited supply of newer varieties with
to Ksh.1 per piece). Because of disease related to avocado production and up- more market potential
Exporting
and neglect of trees, the fruit produced grading by government or private sector  Limited incentives among mango produ-
was poor quality. Most farmers grew  Limited information Small
on how to prevent cers because of low prices and erratic
Fresh
Fuerte avocados, which are in demand in
Retailing disease
Fruit Retailers
Retail Supermarkets
market
Shops
the Middle Eastern markets, but less so in  View among producers that avocados do
European markets not have good potential as a cash crop. A mango producer group leader in Masii
Informal
describes the problem of linking produ-
Exporters
small-
Resulting problems in the supply chain in-
Wholesaling Mango Wholesale
markets
scale
cers to buyers:
Process
clude: Mango production in the areas ofLarge
Eastern
Scale ors
 Supply of low quality fruit by producers province covered by the study isProcessors
spread
Brokers/Traders
… if you get a good buyer like the one I
Medium Scale Growers

Small Scale Growers
(diseased or damaged due to “climb
Processing out. Marketing is a key problem facing said [Mombassa exporter] then it’s a
and shake” picking practices) producers. Farmers have few market out- challenge because they [the farmers]
<400 acres
<50 acres
 Supply of ungraded fruit lets and are dependent on brokers who cannot meet the required standards.
Assembly
 Supply of Fuerte avocados, a variety come to their farms to pick the fruit. They When you try to tell them to meet those
more vulnerable to disease and with buy at low prices, only buy the best fruit, standards, it’s when they give you
Grading things like, "we cannot afford these
little demand in the European export and do not sell on contracts. In general,
chemicals, and we can’t afford these
market the mango farmers in these areas lack standards because we have no money.”
 Limited incentives for Production
producers to up- bargaining power with brokers and feel So you get a challenge because you are
grade exploited by them. torn between two forces one is the ex-
because pectation from farmers, which they ex-
Input Supply Chemical
of low Avocado: “I believe there is a Key problems
Private in the mango
Stockists supply pect you to find them good market and
farm way they [brokers] talk to these chain include:
Nurserie
KARI when a buyer who needs standards
Nurseri
gate exporters in aExtension
way that we do  Droughts and lackes of irrigation Product
comes he finds that farmers cannot
F- meet the standards you see the two
prices not…. So they could even buy  Disease (parast MOA Research PEA
paid by from us at a very R&Dlow price telling atal) (HCDA) Uof N, forces. (a lead farmer, Masii)
us that the market is bad. Brokers
 High price of chemicals to improve Utaali
K
brokers quality

created a barrier between the ex- Passion fruit
porters and us, so we as farmers  Lack of credit to help cover costs of
Passion fruit is widely produced, but
Wastage became like blind men because complying with international stand-
of lower we could not tell what the export- treated primarily as a garden crop for
LEGEND
22 er had to say.”
From: Kula, Olaf, “Activity Status Report: Holding Hands with Folded Arms: Upgrading Kenya Tree Fruit Value= Chains”.
Domestic Deloitte, Touche, Tomatsu,
Final Market
No date. -Kandara Wholesale
avocado grower Fresh
markets = Participant in
the Value Chain
Brokers
= Participant in the
Value Chain, broken Exporting
= Transactions = Market functions
line indicates
skipped functions
= Services not captured in = Pillar Three
value chain Industry Leaders
home consumption. While there is unmet Another
demand for passion fruit in domestic, re- challenge is
gional, and international markets, the vul- the seasonal
nerability of passion fruit to disease and nature of
the lack of technical know-how to man- the market
age and prevent disease have limited the for fresh
ability of farmers to produce the volumes passion
demanded. fruit, with
fluctuating
The low volumes produced relate to a dis- demand
ease that attacks the roots of vines, re- during the year. Passion fruit has poten- 3. GOVERNANCE PATTERNS IN THE
duces yields, and kills the plants. Efforts tial for year-round production if managed TREE FRUIT VALUE CHAIN
to grow large volumes of passion fruit on properly, but farmers have little incentive
plantations in Kenya in the 1940s and to increase production due to low de- The relationships among firms in the tree
1950s were thwarted by rapid spread of mand in certain seasons. Currently, there fruit value chain reflect different types of
disease. This disease has periodically is no passion juice processing facility to governance patterns. Dunn and Villeda
wiped out Kenya’s passion fruit root buy passion fruit in the off seasons and (2005), describe three general types of
stock. Waves of disease over the years the fruit rots. governance patterns:
have relegated passion fruit mostly to a  Market relationships characterized by
dispersed garden crop (the risk disease is arms-length transactions and little in-
lessened if plants are spread out). formation exchange between firms.
Firms in market relationships may en-
Because of its vulnerability to disease, gage in repeat transactions, but their in-
passion fruit involves more technical teractions are limited to the exchange of
management than avocados or mangos, goods or services for money.
especially if the aim is to reduce disease  Network relationships characterized by
and have the plant bear fruit all year more extensive information flows
long. Grafting purple passion onto yellow between firms than in market relation-
passion root stock is one way farmers can ships. Some firms in the chain exert a
reduce the risk of disease. Another issue degree of influence or control over the
for this fruit is the safety of some chemic- operations of other firms in the chain
als that are used to control disease and (Humphrey and Schmitz 2000). Suppli-
keep the plants flowering, especially if ers in network relationships typically
there are residuals on the fruit at time of supply products according to buyers’
sale. specifications, including what is to be
produced, when it should be available, ize some other commodity value chains the exporters, who are primarily Asian, in
and how it should be produced.23 in Kenya -- for example, coffee, tea, and the right way.
 Hierarchical relationships: value-added pyrethrum, for which government mar-
functions that are vertically integrated keting boards still control procurement The Kenya BDS/EAGA project links produ-
under the ownership of a single firm. An and producer prices. As producer groups cers directly to EAGA’s export market
example of a hierarchical relationship in form to link to inputs and markets, and as through the formation of producer
a global value chain would be a retail exporters form associations, the patterns groups. The producer groups have negoti-
chain in one country that manufactures are shifting more toward network rela- ated a Memorandum of Understanding
products in facilities it owns in a differ- tionships. (MOU) with EAGA to supply avocados that
ent country. meet EAGA specifications in return for
Findings from the Qualitative embedded services (spraying, sorting,
All of these governance patterns have Research on Governance Patterns grading, and transport). This direct rela-
trade offs and one is not necessarily bet- tionship marks a significant change in the
ter than another for value chain actors. Avocado avocado value chain in Kandara. In this
Before EAGA/Kenya BDS started to work context, a number of governance issues
The tree fruit value chain is characterized in Kandara, brokers played a dominant related to trust, power asymmetries, so-
by a mix of market and network relation- role in the value chain. The relationship cio-cultural biases, and information flow
ships. The horticulture sector has had between brokers and producers was are noteworthy.
limited government involvement and arms-length, with minimal exchange of
private firms have been active and com- information and no assurance of repeat Most farmers expressed trust in EAGA as
petitive – this differs from the pattern of transactions. Producers do not fully trust an organization, but some were mistrust-
hierarchical relationships that character- brokers to be fair, to come at the right ful of some EAGA staff buyers who have
time to pick the fruit, or to come at all. In continuing relationships with brokers
23
general, they feel an imbalance in the re- (who now source avocados from other
Network relationships can be broken lationship. Brokers have a somewhat areas) and who, they say, have enjoyed
down further into modular governance closer relationship with their buyers. kickbacks from the brokers in the past.
(suppliers provide information but not There is a better flow of information re- These buyers tend to reject a lot of the
process technology), relational gov- garding product specifications and price. avocados from the groups – claiming they
ernance (buyers and suppliers rely on Many brokers have long-standing rela- do not meet grading standards; however,
idiosyncratic, face to face interactions, tionships with wholesale and export buy- farmers believe it is an excuse to reject
relationships often based on trust derived ers, and some even are said to pay kick- their fruit so they can continue to buy
from social and/or ethnic ties, spatial backs to assure access to these markets. from the brokers. Some farmers also ex-
proximity, or reputation) and captive gov- Some producers expressed the feeling pressed mistrust of other producer group
ernance (single buyer and provider of that brokers (who are ethnically similar to members, who they say are continuing to
process technology, asymmetric relation- the producers) have a knack of talking to sell avocados “secretly” to brokers when
ship). they need fast cash. This is in violation of
the MOU. Some growers say they do not older members, primarily retired men. In from an area not far from Kandara, is well
fully trust the reliability of EAGA to spray many cases a man is registered as the respected by all despite her age and
on time, pick up on schedule, or use con- member but his wife is the active group gender. At this point, there appears to be
sistent grading standards. In addition, participant. Payments are made in the cautious optimism on the part of EAGA
during the first year, the growers were name of the man. Older members com- and the farmers about the direct linkage.
not all clear as to how the prices for their mented that young women do not parti-
fruit would be determined – they thought cipate because of time constraints and Mango
the prices would fluctuate according to because they consider farming and farm- Relationships in the mango value chain
the market when they were actually fixed er groups as activities for older women. had not significantly changed as a result
for the season. of the project at the time of the baseline.
In terms of information flows, the biggest Some farmer groups had formed in the
Issues of trust also arise in the relation- complaint was from farmers who believe project areas, but they had not yet forged
ship between industry leaders (in both that they did not have good information new linkages to input, extension, or
avocados and mangos) and government. on export prices and that both Kenya BDS product markets. Most growers were
Although they communicate with each and EAGA were holding back price in- selling their mangos individually to a
other, their relationship has been de- formation. Some farmers also felt that in- wide array of unknown brokers and
scribed as fragile. This is due in part to formation on the MOU negotiations re- traders – suggesting a value chain gov-
the widely held belief that private sector garding prices between the groups and ernance pattern of “market
leaders look after themselves, while the EAGA did not flow from producer group relationships.”
government looks after smallholders leaders to members. This was an issue
(Kula, no date). Some producers believe during the first year of the project. The qualitative research found a high de-
that EAGA buyers change grading stand- gree of mistrust in brokers among mango
ards based on the volumes they need at In sum, there is a moderate level of trust farmers. Brokers were described as “out-
a particular time, not a standard protocol. in the avocado value chain between siders”, “little known, and “not to be trus-
Kandara producer groups and the export- ted.” Some farmers felt brokers lied to
In terms of power asymmetries, both the er. The balance of power is still with the them about what they were paying other
avocado producers and the brokers feel buyers, but growers are slowly increasing farmers.
that exporters call all the shots in terms their power. Men tend to dominate the Most Mango “… The market pay-
of volumes purchased, prices, grades, leadership of the producers groups, but has just become bad … for ments
timing, and contract terms. women are active members. While there from things sold outside to
was some sense among members that people who are strangers…
In socio-cultural terms, the producer they do not know everything that is going it is like they are mixed up
groups and brokers are ethnically homo- on in their groups, most expressed satis- and one can detect some
geneous but the exporters are primarily faction with the groups and how their fishy business in them as it
Asian. Age seems to be a factor – the pro- leaders managed them. Moreover, the is not straight.”
-Mango Farmer, Masii
ducer groups tend to be dominated by Kenya BDS staff person, a young woman
brokers are made in cash; some farmers political and economic reasons, leaving
expressed concerns that the brokers Brokers and other buyers do not offer in- the farmers with limited alternative mar-
might come back and steal it from them. formation to farmers on standards re- kets for their passion fruit. Around this
Larger farmers tended to have better quired. They do not reveal whom they are same time, a passion fruit processing
communication and more trust with selling to for fear producers will develop plant located in Thika also closed down,
brokers, and even direct links to export- direct links. Producer group leaders compounding the problem. Since then,
ers whom they trusted. Several farmers provide information on mango buyers farmers have sold their passion fruit to
talked about the challenge of trusting and prices to group members and the two main outlets: (1) the HCDA, a govern-
other producer group members to stick to groups provide a good forum for informa- ment sponsored organization promoting
certain agreed upon prices. Because each tion exchange. In general, however, pro- horticulture development and buying up
farmer sells from his/her own homestead ducers have barriers in gaining market in- horticulture products that are sold in ex-
and because there are no central collec- formation (prices brokers are paying) due port markets; and (2) Ugandan brokers
tion points for mangos, prices are negoti- to their geographic isolation. The fact who cross the western border into Kenya
ated one on one, with little communica- that brokers come to individual farms area every week to buy from farmers in
tion between farmers. Farmers expressed with their own pickers to buy mangos re- the El Doret area.
confidence that the leaders of their produces information flows and the negotiat-
ducer groups would represent their in- ing power of farmers. Thus, the governance pattern for at least
terests. part of the value chain (those selling to
In general, there is a low level of trust in Ugandan brokers) is characterized by a
In terms of power asymmetries, mango the value chain and the balance of shift from a somewhat ‘captive’ network
farmers expressed the view that brokers power, from the perspective of farmers, is towards a more open market relationship
dictate prices, when they buy, how much with buyers, mostly brokers. The lack of (moving to the left on the continuum in
they buy, and what they buy. Timing is central collection points for mangos and Figure 1). At the time of the baseline, pro-
especially important for mango farmers, the dispersed settlement patterns of ject activities were just beginning and
as over-ripe or under-ripe fruit commands farmers is a key constraint to the flow of had not affected this pattern.
a lower price. Brokers control the timing. information.
In terms of trust, the research found a
In terms of socio-cultural biases, one of Passion Fruit general sense of trust and cooperation
the biggest issues with mango farmers is The farmers interviewed in the qualitative among producers and between producers
that brokers are “outsiders.” Men appear research were already growing relatively and brokers. While women producers had
to dominate the producer groups while large volumes of passion fruit. In previous a great deal of trust in the Ugandan
women play an active role in mango pro- years they sold it to a large-scale brokers, they did not trust input suppliers
duction and sales on their farms. Farmers farmer/politician who had an export firm to sell them safe and effective chemicals.
in the project areas are from the same selling flowers and passion fruit in
ethnic groups, while exporters are European markets. A few years ago his There are power asymmetries in the
primarily Asians. business closed for a combination of value chain. Women do much of the pro-
duction One social divide in this value chain cept that agro-stockists supplying chem-
and are
Avocado (Persea Americana) ori- is gender. Men dominate the formal icals are not very active or knowledge-
ginated in south-central Mexico,
respons- activities of the passion fruit able about passion fruit). Overall, there
sometime between 7,000 and 5,000
ible for B.C. But it was several millennia be- groups, while women do most of seems to be a fairly good balance of
selling fore this wild variety was cultivated. the production and selling. Men go power within this chain between produ-
passion Archaeologists in Peru have found do- to all the training on passion fruit, cers and buyers.
fruit from mesticated avocado seeds buried grafting, and application of chemic-
their with Incan mummies dating back to als, even through it is women who
farms or 750 B.C. and there is evidence that do all the work related to passion 4. INCENTIVES AND DISINCENTIVES
central avocados were cultivated in Mexico fruit. Women expressed fear and TO UPGRADING
collection as early as 500 B.C. lack of understanding about chem-
points. (from: www.avocados.org) icals to prevent disease – historic- Upgrading can be defined as “innovation
The qual- ally a major problem in that area. that increases value added” (Giuliani,
itative re- They also expressed considerable Pietrobelli, and Rabellotti 2004) and can
search found that women had organized lack of trust in small input suppliers in the take five forms:
informal groups to sell to Ugandan area to direct them to safe and effective
brokers, and had good bargaining power chemicals. 1. Process upgrading: an increase in
on prices during the high demand sea- production efficiency, resulting in
son. In previous years, the farmer/politi- Information flows on price have been fa- either a) greater output for the same
cian played a very powerful role in pas- cilitated by mobile phone contact with level of inputs or b) the same level of
sion exports, but his departure left a va- buyers and by HCDA posting their prices output for fewer inputs. Process up-
cuum. on a weekly basis. This price information grading can involve improved organiz-
flows to farmers through the farmer ation of the production process or im-
In terms of socio-cultural issues, the groups. Information on export certifica- proved technology.
Ugandan brokers are all Muslim men tion standards, approved chemicals, and
while the producers are mainly Christian improved production techniques flow to 2. Product upgrading: a qualitative im-
women, but this cultural difference has men through training but they do not provement in the product that makes
not affected the quality of their relation- transfer it to women, who are the main it more desirable to consumers and
ship, which appears to be excellent. The producers. The research found in some commands a higher unit price;
Ugandan brokers even bring Ghanaian cases information that group leaders had
cloth to the women farmers from Kam- did not get transferred to other group
3. Functional upgrading: the entry of
pala. They are in frequent contact with members.
a firm into a new, higher value-added
each other via mobile phone and meet
function in the value chain; functional
once a week on the farm of the woman In general, there appears to be a high
upgrading moves the firm closer to
who leads the informal group (and offers level of trust and cooperation among act-
the final consumer and positions it to
her farm as the central collection point). ors in the passion fruit value chain (ex-
receive a higher unit price for the Opportunities for upgrading in the avo- grading is that newly planted Hass avo-
product; cado industry can be pursued in several cado trees take 4-5 years before they
ways: by improving quality; promoting fruit. An alternative approach is to graft
4. Inter-chain upgrading: the entry of improved varieties; expanding market Hass avocados onto Fuerte avocado trees
a firm into a new and more lucrative outlets; and complying with certification (most avocado producers in Kandara
marketing channel in the value chain standards. have Fuerte trees that they planted in the
such as moving from the domestic to 1980s and 1990). However, the grafting
the export market for the same Higher quality fruit can be produced by process requires technical expertise and
product; reducing the incidence of antracnose, a diligent care – proper grafting techniques,
disease that leaves cuts and black spots spraying, fertilizer application, pruning,
5. Inter-sectoral upgrading: using on the skin of Fuerte avocados (the vari- picking the right way, and picking at the
knowledge gained in one value chain ety traditionally produced in Kenya). This right time. EAGA and Kenya BDS have fa-
to move into a new value chain or disease is prevented through a spraying cilitated this process quite successfully in
product line, involving a completely regime that must be carried out at appro- Kandara, and many farmers are now suc-
different product or service. priate intervals during the growing sea- cessfully producing Hass avocados.
son. The key incentives are that this will
Both the Kenya BDS and Fintrac HDC pro- increase the yield of high-grade fruits. Avocado upgrading also can involve
jects facilitate activities to promote up- The disincentive is that it takes time, ex- selling to new market outlets. In Kandara,
grading of tree fruits, primarily product, pertise, and money to buy chemicals, through the EAGA sub-project, producers
process, and inter-chain upgrading. The spraying machines, and hired labor. If the link directly to exporters through produ-
aim is to improve the capacity of small- spraying is not done at the right time, the cer groups. The groups sign agreements
holders to respond to changing market investment is lost. Moreover, Fuerte avo- to sell exclusively to the exporter in ex-
demand and increase rural incomes. The cados get a lower price than Hass avoca- change for embedded services, including
baseline research identified specific dos. spraying, assembly, grading, and a cent-
forms of upgrading in the tree fruit value ralized payment system managed by the
chain and the views of producers and Another form of avocado upgrading is to group leaders. Each producer is respons-
other actors in the value chain on the in- grow an improved variety of fruit, spe- ible for picking their own fruit and deliv-
centives and disincentives to upgrade. cifically, Hass avocados. The incentive to ering it to a collection point. Farmers are
These findings are discussed below and upgrade to Hass is that it is disease res- paid through their groups, with deduc-
will be compared to findings in the istant, easy to ship, and in high demand tions made for the costs of spraying at
second round of qualitative research on in the export market. It is the most popu- the time of payment. There are a number
changes in upgrading and how they are lar and common variety of avocado in the of incentives for this direct linkage. Pro-
related to project interventions. world and grows well in Kenya. It is more ducers have a secured market and a ne-
tailored to the export market and com- gotiated contract price. The prices they
Upgrading Avocados mands a higher price than Fuerte or other receive are higher than those paid by
varieties. A constraint to this form of up- brokers. Farmers have access to spraying
on credit, and access to training in im- to pay school fees). Moreover, brokers ments are quite stringent - they require
proved production techniques. The spray- used to provide labor for picking – a func- soil and water testing, exclusive use of
ing service provided through the exporter tion that farmers now have to assume in- approved chemicals, proper application of
saves the farmers labor time, organiza- dividual responsibility for. chemicals, use of protective clothing,
tional responsibility, and hassle. They are proper storage, documentation of chem-
able to access the expertise of EAGA and Another form of inter-chain upgrading in- ical use, storage and handling, and tests
it does not require an up front capital out- volves selling to an avocado oil pro- for chemical residuals. Training in EUREP-
lay. The exporter is assured high quality cessing plant that will be opened in the GAP standards was ongoing in Kenya at
fruit and the use of approved chemicals. Kandara area. This plant was built years the time of the baseline, in anticipation of
ago but closed down when the avocado the introduction of these standards
There are also some constraints to this market collapsed. With the revival of avo- (which are actually voluntary) in 2005.
direct market linkage. Organizing the cados in the area, the original owners For producers, the incentive to comply
farmer groups can be challenging and now have plans to reopen this factory. with these standards relates largely to ac-
time consuming. EAGA is used to This will enable avocado producers to sell cessing export markets; however, these
sourcing crops from small holders with their whole crops, even the lower grade practices also improve productivity and
shorter production cycles (e.g., French fruit that does not meet buyer standards. environmental and occupational safety
beans) and there is ongoing pressure The prices would be low, but the farmer and produce higher quality fruit for do-
from management to see more immedi- would at least get something. The ad- mestic sale. For exporters, smallholder
ate bottom line results than is possible vantage of this for producers is that it di- compliance is challenging to monitor and
with avocados. For farmers, the prices versifies the market; for exporters, farm- document, but the incentive is to main-
are negotiated ahead of time, and may ers are likely to stay with the crop even if tain and/or improve competitiveness in
be lower than the market price. They run they have periodic problems in quality or export markets. The main disincentives
the risk that EAGA may not take all of marketing because they will at least get to producer compliance are lack of capital
their fruit, yet their MOU with EAGA limits something. It is not certain that the plant to finance costs associated with regular
their options for selling to others and de- will open, and farmers are a bit unsure soil and water testing, building and main-
creases access to these buyers. There is how their avocados would be delivered to taining structures for collection and stor-
also the risk that EAGA may miss their the factory (and the cost). age, and buying capital equipment (such
spraying or pick up schedules, with asso- as charcoal coolers) and protective cloth-
ciated risks and costs to farmers. Produ- Upgrading can also include meeting certi- ing. The risk to all is if, after all the effort
cer group leaders subsidize the process, fication standards required in export mar- and investment, the standards are not
and do not get paid. While farmers have kets, specifically those being introduced really enforced in the export markets. In
issues with the brokers, some of them by EUREPGAP. This process involves a this case, other countries that do not
have played a support role that some variety of actors on the value chain – comply may beat them out in the market.
farmers miss (one farmer referred to the farmers, input suppliers, extension work-
fact that a broker used to give him ad- ers, producer groups, exporters, import-
vance payments when he needed money ers, and auditors. The EUREPGAP require-
Upgrading Mangos
Like avocados, mangos can be upgraded
by introducing new varieties, improving
production techniques, improving busi-
ness practices, complying with certifica-
tion standards, and linking producers
more directly to exporters.
Several new varieties of mango have

good market potential in both export and
domestic markets. Upgrading to these
varieties can give producers a competit-
Mangos originated in India and
have been cultivated there for over
ive edge. However, it takes four years for
4,000 years. Mangos are grown in the new trees to bear fruit. Because the fu-
tropical and sub-tropical lowlands of ture demand for mangos is uncertain and
the world, and consumed throughout competition from other countries is high
the world. Persian sailors introduced (more mangos are produced in the world
mangos into East Africa in the 10th than any other fruit) farmers take a risk
century AD. They subsequently in investing in new trees. Grafting can
spread to West Africa and South shortcut this process but requires tech-
America in the 17th century by the nical input. Another risk of growing new
Portuguese. Mangos are rich in vit- varieties is that some of them (Apple and
amins A and C and are the most con-
Sensation) are very sensitive to the
sumed fruit in the world.
(from: www.mangos.com) amount of rain that falls in the mango
growing areas of Eastern province.
Improved production techniques for man-

gos involve spraying, fertilization, clean-
ing out brush and debris under the trees,
and pruning the trees into the shape of
an umbrella. The key incentive for adopt-
ing these techniques is higher quality
fruit with better market potential. The
constraints to adopting these practices
include limited access to chemicals due
to their high cost, limited knowledge of
proper chemical use on mango trees by viders who are willing to provide services improved quality and safety in the do-
farmers, lack of training geared to the ap- on a contract basis and get paid from ex- mestic and non-certified export markets.
propriate learning level of farmers, and porters or producer groups when the crop Disincentives to this form of upgrading
lack of service providers who are trained is sold. The incentive to farmers is that it are the expense and lack of equipment
to do the spraying. Farmers face the risk would give them a more reliable market, for testing, the difficulty of getting all pro-
of losing their investment in chemicals fairer/negotiated prices, and allow for ducer group members to comply with
and labor due to lack of markets, low payments directly to their accounts (cash standards (where small scale producers
yields, or bad weather (mangos fall off payments are risky). They would have ac- link to exporters through producer
the trees if they dry out in a drought). An- cess to expertise related to chemicals groups), and the expense that exporters
other constraint is that stockists often and spraying. Exporters would be assured must cover to follow up/audit smallholder
change the chemicals they supply, and higher quality fruit and farmers higher compliance.
farmers are unsure of which new ones prices with these inputs. Input suppliers
will work. and service providers would have a Upgrading Passion Fruit
steady clientele. The disincentive is that Passion fruit can be upgraded in several
Selling to new market outlets is another the exporters are reluctant to sign con- ways: by planting grafted passion fruit
form of upgrading. For mango producers tracts, preferring verbal agreements (that (with purple passion fruit grafted onto
this could involve selling to mango pro- the farmers don’t trust). SITE has found it disease-resistant yellow passion fruit root
cessing plants (dried mango) which challenging to identify exporters willing stock); by improving production tech-
would diversify their sales and provide a to link to the farmer groups, or input sup- niques, by selling to new market outlets,
market for the lower grade fruit that is re- pliers willing to take the risk of providing and by meeting certification standards.
jected by other buyers. The main disin- inputs and services in advance of pay-
centive is that no processing plant exists ments. A challenge for all is organizing a There are many incentives for farmers to
at this time near the project area, al- central collection point for exporters to plant grafted seedlings. It is a way to re-
though farmers had discussed plans by procure the fruit. Brokers go from farm to duce the chance of disease: grafted
some investors to build such a plant. An- farm, and play a role in picking and trans- plants are hardier because the roots are
other form of upgrading is to create a dir- porting the fruit that neither the export- disease resistant. Because they are dis-
ect link between growers and mango ex- ers nor farmers can easily substitute for. ease resistant, they can be planted closer
porters – a model that the SITE sub-pro- together and thus cultivated and harves-
ject is promoting among mango farmers Issues related to meeting certification ted more easily. While some chemicals
in its project areas. SITE is linking mango standards are similar for mangos and are still beneficial, farmers can avoid the
farmer groups to training in mango farm- avocados. However, it is important to use of heavy chemical sprays required to
ing as a business and process upgrading. note that Kenya mangos are not gener- resist the disease on normal purple pas-
They also are working to identify mango ally sold in EUREPGAP and other certified sion fruit. These plants are safer and
exporters who are willing to develop sup- markets at present. The incentive to friendlier to the environment, live longer,
ply contracts with the farmer groups; and meet these standards would be better ac- and have higher yields.
input suppliers and spraying service process to these export markets, but also
tural research and demonstration production processes include seasonal
To date, Purple passion fruit ori- centers and private nurseries to pro- fluctuations in the markets, price fluctu-
however, ginated in southern Brazil and is mote the production of grafted seed- ations, and the opportunity costs of shift-
grown in other parts of Latin Amer-
few farm- lings. They are encouraging farmer ing from other crops (especially wheat).
ica, Hawaii, Australia, New Zealand,
ers have Asia, and Africa. Passion fruit was training in passion cultivation, and Seasonal gluts of passion fruit on the
planted introduced into Kenya and Uganda demonstration plots with grafted market are a disincentive to invest in im-
grafted in the 1930s and originally was seedlings. proved production techniques that in-
seed- grown on plantations, but disease crease volumes.
lings, in spread quickly in the plantations Improved production techniques for
part be- and it became relegated to a passion fruit, in addition to grafting, Market Outlets
cause garden crop. Passion fruit has many include better planting techniques, Passion fruit farmers participating in the
they are uses – the pulp is used in juices and chemical application, and grading study currently sell to HCDA and Ugandan
not dessert sauces, juice is given as a
and sorting. The incentive for farmers brokers. HCDA payments are guaranteed,
digestive stimulant and treatment
widely to improve production techniques in- the prices are fixed and favorable (al-
for gastric cancer; chemicals ex-
available. tracted from air-dried passion fruit clude reduced disease and increased though viewed as unfavorable relative to
The cost leaves are used as a sedative or volume of production, especially of high-season prices paid by Ugandan
of produ- tranquilizer; seeds are pressed to higher-grade fruits. Constraints to traders). Farmers also can access training
cing produce a fragrant moisturizing oil. this form of upgrading relate to in production techniques from HCDA. The
these -Morton, 1987 gender discrepancies in access to main constraint is that payments from
seedlings training: the qualitative research HCDA are sometimes delayed. Moreover,
is still found that men received training in larger farmers have an advantage over
high and channels to distribute them passion fruit cultivation, although women smaller growers because they have
commercially have yet to be established. do most of the work. Women have little means of transporting their fruit to the
Moreover, most farmers do not know how knowledge of what chemicals to use and collection points. Ugandan brokers pay
to care for grafted plants, which require how to apply them; they also expressed good prices in cash, are regular custom-
technical know how and proper planting fears of the side-effects. Other con- ers, and have direct and regular contact
and care to succeed. There has been lim- straints include lack of manure (blamed with women farmers through mobile
ited experience with grafted seedlings to on problems in the dairy industry) and phones. They come directly to the women
date in Kenya. Farmers interviewed also sufficient labor. The use of the wrong farmers to buy. The good relationship
mentioned lack of sufficient capital to buy chemicals increases human health risks between the women farmers and
the seedlings, wire, and sprays and to and the costs of production, and reduces Ugandan brokers is a striking contrast to
hire labor. They also mentioned the im- the competitiveness of Kenyan passion the relationships between avocado and
portance of certification as an incentive – fruit when they do not meet the stand- mango farmers and brokers. The main
without this, it is impossible to determine ards in export markets regarding chemic- constraint for passion fruit farmers is the
the quality of the root stock ahead of al residuals (MRLs). Other risks affecting seasonal demand for passion fruit in the
time. Fintrac HDC is working with agricul- farmers’ decision to upgrade passion fruit Uganda market, and lack of a market for
their lower grade fruits because there is to unmet demand in European markets though men tend to dominate the meet-
no passion fruit processing plant.24 A pre- for fresh passion fruit. The constraints are ing discussions and leadership.
liminary feasibility assessment for a pas- the expense, lack of knowledge, and diffi-
sion fruit processing plant by Fintrac HDC culty of certifying individual smallholders Farmers commented on the reasons (in-
suggests that it may not be a viable in- (the plan described by respondents in incentives) that they joined, and what they
vestment. For the plant to compete on terviews is for a small subset of group see as the potential benefits of the
price with passion fruit pulp from South members to be randomly selected and groups. They see the groups as a way to
Africa (currently sold in Kenya and in out- audited, and if they pass, the whole access training, finance, information, and
side markets), the prices that would have group would be certified). advice about growing tree fruits. They
to be paid to Kenyan farmers would be can provide a forum for learning good
extremely low, probably too low to be an 5. INTER-FIRM COOPERATION farming techniques, new production tech-
incentive for farmers to produce the niques, and to encourage more business
volumes required to make it profitable. Horizontal Cooperation through Pro- oriented farming. They are seen as a fo-
ducer Groups cal point to exchange information on
The most promising potential for expand- All of the sub-projects in the study pro- prices, buyers, and markets and obtain
ing market outlets for passion fruit ap- mote the formation of producer groups as information on market demand (varieties
pears to be export markets – which re- part of their strategy to link smallholders and quality of fruit demanded). Farmers
quire farmers to meet certification stand- to input, service, and product markets. feel they can improve access to better
ards. As with avocados and mangos, this The EAGA and Just Juice sub-projects are markets for tree fruits and help farmers
involves using certified chemicals, follow- designed to link avocado and passion coordinate marketing activities so they
ing specific handling and storage prac- fruit farmers directly to an exporter and have less chance of being cheated. They
tices, wearing protective clothing, testing spraying services. The SITE and KADI sub- see the potential for using the groups to
soil and water, and documenting chemic- projects link mango producer groups to access market for other crops, in addition
al use and product handling. The total up- training, inputs, and extension services. to tree fruits. They see that groups have
front cost of certification is around SITE also is trying to link the groups to potential for giving them ‘one voice’ and
$1,000, with a lower level of recurrent mango exporters and inputs suppliers. more negotiating power with buyers.
costs. The incentive for farmers to do this Fintrac HDC is working through farmer They provide a link to the outside world
is that they can continue to sell to HCDA groups to promote grafted passion fruit and a way to reduce isolation and devel-
and other buyers who supply exporters through demonstration plots and links to op a sense of belonging. Some commen-
and thus build up relationships with ex- training resources. At the time of the ted that they provide a renewed sense of
porters to markets other than Uganda qualitative research, producer groups had importance for retired men.
and Rwanda. Input suppliers can benefit been formed, but most (except the avo-
by selling more chemicals and providing cado groups) were still at an early stage Mango and passion fruit farmers further
spraying services. Exporters can respond in their actual activities. All of the groups see the groups as a way of increasing
24
have both women and men members, al- their bargaining power with brokers – by
Ugandan traders are viewed favorably because they sharing information and discussing prices
take all the fruit.
ahead of time. Women group members as reflected in the bottom line, is still not farmer commented that brokers now
near Eldoret have formed an informal in. steer away from their area when buying
group that sells to brokers from a central mangos because the group members
collection point. For input suppliers, farmer groups are agreed to a minimum price. When the
seen as an efficient vehicle for providing brokers found it difficult to negotiate for
Avocado groups have played a formal information on product choices and lower prices, they abandoned the area,
role in negotiating specific MOUs product use, and for promoting their leaving the farmers with over-ripe fruit
between the groups and EAGA about sup- products. and no market at all. Another mango
plying avocados. The groups are respons- farmer commented on how difficult it was
ible for keeping records of volumes sup- Respondents also commented on the po- to get group members to stick to prices
plied and distributing payments to mem- tential constraints, or downside, of group they agree upon with other members. If
bers and, with Kenya BDS support, have membership. When farmers sell exclus- they need money quickly or their fruit is
done this well. The groups are formally ively through producer groups, they lose over-ripe, they are forced to accept
registered and can play a formal role in autonomy to negotiate as individuals. whatever brokers offer.
negotiating legal agreements. They become dependent on the group for
their market link, as their relationship
The exporters we interviewed saw produ- with brokers or other buyers is cut off.
cer groups as a way to reduce the trans- Some farmers have a competitive ad-
action cost of dealing with small produ- vantage, and they lose this as a member
cers and comprise a more efficient supply of the group. Leaders do not get paid,
chain than brokers. If successful relation- and often lose motivation or become
ships can be established, it is a way for domineering. People can spend inordin-
them to access larger volumes of fruit. ate amounts of unproductive time attend-
Over time, the exporters see smallholder ing group meetings – and if the group
producer groups as having potential for does not deliver to members, it is often
offering a more reliable supply of fruit people who have time on their hands who
and better quality fruit. This will help ex- end up participating. Another issue is that
porters plan ahead for their uplift, which there may be social pressure to stay in
makes them more competitive. Group re- the group even if someone doesn’t want
cords can help exporters comply with to. A group issue raised by passion farm-
traceability requirements. It should be ers is that the upgrading approach pro-
noted, however, that the exporters inter- moted by the group in some cases may
viewed said that there was not full con- be more appropriate for larger farmers
sensus in their firms about sourcing dir- (e.g., certification requirements). Small
ectly from producer groups. The verdict, farmers may be adopting a model that is
not cost effective for them. One mango
Role ofTable 5: Survey Respondents by Location lated to ate direct links between producer groups
the Pro- Parti-
passion and exporters, and the provision of spray-
jects in Location Control Total fruit – mar- ing services by exporters. The most vis-
cipant
Organiz- keting and ible progress at the time of the qualitat-
ing Pro- EAGA avocado (Central "merry-go- ive research, as described in earlier sec-
250 250 500
ducer Province) round" sav- tions, has been with avocado producers.
SITE mango
Groups ings One challenge of this approach to pro-
(Eastern and Central 350 349 699
Field ob- provinces)
schemes -- moting vertical cooperation has been ne-
serva- KADI mango gets done gotiating a Memorandum of Understand-
tions dur- 70 -0- 70 outside the ing that all parties agree to. Wrinkles in
(Eastern Province)
ing the Fintrac passion fruit group the initial agreement and changes made
qualitat- 206 173 379 structure in the agreement during the first year
(Rift Valley Province)
ive re- Just Juice passion fruit through rankled some of the smallholders. Anoth-
148 151 299
search (Eastern Province) other in- er challenge in the first year was the
found the All sites
1,024 923 1,947
formal EAGA did not spray on time, which re-
avocado grouping of duced the crop yields and quality. Small-
groups to be highly organized and fo- women. The groups play a limited role at holders were reluctant to agree to pay for
cused in their activities. The groups are present, mostly related to training. Non- spraying as part of their agreement in the
strict and disciplined. Leaders mentioned group members also participate in the second year, given the bad experience in
that some members dropped out because training, benefit from the demonstration the first year. They also experienced
of the stringent requirements for attend- plots, and attend sales date meetings. problems in delayed pick ups (with asso-
ance and participation. Only one member ciated costs of hauling their fruit back
of each family can be a member, and this Horizontal Cooperation through As- and forth to the EAGA collection point).
is usually in the name of the man, even if sociations of Exporters
the woman is the main farmer and parti- Despite discussions among some export B. SMALLHOLDER PRODUCERS OF
cipant. The mango groups are much lar- companies about pursuing common in- AVOCADOS, MANGOES, AND
ger (over 100 members in some, com- terests through an association, there is PASSION FRUIT
pared to 25-30 in the avocado and pas- currently no functioning association and
sion fruit groups), but many members the exporters do not coordinate formally The baseline survey covered 1,947 small-
were inactive. They appeared to be less with one another. holders located at five different sites
tightly organized and had been less suc- (Table 5).
cessful in linking directly to exporters or The majority of survey respondents (79
extension services. They tend to be dom- Vertical Cooperation through Embed- percent) were household heads and
inated by larger farmers, mostly men. ded Services nearly all (94 percent) managed the farm
The passion fruit groups were well organ- Vertical cooperation is promoted by the in question. Ninety percent of the
ized, although much of the activity re- Kenya BDS project in its efforts to facilit- sampled households (1,748) were headed
by men; ten percent (199) were headed sion fruit sites) varied among the three
by women. These percentages differed tree fruits, and also between program
little among sites or between the parti- participants and control group members
cipant and control groups. Additional (Table 6). There was wide dispersion
characteristics of the households covered around all the averages. The survey thus
by the survey are discussed in Section 3, covers a considerable range of farm
below. scales, from very small to rather large.
Mango farms tend to have far more trees

1. ENTERPRISE-LEVEL FINDINGS (160 on average) than avocado farms (27
trees), while the average number of pas-
The baseline survey collected data on a sion fruit vines is higher still. The
wide range of topics. Several important sampled farms that are
features of the MSEs surveyed are ana- participating in the Fintrac project have
lyzed in this section: scale of enterprise; far more vines (1,252 on average) than
production; productivity; participation in the ones at Kenya BDS’s Just Juice inter-
trade; use of hired labor; innovations and vention site (872). At both passion fruit
investments; participation in a producer sites, there were large numbers of imma-
group; and sources of useful technical ad- ture vines.
vice, information, or training related to
tree crops. The analysis determines b. Production and
baseline values for these variables and productivity
makes comparisons between participant
and control groups. At the end of this sec-
tion, we discuss the relationship between
these enterprise variables and both so-
cio-economic status (measured by asset
score) and the gender of the farmer.
a. Scale of enterprise
The average number of trees of the tar-
geted fruit (that is, avocado trees in the
avocado sites, mango trees in the mango
sites, and passion fruit vines25 in the pas-
25
Passion fruit actually grows on vines rather than trees.
Table 6. Number of Trees/Vines per SME by Intervention: Average Fruit production (measured in pieces of
and Distribution fruit harvested in the past year for avo-
EAGA SITE KADI Fintrac HDC Just Juice cado and mango and in kilograms for
passion fruit) also varied widely from
Total
Participant
Total
Total
Participant
Total
Participant
Total
Control
Control
Control
Control
Control
Participant
Participant
farm to farm. In the EAGA avocado area
(see Table 7), smallholders averaged
around 10,000 pieces per year, with pro-
gram participants producing more than
twice as much on average as control
25 25 50 35 34 69 17 14 15 29
BASE: Total Sample 0 0 0 0 9 9 70 0 70 206 3 379 8 1 9
group farmers. Participants at the SITE
mango area averaged about 35,000
pieces per year, again more than twice
14 20
1-9 558 3 2 2 4 0 0 0 0 0 0 0 0 0 the production of control group members.
15 10 At the KADI mango site, average produc-
10-19 82 75 7 50 51 1 3 0 3 0 0 0 1 9 10 tion among participants was much lower
20-29 45 16 61 29 52 81 4 0 4 0 1 1 0 10 10 (11,000 pieces) than in the SITE area.
13 Passion fruit production (see Table 8) was
30-49 48 7 55 55 77 2 14 0 14 2 5 7 1 11 12 very high among Fintrac HDC project par-
17 ticipants and controls, averaging 353,000
50-99 13 2 15 83 91 4 17 0 17 6 31 37 14 34 48 kilograms. In the area covered by the
12 20 13 13 21 Kenya BDS/Just Juice intervention, aver-
100+ 7 2 9 6 76 2 32 0 32 198 5 333 1 87 8 age production was far smaller, 67,000
12 10 10 10 45 37 19 28 kg. per year. It is clear that the Fintrac
Mean 25 12 18 0 82 1 9 0 9 961 3 730 6 2 2 HDC project deals with an entirely differ-
15 15 16 12 12 1,25 81 1,10 87 24 64 ent size class of farms from the Kenya
Standard Deviation 34 16 27 9 9 0 5 0 5 2 1 2 2 4 3 BDS interventions.
The distribution of avocado farms by an-

nual production level was relatively even.
Modal production was in the range of
10,000-15,000 pieces; eight percent of
the farms sampled produced fewer than
1,000 pieces while nine percent produced
more than 30,000.
The distribution of production among Table 7. Pieces of Avocado and Mango Harvested in Past Year: Average and Dis-farm-
farms in the SITE mango intervention tribution ers
area was more slanted toward the larger EAGA Avocados SITE Mangos who KADI Mangos
farms. Seven percent of farms produced (pieces) (pieces) are (pieces)
less than 1,000 pieces (including seven get-
Participant
Participant
Participant
that had no production) while fifteen per- ting
Control
Control
Control
Total
Total
Total
cent harvested more than 30,000 man- con-
gos in the previous year. KADI mango sider-
participants clustered in the 3,000- ably
0 0 0 0 7 0 7
10,000 production range. higher
1-1,000 21 35 56 18 25 43 3 0 3
yields,
1,001-2,000 13 43 56 24 56 80 2 0 2
In the Fintrac HDC passion fruit area, the run-
distribution of production was highly un- 2,001-3,000 20 50 70 31 49 80 8 0 8 ning
equal, with 77 farms not yet in production 3,001-5,000 35 54 89 38 43 81 13 0 13 from
and 133 farms producing more than 5,001-7,000 27 14 41 32 34 66 15 0 15 300 to
100,000 kilograms of fruit per annum. Av- 7,001-10,000 28 27 55 39 40 79 10 0 10 more
erage production on participating small- 10,001-15,000 38 10 48 30 36 66 8 0 8 than
holdings was more than twice as great as 15,001-20,000 23 8 31 27 21 48 3 0 3 1,000
on control farms. The Just Juice area also 20,001-30,000 23 4 27 24 20 44 4 0 4 man-
had a significant number of non-produ- 30,001 – 50,000 14 3 17 25 14 39 2 0 2 gos
cing farms but also had fewer large pro- 50,001 – 70,000 4 0 4 19 4 23 1 0 1 per
ducers, hence much less dispersion 70,001 –100,000 2 1 3 13 3 16 0 0 0 tree.
around the mean production level. 100,001+ 2 1 3 23 4 27 1 0 1 Pos-
Total 250 250 500 350 349 699 70 0 70 sible
The productivity of the avocado and Average in 000's 13 6 10 35 14 24 11 0 11 ex-
mango smallholdings in the survey (Table Standard Deviation 17 13 15 147 56 111 16 0 16 plana-
9) is defined as annual production per NOTE: Average production is the average per producing farm. tions
producing tree or vine. In the EAGA avo- for
cado area, modal productivity is in the these
401-800 range for both participants and super-
controls. However, participants have cipants in the SITE area average 300 ior
slightly higher productivity on average mangos per tree, versus 150 for both yields include better tree stocks, better
(600 pieces per tree) than controls (500). control group farmers and participants in cultivation techniques, and greater ma-
the KADI area. The higher average for turity of the trees.
The SITE mango area is more productive SITE area participants is accounted for by
on average than the KADI area. Parti- the presence of a significant number of
Fintrac HDC passion fruit participants mango sales. In the KADI area, earnings
(see Table 10) attain productivity levels were much lower: average mango sales
(500 kg. of fruit per vine on average) that were only Ksh 6,000 (US$76) and no
are two and a half times as high as Just farmer earned more than Ksh 20,000
Juice participants (200 kg.). Fintrac HDC from this source.
controls also have relatively high pro-
ductivity, while Just Juice controls have Significant numbers of passion fruit grow-
the same average productivity as Just ers did not sell any fruit because their
Juice participants. vines were still immature. Fintrac HDC
project participants averaged Ksh 11,000
c. Participation in trade in sales; 15 project participants earned in
Many Kenyan farmers grow tree fruit for the Ksh 20,000-90,000 range and three
use at home and limited sales to the local made more than Ksh 100,000 selling pas-
market, but few participate in supplying sion fruit. Control group members in the
the export trade. In our sample, 97-98 Fintrac HDC project areas did consider-
percent of avocado and mango growers ably less well, averaging Ksh 5,000 in
earned some income from tree fruit sales, passion fruit sales. Passion fruit growers
but nearly all of them earned only very in the Just Juice intervention area also av-
small amounts (see Table 11). In the eraged just Ksh 5,000 in sales, with little
EAGA intervention area, 84 percent of difference between participants and con-
farmers earned less than Ksh 10,000 trols. No one in this area earned over Ksh
(US$127) per year from avocado sales 20,000 from passion fruit sales.
and the average for all growers was only
Ksh 8,000 (US$101). Just two farmers,
both program participants, earned more
than Ksh 100,000 (US$1,266) from avo-
cado sales in the year preceding the sur-
vey.
SITE mango farmers had a similar pattern

of crop marketing, although they earned
slightly more on average (Ksh 11,000 or
US$ 139) and had more farmers who
earned Ksh 10,000-40,000. One SITE par-
ticipant and one control group member
earned more than Ksh 100,000 from
Table 8. Kilograms of Passion Fruit Harvested in Past Year: Farmers were also asked to whom they
Average and sold the fruit that they marketed.26 Table
Distribution 12 shows the marketing channels used
Fintrac Passion Just Juice Passion by each of the groups surveyed. It indic-
(Kgs) (Kgs) ates that the farmers surveyed use a
Parti-
Control Total
Parti-
Control Total
range of marketing channels, no one of
cipant cipant which accounts for more than 30 percent
0 27 50 77 37 15 52 of total sales for any of the groups identi-
1-1,000 0 0 0 11 21 32 fied in the table. Farmers commonly sell
1,001-2,000 0 0 0 5 8 13 their fruit either to local traders, who sup-
2,001-3,000 1 0 1 7 4 11 ply nearby markets, or to brokers, who
3,001-5,000 4 1 5 13 14 27 then resell it in national, regional, or ex-
5,001-7,000 2 2 4 9 9 18 port markets. Direct sales to consumers
7,001-10,000 5 7 12 8 14 22 are also important for mango and passion
10,001- fruit growers, but not for avocado produ-
15,000 6 6 12 7 14 21 cers. Direct sales to supermarkets and
15,001- processing plants are relatively uncom-
20,000 5 4 9 7 11 18 mon. Marketing patterns for program par-
20,001- ticipants and control group members at
30,000 12 10 22 6 15 21 any particular program site are generally
30,001 – similar. However, there is a notable differ-
50,000 11 23 34 15 10 25 ence between passion fruit growers in the
50,001 – Fintrac HDC areas, who sell substantial
70,000 14 13 27 3 5 8 amounts to the regional East African mar-
70,001 – ket, and growers in the Just Juice area,
100,000 29 14 43 6 7 13
who have not tapped the regional market
100,001+ 90 43 133 14 4 18 and rely heavily on sales to local traders
Total 206 173 379 148 151 299 and consumers.
Average in Farmers sell their fruit either on a spot
000's 462 195 353 43 27 35
basis for the prevailing market price or
Standard Devi-
ation 1,694 751 1,394 72 61 67 26
Respondents were asked to name which of a wide
range of possible customers they sold fruit to. In addition
to those listed in Table 13, other types of customer, which
did not account for as much as five percent of sales for
any class of producer, were the HCDA, fruit processing
factories, Top Notch, schools, and hospitals.
under long-term contracts at fixed prices. percent of them rated their contractors
Table 13 shows that contract sales pre- as “very” or “fairly” reliable.
dominated among participating avocado
growers in the EAGA project area as well d. Use of hired labor
as among both participants and controls Enterprise development programs often
in the Fintrac HDC passion fruit areas. rank employment creation as one of their
They were less important for participants principal objectives, but micro and small
in the two mango projects and relatively enterprises frequently rely heavily on
insignificant for Just Juice participants and family labor and hire few workers from
for all control groups except those in the outside the family circle. This pattern
Fintrac HDC areas. Broadly, contracts are holds for most but not all of the Kenyan
associated with brokered sales to nation- fruit growers in our survey sample, as
al, regional, and export markets, while Table 14 shows.
spot sales characterize sales to local mar-
kets. Surveyed smallholders vary widely in
their use of hired labor. Overall, more
Mutual trust is needed for a contract sys- than one-third (37 percent) reported us-
tem to work effectively. Contractual ar- ing no hired labor at all in the previous
rangements can give rise to “post-con- year. The percentage of farmers who
tract opportunism” when the spot market used no hired labor was uniformly high
price changes and either the buyer or the (45 percent or more) in all the control
seller reneges on contractual obligations groups surveyed. At the other end of the
in search of a better deal. Farmers who scale, 99 percent of participating KADI
sold fruit under contract during the past mango growers and 96 percent of EAGA
year were asked to rate their contractors avocado participants used at least some
on a four-point scale ranging from “very hired labor. Fintrac-area passion fruit
reliable” to “very unreliable.” Half or growers made especially heavy use of
more of respondents rated their contract- hired labor.
ors as “very” or “fairly” reliable, while
hardly any found their contractors to be
“very unreliable.” Among avocado produ-
cers, satisfaction with contractors was in
the 80-90 percent range. Fintrac HDC
passion fruit growers who sold on con-
tract were somewhat less satisfied; 64
Table 9. Pieces of Avocado and Mango Harvested in Past Year
per Producing Tree
EAGA Avocado SITE Mango KADI Mango
(pieces) (pieces) (pieces)
Parti- Parti- Parti-
cipant Control Total cipant Control Total cipant Control Total
1-100 14 28 42 53 51 104 18 0 18
101-
200 22 21 43 108 166 274 44 0 44
201-
300 25 28 53 82 90 172 5 0 5
301-
400 29 28 57 38 32 70 1 0 1
401-
800 108 117 225 40 10 50 2 0 2
801-
1000 33 23 56 17 0 17 0 0 0
Total 250 250 500 350 349 699 70 0 70
Average 600 500 500 300 200 300 100 0 100
e. Innovations and investments

Respondents were asked two questions and Fintrac HDC passion fruit areas also
about recent innovations and invest- reported replanting.
ments. The first question was whether
they had made any changes in their cul-
tivation or marketing methods for the tar-
geted tree fruit in the past two years. As
Table 15 shows, program participants re-
ported high rates of innovation in all
cases except the Just Fruit passion fruit
intervention. These rates of change in
cultivation methods were much higher
than those reported by the control group,
although in the Fintrac HDC passion fruit
areas controls also had a relatively high
rate of innovation.
Table 15 also shows the percentage of re-

spondents who said they had planted
fruit trees in the past year. High percent-
ages of participants reported planting
trees in the SITE mango area and all pas-
sion fruit areas. Large numbers of control
group smallholders for the SITE mango
Table 10. Kilograms of Passion Fruit Harvested in Past Year per Produ-
cing Vine
Fintrac Passion Just Juice Passion
(Kgs) (Kgs)
Parti- Parti-
cipant Control Total cipant Control Total
0 27 50 77 37 20 57
1-100 23 6 29 68 66 134
101-200 41 36 77 11 33 44
201-300 32 41 73 15 11 26
301-400 28 28 56 6 10 16
401-800 44 10 54 9 11 20
801-1000 1 0 1 1 0 1
1000+ 10 2 12 1 0 1
Total 206 173 379 148 151 299
Average 500 300 400 200 200 200
Table 11. Earnings from Sales of Fruit in Past Year (Kenya shil-
lings)
Fintrac
EAGA SITE KADI Just Juice
HDC
Participant
Participant
Participant
Participant
Participant
Control
Control
Control
Control
Control
Total
Total
Total
Total
Total
0 6 4 10 11 7 18 0 0 0 24 52 76 18 12 30
1-10,000 197 213 410 239 271 510 65 0 65 152 111 263 129 138 267
10,001-20,000 30 24 54 53 48 101 5 0 5 12 7 19 1 1 2
20,001-30,000 7 4 11 13 7 20 0 0 0 2 1 3 0 0 0
30,001-40,000 5 3 8 10 4 14 0 0 0 2 0 2 0 0 0
40,001-50,000 0 2 2 3 5 8 0 0 0 2 0 2 0 0 0
50,001-60,000 2 0 2 7 3 10 0 0 0 5 1 6 0 0 0
60,001-70,000 1 0 1 7 1 8 0 0 0 2 0 2 0 0 0
70,001-80,000 0 0 0 3 1 4 0 0 0 1 1 2 0 0 0
80,001-90,000 0 0 0 2 1 3 0 0 0 1 0 1 0 0 0
90,001-100,00 0 0 0 1 0 1 0 0 0 0 0 0 0 0 0
100,001-150,000 1 0 1 1 1 2 0 0 0 1 0 1 0 0 0
150,001+ 1 0 1 0 0 0 0 0 0 2 0 2 0 0 0
Total 250 250 500 350 349 699 70 0 70 206 173 379 148 151 299
Average (000's) 9 6 8 12 9 11 6 0 6 11 5 8 4 5 5
NOTE: Average sales the average for all farms, regardless of whether they sold
fruit.
Table 12. Sales by Type of Customer and Interventions (percent of total
sales)
Fintrac
EAGA SITE KADI HDC Just Juice
Avocados Mango Mango Passion Passion
Participant
Participant
Participant
Participant
Participant
Control
Control
Control
Control
Control
Total
Total
Total
Total
Total
HCDA 3 n 2 1 - n - - - 13 3 9 1 - n
Exporters (direct) 4 1 2 3- 1- - - 3- 2 5 3 4
Subtotal direct sales
to exporters 7 1 4 4 0 1 0 0 0 16 3 11 6 3 4
Broker selling to exporter 7 20 13 12 1 7 1 - 1 12 6 9 18 21 20
Brokers selling to Uganda
/Rwanda/Other regional mar-
ket 1 n 1 1 n n - - - 28 23 27 1 1 1
Broker selling in Kenya 17 24 20 16 11 14 23 - 23 9 13 10 8 10 9
Broker (Do not know
whom selling to) 18 20 19 14 12 13 6 - 6 12 19 15 9 14 11
Local traders selling
in Kenya (from nearby towns) 18 26 22 22 22 22 21 - 21 10 13 11 22 31 26
SubTable
total13: Percentage
sales to of 61
Contract Sales
90 75 65 46 56 50 0 50 71 74 71 58 77 67
brokers & traders PARTICIPANT CONTROL TOTAL
Consumers 5 5 5 17 29
(%) 23 (%)
21 - 21 10 17 12(%)
21 19 20
Wholesale markets/ whole- 2 3 2 4 5 5 6 - 6 1 1 1 2 - 1
EAGA Avocado
saler 80 12 46
Shop / Supermarket n - n 4 11 8 9 - 9 1 2 2 1 n 1
SITE Mango
Masii mango 47
24 - 13 1 - n 13
- - - n 1 30
1 7 - 4
Fruit processor
KADI Mango factory 1 1 1 - - - 3 - 3 n - n 1 1 1
Another farmer 23 - n
n 3 7 5 -0-
10 - 10 1 2 123 3 - 2
Top notch
Fintrac Passion n - n 1 n 1 - - - - - - - - -
61 74 64
Schools - - - 1 2 1 - - - - - - 1 - n
Just Juice Passion
Hospitals -6 - - n - n 19- - - - - - 38- - -
Subtotal direct sales 32 9 21 31 54 43 49 0 49 13 23 17 36 20 29
to other domestic buyers
10 10 10 10 10 10 10 10 10 10 10 10 10 10
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
Table 14: Distribution of Respondents by Amount of Hired Labor
f. Use of inputs Used
The survey included questions about the EAGA SITE KADI Fintrac HDC Just Juice
use of water, fertilizer, and insecticide or
Participant
Total
Total
Participant
Total
Participant
Total
Total
Control
Control
Control
Control
Control
Participant
Participant
fungicide sprays. Water is necessary for
fruit tree cultivation, but very few of the
smallholders surveyed had access to ir-
rigation. Of the 1,947 farmers in the
sample, only 40 reported using drip irrig-
17 18 10 15 26 13
ation and 135 had sprinkler systems. The
None 11 2 3 9 8 7 1 0 1 50 78 128 63 70 3
remainder relied on hand watering or
10 16
rainfall.
1-5 30 34 64 56 4 0 2 0 2
5 11
16 15 15 30
6-10 39 16 55 48 29 77 5 0 5
14 22 85 11 16
A total of 798 farmers, 41 percent of the
11-15 30 6 36 42 17 59 4 0 4
2 6 45 15 20
total sample, reported spending money
16-20 43 3 46 29 13 42 5 0 5
3 8 59 7 16
to purchase fertilizer for use on their fruit
21-25 23 3 26 19 12 31 11 0 11
2 3 16 3 9
trees. Expenditure by these farmers aver-
26-40 33 7 40 15 13 28 18 0 18
5 14 95 14 19
aged Ksh 3,400 (US$43). On average,
41+ 41 9 50 32 3 35 24 0 24 118
64 182 40 16 56
program participants spent more than
Total 25 25 50 35 34 69 15 29
twice as much on fertilizer as control
0 0 0 0 9 9 70 0 70 206 173 379 148 1 9
group members (see Table 16).
Average days used 34 6 20 16 5 10 68 0 68 198 375 278 53 17 35
A larger number of respondents (1,388 or
71 percent of the total) said they had pur- nearly all program participants are mem-
chased sprays for use on the targeted bers of such groups. In sharp contrast,
fruit trees during the past year. The aver- hardly any control group members belong
age expenditure reported was the same to producer groups. This dichotomy is
as for fertilizer, Ksh 3,400. Again, pro- shown in Table 18.
gram participants spent more than twice
as much as control group members (see
Table 17).
g. Participation in producer groups

Since all the programs in the study ap-
proach smallholders through tree fruit
producer groups, it is not surprising that
The tree fruit producer groups to which
participants belong appear to be active
organizations. More than 60 percent of
members report having attended five or
more producer group meetings in the
past six months. When producer group
members were asked to rate the useful-
ness of these organizations on a three-
point scale, 66 percent said they were
“very useful,” 32 percent characterized
them as “fairly useful,” and only two per-
cent found them to be “not at all useful.”
Table 15: Changes in Cultivation Methods and Planting of Fruit Trees (%

answering yes)
Change in cultivation methods in past two years -- % answering yes
Participant Control Total
EAGA 88 4 46
SITE 72 38 55
KADI 50 -0- 50
Fintrac HDC 73 51 63
Just Juice 36 20 28
Fruit tree planting in past year -- % answering yes

Participant Control Total
EAGA 12 15 14
SITE 53 61 57
KADI 31 -0- 31
Fintrac HDC 68 64 66
Just Juice 68 32 49
h. Sources of Useful Technical Ad- spondents reported attending at least
vice, Information, or Training Re- one producer group in the past six
lated to Tree Fruit Production months, compared to 45 percent of men.
Finally, respondents were asked to say Sixty-four percent of women who atten-
where they obtained advice, information, ded producer group meetings said they
or training that assisted them in tree fruit went to five or more such meetings, while
cultivation. Their spontaneous responses 59 percent of male producer group mem-
were coded on a grid by the enumerator. bers reported attending five or more
The sources mentioned most often, in or- meetings in the past six months.
der of frequency, were: non-official exten-
sion agents; Ministry of Agriculture/KARI j. Differences by Socio-economic
extension agents; neighbors, family, and Status
friends; producer groups of which the re-
spondent is a member; the USAID project
in question; seminars and meetings; local
leaders; buyers of the fruit; and nurser-
ies.
i. Differences by Gender of Farmer

Women manage 24 percent of the small-
holdings in our sample (469 out of
1,947). Their holdings tend to be smaller
than those managed by men, with 142
fruit trees on average, versus 258 for
men. Woman-managed farms also have
lower average levels of production, pro-
ductivity, and sales. However, they hire
more labor than farms managed by men:
134 person/days per year on average
compared to 97. There was little differ-
ence between men and women in sales
by type of customer.
Female proprietors participated in produ-

cer groups as actively as male propriet-
ors. Forty-nine percent of women re-
As dis-Table 16: Average Expenditure on Fertilizer Used on k. Summary of Enterprise-level Find-
cussed Targeted Fruit Trees in Past Year (Ksh) The data ings
else- PARTICIPANT CONTROL TOTAL
on sales by
where in type of cus-
this re-EAGA 2,600 900 2,100 tomer show
port, theSITE 1,900 1,100 1,700 that farm-
entire KADI 11,800 -0- 11,800 ers in the
sample Fintrac HDC 7,000 3,700 5,600 lowest as-
was di-Just Juice 3,800 1,500 2,700 set score
vided category
Total Sample 4,300 2,000 3,400
into eight sell more to
socio- brokers
economic than those
groups based on household asset scores. in the highest asset score category (78
Unsurprisingly, these differences in percent of sales compared to 62 percent
household wealth correlated with several of sales). The lowest asset score group
differences among tree fruit enterprises sells also less to HCDA and exporters (dir-
(Table 19). Some of these differences ect) than the highest asset score group
were very large. For example, the aver- (two percent compared to eight percent).
age number of trees owned by farmers in
the highest asset score group was more Fruit sales are higher for the participant
than 20 times as the average for the low- group compared to controls across wealth
est asset score group while the disparity categories. In looking at
in production was more than 40:1. The the use of hired labor,
disparity in sales was much smaller, poorer participant groups Table 17: Average Expenditure on Sprays
about three-to-one. The relationship use more hired labor than Used on Targeted Fruit Trees in Past Year
between household wealth and the poorer controls, while (Ksh)
amount of labor hired to cultivate tree wealthier participant PARTICIPANT CONTROL TOTAL
fruit was more complex. The richest groups use less hired
group of farmers (Group 8) hired the labor than wealthier con- EAGA 1,400 800 1,300
SITE 4,100 1,000 3,000
largest amount of labor, but the next two trols.
KADI 2,100 -0- 2,100
groups down (Groups 6 and 7) hired relat-
Fintrac
ively little. The distribution then hit a HDC
5,100 3,600 4,400
second peak at Group 4 before declining Just Juice 5,800 2,900 4,300
sharply to a very low level of labor use by
Group 1. Total
4,200 2,000 3,400
Sample
Table 18: Percentage of Respondents who are We surveyed five interventions intended
Members of a Producer Group to promote upgrading and raise pro-
PARTI- CON-
TOTAL
ductivity and income from tree fruit
CIPANT TROL among smallholder producers of avocado,
EAGA 99 0 50 mango, and passion fruit. The 1,947
SITE 100 0 50 MSEs included in the survey cultivated
KADI 93 93 varying numbers of trees/vines, with avo-
Fintrac 98 2 54 cado holdings the smallest and passion
Just Juice 92 4 47 fruit the largest. For each fruit, the range
of holding sizes was wide. In each case,
Total
98 1 52 production and productivity were higher
Sample
for program participants than for con-
trols.27 It is unclear whether the differ-
ence reflects selection bias or early im-
pacts of program participation. Between
the two passion fruit sites, Fintrac HDC
definitely works with larger, more pro-
ductive farmers than Just Juice.
Nearly all the farms in the survey sell tree

fruit, primarily
Table 19: Selected Enterprise Data by Asset Score Group (En- through traders of
tire Sample) different sorts, but
Asset Number Annual Annual Person/d Avg # produ- most earned only
score of trees pro- Sales ays of cer group small amounts from
group duction (Ksh hired meeting at- these sales. Con-
(000) 000) labor in tended in tract sales have be-
past year past 6 come dominant and
months relatively well ac-
8 677 244 14,800 551 2.4 cepted for EAGA
(highest) avocado parti-
7 385 189 8,700 136 2.3 cipants as well as
6 290 69 8,000 80 2.0 both participants
5 176 40 6,200 230 2.2
and controls in the
4 133 37 5,400 400 2.1
3 91 30 6,400 160 27 2.5
With the exception of Just Juice passion fruit, where
2 69 17 4,200 180 2.1and controls had similar productivity levels.
participants
1 (low- 30 6 4,900 80 1.6
est)
Overall 230 76 7,200 106 2.2

mean
Fintrac HDC passion Table 20: Average Number of Household Members, Earning Members, and In the overall
fruit areas; remain- Earner/Dependent Ratios by Intervention sample, house-
ing groups sold Total Members Earning members
Earner/Dependent Ra- hold size aver-
their fruit predomin- tios ages 6.1 mem-
antly in spot mar- Parti- Con- Total Parti- Con- Total Parti- Con- Total bers, including
kets. cipant trol cipants trol cipant trol two earning
EAGA 5.6 5.6 5.7 1.9 1.9 1.9 .34 .34 .33 members.
Hired labor was SITE 6.2 7.1 6.6 2.2 2.5 2.2 .35 .35 .35 Household size
KADI 5.9 -0- 5.9 2.6 -0- 2.6 .44 .44
used fairly extens- is lowest for the
Fintrac
ively by richer farm- 6.9 6.4 6.6 1.9 1.7 1.8 .28 .27 .27 Just Juice sample
HDC
ers, while poorer JUST JUICE 5.1 4.3 4.7 1.8 1.7 1.7 .35 .40 .36
at 4.7 and
farmers reliedTOTAL 6.0 6.1 6.1 2.0 2.1 2.0 .33 .34 .33 highest for SITE
primarily on family and Fintrac HDC
labor. Woman-man- at 6.6 members
aged farms tended ers looked to a wide range of sources for (Table 20). 28
to hire more labor than comparable farms useful technical advice, information, or Overall, there is little difference between
managed by men. training. the participant and control groups. Com-
paring to figures from the 1997 Kenya
Producer group membership was almost Welfare Monitoring Survey, the sample
ubiquitous among program participants, 2. HOUSEHOLD-LEVEL FINDINGS households appear to be slightly larger
both male and female. Moreover, nearly than for the general rural population –
all of the farmers who belonged to produ- Our second hypothesis is that greater in- which is 5.0 members, but similar to the
cer groups characterized them as either tegration of smallholder MSEs into the size of poor households. In Central
very or fairly useful. tree fruit value chain will contribute to Province, poor households have an aver-
improved enterprise performance and age of 5.3 members (EAGA and Just Juice
Few farmers had access to irrigation and household well-being. Household well-be- have 5.1 and 5.6 respectively); in Eastern
less than one-half purchased fertilizer for ing is considered here in terms of diversi- Province the figure is 6.0 members (SITE
use on their fruit trees. A larger number fication of household income sources, in- and KADI have 6.2 and 5.9 respectively);
said they had bought pesticide or fungi- crease in household consumption ex- and in Rift Valley, poor households have
cide sprays. penditures, and increased household as- an average of 5.5 members (the Fintrac
sets. A profile of households and their
Considerable numbers of respondents baseline status on these household level
had instituted improved cultivation or variables is presented below.
28
A small number of households in the sample (18) re-
marketing methods in the previous two fused to provide information on the number of household
years. Large numbers in some areas had members – according to the survey team, this is not un-
a. Household Size and Economic usual in some parts of Kenya particularly where there are
planted fruit trees in the past year. Farm- Activity young children.
households are even larger at 6.9 mem- Among households in the sample, 78 per- Overall, 43 percent of respondent house-
bers). 29 cent have one or more children in school holds had at least one member with
(Table 21). This ranges from a low of 63 salaried income.30 Across interventions,
The number of earning members aver- percent in the EAGA sample to a high of this ranged from 27 percent for Just Juice,
ages around two across categories in the 90 percent in the KADI sample. Differ- 33 percent for EAGA, 42 percent for Fin-
sample. The average earner dependent ences between participants and controls trac HDC, 50 percent for KADI, and 58
ratio in the overall sample is .33 (one emerge for EAGA and Just Juice -- EAGA percent for SITE. The most notable differ-
earner for every three household mem- participant households have fewer school ences between participants and controls
bers). This ratio is similar for the test and children and Just Juice participant house- were for EAGA and Just Juice – in both
control samples and across interventions, holds have more schoolchildren com- cases, participants had more households
fruit, gender of household head, and pared to controls. The average number of with salaried income than controls. (Table
household asset score. While earner- de- children in school is 2.27 overall and low- 22)
pendent ratios have been shown to be a est for Just Juice (1.43) and highest for
determinant of poverty and vulnerability Fintrac HDC (3.08). These averages are
in Kenya, the similar ratios across wealth similar for both participants and controls.
levels
(indic- Table 21: Respondent Households by Children in School and Intervention
ated % HH WITH CHILDREN IN AVERAGE NUMBER CHIL-
by as- SCHOOL DREN IN SCHOOL
set Parti- Parti- Control Total
scores) Control Total
cipant cipant
(%) (%)
in the (%)
EAGA 56 71 63 1.58 1.79 1.69
SITE 83 92 87 2.31 2.65 2.48
KADI 90 90 2.77 -0- 2.77
Fintrac HDC 89 89 89 3.10 3.07 3.08
JUST UICE 70 53 61 1.37 1.28 1.43
TOTAL 76 78 78 2.27 2.26 2.24
sample suggests other factors may be

more important. Salaried income represents a regular and
steady source of household income that 30
A priori, this percentage seems surprisingly high. Re-
can help rural households smooth sea- search International believes that the high frequency of
29
According to the USAID strategy paper, the average sonal farm incomes. It is associated with salaried workers may be attributable to the framing of the
number of household members for a farm household is reduced vulnerability of rural households. question, which categorized weekly and monthly wages
6.8 members. as salaries.
Table 23: Percent Distribution of house-
holds by number of income sources
PARTICI- CON-
PANT TROL TOTAL
BASE: Total
Sample 1024 923 1947
Total num-
ber income
sources
1 (Lowest) .01 .01 .01
2 .17 .12 .14
3 .34 .32 .33
4 .25 .36 .30
5 .15 .15 .15
6 .05 .04 .05
7 .02 .00 .01
8 (Highest) .00 .00 .00
Average #
income
sources 3.6 3.6 3.6
Table
Table
25:22: Sources
Respondent
of household
Households by have five or more sources and sale of vegetables (29 percent).
b. income
Members (Totalwith
sample)
Salaried Employment of income. Mango farmers Within the sample, 27 percent of house-
Sources Source and of Intervention
income % house- have a slightly higher aver- holds had income from business activit-
of holds
% HH with salar- withAverage age number of income ies, 12 percent from farm labor, and five
House- ied workersincome number sources, but otherwise percent from non-farm labor. Only five
hold In- from salaried there are few differences percent of respondents had income from
sourceworkers
come across interventions, fruits, remittances (Table 25).
(%)
The sur- (%)Participant or between participants
Participant
(%)Control
Tree fruit production 96
(%)Total
vey gen-and sale and controls (Tables 23 and While earnings from tree fruit sales are
Control
Total
erated Livestock production 56 24). not high in absolute terms (see Enterprise
data fromand sale section B.1.c), this income plays an im-
respond- Cereals and tubers pro- 42 The most frequently repor- portant role as a source of household in-
ents onduction and sale % % % ted sources of income income. Almost half of all households
their OtherEAGA
fruits production
40 27 40 .5 .3 .4
33 clude production and sale ranked tree fruit as the number one
sources and sale 1 6 4 of tree fruits (96 percent)31; source of income (Table 26). Tree fruit in-
of house-SalariedSITElabor 49 67 30 .6 .9 .8
58 production and sale of live- come is more important for participants
hold in-Vegetables production 29 6 9 2 stock (56 percent); produc- (61 percent ranked it number one) than
come, and sale
KADI 50 50 .7 -0 -0 tion and sale of cereals for controls (32 percent ranked it number
the most Business activities 27 0 - - and tubers (42 percent); one). Tree fruit income is especially im-
import- Farm labor
Fin- 42 42 12
42 production and sale of oth- portant for Fintrac HDC participants,
.5 .4 .5
ant Non-farm
tractlabor 5 er fruits (40 percent); and three-fourths of whom ranked passion
2 7 0
Remittance
HDC 5 fruit income as their number
sources production
Investment/interest
JUST 33 22 5 .4 .3 .4
27 Table 24: Average
of house- one source. The im-
JUICE 9 7 3 Number of Household
hold in- Table 26: Percent of Households Ranking Tree Fruit In- portance of tree
come, come Income Sources by fruits as a source of
TOTAL 43 44 44 .5 .6 #1.6 or #2
and the PARTI- InterventionTOTAL
CONTROL income for farmers
7 2 0
relative CIPANT (%) (%) is further supported
PARTICIPANT
(%) by the number of re-
CONTROL
position
TOTAL
of tree fruit income in the households. % Household ranking tree fruit income #1 spondent house-
EAGA 58 37 48 holds ranking it
The findings show that households in SITE 62 27 44 number two in im-
the sample are quite diversified in their KADI 50 N/a N/a portance – 22 per-
Fintrac HDC 77 46 63 cent of participants,
sources of income. On average, they 3. 3. 33 3.
JUST JUICE 45 21
EAGA
have 3.6 sources. Very few households TOTAL and 30 percent of
31
The survey supervisor 61
believes that this 32 may7be 6 476
number
have only one source of income (one % Households ranking tree fruit income #2 3. 4. 3. controls (Table 26).
overstated because the subject of the surveySITE
was tree fruit Overall, 35 percent
percent) and 21 percent of households EAGA 25 29 6 0 8
27
production. 3. -0 3.
SITE 20 28
KADI 24
KADI 24 -0- 8 - 248
Fintrac HDC 15 Fintrac
28 3. 3. 3.
21
HDC 5 4 4
JUST JUICE 30 36 33
JUST 3. 3. 3.
TOTAL 22 30 26
JUICE 8 2 5
of households estimated that tree fruits than Ksh. 3,000 (US$40) and 40 percent sets, including productive assets. Based
comprised more than half their household have average monthly per capita ex- on these scores, the households were
incomes – 48 percent of participant penditure less than Ksh. 2,000 (US$27) classified into eight wealth categories
households and 21 percent of control (Table 30). This suggests that a signific- providing a socioeconomic profile of parti-
households (Table 27). Income from pas- ant number of households are under the cipants and non-participants. 32
sion fruit was particularly important for $1/day poverty line (39 percent of parti-
Fintrac HDC participants (Table 28). cipants and 41 percent of controls). By The findings show that there are more
comparison, 26.5 percent Table 27: Proportion of Household controls than
c. Household Consumption of the Kenyan population Income from Tree Fruit (Estim- participants in
Expenditure was under the $1/day ated) the lowest asset
Consumption expenditure as defined in poverty line in 1999 PARTI- CON- TOTA categories (1
the study includes the estimated value of (UNDP, 2001). At the same CIPANT TROL L and 2) and
food grown at home, education expendit- time, many of the respond- (%) (%) (%) more parti-
ures, and all other cash expenditures ent households have relat- 1- cipants than
21 43 32
over the past month. From these ques- ively high levels 24% con-
tions, total monthly consumption ex- of consumption trols
Table 28:25-
Proportion
30 of Household
36 Income
33 from Tree
penditures and per capita monthly con- (15 percent have 49% in the
Fruit (Estimated) by Intervention
sumption expenditures were calculated average monthly 50-
EAGA SITE38 KADI 17Fintrac28Just Juice
for each respondent household. House- per capita ex- 74%
HDC
hold consumption expenditure generally penditures over 75-
9 3 6
Control
Total
Total
Control
Participant
Participant
Participant
Total
Total
Participant
Control
Control
Control
Participant
is preferred over income as a measure of Ksh.10,000. 99%
Total
household poverty. (US$138)). 100% 1 1 1
Total 100 100 100
Monthly per capita expenditure averages d. Household
Ksh. 5,800 for the overall sample (Table Assets 0-
29). Consumption is higher for parti- Assets can be a 24% 20 31 26 19 50 35 30 0 30 14 17 25 36 57 46
cipants (Ksh 6,500) than controls (Ksh. proxy measure 25-
5,200). The table further shows much for household 49% 39 46 42 31 36 34 34 0 34 15 25 19 32 29 31
lower expenditures in the EAGA sample, wealth or eco- 50-
especially among controls (Ksh. 2,900), nomic status. 74% 37 20 29 43 13 28 33 0 33 42 26 35 23 13 18
and much higher expenditure among Respondents in 75-
the study were 99% 2 3 3 5 1 3 3 0 3 27 8 18 9 2 5
KADI participants (Ksh. 16,000).
asked a set of 1
00% 1
0 1 1 0 n 0 0 0 2 3 3 0 0 0
yes/no questions
Half of the sample households have aver- on a long list of 10 10 10 10 10 10 10 10 10 10 10 10 10
100 0 0 0 0 0 32 A 0 description
0 0 0of the 0 methodology
0 0 100 used 0 to develop the
age monthly per capita expenditures less household as-
asset scores in included in Annex C.
highest two asset categories (Table 31). Table 29: Average Monthly Consumption Expenditure per Capita e.
Comparing interventions, EAGA respond- by Intervention (Ksh) Dif-
ents generally have lower asset scores Intervention PARTI- CONTROL TOTAL fer-
than other groups; Fintrac HDC and Just CIPANT (Ksh) (Ksh) ences
Juice respondents have higher asset (Ksh) by
scores (Table 32). The data shows a lar- EAGA 4,900 2,900 3,900
ger proportion of EAGA respondents in SITE 5.200 4,800 5,000
the lowest asset category and a smaller KADI 16,000 -0- 16,000
FINTRAC 7,100 7,500 7,300
proportion of respondents in the highest
JUST JUICE 6,600 6,900 6,800
asset category compared to other inter-
TOTAL SAMPLE 6,500 5,200 5,800
vention groups. The data further shows a
larger proportion of Fintrac HDC and Just There are slightly more participants than gender of household head
Juice respondents in the highest asset controls in the lower size landholding Years of research in Kenya have shown that wo-
category compared to other respondent groups (two acres and below) and more
groups; and fewer respondents in the men headed households are more vulnerable
controls than participants in the higher than those headed by men. They tend to have
lowest asset categories. KADI and SITE size landholding groups (over six acres)
are in between. more limited access to productive resources
(Table 33). Comparing landholding across (land, labor, credit, technology, extension ser-
interventions, a higher proportion of re-
Landholding is another important house- vices, and other productive resources), and less
spondents in the EAGA and Just Juice
hold asset. It is important to note, samples – which are located in Central
extensive social networks. Many Kenyan wo-
however, that the sample covers a num- province -- are in the smaller landholding men who head households have both time and
ber of ecological zones, affecting the type size groups (Table 34). Given smaller av- spatial constraints, which combine to limit their
of land, its productivity, and average erage landholding size in these areas, mobility and participation in market activities
landholding size. For example, landhold- this is not unexpected. A higher propor- beyond a certain range of their homes. Women
ings in the fertile and populous areas of tion of the SITE, KADI, and Fintrac HDC who head households in geographically isolated
Central province (EAGA, Just Juice) typic- participants are in the larger landholding areas are particularly vulnerable. In general,
ally are smaller than landholdings in the size groups. This also is likely to be re- woman-headed households have fewer business
more arid areas of Eastern or Rift Valley lated to their geographic location (East-
provinces (SITE, KADI, Fintrac). Given this resources to draw upon than men headed house-
ern and Rift Valley provinces). holds, but are more dependent on these activit-
variation, the figures presented are not a
good comparable measure of wealth ies to meet their household needs (Ruth-Aspaas,
across interventions in the sample, but 2003). In light of these constraints, increased
more useful in comparing respondents participation in on-farm cash crop activities
within intervention samples. such as tree fruit production and sales has po-
tential to reduce the vulnerability of households
headed by women.
Table 30: Distribution of Re-
spondent
Households by Monthly Con-
sumption producers in Central province (Table sorb the risks associated with participat-
Women Expenditure per Capita 35). ing in the projects.
comprise Parti- Con- T
24 per- cipant trol otal As a point of comparison, Kenya’s The baseline research reveals some dif-
cent of 1 1997 Welfare Monitoring Survey ferences between men- and women-
the farm-BASE: Total ,94 (Government of Kenya, 2000) found headed households (Table 36). On aver-
ers in theSample 1,024 923 7 a significantly higher proportion of age, women-headed households have
sample None 1 1 1 women-headed households in rural fewer members, but earner- dependent
but head 1,000 and be- areas of Kenya – overall, 31.2 per- ratios are about the same as for men-
only tenlow 23 21 22 cent of poor rural households and headed households. A smaller proportion
percent 1,001 and 25.5 percent of non-poor rural of women-headed households have salar-
of the2,000 15 19 17 households included in this survey ied members and children in school than
house- 2,001 and were headed by women. The pro- men-headed households. In terms of pro-
holds –3,000 10 10 10 portion of poor rural households ductive resources, the average landhold-
12 per- 3,001 and headed by women in Eastern ing size for women- headed households is
cent in4,000 7 10 9 province was 33.6 percent, in Cent- almost half that of men, 6.3 compared to
the test 4,001 and ral province 32.3 percent, and in 11.5 acres. Average consumption ex-
sample 5,000 7 8 7 Rift Valley province 22.5 percent. penditure per capita per month is lower
and eight 5,001 and for participant women-headed house-
percent 6,000 7 5 6 This finding suggests that the tree holds than similar men headed house-
in the 6,001 and fruit projects may have reached a holds (Ksh. 4,700 compared to Ksh.
control 7,000 4 4 4 lower proportion of women headed 6,700). However, among controls con-
sample. 7,001 and households than in the general pop- sumption is higher for women- than for
The 8,000 3 4 3 ulation. One reason may be that men-headed households (Ksh 7,800 com-
largest 8,001 and project outreach strategies target pared to Ksh 4,900). This suggests that
share of 9,000 3 3 3 women farmers, but not explicitly the woman-headed households reached
women- 9,001 and women household heads, who are by the projects are poorer. A higher pro-
1 2 2
headed 10,000 harder to reach and recruit given portion of women-headed households is
house- 10,001 and their geographic and social isola- below the Ksh 2,000/day consumption
15,000 9 7 8 tion. Another reason may be that
holds, 19 figure, but the difference is slight (42 per-
15,001 and women who head households have
percent, cent compared to 40 percent) and may
20,000 4 2 3
is among less time available to participate in not be significant. A larger proportion of
20,001 and
avocado 50,000 producer groups and fewer labor re- women- compared to men-headed house-
5 2 4
sources to upgrade tree fruit pro- holds is in the lowest asset group (13 per-
50,000+ 1 1 1 duction. Given their vulnerability, cent of woman-headed households com-
Total 100 100 100 they may have less capacity to ab- pared to six percent of man-headed
house- Table 31: Distribution of Respondent less than Ksh.2,000 per
holds), Households by Asset Score month is highest for the
indicat- PARTICIPANT CONTROL TOTAL bottom two wealth groups
ing theirBASE: Total (68 percent and 53 percent
1,024 923 1,947
greater Sample respectively).
vulnerab- 1 (Low-
5 10 8
ility. This est)
gender 2 13 17 15
gap is 3 18 14 16
larger 4 13 14 14
among 5 10 10 10
controls 6 13 14 13
than par- 7 14 12 13
ticipants. 8
14 9 12
There is(Highest)
little dif-Total 100 100 100
ference Mean 4.7 4.3 4.5
between
women- and men-headed households in
the proportion of households in the top
asset category.
f. Differences by Socio-economic
Status
Earner-dependent ratios, average number
of income sources, and the importance of
tree fruit income within households are
similar across wealth levels (Table 37).
However, as expected, consumption ex-
penditures increase with wealth level. Av-
erage monthly per capita consumption
expenditure rises steadily from Ksh.3,000
for households in the bottom asset cat-
egory, to Ksh.11,200 for households in
the top asset category. The proportion of
households with per capita expenditures
g. Intra-household Issues
A number of intra-household factors can
affect the supply response of farmers in
the tree fruit value chain and provide a
context for understanding impacts. With-
in households, who assumes responsibil-
ity for doing the work, who makes man-
agement decisions, and who controls the
earnings can all influence decisions re-
garding the investment of time and re-
sources in tree fruits. The tree fruit value
chain is characterized by a gender differ-
ences
Table 32: Distribution of Respondent Households by Asset Score in all
and Intervention these
areas.
Fintrac
EAGA SITE KADI Just Juice
HDC
Participant
Participant
Participant
Participant
Participant
Control
Control
Control
Control
Control
Total
Total
Total
Total
Total
BASE: Total 35 20 14 15 29
Sample 250 250 500 0 349 699 70 0 706 173 379 8 1 9
1 9 28 19 6 6 6 11 0 110 2 1 1 1 1
2 23 28 25 15 15 15 14 0 144 9 6 5 11 8
3 17 17 17 26 17 21 11 0 11
10 8 9 15 11 13
4 12 8 10 14 17 16 20 0 20
13 13 13 13 15 14
5 12 6 9 9 13 11 4 0 4
11 9 10 14 11 12
6 10 7 8 10 14 12 13 0 13
18 16 17 16 21 18
7 8 4 6 11 12 11 10 0 10
22 20 21 19 19 19
8 10 2 6 9 7 8 16 0 16
21 23 22 19 12 15
10 10 10 10 10 10
Total 100 100 100 0 100 100 0 0 100 0 100 100 0 0 0
Mean 4 2.8 3.4 4.2 4.4 4.3 4.5 0 4.5 5.8 5.6 5.7 5.5 5.2 5.3
Passion fruit is mostly a woman’s crop vest season for other crops. Women’s income from the passion goes to ‘obvious
and women make most of the decisions “ownership” of passion fruit is reflected in things’ like school fees and paying off
regarding production and sale. Men de- the comment of a farmer in Eldoret who loans. If there is a surplus after covering
cide on the initial investment; but women accused a man of “stealing his wife’s fruit these expenses, husbands and wives
make most other decisions. They also do and selling it for drink”. Women are the jointly negotiate the use of the surplus.
most of the work. Men dig holes for the first to receive income from sales to
plants and install the poles and wire (one- brokers, and generally control this in- For avocados, men pick and prune the
off tasks) while women maintain the come. According to one woman farmer, trees and women do most other mainten-
vines, ance work. If there is a shortage of
harvest Table 34: Distribution of Respondent Households by Size of Landholding male labor within the family, wo-
the fruit,by Intervention men will hire male labor for picking
and play EAGA SITE KADI Fintrac HDC Just Juice fruit, but they also face labor short-
a key role ages. Although women do most of
Participant
Participant
Participant
Participant
Participant
Control
Control
Control
Control
Control
in sales the work cultivating avocados,
Total
Total
Total
Total
Total
to group payments are in the name of
brokers. their husbands or sons. Women tra-
Men play ditionally are responsible for nego-
BASE: Total
a role inSample 250 250 500 350 349 699 70 0 70 206 173 379 148 151 299 tiating with brokers, so they may
selling to Upto 0.5 acre 1 2 1 1 - n 1 - 1 -1 - 1 5 15 11 lose some control over avocado in-
export- 0.51 to 1.0 come in working through groups.
ers. Wo-acre 18 7 13 5 n 3 3 - 3 6 1 4 26 22 24 However, when asked about de-
men 1.1 to 2.0 cision-making, both men and wo-
some- acres 38 36 37 10 2 6 9 - 9 13 12 12 18 25 21 men group members said that hus-
times 2.1 to 4.0 bands and wives make decisions
hire maleacres 29 36 33 25 15 20 19 - 19 17 13 15 30 19 24 together about money earned from
labor for 4.1 to 6.0 avocados.
picking, acres 8 12 10 20 13 17 34 - 34 21 17 20 13 11 12
although 6.1 to 10.0
they of-acres 5 6 6 17 27 22 14 - 14 13 23 18 7 4 6
ten face 10.1 to 20.0
labor acres 2 n 1 12 19 16 14 - 14 14 15 14 1 2 1
short- 20.1 + acres - - - 11 23 17 6 - 6 16 17 16 - 1 n
ages, es-Total 100 100 100 100 100 100 99 - 99 99 99 99 99 99 99
pecially Average (acres) 2.6 2.8 2.7 7.3 10.8 9 6.9 0 6.9 8.1 9.1 8.6 2.7 2.5 2.6
during Standard Devi-
the har-ation 2.4 2 2.2 6 6.4 6.4 5.3 0 5.3 6.7 6.5 6.6 2.3 3 2.7
Table 33: Distribution of Respondent In
Households by Size of Landholding
PARTICIPANT CONTROL TOTAL
BASE: Total
1,024 923 1,947
Sample
Upto 0.5
1 4 2
acre
0.51 to
11 6 9
1.0 acre
1.1 to 2.0
18 17 18
acres
2.1 to 4.0
25 21 23
acres
4.1 to 6.0
17 13 15
acres
6.1 to
12 17 14
10.0 acres
10.1 to
8 11 9
20.0 acres
20.1 +
7 12 9
acres
Total 100 100 100
Average
2.9 2.8 2.9
(acres)
Standard
1.6 1.6 1.6
Deviation
mango farming, men appear to play a
more central role in crop management
and upgrading activities. Brokers are re-
sponsible for picking mangos (the trees
are tall) and factor this into their prices.
Households sometimes hire outside labor
if there is excess production or they are
involved in non-broker sales. Men and
women make joint decisions at the farm
level, and farmers say that payments
[from brokers] may be received by any-
one in the household, even children.
Household members generally discuss
ahead of time a ‘floor’ price they will not
sell below. Once income is in household,
men and women decide jointly on ‘big’
expenditures, while other expenditures
are routine and don’t require discussion.
Although women play a secondary role in
mango farming in some households, they
said they have a fairly good idea of how
much money is coming in.
In general, decisions to upgrade, join pro-

ducer groups, and use income from tree
fruits are joint decisions within families,
with women and men both playing a role.
Spouses discuss these decisions and gen-
erally agree together what to do. Realloc-
ation of significant household capital and
labor to tree fruits does not appear to be
a key household issue at this point, per-
haps because major upgrading activities
are still getting underway.
Table 35: Percentage of Woman- pendent ratios between men and wo-
h. Summary of Household-level Find- headed Households in Study Sample men headed households or wealth
ings PARTI- CON- TOTA level (as indicated by asset scores)
Household size in the sample is large rel- CIPANT TROL L suggesting this may not be a major
ative to the total population, but about EAGA 20 17 19 determinant of vulnerability for house-
average relative to poor rural households. SITE 10 3 6 holds in the sample.
The number of earning members in KADI 16 -0- 16
Fintrac The asset scores and consumption ex-
households and the number of household 9 4 7
HDC
income sources suggest an active work- penditure data show a significant
JUST
ing population among respondents. There 9 8 8 number of poor households in the
JUICE
are no major differences in earner de- TOTAL 12 8 10
sample, in both the participant and
Table 36: Selected Household Data by Gender of Household Head

Women Headed Men Headed House- Total Households
Household holds
Parti- Con- Tota Parti- Con- Tota Parti- Con- Tota
cipant trol l cipant trol l cipant trol l
Average no.
household 5.7 5.6 5.7 6.1 6.1 6.1 6.0 6.1 6.1
members
Average no.
earning mem- 2.1 2.1 2.1 2.0 2.1 2.0 2.0 2.1 2.0
bers
Earner depend-
.37 .38 .38 .33 .33 .33 .34 .34 .34
ent ratios
% HH w/ salar-
40 33 38 43 45 44 43 44 44
ied workers
% HH w/ chil-
65 68 66 78 81 79 76 80 78
dren in school
Average size of
1
landholding 5.9 6.9 6.3 9.7 13.4 9.3 12.9 11
1.5
(acres)
Average no. in-
3.6 3.6 3.6 3.6 3.6 3.6 3.6 3.6 3.6
come sources
% HH ranking
tree fruit in- 66 33 54 60 32 46 61 32 47
come #1
Average con- 5 5 5
sumption ex- 4,700 7,800 ,80 6,700 4,900 ,80 6,500 5,200 ,80
penditure 0 0 0
% HH with per
capita con-
sumption ex- 42 42 42 39 41 40 39 41 40
penditure
control groups. This sug- Table 37: Selected household data by asset score group (total
gests the projects are insample)
volving poor households Over-
and, thus, have potential 1 2 3 4 5 6 7 8 all
for direct impact on their in- Mean
come from tree fruits. The Average
sample also includes less earner/de- .32 .34 .37 .35 .33 .32 .32 .34 .34
pendent ratios
poor households, which
Average num-
should provide a good basis ber income 3.3 3.5 3.5 3.5 3.6 3.8 3.8 4.2 3.6
for comparing impacts sources
across poverty groups at % HH ranking
the end line. tree fruit #1
57 50 46 41 49 46 45 50 47
source of in-
Households are quite diver- come
sified in their sources of in- % HH income
come and tree fruits are an from tree
41.8 41.0 39.5 36.6 41.9 37.8 37.2 40.5 39.4
important source. While fruits (aver-
these figures may reflect an age)
upward bias in some re- Average
monthly per
spondents who associated
capita con- 3,00 3,30 5,30 4,70 4,90 6,70 7,40 11,2
the study with the tree fruit 5,800
sumption ex- 0 0 0 0 0 0 0 00
projects, it suggests the im- penditure
portance of relatively small (Ksh)
amounts of cash income for in the impact of the projects. Producer
rural households.33 groups appear to be an effective means
of reaching women and poorer tree fruit
Gender differences in the division of labor farmers.
related to tree fruit production, the con-
trol of tree fruit income, and access to
productive resources are likely to play out
33
The survey teams were careful NOT to tell those inter-
viewed that the study concerned the impact of the tree
fruit projects. However, given the number of questions
about tree fruit, it would be natural for respondents to as-
sociate the study with the USAID projects, which are rel-
atively high profile.
IV. CONCLUSIONS OF THE BASELINE STUDY AND IMPLICATIONS FOR ROUND TWO OF THE
IMPACT ASSESSMENT
This impact assessment is designed to Broadly speaking, the baseline research Vertical links to higher-value markets
test the hypotheses enunciated on pages shows that smallholders are part of the provide critical incentives for tree fruit
16-18, above. The general hypothesis is tree fruit value chain, but they occupy a producers to upgrade. So far, only one of
that the activities of the Kenya BDS and low position within that chain. They are the interventions studied – the EAGA avo-
Fintrac HDC projects can be effective in numerous and active producers, but their cado intervention – has begun to realize
opening up opportunities for smallholder productivity is low and they sell much of this potential by forging a direct link from
MSEs in local, regional, and global mar- their produce under unfavorable condi- farmers to the European market. This has
kets and in improving the competitive- tions. Income from tree fruits plays an im- required considerable ‘hand holding’ by
ness of Kenya’s tree fruit value chains. portant role as a source of household in- Kenya BDS and has involved an exporter
More specifically, the impact assessment come, especially for the poorer farmers, that has received other support from
hopes to identify and measure improve- but income from tree fruit and total USAID to help prepare smallholders to
ments that lead to greater integration of household income are both very low in meet EUREPGAP standards.
smallholder MSEs into the value chain, to most cases.
improved enterprise performance and Brokers remain alive and well in all three
household well-being, and to improved To varying degrees, the five interventions fruit value chains and continue to be im-
competitiveness and growth of the tar- included in the baseline study succeed in portant marketing channels for many
geted value chains. reaching low-income farmers. This means farmers.
that there is potential for direct impact by
The study is longitudinal and has two raising their incomes through the pro- It is too soon to tell whether the project
stages. This report covers the first, or jects. activities will result in "sustainable solu-
baseline, stage. It examines the condition tions" to the recurrent needs of tree fruit
of the value chain and a sample of small- Producer groups are an important part of producers. This includes both embedded
holders and others involved in the pro- this potential because they provide hori- and stand-alone solutions/services that
duction and marketing of tree fruit at an zontal linkages that give poor farmers a provide inputs, TA, or market access. In
early stage of implementation of the two chance to link to export markets – some cases changes might take place
projects. The main purpose of the something they have very little opportun- due to direct provision by the projects but
baseline study has been to establish a ity to do by other means. The projects it remains to be seen if embedded ser-
standard against which change can be have been instrumental in organizing and vice arrangements, the commercializa-
measured two years later. strengthening tree fruit producer groups. tion of nursery and extension services, or
the “network broker” concept of EAGA
and Kenya BDS will last once the project at the specific services/solutions under-
activities end. taken during the course of the projects.
Time for these activities should be in-
Annex B (below) provides a checklist of cluded in the budget for the second-
issues that should be addressed in the round assessment. Provision should also
follow-up survey. A general point is that be made for reviewing the monitoring
while scrupulous efforts were made to se- data that are being collected by both pro-
lect control group samples for the jects.
baseline survey that were comparable to
the participant samples, the findings
cited in this report show that at the time
of the survey the participants as a group
were significantly better-off and more
productive -- in many cases and in sever-
al ways -- than the controls. When each
group is resurveyed two years hence,
care will need to be taken in analyzing
the results to ensure that differences in
household wealth and other mediating
variables are taken into account in de-
termining the impact of the programs.
In the second round it will be crucial to in-

clude a careful review and documentation
of the interventions – that is, the project
facilitation activities of Kenya BDS and
Fintrac HDC as well as the specific sub-
project activities. One reason for this is
that the activities are very different, mak-
ing it important to be clear about what
we are comparing. Another reason is that
the scopes of the activities and the ap-
proaches they take are likely to evolve
over time. Finally, it will be important to
provide a close analysis of the commer-
cialization issue, including a careful look
REFERENCES
Dunn, Elizabeth and Lillian Villeda. 2005. “Weaving Micro and Small Enterprises
into Global Value Chains: The Case of Guatemalan Textile Handicrafts.” AMAP pa-
per.
FKAB Feldt Consulting. 2001. “Sector Study of the Horticultural Export Sector of
Kenya. A Study Made on Behalf of USAID, Kenya.”
Government of Kenya. 2000. Kenya Welfare Monitoring Survey.
Government of Kenya. n.d. Kenya Interim Poverty Reduction Strategy Paper 2000
– 2003. Prepared by the Government of Kenya. World Bank (no Date)
Giuliani, Elisa, Carlo Pietrobelli and Roberta Rabellotti, 2004. "Upgrading in global
value chains: lessons from Latin American clusters," Working Paper 72, SEMEQ
Department, Faculty of Economics, University of Eastern Piedmont.
HCDA. 2003. Export Statistics for Fruits, Flowers, and Vegetables for 2003.
Horticulture Crops Development Authority. 2003. Export Statistics for Fresh

Fruits, Flowers and Vegetables for the Year 2003
Humphrey, John and Hubert Schmitz. 2003. “Chain Governance and Upgrading:
Taking Stock.” In Hubert Schmitz, ed. Local Enterprises in the Global Economy:
Issues of Governance and Upgrading.
Jaffee, Steven. 2003. “From Challenge to Opportunity: Transforming Kenya’s

Fresh Vegetable Trade in the Context of Emerging Food Safety and Other Stand-
ards in Europe.” Agriculture and Rural Development Paper 1. World Bank.
Jaffee, Steven, and Spencer Henson. 2004. “Standards and Agro-Food Exports
from Developing Countries: Rebalancing the Debate.” World Bank Policy Re-
search Working Paper 3348.
Kula, Olaf, “Activity Status Report: Holding Hands with Folded Arms: Upgrading
Kenya
Tree Fruit Value Chains”. Deloitte, Touche, Tomatsu, No date.
Lipton, Michael. 1994. “Growing Points in Poverty Research: Labour Issues.” Insti-
tute of Development Studies, Sussex University.
McCulloch, Neil and Masako Ota. N.d. “Export Horticulture and Poverty in Kenya.”
Institute of Development Studies, University of Sussex.
Minot, Nicholas and Margaret Ngigi. 2003. “Are Horticultural Exports a Replicable
Success Story? Evidence from Kenya and Cote d’Ivoire.” Paper presented at the
InWEnt, IFPRI, NEPAD, CTA Conference, “Successes in African Agriculture,” Pre-
toria, December 1-3.
86
Morton, Julia. 1987. Fruits of Warm Climates. (chapter on passion fruit. p. 320–
328). Miami, FL
Muendo, Kavoi Mutuku, David Tschirley, and Michael T. Weber. 2004. “Improving
Kenya's Domestic Horticultural Production and Marketing System: Current Com-
petitiveness, Forces of Change, and Challenges for the Future. Volume II: Horti-
cultural Marketing.” Nairobi: Tegemeo Institute of Agricultural Policy and Devel-
opment, Egerton University.
Ruth-Aspaas, Helen. 2003. “Heading household and heading businesses: Rural

Kenyan women in the informal sector.” The Professional Geographer, Volume 50,
Issue 2. pp. 192-204.
Tschirley, David and Mary Mathenge. 2003. “Developing Income Proxy Models for
Use by the USAID Mission in Kenya: A Technical Report.” Tegemeo Working Paper
7. Nairobi: Tegemeo Institute of Agricultural Policy and Development.
United Nations Development Programme, 2001. Human Development Report

2001. Making
USAID/Kenya. 2000. Integrated Strategy Plan 2001-2005.
United States Department of Agriculture, 2005. “World Horticulture Trade and

US Export Opportunities. March (www.fas.usda.gov)
Wallis, William. 2004. “African Farmers Dig in to Comply with EU Food Rules.” Fin-
ancial Times, April 7.
87
ANNEX A. DESCRIPTION OF KENYA BDS AND FINTRAC
HDC PROJECT ACTIVITIES
KENYA BDS
Value chain Type of activity Geographic Target Participa-

area group tion vari-
Partner ables
Avocados Form producer groups Central province Small Smallholders
to engage in forward holder who used
and backward linkages EAGA avocado embedded
farmers service
Facilitate improved pro-
vision of inputs (agro- Lead firm Smallholders
chemicals and seed- providing selling avo-
lings), equipment, and contracts cados to
extension services in to produ- EAGA
crop husbandry through cers and
embedded service ar- embed- EAGA
rangement with lead ded ser-
firm vices
Facilitate the establish

of collection sites in col-
laboration with produ-
cers and buyers
Facilitate improved
transport of goods to
Nairobi
Mangos Form producer groups Eastern and Small Smallholders

to engage in forward Central province holder who join pro-
and backward linkages (Machakos, Mak- mango ducer groups
ueni, Murang’a producers Smallholders
districts) who use ex-
Link producer groups to Buyers tension ser-
existing commercial SITE vices
providers of extension Business Smallholders
services service who sell to
providers buyers ID’s
Facilitate market link- (Private by SITE
ages between producer extension
groups and multiple agents)
buyers
Mangos Provide on-farm training Coast Province Small Private ex-

of trainers for unem- (Lamu and Tana holder tension of-
ployed extension of- River districts) mango ficers who
ficers producers participate in
Coastal Devel- training
Creation of Farmer Led opment Author- Business
Extension Teams (made ity (CDA) service Extension
up of lead farmers and providers agents/lead
88
KENYA BDS

Partner ables
extension agents) to (farmer farmers who
provide commercially led exten- join FLET
viable extension ser- sion
vices teams; Farmers who
financial access fin-
Launch information services) ance through
campaign to increase revolving
awareness of value of fund
extension services
Establish revolving fund

to finance adoption of
good agricultural prac-
tices
Mangos Form producer groups Coast province Small Smallholders

to engage in forward Watamu/Msa- holder who join pro-
and backward linkages baha region, Ma- mango ducer groups
lindi producers Extension
Facilitate market link- workers who
age to buyers Kenya Gatsby Extension participate in
Trust, KARI, service training
Develop market inform- KWETU providers Buyers who
ation data base (private participate in
extension brokerage
Facilitate brokerage agents) worKshops
worKshops
Train private extension

workers mango hus-
bandry and business
management
Information campaign
to sensitize producer
groups to value of ex-
tension services
Mangos Information campaign Eastern Province Small Nursery op-

to sensitize farmers on Mbeere and holder erators who
nursery development Machakos dis- mango participate in
and benefits of nursery tricts producers training
seedlings Service pro-
KADI - Catholic Input sup- viders who
Train nursery operators Dioceses of pliers participate in
in mango husbandry Embu (seedling technical
and business manage- suppliers, training
ment extension
services)
Train extension service
89
KENYA BDS

Partner ables
providers on grafting,
budding, top working
Passion fruit Form producer groups Eastern province Small Small hold-
Mangos to engage in forward (Embu and Meru holder ers in produ-
and backward linkages districts) passion cer groups
fruit farm- Small hold-
Just Juice, ers ers selling to
Facilitate market link- KARI, KMEPP market out-
age to Greenlands Greenlands Buyers lets
Agrocprodu- Buyers
Facilitate the establish- cers LTd
ment of a private pas-
sion fruit nursery
(including backward
linkages with farmers)
Tree fruits (gen- Develop a system for Karatina, Mur- Small Small holder
eral) collecting and posting ang’a holder who particip-
market information on Embu tree fruit ate in SMS
electronic data base producers training
KACE Small hold-
Build awareness of Buyers ers who re-
availability and use of quest market
ICT for market informa- Business information
tion service Small hold-
providers ers who
Train farmers on use of (market made a bid
SMS informa- Buyers who
tion ser- use market
Establish three rural vices) information
trading floors/market in- Entrepren-
formation points eurs who
purchased a
franchise
Tree fruits (gen- Train agrochemical Eastern and Input sup- Stockists
eral) stockists in advisory Central province pliers who particip-
services related to prop- (stockists ate in train-
er storage, labeling, Ideal Business who can ing
transport, handling, re- Link, Ltd. provide Stockists
packing, and adultera- technical who access
tion advice to credit facility
their MSE Agrochemic-
Train agrochemical clients) al distribut-
stockists in business ors who sell
management to stockists
Raise awareness of rur-

al farmers on safe ap-
plication of agrochemic-
als
90
KENYA BDS

Partner ables
Establish credit facility
link between agrochem-
ical distributors and
stockists
Develop monitoring sys-

tem to inform manufac-
turers and stockists on
consumer trends
FINTRAC HORTICULTURE DEVELOPMENT CENTER

Partner ables
Passion Fruit Establish 30-40 demon- Central, Smallhold- Smallhold-
stration plots and offer Western, and ers and ers attend-
training in: Rift Provinces smallhold- ing the
er associ- training
Grafting techniques KARI ations cur- Associ-
Planting techniques rently or ations with
Seedling production with po- demonstra-
Pruning techniques tential for tion plots
Disease management growing
New products (Jumbo passion
passion fruit) fruit
Private
nurseries
Passion fruit Development of commer- Central, Commer-
cial nurseries Western, and cial nursery
Rift Provinces operators
Smallhold-
ers pur-
chasing
products
from nurs-
eries
Passion fruit Search for investor in Central,
passion fruit processing Western, and
plant (to buy up passion Rift Provinces
fruit produced by small-
holders)
91
ANNEX B: ISSUES TO FOLLOW UP IN ROUND TWO
Avocado
- Whether avocado production increases and quality improves

- In terms of trust between growers and EAGA – whether MOUs are worked out;
whether agreements are abided by
- In terms of dynamics of participation of the poor: Whether smaller scale grow-
ers are included in or left out of groups
- Whether those left out benefit through spin offs -- increased market demand,
increased prices, and increased employment
- Exporter preference for dealing with groups vs. brokers vs. smaller number of
larger growers
- Whether Kenyan avocados become more competitive in export markets
- For exporters, whether the shift from spot markets to retail markets continues
to benefit the industry
- Whether the avocado oil processing plant opens and the extent to which it pro-
motes increased production and increased income
- Whether the groups continue and relationship between groups and EAGA con-
tinues as KBDS pulls out
- Whether other exporters enter into the avocado value chain and offer competi-
tion to EAGA
- Whether the model of direct links between exporters and producer groups –
marketing links and embedded services – is expanded by EAGA with Avocado
and other horticulture; whether it is replicated by other exporters working with
smallholders involved in the avocado or other horticulture value chains.
- Whether groups and group members expand their activities into other export
crops (and link to exporters through group mechanisms) [intersectoral upgrad-
ing].
- Whether brokers continue to play a role in value chain and, if so, what it is
- How mobile phones are used for marketing purposes
- Whether greater levels of trust develop between smallholders and exporters
- The extent to which cell phones are used to improve information flows within
the value chain.
Mango
- Whether direct market links to exporters can be established

- Whether some system for providing embedded services (inputs) evolves
- Whether farmers expand production, introduce new varieties, improve quality
- Whether collection points for mangos are established and used
- Whether smaller growers continue to participate in groups and benefit; or large
farmers dominate
- Whether some kind of mango processing factory is established and how this im-
pacts mango production and farmer incomes
- Whether Kenyan mangos become more competitive in export markets
- How cell phones effect relations with brokers/marketing links; the extent to
which they promote direct market links for farmers
- What other role the farmer groups play in helping farmers
92
- The extent to which cell phones are used to improve information flows within
the value chain.
Passion fruit
- Whether farmers continue to plant and harvest passion fruit

- Whether grafted seedlings become more available and are adopted by growers
- Whether vibrant (albeit seasonal) Uganda market continues
- Whether processing plant established and if so how this affects production
volumes and farmer incomes
- Whether Kenyan Passion fruit stays competitive – or becomes more competitive
-- in export markets
- How passion fruit piggy backs on to other export crops (or how other horticul-
ture crops piggy back on passion)
- The extent to which cell phones are used to improve information flows
within the value chain.
Cross cutting issues
- Assessment of the types of activities in which they are involved and the
extent/degree of management responsibility that the group assumes for particu-
lar activities vis-à-vis other value chain actors.
93
ANNEX C: CALCULATION OF THE ASSET SCORE GROUP-
INGS
1. The following attributes were selected to indicate a household’s standard of
living:
• Whether farm land owned or rented

• Size of land holding
• Material of house wall
• Roofing material
• Presence of a two-story house
• Presence of a domestic worker
• Number of tables
• Source of drinking water
• Agricultural assets owned
• Cooking arrangements
• Presence of deep freeze
• Consumer durables owned
• Vehicles owned
• Toilet arrangements
• Main floor material of the house
• Number of fruit trees
The philosophy guiding this procedure is that no single question or answer on its
own can adequately summarize a reality as complex as standard of living.
Hence, many variables were used in conjunction with each other to contribute to
the discovery of a single measure of standard of living.
2. The attributes listed above were than factorized into a single factor using
SPSS to find a single unifying theme. The output of this factorization was that
each respondent received a single factor score based on his/her responses to
questions about ALL the listed attributes. Each respondent’s factor score de-
scribes the respondent’s relationship to the extracted theme. Thus the respond-
ents’ factor scores form a continuum from the lowest value to the highest value.
This factor is a measure of the overall living standard of an individual. Factor

analysis is the best mathematical tool available for summarizing a complex real-
ity. While every variable used contributes to the overall factor, the fact that
many are used means that the overall dimension is unlikely to be a reflection of
one item only; rather, the factor is a summarized version of the single reality un-
derlying different response patterns and levels in the constituent measures. This
is precisely what we are looking for as a measure of living standard.
The continuum of values was then cut into eight roughly equal groups, one being
the group with the lowest values and eight being the group with the highest val-
ues. The eight groups are not exactly equal in size because we wanted to
achieve a good spread of each farmer group across several asset score groups.
94
The factor was then cross-tabulated against attributes that were outside those
listed above to determine whether the theme described by the single factor
really was a measure of living standard. The theme described by the single
factor did correlate with the distributions of these other, unused validation vari-
ables (and this is reflected wherever we see the asset score cited in the report)
and thus the factor was deemed suitable to describe the standard of living.
3. Following this process, a smaller “predictor” set of attributes was identified.

The reason for the predictor questionnaire is to avoid having to ask all the same
questions in the follow-up survey, thus reducing the time that will have to be
spent in the field. To arrive at a short set of attributes that would predict the ori-
ginal factor, we did a step-wise regression, which means we took the factor that
was created and made it a dependent variable, and looked for the attributes
from the entire set that best predict the original factor. Step-wise regression is a
technique that cycles through all the attributes supplied and creates models of
prediction using first one, then two, then three variables and so on until all the
supplied variables are used. From these models, one was selected that gave a
sufficiently accurate prediction of the factor value while consisting of the maxim-
um number of questions that we judged to be possible to include in future sur-
veys.
4. Because the predictor scores do not match the factor scores exactly (al-
though they are extremely close), it follows that the “predictor” groups cannot
be exactly the same as the original “factor” groups. Thus, the data were run
again using the predictor attributes, and this is what is then used in analysis for
this round of the survey and will be used again in the next round to ensure com-
parability between the two rounds. These are the asset score groupings that are
shown in this report.
95

MR 33 - Kenya Impact Assessment: Baseline Research Report

Загружено:

Сведения о документе

Оригинальное название

Авторское право

Доступные форматы

Поделиться этим документом

Поделиться или встроить документ

Параметры публикации

Этот документ был вам полезен?

Это неприемлемый материал?

Авторское право:

Доступные форматы

MR 33 - Kenya Impact Assessment: Baseline Research Report

Загружено:

Авторское право:

Доступные форматы

Assessing the Impact of Kenya BDS

For more information on AMAP and related publica-

Accelerated Microenterprise Advancement Pro-

Donald Snodgrass, Ph.D., Institute Fellow Emerit-

Jennefer Sebstad is a consultant working on devel-

Action for Enterprise (AFE) is a non-profit organiz-

Baseline Research Report

Don Snodgrass and Jennefer Sebstad

LIST OF TABLES AND FIGURES..............................5

LIST OF ACRONYMS AND ABBREVIATIONS.........7

I. INTRODUCTIONS DEVELOPMENT ...................12

II. DESIGN OF THE IMPACT..................................22

III. BASELINE RESEARCH FINDINGS...................30

IV. CONCLUSIONS OF THE BASELINE STUDY AND

ANNEX A. DESCRIPTION OF KENYA BDS AND

ANNEX B. ISSUES TO FOLLOW UP IN ROUND

ANNEX C. CALCULATION OF THE ASSET SCORE

Table 1: Value of Horticulture Exports

Figure 1: Activities under BDS Tree Fruit Project

During this time, project staff

Training in EUREPGAP certifica-

This part of the impact assess-

Facilitate integration Market Access

Markets Provision of sus- Improved and sustainable mar- Survey

Medium Scale Growers

Several new varieties of mango have

Improved production techniques for man-

Mango farms tend to have far more trees

The distribution of avocado farms by an-

SITE mango farmers had a similar pattern

e. Innovations and investments

Table 15 also shows the percentage of re-

g. Participation in producer groups

Table 15: Changes in Cultivation Methods and Planting of Fruit Trees (%

Fruit tree planting in past year -- % answering yes

i. Differences by Gender of Farmer

Female proprietors participated in produ-

Nearly all the farms in the survey sell tree

Overall 230 76 7,200 106 2.2

sample suggests other factors may be

In general, decisions to upgrade, join pro-

Table 36: Selected Household Data by Gender of Household Head

In the second round it will be crucial to in-

Government of Kenya. 2000. Kenya Welfare Monitoring Survey.

Horticulture Crops Development Authority. 2003. Export Statistics for Fresh

Jaffee, Steven. 2003. “From Challenge to Opportunity: Transforming Kenya’s

Ruth-Aspaas, Helen. 2003. “Heading household and heading businesses: Rural

United Nations Development Programme, 2001. Human Development Report

USAID/Kenya. 2000. Integrated Strategy Plan 2001-2005.

United States Department of Agriculture, 2005. “World Horticulture Trade and

Value chain Type of activity Geographic Target Participa-

Facilitate the establish

Mangos Form producer groups Eastern and Small Smallholders

Mangos Provide on-farm training Coast Province Small Private ex-

Value chain Type of activity Geographic Target Participa-

Establish revolving fund

Mangos Form producer groups Coast province Small Smallholders

Train private extension

Mangos Information campaign Eastern Province Small Nursery op-

Value chain Type of activity Geographic Target Participa-

Raise awareness of rur-

Value chain Type of activity Geographic Target Participa-

Develop monitoring sys-