AGRIBUSINESS AGENT PROGRAMME, ZIMBABWE
compiled by Sarah Sumbureru, WISE representative, Zimbabwe
Jean Francois Legrand and Phillip Christiannsen
Physical Address: 8 Ross Avenue, Belgravia, Harare, ZIMBABWE
Postal Address: P O Box HG 937, Highlands, Harare, ZIMBABWE
The information contained in this case study was derived from one interview with J F
Legrand and P Christiannsen, review of documents provided by them and DFID's ED
staff in the Central Africa Office, Harare (in particular the Output to Purpose Review of
October 1999, attached), and a subsequent visit to Phillip Christiannsen‟s office located
in Harare, to obtain further facts and documents and review earlier drafts of these notes.
OVERVIEW: This case study is an example of how important baseline information can
be in guiding activity, and how potential for development and impact on poverty may be
missed without it. In AGENT III CARE will be taking a proactive approach to impact
assessment, and drawing on a wide selection of IA methods across the range of
quantitative, qualitative and participatory methods.
Three-year funding for the expansion of the Agribusiness Entrepreneur Network and
Training (AGENT) Programme was approved by DFID in December 1996 and activities
under the expanded programme commenced in February 1997 As the title suggests,
AGENT II was an extension of a pilot phase that began in 1995 and operated with
agents involved in the sale of agri-inputs to smallholder farming households in Masvingo
and Midlands Provinces.
The purpose of AGENT II was to improve access to agri-inputs for smallholder farmers
by the establishment of independent, self-sustainable agents in the communal areas of
Mashonaland East and Central Provinces.
The project is implemented by CARE International in partnership with the relevant
Government agencies (Ministry of Agriculture and Agritex), private sector input supply
companies, small independent storekeepers and rural business persons.
The pilot phase of AGENT had been co-funded by the Governments of Austria (in the
Districts of Gutu and Masvingo in Masvingo Province) and Australia (in Chivi District in
Masvingo Province). In addition to DFID's funding for expansion of AGENT II into
Mashonaland East and Central, the Austrian and Dutch Governments contributed to the
programme in several Districts in Manicaland Province.
DFID carried out an Output to Purpose Review in October 1999, with the project due for
completion in March 2000. A Project Completion Report was carried out in June 2000,
when a no-cost extension was approved up to June 2001. CIDA has approved funding
for a further three year phase, during which AGENT will be operationalised in Masvingo,
Manicaland and Midlands Provinces, and it was agreed that the remaining DFID funds
(approx £60,000) would be used to cover the cost of the input supply fund (ISF)
component of the extended project.
An important aim of the third phase will be to achieve localisation of the AGENT
programme, by encouraging the private sector engaged in wholesale agricultural
supplies to take over its future development. This is considered achievable since the
programme now has good relations with all of the country's main suppliers, and
individual agents have built up their own sustainable links with these suppliers at local
2 OBJECTIVES AND ACHIEVEMENTS OF THE AGENT PROGRAMME
The goal of AGENT is to improve access to agri-inputs for smallholder communal
farmers through the establishment of a network of agri-business agents.
The main focus of the AGENT programme is to link large input supply companies to
small rural traders (or “Agents”). This has been achieved and is now continuing after
the withdrawals of initial DFID project guarantees. In some cases, new relationships
and the growth of the network has occurred without any form of guarantee. This has
been based on the natural development of long-term viable business partnerships
between wholesalers and rural retailers. The programme has had a positive
demonstration effect and thus it‟s impact can also be noted in areas outside the project
2.2 Summary of Achievements
The project experienced a number of fairly substantial revisions to its basic procedures
since it moved into Phase II. These changes affected performance, and key elements
such as the importance of Agent selection, have become more critical. Early
deficiencies in this screening stage led to the inclusion of a number of traders who did
not repay on a timely basis, or who were not fully committed to goals of the programme.
The high instance of credit default early on in the programme life, reflects a common
problem where clients perceive NGO‟s as social and welfare-oriented, rather than
developmental and commercial, and their credit activities more akin to grants.
However, by handing defaulters over to lawyers and actively pursuing debtors, CARE
has slowly established a tougher stance, one that indicates to agents that the
programme is serious and businesslike.
Despite these and other early difficulties, and the wider, economic and political situation
in the third year, AGENT II increased its retention of selected traders by 99, and is now
seeing an increasing number of rural traders in agriproducts emerging as a result of the
increased demand from small holders, thus showing its capacity for generating a
multiplier effect in rural economies.
The fact that no detailed baseline survey of a sample community was carried out before
or after the implementation of AGENT II, combined with the difficulty of attributing
effects (given the changes in market prices, weather and seed types), made it very
difficult, in the OPR, to comment specifically on the possible impact of the programme
on overall agricultural development in the target districts.
However, the OPR states the programme did make considerable achievements:
99 Agents were established as independent traders (against a target of 125; though
151 were selected)
Overall farmer incomes seem to have improved (but stronger poverty criteria is
needed to assess this impact; referred to below under Planned Impact Assements at
AGENT II largely met its objectives of leveraging the private sector into input supply
for small holder farmers. In respect of the facilitated sales to guarantee ratio, the
project facilitated (cumulatively) roughly seven times the value of the original ISF
Given the prevailing economic situation the project has made good progress towards
being sustainable at the AGENT level due to the commercial nature of the
supplier/AGENT relationship. At present 70% of selected AGENTS have
independent commercial links with suppliers.
Overall the range of products offered by AGENTS has increased substantially,
encouraged by promotion activities
There appears to have been a demonstration effect which has led to an increase in
rural traders establishing similar agencies, ie outside of the AGENT programme,
when they can see that small holders are interested in improving the quality and
quantity of their inputs
There is also expected to have been a multiplier effect in rural economies, as a
result of increased trading and agricultural production taking place locally
Few of the above achievements have actually been assessed, however, and the
principal recommendation to emerge from the OPR, which has been enthusiastically
taken up by CARE and AGENT staff, as well as requested by CIDA, is for a series of
impact studies to be carried out at the various levels of AGENT's operations.
3.0 IMPACT ASSESSMENT APPROACHES AND METHODOLOGIES
As a result of the OPR and general change in development practice regarding the
importance of ongoing assessment of impact, and a desire to build methods of ongoing
impact assessment into the design of projects from the start, CARE and AGENT staff
are now entering AGENT III with commitment to conduct of a series of studies. These
3.1 Beneficiary Assessment (Agents and their small holder customers):
Although the project has been largely successful to date in establishing a network of
agri-business agents, as mentioned above there is limited data about the socio-
economic and gender profile of the clients. Monitoring mechanisms will be put in place
to establish the extent to which coverage of agricultural inputs is being increased
through this project, both in terms of geographical reach and poverty reach. The overall
impact will be assessed through a mixture of client profiling at the stores themselves
(client interviews etc.) as well as household interviews in communities surrounding the
Wealth ranking has already been carried out in fifteen communities , followed by
interviews with representatives from the different wealth ranking groups. This is
expected to establish:
the extent to which different wealth groups are able to access agri-inputs
the constraints faced in accessing agri-inputs
whether the types of inputs sold in the stores meet the agricultural needs and
opportunities for diversifying the product range.
Such an Impact Assessment (IA) study is not only illuminating the impact of the
programme but also providing pointers for where the project should go in the future. For
example, should Agent be looking to replicate itself in poorer regions of the country i.e.
widen geographical coverage, or should it be attempting to deepen its reach to poorer
farmers within the areas it is already working? The information is also helping agents
know the needs of their target markets and vary their supplies and services accordingly.
In this way impact assessment is being used as a market research tool.
The IA study is also expected to reveal other constraints to rural livelihoods, an
approach towards which CARE International is moving. A qualitative approach to a
beneficiary assessment which looks at some case study households and/or
communities can also reveal the extent to which factors such as HIV, rural transport,
and rural credit affect rural livelihoods. These issues can then be addressed through
Agent III or complimentary initiatives. It is also important for the study to address the
needs of more vulnerable farmers such as the elderly and female heads of households,
The impact assessment will also provide data to determine whether the form of
marketing of certain inputs is an impediment to poorer and female farmers, particularly
the selling of fertiliser in 50 Kg bags. Although some agents admitted (during the OPR)
that they did break bulk, some agents said they were reluctant to do so because of the
accusations of tampering the product.
The OPR team noted that AGENT II did not provide any credit for farmers although
some agents stated during the OPR that they had flexible payment arrangements with
some of their more trusted customers. If the client profiling of the Impact Assessment
reveals that poorer farmers are not benefiting, CARE can investigate as to how this
constraint might be alleviated in AGENT III.
CARE staff have also indicated that they plan to include in future impact assessment
work, a longitudinal study of the impact of AGENT on both Agents and customers.
3.2 Economic Assessment (Agents and the communities they serve)
In the various meetings held with agents and farmers, during the OPR, it was clear that
the programme had significantly contributed to inputs being available at competitive
prices, especially if the very high transport cost experienced by farmers travelling to
town to purchase their own supplies was included. There were also indications that the
success of some traders had caused other non-project traders to also start stocking
inputs. This, if it generates real and genuine competition will also serve to keep prices in
check and make an additional and useful direct impact on rural incomes.
The fact that some wholesalers have expanded their provision of inputs, through their
own „Agent‟ networks (based on the CARE model) to rural traders has already been
described as the demonstration effect. How much this would have occurred without
AGENT is unknown but clearly the programme showed that money could be made from
delivering agricultural input supplies in rural areas. The attribution of this phenomenon
purely to AGENT may not be incorrect because in the past some suppliers have
expanded their network only to be caught up in the credit default and confusion caused
by government guaranteed loans through the AFC. The interesting aspect now is that
the current system is not based on farmer credit but on cash sales. The programme is
however based on supplier credit and this still creates a risk of Agent default.
However, The calculation of a realistic economic return on the investment was
impossible given that there was little or no indication of the actual income being derived
by the individual agents. Even if this information were available, the analysis would be
difficult because one of the key costs will be the funds lost due to default by non-
performing agents. The benefits gained by individual farmers clearly make a significant
return to economic benefits and the use of purchased inputs have previously been
shown to have a positive return.
Any detailed economic return evaluation would have to try and determine the “without
case” scenario and this would involve detailed work on the actual costs incurred in
travel to the nearest supply point and the costs incurred in transporting the inputs back
to the homestead. In this regard, one aspect of the study would be the role of
homestead to road transport in barrows, scotch-carts, etc. The Impact Assessment
includes analysis of the 'without case' scenario, looking at sample farmers' costs across
3.3 Other research studies in hand or planned:
The OPR showed the need for an improved Management Information System. This has
now been put in place and, as a result, work has commenced on providing more
detailed information about Agents' operations making possible. As a result the following
studies and analyses will become possible:
i) Feasibility analysis to determine potential for marketing of smallholders'
outputs through the AGENT network:
In some project areas smallholders' outputs have already been marketed through the
Agent network, and this service is significant in several ways because it:
provides an important source of cash income for poor farmers for small amounts of
grain (which normally could never be encashed at the market rate);
provides an attractive diversification to Agents‟ businesses by bulking up sorted
outputs to GMB; and
provides the incentive for the private sector to take up, and possibly further invest in,
It is therefore felt that output marketing could become an important cornerstone for any
development of the AGENT III programme, and CARE is undertaking a feasibility
analysis of output sales in AGENT III. This analysis will include investigation of the risks
to open pricing.
ii) Mapping of distribution and coverage of Agents in existing areas:
The aim of this study is to more systematically record and present the basis for
selecting Agents, analysis of the geographical coverage of existing traders and the
impact on possible competition in the business centres selected, as this has real
consequences in terms of an expansion and/or consolidation strategy. If the area
where an Agent is required has been identified, what should happen when that Agent is
dropped or goes out of business (in case of withdrawal or insolvency)?
The following methods will be included in the study:
mapped and physical distribution of agents (possibly GPS data) and population
mapped distribution of other non-Agent input supply points - the possible competition
from GMB should be incorporated in this information base as should information on
Cotco input programme. Sites of existing commercial outlets would also be of
in a few selected locations, description and analysis of the interaction between
Agent and non-Agent traders
analysis of turnover and sales of an updated distribution map to determine if there
are distinct catchment areas.
participatory interviews with agents and their customers would also prove useful in
determining catchment areas and supply point interaction.
The above analysis is expected to reveal where additional agents should be sought in
order to achieve optional coverage and the provision of adequate services to
smallholders. This distribution and density analysis of the existing input supply points
will be done in a cross-location analysis that examines how optimal distribution and
catchment area is affected by:
distance from formal large-scale outlets;
presence or absence of cash cropping.
If there are recommendations about the possible incorporation of a marketing function
for agents (see above ii) , then the above study can also be used to look at the
distribution of marketing points, GMB intake points, private buyers and grain traders.
iii) Identification of characteristics of Agents
The programme now has some 250-300 agents in all its project areas and analysis of a
full data base of the characteristics of this large and diverse group should contribute to
an understanding of what makes a successful Agent. This ongoing observation could
then be used in the future selection of agents and to help target those most likely to be
The type of data to be collected includes:
personal characteristics (age, sex, religion, education, etc.);
level of other business activities (general dealer, bottle store, transport);
facilities available (own store or rented facility);
previous history (employment record, training received).
iv) Identification of Agents' Clients' characteristics
An indication of the different income and resource levels of the different Agents'
customers would be useful on an ongoing basis. Linked to the information obtained
through the impact assessement described at 3.1, the detailed work on the benefits
gained by different customers would enable the variable impact and benefits of the
programme to be calculated This will form part of the longitudinal study referred to
v) Tracking and Analyis of Financial performance of Agents
The differing margins being offered to the Agents by the various suppliers will be better
analysed regularly to help determine the interaction between:
turnover and profits;
credit limits, credit costs and overall income;
effects of consignment or deposit requirements;
contribution agency business makes to overall turnover and income.
vi) Assessing Agents' satisfaction with CARE services
There will be regular review of the opinions of the Agents about the programme either
annually or six-monthly. Areas CARE will focus on include:
the role of project training;
their future training needs;
timing of input supplies
vii) Consideration of whether to intensify or expand the programme
It is expected that improved MIS will be able to guide the answer to this question. If, as
CARE claimed in the OPR, there is an optimal distribution pattern of agents to target
group farmers, then clearly more investigation needs to be carried out to determine
what actually exists on the ground in all the project areas. This analysis should include
all existing rural traders (those selected by CARE, incorporated under the on-going
supplier network expansion – the demonstration effect - and those already operating in
rural areas). Coupled to the distribution of the traders, must be the purchasing power of
The proposed analysis of the existing, and required, number of agents could provide the
scope of this “intensification” phase of the programme. The evaluation could also
indicate if the existing network provides more than adequate coverage and this could, in
turn, suggest the way forward would be better served by an “expansion” of AGENT II
into completely new areas.
viii) Regional and Sectoral Analyses
Besides looking at potential for expanding the AGENT programme further in
Zimbabwe,CARE is also considering whether it can be applied in other countries in the
region, and in which agricultural sub-sectors.
All the above proposed areas of study indicate the importance of building impact
assessment design and methodology into the project from the start.
CARE and AGENT staff believe the Impact Assessment and research studies outlined
above should also contribute to discussions on how the programme should proceed in a
possible subsequent phase. This case study has been an example of how important
baseline information can be in guiding activity, and how potential for development and
poverty impact may be missed without it.
In AGENT III CARE will be taking a proactive approach to impact assessment, and
drawing on a wide selection of IA methods across the range of quantitative, qualitative
and participatory methods. These include:
construction of baselines for all types of research amongst primary (Agents their
clients, households and communities) and some secondary stakeholders (wholesale
longitudingal studies across all these stakeholders, including quantitative and
structured and semi-structured interviews, supported by focus groups
wealth ranking, mapping and other participatory techniques
application of CARE's household livelihood security analytical framework
5.1 DFID Output to Purpose Review (OPR), October 1999
5.2 DFID Project Completion Report, June 2000