Issue Brief No 1, March 2017

Improving market access for

smallholder farmers: What What Works
works in out-grower schemes in SME
– evidence from Timor-Leste Development
1. Key findings The What Works in SME
Development Series is pre-
• Smallholder farmers often lack access to profitable, value-added senting key findings of inter-
markets. In the absence of critical supporting functions – such as ventions promoting small
infrastructure and service provision – farmers struggle to shift from and medium enterprises as
subsistence and barter to more productive forms of exchange. a means to create more and
better jobs. It covers ILO
• Contract farming models – where buyers and farmers enter into a programs as well as inter-
forward agreement for agricultural production – is one way of linking ventions of other agencies
smallholders to value chains. Such ‘out-grower’ models can result in using ILO products.
improved access to technical assistance and inputs such as hybrid
seeds, as well as a secured market and stable prices. The main objective of the
new Series is to increase
• 30% of the farmers participating in the scheme increased their income the take up of effective SME
by 14 % or more and rose above the national poverty line. These results programs by leading actors
point to the potential for large income gains and improvement in the in this field. The issue briefs
productivity of self-employment. However, the sustainability of results target ILO constituents,
should not be taken for granted. Impact faded away as half of the other policy makers, deve-
farmers dropped from the out-grower scheme. lopment practitioners, and
the private sector presen-
• Interventions can maximize their chances of success by a) ensuring ting key evidence at a glance.
contract farming does not overload support to small holders in the Preference is given to rigo-
early stages of the scheme, b) by ensuring constant information flows rous quantitative research,
between buyers and suppliers, and encouraging transparent setting but the Series also covers
of price ceiling and floors, and c) being open to adapt and change the other research approaches
scheme over time to arrive at a commercially viable model. contributing to more evi-
dence on what works and
2. The challenge what does not work.

Across the developing world, people rely on agriculture to generate income The What Works Series is
and support livelihoods. For those in poverty, however, farming is often coordinated by the SME
characterized by low-value added activities – such as subsistence, barter or Unit of the International La-
selling in local markets. Smallholders find themselves ‘locked out’ of more bour Organization, for more
lucrative markets serving regional, capital, or export value chains. The reasons info see www.ilo.org/sme
for this are due to a lack of economies of scale, low awareness of market
demand – as well as insufficient knowledge about 3. The program and evaluation design
production practices required to meet this demand
– and poor infrastructure leading to high transaction BOSS and Josephina Farm jointly identified a group
costs. of 45 farmers (16 women and 29 men) who had
been growing ‘traditional’ produce such as tomatoes,
The situation is no different in Timor-Leste where, cauliflower and cucumber for home consumption
despite recent and rapid oil-fueled growth, poverty and local selling. Josephina Farm provided the
has remained acute with half the population living on farmers with a one-day on-site training on good farm
less than a dollar a day. Three quarters of the poor management practice, including about irrigation and
live in rural areas, and most of them are farmers. seed beds. Based on end-market demand, Josephina
While demand for vegetable products in the country’s Farm then reached a agreement with farmers on what
capital Dili is growing, it is largely satisfied by rising vegetables to grow, and provided a verbal buy-back
imports. Domestic production of vegetables is small- guarantee for a specified quantity, and a minimum
scale, scattered, and lacking in supporting services. quality, of produce. Josephina Farm also provided
The country’s first dedicated agricultural input improved seeds to farmers to encourage them to
supplier, for example, only opened in 2012. grow new varieties such as zucchini and coriander.
During the growing season, farmers received at least
The ILO Business Opportunities and Support Services one follow-up monitoring visit from the company to
(BOSS) partnered with Josephina Farm, a company receive advice on horticulture production, including
supplying organic vegetable produce for onwards seed management, planting and watering.
retail in the capital1. BOSS supported Josephina Farm
to set up an out-grower scheme to source produce
from farmers in Ainaro – one of the poorest districts
in Timor-Leste, but with an altitude, climate, soil
quality favorable for horticulture production. The
arrangement was designed to help the company
secure a reliable, low-cost supply of vegetable
produce, and give smallholders access a higher
value-added market to increase their profitability
and on-farm productivity.

The development impact logic of contract farming

models is based on an assumption that switching
from producing for own-consumption (where family
food needs are satisfied first, with only the excess
being sold) to producing for commercial markets
(where market needs are satisfied first, and farmers
may have to purchase staple crops to feed their
family) results in net positive outcomes. It also based
on the assumption that once the economic benefits The farmers received a further 2 days of training from
have been proven – and farmers and the firm realize Timor-Leste’s Institute of Business Support (IADE)
economic gains - the arrangement will be sustained. , the government ministry where the BOSS project
In the real-world of dynamic markets, however, was embedded, to explain the concept of contract
success cannot be taken for granted, and there are farming and simple marketing and business
no guarantees that outcomes will endure. principles.

The research presented here aims to provide new BOSS did not directly interact with the farmers, but
evidence on the different effects of a contract supported Josephina Farm’s business model in a
farming model in Timor-Leste - from food security to number of ways. BOSS carried out market research
production, market access to profitability - with the for Josephina Farm, and paid for an international
ultimate aim of exploring the critical success factors expert to undertake a technical assessment of the
influencing the sustainability of out-grower schemes. company’s post-harvesting handling practices. BOSS

1 Josephina Farm has three regular supermarket clients in Dili, with occasional buyers including restaurants.
also provided financial support for the construction group of farmers not part of the intervention. The
of a compost production facility, a simple cool room 36 comparison farmers were selected from nearby
facility, equipment for outdoor production and geographic locations, based on matching criteria4.
plastic to create protective tunnels. BOSS organized Comparing the treatment with the control groups
business match-making events for Josephina Farm, helped account for changes not caused by the out-
and funded the production of marketing material and grower model, such as rainfall. A single survey was
TV commercials. BOSS also supported two farmers administered to both groups of farmers, asking them
to undertake study programs in Bali to develop about current practice (in 2014) and past practice
knowledge on indoor and outdoor horticulture. (prior to 2012).

Following the initial out-grower pilot, the company 4. What we found

replicated the same arrangement with a further
three farmer groups, reaching a total of 125 farmers Farm management
in Ainaro and Ermera Districts. Farmers involved in the intervention adopted
better horticulture production techniques than
The model is mapped onto a simplified value chain those not involved. Among the farmers in the out-
as follows: grower scheme, the use of good farm management
(including garden beds, seed nurseries and rain
protection tunnels) averaged 61%, up from a pre-
intervention figure of 44%. This compares to 31% in
Farmers
the comparison group.
BOSS brokers
farm-firm relationship
In the absence of an effective system of public
and provides a range
of grant and technical Distributor or private extension agents, embedded technical
support to support from the company to farmers was critical
Josephina Farm in plugging a knowledge and skills gaps prohibiting
higher productivity practices.
Retailers
Product
Money Selling patterns and food security
Training and inputs
A clear behavior change took place with a shift
Technical and financial
assistance from BOSS Consumers from selling to the local market to selling through
Josephina Farm. Prior to the intervention, farmers
sold an average of 69% of their production at local
markets. After the intervention, this fell to 17%, with
The study took place two years after the intervention 65% being sold instead to the company. Farmers
began2. This raised challenges of possible recall bias reported that the production of food for household
– since farmers may not be able to remember their consumption slightly decreased (from 16% of
practice prior to entering to the out-grower scheme harvest pre-intervention to 11% post intervention).
– but had the benefit of being able to capture the However, the majority of farmers in the treatment
sustainability of the model over three growing group reported having enough food for their families
seasons. during the last year. While 58% of farmers reported
no change in food security, 30% reported a slight
Researchers took a random sample of 33 out-grower improvement, and 10% a slight deterioration. The
farmers across the 4 groups in 2 districts3. To isolate study found an overall mixed effect on food security
the effect of participation in the out-grower model of switching from subsistence to market-selling,
on farm management practice, productivity and which would warrant further investigation.
profits, researchers compared outcomes with a

2  The impact assessment took place in 2014. The intervention began with the first farmer group in 2012. No useable baselines were
collected at the time of intervention.
3  Research was led by the BOSS M&E officer, using local enumerators from the IADE team (who had not previously been involved in the
intervention). The research design was supported by an international expert.
4  Four pre-qualifying criteria were used to check whether the farmers were eligible to be in the comparison group: proximity to a road-
head, landholding size, extent of market-selling, and irrigation method.
Poverty and employment impact comparison group improved the productivity of their
Compared to the non-participating farmers, those self-employment.
in the out-grower scheme recorded net attributable
income increases of $274 per farm per year. Over the Continuity of service provision
same period, farmers in the comparison group had 90% of out-grower farmers reported receiving seeds
net income increases of just $31 per farm per year. from Josephina Farm, although some did not receive
The poverty impact was likely to be significant, since sufficient quantities. In the first growing season,
71% of the farmers supported by Josephina Farm the technical assistance, inputs and wider services
were estimated to be living below of the national (including collection) provided by Josephina Farm
poverty line ($0.88 daily per capita consumption). to the farmers was rated highly. However, in the
second year of operations, complaints grew: Farmers
The researchers used a productive employment reported delayed collections and significant decreased
methodology (see www.ilo.org/thelab for guidance company time spent monitoring production in the
on measuring changes in productive employment) field. This was largely due to the company taking on
to measure whether an increase in income was more contract farmers, and without a commensurate
significant for a farmer, relative to their existing rise in Josephina Farm’s staffing levels, company
earnings and number of dependents. resources became stretched.

Ability to consistently meet orders

During the first season, Josephina Farm was able
to collect the expected volume of vegetables from
its out-grower farmers. While some lower grade
produce was rejected, the overall quality was good.
Both quality and quantity suffered from year two,
with some farmers switching to ‘traditional’ varieties,
and others beginning to side-sell and not meeting
their obligations to the company. This came despite
real income gains from farmers participating in
the scheme. The researchers found no correlation
between drop-out farmers and their profits: in other
words, those farmers stopping selling to Josephina
Farm were equally likely to have been benefitting
from the arrangement as those farmers continuing
to sell. By the third growing season, the attrition
rate was 47%, meaning just 59 (out of 125) farmers
continued to be part of the out-grower scheme.
BOSS used a proxy based on Timor Leste’s poverty
gap – which is a World Bank figure that estimates 5. Policy recommendations
how far, on average, poor people’s incomes are from
the poverty line. In Timor-Leste, the poverty gap was Smallholder out-grower schemes are inherently
14% - meaning the average family needs to earn 14% fragile. Progress is not linear, and success in one year
more to graduate from poverty. For 30% of farmers, is not a predictor of the continued success in the
the additional income from horticulture farming next. Indeed, arrangements can quickly unravel due
resulted in an overall household income increase of to a number of factors, including prevailing market
greater than 14% - and therefore they can be said to conditions such as prices, or exogenous factors such
have significantly improved the productivity of their as droughts.
self-employment5. While remainder of the farmers
recorded income increases from participating in the Interventions need to deploy models and methods
out-grower scheme, this represented less than a 14% that understand smallholders not only as rational
rise in their household income. No farmers in the economic actors, but also as humans shaped by

5  Note that this does not mean that 30% of out-growers moved from being ‘poor’ to ‘not poor’, since the 14.2% figure is a national average
– not specific to the out-grower households. The benefit of using this proxy is that is allows poverty impact to be estimated where detailed
and expensive household surveys (to capture consumptions and expenditure) are not possible.
their particular culture, context and history. Intervention teams need to
know not just ‘what’ happened, but ‘why’. Research that involves quantitative
analysis into commercial viability and qualitative methods such as interviews
is therefore essential to provide actionable insight. Despite positive income
gains, the contract farming model in Timor-Leste ultimately failed to scale or
sustain due to complex reasons. Focus group discussions with farmers raised
a number of issues such as poor information flows (changing market prices
were not effectively communicated to farmers) as well as low motivation
levels due to the perceived decrease in support from the company. The
absence of agreed-on price ceiling and floors between the company buyer
and the farmer produce proved to be an issue when market prices fell. While
the contract farming remained a verbal, rather than written agreement, it
is doubtful this would have made any difference given the weak nature of
contract enforcement in Timor-Leste, and the fact that transaction completion
is based more on trust than legalese.

In future, contract farming interventions should be careful to not overload

too much support onto smallholders. Such ‘quick wins’ come at the expense
of longer-term sustainability. Once key elements such as technical assistance
– whether provided by the donor project or the company – are inevitably
down-sized in subsequent years (as the model scales up), it risks undermining
earlier gains.

From the company perspective, working towards a profitable business model

to secure a supply chain is a process, not an event. Interventions need to be
open to pivoting and supporting the adaptation of models, or the adoption of
new ones. While continuing with a more limited out-grower scheme, Josephina
Farm also began to experiment with greater backward integration. Instead
of sourcing from smallholders, the company utilized three greenhouses to
produce their own vegetables, using hired farm laborer rather than solely
depending on farmers working their land. Production and quality can now be
more cost-effectively monitored and controlled, at the same time as creating
new jobs.

6. Further readings
Major, Annie and Ripley, Matt (2015), ‘The BOSS project in Timor-Leste; Thin
markers, thick impact?’

The BOSS Project (2016), ‘Horticulture Sector Intervention Report’

Note on Impact Evaluation Technique: The methodology applied for this

research was a quasi-experiment meaning that a control group was matched
using nearby farmers with similar characteristics that did not participate in
the scheme. Impact is the difference in outcomes between the treatment
and control group. Baselines were reconstructed on a recall basis. As this
method cannot fully exclude differences between the two groups results
are of medium reliability and cannot be generalized for other contexts or
countries.

Author: Matt Ripley, Senior Advisor The Lab, ILO (www.ilo.org/thelab)

Sponsors: Irish Aid, New Zealand Aid Programme, Swiss Confederation – State
Secretariat for Economic Affairs